Microelectronic integration process data management method and system

Abstract

The invention discloses a microelectronic integration process data management method and system, and relates to the field of microelectronic integration technology.The method comprises the steps that firstly, parameter data of all key processing links of a microelectronic integration packaging product are collected through various sensors; preprocessing the data, and converting the data into a packaged data unit for storage; carrying out stability quantitative evaluation on the production path of the microelectronic integration process, and establishing a microelectronic integration process data management model; carrying out process flow association, determining a process dependency relationship, and carrying out abnormal problem analysis and tracing positioning to obtain microelectronic integration process association data; and integrating and packaging the packaging data unit and the associated data of the microelectronic integration process to obtain the data management system of the microelectronic integration process. The problem that in the prior art, centralized systematic management is not carried out on microelectronic integration process data, and parameters of related process steps are not managed, so that accurate product abnormal process step tracing cannot be carried out is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of microelectronic integrated process, and particularly relates to a microelectronic integrated process data management method and system. BACKGROUND

[0002] The statements in this section merely provide background information related to the present disclosure and can not constitute the prior art.

[0003] The existing microelectronic integrated process has many process flows, especially the processing flow of high-density packaging products has a large number, and lacks data management and monitoring of the process flow, which often causes errors in process parameters in the microelectronic integrated process steps, resulting in a large number of defective products, and ultimately causing the entire process flow line to be replaced or checked and corrected, resulting in great delay in microelectronic integrated process production time and efficiency.

[0004] Therefore, a method and system capable of systematically managing the entire microelectronic integrated process data to monitor the microelectronic integrated process flow are needed in the market. SUMMARY

[0005] The present application aims at: in view of the problem that the prior art does not centrally and systematically manage the microelectronic integrated process data, and does not manage the parameters of the related process steps, resulting in the inability to accurately trace the abnormal process steps of the product, a microelectronic integrated process data management method and system are provided, which solve the above problems.

[0006] The technical scheme of the present application is as follows: Step S1: collecting parameter data of each key processing link of microelectronic integrated packaging products through various sensors; Step S2: data preprocessing is performed on the collected parameter data of each key processing link of the packaging products, and the data is converted into packaging data units for storage; Step S3: performing stability quantitative evaluation on the production path of the microelectronic integrated process, and establishing a microelectronic integrated process data management model; Step S4: correlating the process flow according to the microelectronic integrated process data management model, establishing the process dependency relationship to analyze and trace the abnormal problems, and obtaining microelectronic integrated process correlation data; Step S5: integrating and packaging the packaging data units and the microelectronic integrated process correlation data to obtain a microelectronic integrated process data management system.

[0007] Preferably, the step of collecting parameter data for each key processing stage of the packaged product using various sensors includes: collecting real-time temperature changes of the heating platform, etching reaction chamber, and metallization precipitation furnace of the substrate bonding equipment using high-precision contact or non-contact temperature sensors; collecting the actual temperature during the substrate bonding process using thermocouple sensors; collecting temperature changes of high-temperature process stages in the microelectronic integration process using special high-temperature resistant temperature sensors; installing pressure sensors in gas and liquid delivery pipelines and vacuum chambers to collect gas pressure and vacuum level, and providing real-time feedback on pressure changes; installing flow sensors in gas and liquid delivery systems to collect the flow rates of various process gases and chemical reagents; installing displacement sensors on substrate handling equipment and lithography machines to collect the position and moving distance of the substrate in real time; and collecting light intensity and wavelength parameters of photodiodes and spectrometers during the microelectronic integration process using optical sensors.

[0008] Preferably, the collection of parameter data for each key processing stage of the packaged product through various sensors also includes: collecting the morphology, defects, and thickness of the substrate surface using optical inspection equipment; collecting the quality and performance of the chip using electrical inspection equipment; and collecting specific packaged product structure and thermal performance data using X-ray inspection equipment and infrared thermal imagers.

[0009] Preferably, the step of preprocessing the parameter data of each key processing stage of the packaged product and converting it into packaged data unit storage includes: cleaning the parameter data of each key processing stage of the packaged product; removing noise data collected by sensors using filtering algorithms, selecting different filtering methods for different types of sensor data, such as median filtering / mean filtering to remove high-frequency noise for temperature sensor data, and low-pass filtering to remove noise caused by pressure fluctuations for pressure sensor data; filling missing values ​​that occur during data acquisition using interpolation methods, selecting one of linear interpolation, cubic spline interpolation, and nearest neighbor interpolation methods to handle missing values ​​according to the characteristics and distribution of the data; identifying outliers through statistical analysis methods and processing them, calculating the mean, standard deviation, and median of the data, and deleting or replacing the outliers with the mean or median.

[0010] Preferably, the data cleaning of the collected parameter data of each key process link of the packaged product further comprises: data standardization and normalization of the parameter data of each key process link of the packaged product after data cleaning; converting the parameter data of each key process link of the packaged product after data cleaning into a standard normal distribution with zero mean and unit variance, eliminating the dimensional differences between different sensor data, and improving the accuracy and comparability of data analysis; mapping the data to a specific interval to allow different data to be normalized on the same scale for comparison.

[0011] Preferably, the data standardization and normalization of the parameter data of each key process link of the packaged product after data cleaning further comprises: data fusion, labeling and data conversion storage of the parameter data of each key process link of the packaged product after data standardization and normalization; multi-sensor data fusion, time series data fusion, and spatial data fusion of the parameter data of each key process link of the packaged product after data standardization and normalization; using weighted average method, Kalman filter, and fuzzy logic method to fuse data from different sensors and give different weights according to the accuracy, reliability and correlation of different sensors; using moving average method, exponential smoothing method, autoregressive moving average model, and difference integrated moving average autoregressive model to smooth and predict time series data and remove short-term fluctuations; using interpolation method, Kriging interpolation method, and inverse distance weighting method to interpolate and fit spatial data; automatically labeling the parameter data of each key process link of the packaged product according to process requirements and quality standards; establishing a data warehouse and storing and managing the collected parameter data of each key process link of the packaged product after preprocessing using a relational database.

[0012] Preferably, the step of establishing a microelectronic integrated process data management model by quantitatively evaluating the stability of the production path of the microelectronic integrated process comprises: analyzing the principal components in the production path of the microelectronic integrated process, screening the influence of parameters in different sections on the final quality, analyzing the correlation between parameters within the sample through scatter matrix diagram, eliminating strong correlation factors, and aggregating the analyzed factors; finding the aggregation of the analyzed factors in different dimensions through commonality analysis; quantifying the influence of different analyzed factors on quality through decision tree analysis, screening the best parameter combination, confirming the optimal production path, and establishing the microelectronic integrated process data management model.

[0013] Preferably, the process flow correlation according to the microelectronic integrated process data management model is established, the process dependency relationship is determined for abnormal problem analysis and trace positioning, and the microelectronic integrated process correlation data is obtained, comprising: determining the best process parameter combination by analyzing the process data of the microelectronic integrated process data management model; predicting the performance index of the product by using the microelectronic integrated process data management model, estimating the potential problems of the performance index of the product and adjusting and correcting the best parameter combination of the microelectronic integrated process data management model; combining the microelectronic integrated process data management model to study the variable correlation between different process steps through historical data, determine the key factors affecting product performance, and perform cluster analysis on different batches of products to perform trace positioning of defective products.

[0014] Preferably, the packaging data unit and the microelectronic integrated process correlation data are integrated and packaged to obtain a microelectronic integrated process data management system, comprising: integrating and packaging the packaging data unit and the microelectronic integrated process correlation data to establish a binary tree-shaped multi-layer microelectronic integrated process data management data system; the first layer tree model of the binary tree-shaped multi-layer microelectronic integrated process data management data system is a single packaging data unit, which is a point step of the microelectronic integrated process flow; the second layer tree model of the binary tree-shaped multi-layer microelectronic integrated process data management data system is the microelectronic integrated process correlation data, which is a segmented step of the microelectronic integrated process flow; the remaining layer tree model of the binary tree-shaped multi-layer microelectronic integrated process data management data system is a multi-segmented second layer tree model set, which is a set of segmented steps of the microelectronic integrated process flow.

[0015] The application also provides a microelectronic integrated process data management system capable of realizing the microelectronic integrated process data management method, comprising: The data acquisition unit acquires parameter data of each key processing link of the packaged product through various sensors; The data processing unit pre-processes the acquired parameter data of each key processing link of the packaged product; The data storage unit establishes a data warehouse and stores and manages the pre-processed parameter data of each key processing link of the packaged product using a relational database; The model establishment unit quantitatively evaluates the stability of the production path of the microelectronic integrated process and establishes a microelectronic integrated process data management model; The abnormal analysis unit correlates the process flow according to the microelectronic integrated process data management model, determines the process dependency relationship for abnormal problem analysis and trace positioning; The data packaging unit integrates and packages the packaging product data unit and the microelectronic integrated process related data.

[0016] Compared with the prior art, the present application has the following beneficial effects: The present application collects various parameters in the microelectronic integrated process through various sensors, realizes full coverage of the microelectronic integrated process steps, realizes correction by discovering processing step abnormalities in the first time through sensor data, can reduce the loss in the microelectronic integrated process to the minimum, then realizes format conversion and storage of various parameters through data cleaning, standardization and normalization, data fusion and labeling and data conversion, realizes common storage of different parameters, can save more storage resources, establishes a microelectronic integrated process data management model according to the parameters, correlates the entire microelectronic integrated process, realizes analysis and tracing positioning of abnormal problems, and establishes a systematic microelectronic integrated process data management system. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the drawings needed in the specific embodiments or prior art description will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale.

[0018] Figure 1 A flowchart of a microelectronic integrated process data management method provided by an embodiment of the present application is shown in the figure. Figure 2 A block diagram of a microelectronic integrated process data management system provided by an embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0019] It should be noted that the terms "first" and "second" and the like such relational terms are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. Without more limitations, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or equipment including the element.

[0020] The features and performances of the present application will be further described in detail below in combination with embodiments.

[0021] Embodiment one The existing microelectronic integrated process has many process flows, especially the processing flow of high-density packaging products has many processes, and lacks data management and monitoring of the process flow, which leads to errors in process parameters in the microelectronic integrated process steps, resulting in a large number of defective products, and ultimately leading to the need to replace or check and correct the entire process flow line, resulting in great delay in microelectronic integrated process production time and efficiency, so a method and system for systematically managing the entire microelectronic integrated process data to monitor the microelectronic integrated process flow is needed in the market.

[0022] The present application aims to provide a microelectronic integrated process data management method and system, which solves the problem that the prior art does not centrally and systematically manage microelectronic integrated process data and does not manage the parameters of related process steps, resulting in the inability to accurately trace abnormal process steps of products.

[0023] As shown in Figure 1 The embodiment of the present application provides a microelectronic integrated process data management method, which comprises the following steps: S1: Collecting parameter data of each key processing link of packaging products in the microelectronic integrated process by various sensors.

[0024] In the embodiment of the present application, high-precision contact or non-contact temperature sensors are used to collect the real-time temperature change conditions of the heating platform of the substrate bonding device, the etching reaction cavity and the metallization deposition furnace; thermocouple sensors are used to collect the actual temperature during the substrate bonding process; special high-temperature resistant temperature sensors are used to collect the temperature change conditions of the high-temperature process link in the microelectronic integrated process; pressure sensors are installed in the gas and liquid delivery pipelines and vacuum cavities to collect gas pressure and vacuum degree and real-time feedback pressure change conditions; flow sensors are installed in the gas and liquid delivery system to collect the flow of various process gases and chemical reagents; displacement sensors are installed on the substrate handling device and the photolithography machine to real-time collect the position and movement distance of the substrate; optical sensors are used to collect the light intensity and wavelength parameters of the photodiode and spectrometer in the microelectronic integrated process; optical detection equipment is used to collect the topography, defects and thickness of the substrate surface; electrical detection equipment is used to collect the quality and performance of the chip; X-ray detection equipment and infrared thermal imaging equipment are used to collect specific substrate structure and thermal performance data; high-precision contact or non-contact temperature sensors are arranged at positions such as the heating platform of the substrate bonding device, the etching reaction cavity and the metallization deposition furnace in each key link of packaging product processing. For example, thermocouple sensors are used to measure the actual temperature during the bonding process to ensure that the temperature is within a suitable range to ensure the bonding quality. For high-temperature process links, special high-temperature resistant temperature sensors can be selected to ensure the accuracy and stability of the data; Pressure sensors: Install pressure sensors at gas delivery pipelines, vacuum chambers, etc. to monitor gas pressure and vacuum degree in real time. For example, in the chemical vapor deposition (CVD) process, accurate control of the pressure in the reaction chamber is crucial to the quality of the deposited layer. Pressure sensors can provide timely feedback on pressure changes to adjust process parameters; Flow sensors: Install flow sensors in gas and liquid delivery systems to monitor the flow of various process gases and chemical reagents. For example, in the etching process, accurate control of the flow of etching gas can ensure etching rate and uniformity. Flow sensors can provide real-time flow data for accurate flow adjustment; Displacement sensors: Install displacement sensors on substrate handling equipment, lithography machines, etc. to monitor the position and movement distance of the substrate. This is very important to ensure the accurate position and alignment accuracy of the substrate in each process. For example, in the lithography process, accurate displacement control can ensure accurate pattern transfer; Optical sensors: Use optical sensors such as photodiodes, spectrometers, etc. to monitor light intensity, wavelength, etc. during the process. For example, in the lithography machine, optical sensors can monitor the intensity and stability of the exposure light source to ensure lithography quality. In the substrate detection link, optical sensors can detect defects and topography on the substrate surface; Use optical detection equipment such as microscopes, interferometers, scanning electron microscopes (SEM), etc. to collect data such as substrate surface topography, defects, thickness, etc. For example, use a microscope to observe the interface after substrate bonding to detect whether there are bubbles, cracks, etc. defects; Use an interferometer to measure the flatness and thickness of the substrate; Use SEM to observe the via morphology and sidewall roughness after etching; Use electrical detection equipment such as probe stations, testers, etc. to collect electrical performance data of chips such as resistance, capacitance, inductance, leakage current, etc. These data can reflect the quality and performance of the chip, and provide important basis for process optimization; Combine advanced detection equipment such as X-ray detection equipment, infrared thermal imager, etc. to collect more detailed structure and thermal performance data. For example, use X-ray detection equipment to analyze the thickness and uniformity of the metallization filling layer; Use infrared thermal imager to monitor the temperature distribution of the chip during work to evaluate the heat dissipation performance.

[0025] S2: Data preprocessing is performed on the collected parameter data of each key processing link of the packaged product, and is converted into a packaged data unit for storage.

[0026] In the embodiments of the present application, the parameter data of each key processing link of the collected packaging products is subjected to data cleaning; a filtering algorithm is used to remove the noise data collected by the sensor; for different types of sensor data, different filtering methods are selected; for temperature sensor data, median filtering / mean filtering is used to remove high-frequency noise; for pressure sensor data, low-pass filtering is used to remove noise caused by pressure fluctuations; for missing values occurring in the data collection process, interpolation method is used to fill in; according to the characteristics and distribution of the data, one of the linear interpolation, cubic spline interpolation and nearest neighbor interpolation is selected to process the missing values; through statistical analysis method, the abnormal values are identified and processed, the mean, standard deviation and median of the data are calculated, the data points exceeding the specified range are regarded as abnormal values and the abnormal values are deleted or replaced by the mean or median; the parameter data of each key processing link of the packaging products after data cleaning is subjected to data standardization and normalization; the parameter data of each key processing link of the packaging products after data cleaning is converted into a standard normal distribution with zero mean and unit variance, the dimensional difference between different sensor data is eliminated, and the accuracy and comparability of data analysis are improved; the data is mapped to a specific interval, and different data is normalized for comparison on the same scale; the parameter data of each key processing link of the packaging products after data standardization and normalization is subjected to data fusion, labeling and data conversion storage; the parameter data of each key processing link of the packaging products after data standardization and normalization is subjected to multi-sensor data fusion, time series data fusion and spatial data fusion; the weighted average method, Kalman filtering and fuzzy logic method are used to fuse the data of different sensors, and different weights are given according to the accuracy, reliability and correlation of different sensors; the moving average method, exponential smoothing method, autoregressive moving average model and difference integrated moving average autoregressive model are used to smooth and predict the time series data, and remove short-term fluctuations; interpolation method, Kriging interpolation method and inverse distance weighting method are used to interpolate and fit the spatial data; the parameter data of each key processing link of the packaging products is automatically labeled according to the process requirements and quality standards; a data warehouse is established, and the parameter data of each key processing link of the packaging products collected after preprocessing is stored and managed by using a relational database. A filtering algorithm is used to remove the noise data collected by the sensor. For different types of sensor data, appropriate filtering methods can be selected. For example, for temperature sensor data, median filtering or mean filtering can be used to remove high-frequency noise; for pressure sensor data, low-pass filtering can be used to remove noise caused by pressure fluctuations; For missing values occurring in the data collection process, interpolation method can be used to fill in. Common interpolation methods include linear interpolation, cubic spline interpolation, nearest neighbor interpolation, etc. According to the characteristics and distribution of the data, appropriate interpolation method is selected to process the missing values; Identify and process outliers through statistical analysis methods. Calculate statistical quantities such as mean, standard deviation, median, etc. of the data, and consider data points outside a certain range as outliers. For outliers, methods such as deletion, replacement with mean or median, etc. can be used for processing; Convert data to standard normal distribution with zero mean and unit variance. Standardization can eliminate the dimensional differences between different sensor data, improve the accuracy and comparability of data analysis. For example, for data of different physical quantities such as temperature, pressure, flow, etc., standardization methods can be used for processing; Map data to a specific interval, such as [0,1] or [-1,1]. Normalization can make different data comparable on the same scale, facilitating subsequent data analysis and model training. For example, for the electrical performance parameters of chips, normalization methods can be used for processing to better evaluate the performance of the chips; Multi-sensor data fusion: fuse data from different sensors to obtain more comprehensive process information. Weighted average method, Kalman filter, fuzzy logic, etc. can be used for multi-sensor data fusion. According to the accuracy, reliability and correlation of different sensors, different weights are given to improve the accuracy and reliability of the fused data; For data collected by the same sensor at different time points, perform time series analysis and fusion. Moving average method, exponential smoothing method, autoregressive moving average model (ARMA), difference integration moving average autoregressive model (ARIMA), etc. can be used to smooth and predict time series data, highlighting long-term trends and removing short-term fluctuations, providing more accurate prediction information for process optimization; For data related to different positions in microelectronic integrated processes, such as temperature, thickness, electrical performance, etc. of different regions of the substrate, perform spatial data fusion. Interpolation method, Kriging interpolation method, inverse distance weighting method, etc. can be used for interpolation and fitting of spatial data to obtain spatial distribution information of the entire substrate. This helps to analyze the uniformity and consistency of the process and find potential problem areas; Label the collected data for subsequent data analysis and model training. Data can be manually labeled or automatically labeled according to process requirements and quality standards. For example, for data of the substrate bonding process, quality indicators such as bonding strength, interface flatness, bubble number, etc. can be labeled; for data of the via etching process, parameters such as via diameter, depth, perpendicularity, sidewall roughness, etc. can be labeled; For some complex quality indicators, machine learning algorithms can be used for automatic labeling. For example, image recognition algorithms can be used to automatically label defects in substrate surface images collected by optical detection equipment; deep learning algorithms can be used to automatically classify and label electrical performance data of chips; Establish a data warehouse to store and manage the collected raw data and preprocessed data. The data warehouse can be implemented using relational databases, non-relational databases, or distributed file systems, etc. Ensure the security, reliability, and accessibility of the data; Develop data storage and management specifications, including data naming conventions, storage formats, backup strategies, etc. Regularly back up data to prevent data loss. Establish a data access permission management mechanism to ensure the security and confidentiality of the data.

[0027] S3: Quantitative evaluation of the stability of the production path of the microelectronic integrated process, and establishment of a microelectronic integrated process data management model.

[0028] In the embodiment of the present application, by analyzing the principal components in the production path of the microelectronic integrated process, the influence of parameters in different sections on the final quality is screened, the correlation between parameters within the sample is analyzed through scatter matrix analysis, strong correlation factors are eliminated, and the analyzed factors are aggregated; commonality analysis is used to find the aggregation of the analyzed factors in different dimensions; decision tree analysis is used to quantify the influence of different analyzed factors on the quality, screen the best parameter combination, confirm the optimal production path, and establish the microelectronic integrated process data management model.

[0029] S4: According to the microelectronic integrated process data management model, the process flow is associated, the process dependency relationship is established, the abnormal problem is analyzed and traced, and the microelectronic integrated process correlation data is obtained.

[0030] In the embodiment of the present application, by analyzing the process data of the microelectronic integrated process data management model, the best process parameter combination is determined; the performance index of the product is predicted using the microelectronic integrated process data management model, the potential problems of the performance index of the product are estimated, and the best parameter combination of the microelectronic integrated process data management model is adjusted and corrected; the variable correlation between different process steps is determined by combining the microelectronic integrated process data management model with historical data, the key factors affecting the performance of the product are determined, the clustering analysis of different batches of products is performed, and the traceability of the defective products is performed.

[0031] S5: The packaging data unit and the microelectronic integrated process correlation data are integrated and packaged to obtain a microelectronic integrated process data management system.

[0032] In the embodiment of the present application, the packaging data unit and the microelectronic integrated process associated data are integrated and packaged, and a binary tree-shaped multi-layer microelectronic integrated process data management data system is established; the first layer tree-shaped model of the binary tree-shaped multi-layer microelectronic integrated process data management data system is a single packaging data unit, which is a point step of the microelectronic integrated process flow; the second layer tree-shaped model of the binary tree-shaped multi-layer microelectronic integrated process data management data system is microelectronic integrated process associated data, which is a segmented step of the microelectronic integrated process flow; and the remaining layer tree-shaped model of the binary tree-shaped multi-layer microelectronic integrated process data management data system is a multi-segmented second layer tree-shaped model set, which is a set segmented step of the microelectronic integrated process flow.

[0033] As shown in Figure 2 The present application provides a microelectronic integrated process data management system, which comprises: A data acquisition unit acquires parameter data of each key processing link of a packaged product through various sensors; A data processing unit pre-processes the acquired parameter data of each key processing link of the packaged product; A data storage unit establishes a data warehouse and stores and manages the pre-processed acquired parameter data of each key processing link of the packaged product by using a relational database; A model establishing unit quantitatively evaluates the stability of a production path of the microelectronic integrated process and establishes a microelectronic integrated process data management model; An abnormality analysis unit analyzes and traces abnormal problems according to the microelectronic integrated process data management model, establishes a process dependency relationship, and performs abnormality analysis and tracing positioning; A data packaging unit integrates and packages the packaging data unit and the microelectronic integrated process associated data.

[0034] The present application has the following beneficial effects: the present application acquires various parameters in the microelectronic integrated process flow through various sensors, realizes full coverage of the microelectronic integrated process steps, realizes correction by discovering processing step abnormalities in the first time through sensor data, can minimize losses in the microelectronic integrated process flow, and realizes format conversion and storage of various parameters by performing data cleaning, standardization and normalization, data fusion and labeling, and data conversion, realizes common storage of different parameters, can save more storage resources, establishes a microelectronic integrated process data management model according to parameters, correlates the entire microelectronic integrated process flow, realizes analysis and tracing positioning of abnormal problems, and establishes a systematic microelectronic integrated process data management system.

[0035] The above embodiments only express the specific implementation of the present application, which is described in detail and specifically, but cannot be understood as a limitation on the protection scope of the present application. It should be pointed out that, for ordinary skilled persons in the art, without departing from the technical concept of the present application, a number of modifications and improvements can be made, which all belong to the protection scope of the present application.

[0036] This Background section is intended to provide a general overview of the context of the application, the work of the current named inventors, the work of others in the field, and the work of the inventors described in this section to the extent it is not prior art to the application, is not admitted to be prior art by express reference to it in the following disclosure.

Claims

1. A method of managing data for microelectronic integration processes, characterized in that, The method comprises the following steps: Step S1: Collecting parameter data of each key processing link of microelectronic integrated packaging products through various sensors; Step S2: Preprocessing the collected parameter data and converting it into packaging data units for storage; Step S3: Quantitatively evaluating the stability of the production path of the microelectronic integrated process, and establishing a microelectronic integrated process data management model; Step S4: Correlating the process flow according to the microelectronic integrated process data management model, establishing the process dependency relationship, analyzing and tracing the abnormal problems, and obtaining microelectronic integrated process correlation data; Step S5: Integrating and packaging the packaging data units and the microelectronic integrated process correlation data to obtain a microelectronic integrated process data management system.

2. The method of claim 1, wherein, The step S1 comprises: Collecting the real-time temperature change of the heating platform of the substrate bonding equipment and the etching reaction cavity, and the metal deposition furnace through high-precision contact or non-contact temperature sensors; Collecting the actual temperature during the substrate bonding process through a thermocouple sensor; Collecting the temperature change of the high-temperature process link in the microelectronic integrated process through a high-temperature-resistant temperature sensor; Installing pressure sensors in gas and liquid delivery pipelines and vacuum cavities to collect gas pressure and vacuum degree, and feeding back the pressure change in real time; Installing flow sensors in the gas and liquid delivery system to collect the flow of various process gases and chemical reagents; Installing displacement sensors on the substrate handling equipment and the lithography machine to collect the position and movement distance of the substrate in real time; Collecting the light intensity and wavelength parameters of the photodiode and spectrometer in the microelectronic integrated process through an optical sensor.

3. The method of claim 2, wherein, The step S1 further comprises: Collecting the topography, defects, and thickness of the substrate surface using optical detection equipment; Collecting the quality and performance of the chip using electrical detection equipment; Collecting the structure and thermal performance data of the packaging product using X-ray detection equipment and infrared thermal imaging equipment.

4. The method of claim 3, wherein, The step S2 comprises: Data cleaning of the collected parameter data; Using filtering algorithm to remove noise data collected by sensors, for different types of sensor data, different filtering methods are selected, for temperature sensor data, median filtering / mean filtering is used to remove high-frequency noise, for pressure sensor data, low-pass filtering is used to remove noise caused by pressure fluctuation; For missing values occurring in the data collection process, interpolation method is used for filling, according to the characteristics and distribution of the data, one of linear interpolation, cubic spline interpolation and nearest neighbor interpolation is selected for missing value processing; Identify outliers through statistical analysis method and process, calculate the mean, standard deviation and median of the data, and delete the data points exceeding the specified range as outliers or replace them with the mean or median.

5. The method of claim 4, wherein, The step S2 further comprises: Data standardization and normalization of the parameter data after data cleaning; Convert the parameter data after data cleaning into a standard normal distribution with zero mean and unit variance, eliminate the dimensional difference between different sensor data, and improve the accuracy and comparability of data analysis; Map the data to the corresponding interval to compare different data on the same scale.

6. The method of claim 5, wherein, The parameter data after data cleaning is subjected to data standardization and normalization, and further includes: The parameter data after data standardization and normalization is subjected to data fusion, labeling and data transformation storage; The parameter data after data standardization and normalization is subjected to multi-sensor data fusion, time series data fusion and spatial data fusion; Different sensors are fused by using weighted average method, Kalman filtering and fuzzy logic method, and different weights are given according to the accuracy, reliability and correlation of different sensors; Time series data is smoothed and predicted by using moving average method, exponential smoothing method, autoregressive moving average model and difference integrated moving average autoregressive model, and short-term fluctuations are removed; Spatial data is interpolated and fitted by using interpolation method, Kriging interpolation method and inverse distance weighting method; The parameter data is automatically labeled according to process requirements and quality standards; The parameter data after preprocessing is stored and managed by using a relational database.

7. The method of claim 1, wherein, The step S3 includes: By principal component analysis of the production path of the microelectronic integrated process, the influence of parameters in different sections on the final quality is screened, the correlation between parameters in the sample is analyzed by scatter matrix diagram, strong correlation factors are removed, and the analyzed factors are aggregated; By commonality analysis, the aggregation of the analyzed factors in different dimensions is found; By decision tree analysis, the influence degree of different analyzed factors on the quality is quantified, the best parameter combination is screened, the optimal production path is confirmed, and a microelectronic integrated process data management model is established.

8. The method of claim 1, wherein, The step S4 includes: By analyzing the process data of the microelectronic integrated process data management model, the best process parameter combination is determined; The performance index of the product is predicted by using the microelectronic integrated process data management model, potential problems of the performance index of the product are estimated, and the best parameter combination of the microelectronic integrated process data management model is adjusted and corrected; By combining the microelectronic integrated process data management model with historical data, the correlation between variables in different process steps is studied, the key factors affecting the performance of the product are determined, and the clustering analysis of different batches of products is performed to locate the defective products.

9. The method of claim 1, wherein, The step S5 includes: The packaging data unit and the microelectronic integrated process related data are integrated to establish a binary tree-shaped multi-layer microelectronic integrated process data management data system; The first layer tree model of the binary tree-shaped multi-layer microelectronic integrated process data management data system is a single packaging data unit, which is a point step of the microelectronic integrated process flow; The second layer tree model of the binary tree-shaped multi-layer microelectronic integrated process data management data system is the microelectronic integrated process related data, which is a segmented step of the microelectronic integrated process flow; The remaining layer tree models of the binary tree-shaped multi-layer microelectronic integrated process data management data system are a plurality of second layer tree model sets, which are collective segmented steps of the microelectronic integrated process flow.

10. A microelectronic integrated process data management system, characterized by, A microelectronic integrated process data management method according to any one of claims 1-9 is implemented, including: The data acquisition unit acquires parameter data of each key processing link of the packaged product through various sensors. The data processing unit performs data preprocessing on the acquired parameter data. The data storage unit establishes a data warehouse and stores and manages the preprocessed parameter data using a relational database. The model establishment unit quantitatively evaluates the stability of the production path of the microelectronic integrated process and establishes a microelectronic integrated process data management model. The abnormality analysis unit analyzes and traces and locates abnormal problems according to the process flow correlation and the process dependency relationship of the microelectronic integrated process data management model. The data packaging unit integrates and packages the packaged product data unit and the associated data of the microelectronic integrated process.

Images