Semiconductor chip welding method and device

Abstract

The invention discloses a semiconductor chip welding method and device, and the method comprises the steps: carrying out the thermal field scanning of a multi-layer stacked package, obtaining the overall thermal imaging data, and generating a thermal field partition diagram containing a plurality of independent thermal control partitions; performing pulse type backflow heating treatment on each independent thermal control subarea to obtain real-time temperature data; carrying out real-time molten state detection to obtain surface topography data and temperature distribution data; performing atmosphere switching treatment on the independent thermal control subareas in the cooling stage to generate atmosphere switching subareas; performing micro-cycle steady-state detection on the atmosphere switching subareas, and performing local reheating treatment on the abnormal atmosphere switching subareas to generate reheating subareas; and carrying out dynamic cooling treatment on the reheating subarea. According to the technical scheme provided by the invention, through combination of partitioned pulse backflow and real-time monitoring, the welding spot melting state in multi-layer stacked packaging is accurately controlled, and the incomplete backflow defect is effectively prevented.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor chip packaging and welding, and in particular to a semiconductor chip welding method and device. Background Art

[0002] In the context of the continuous iteration and upgrading of high-performance electronic systems and smart terminals, multi-layer stacked packaging (such as Package-on-Package, PoP, or 3D packaging such as three-dimensional silicon through-via TSV) is increasingly becoming a key solution for high-density integration of semiconductor chips. This type of packaging structure usually consists of multiple chips vertically interconnected through micro-bumps or solder balls, and together with the intermediate substrate, it forms a compact stacking form, which can not only integrate more functional units on a limited board area, but also greatly shorten the signal transmission path between chips, and has wide application value in data centers, high-performance computing, and smart mobile devices.

[0003] However, since each layer of chips inside the multi-layer stacked package is different in terms of material, thickness, thermal capacity and heat dissipation path, traditional unified reflow soldering often cannot take into account the uniform melting of all solder joints, and it is easy for the solder joints in some areas to fail to fully melt or wet during the reflow process. This phenomenon of solder joints being only in a "partially melted" or "semi-melted" state can be called "incomplete reflow". It is not easy to detect in time through simple visual inspection or conventional temperature monitoring in the early stage, but it will cause micro-voids or uneven grains to form inside the solder joints, which are prone to cracks or cold solder joints when subjected to temperature cycles or mechanical stress, ultimately reducing the reliability of the entire multi-layer stacked package. Therefore, how to achieve refined zoning control and online detection based on the complexity of temperature field distribution, material thermal capacity differences and local stress concentration of multi-layer stacked structures during reflow soldering has become a key technical challenge to improve welding quality and extend device life. Summary of the invention

[0004] The main purpose of the present invention is to solve the technical problem that in the existing multi-layer stacked package reflow soldering process, it is difficult to perform accurate melting monitoring and zoned temperature control for different chip layers, which leads to "incomplete reflow" of local solder joints and subsequent failure.

[0005] A first aspect of the present invention provides a method for welding a semiconductor chip, the method comprising: Performing thermal field scanning on the multi-layer stacked package to obtain overall thermal imaging data, and generating a thermal field partition schematic diagram including a plurality of independent thermal control partitions according to the overall thermal imaging data; According to the position information of each independent thermal control partition in the thermal field partition schematic diagram, pulse reflux heating treatment is performed on each independent thermal control partition to obtain real-time temperature data of each independent thermal control partition, and pulse reflux parameters of each independent thermal control partition are dynamically adjusted according to the real-time temperature data; During the pulse reflow temperature rise treatment, real-time melting state detection is performed on each of the independent thermal control zones to obtain surface morphology data and temperature distribution data of each of the independent thermal control zones, and the melting state of the solder joints of each of the independent thermal control zones is determined based on the surface morphology data and the temperature distribution data; According to the melting state of the solder joints of each of the independent thermal control partitions, an atmosphere switching process is performed on the independent thermal control partition in the cooling stage to generate an atmosphere switching partition; Performing micro-cycle steady-state detection on the atmosphere switching partition to obtain solder joint surface deformation velocity data and heat distribution data of the atmosphere switching partition, and performing local reheating treatment on the atmosphere switching partition with abnormality according to the solder joint surface deformation velocity data and the heat distribution data to generate a reheating partition; The reheating partition is subjected to dynamic cooling processing, and the cooling rate of the reheating partition is regulated according to the temperature history data of the reheating partition.

[0006] Optionally, the thermal field scanning is performed on the multi-layer stacked package to obtain overall thermal imaging data, and a thermal field partition diagram including multiple independent thermal control partitions is generated according to the overall thermal imaging data, including: Performing multi-angle thermal field scanning on the multi-layer stack package to obtain top thermal imaging data, side thermal imaging data, and bottom thermal imaging data of the multi-layer stack package; Comparing and analyzing the top thermal imaging data, the side thermal imaging data, and the bottom thermal imaging data with a standard temperature curve to generate an initial temperature distribution diagram of the multi-layer stack package; Determine the thermal diffusion coefficient and the expected thermal hysteresis time of each region in the multi-layer stack package according to the initial temperature distribution map, and generate a thermal characteristic distribution map of the multi-layer stack package; Performing critical temperature analysis on the temperature gradient change area in the thermal characteristic distribution map, determining the maximum allowable temperature value and the minimum required temperature value of each area of ​​the multi-layer stack package, and generating a temperature boundary distribution map and temperature threshold data of the multi-layer stack package; A thermal field partition diagram including a plurality of independent thermal control partitions is generated according to the temperature boundary distribution diagram and the temperature threshold data.

[0007] Optionally, the pulse reflux heating process is performed on each independent thermal control partition according to the position information of each independent thermal control partition in the thermal field partition schematic diagram to obtain real-time temperature data of each independent thermal control partition, and the pulse reflux parameters of each independent thermal control partition are dynamically adjusted according to the real-time temperature data, including: According to the position information of each independent thermal control partition in the thermal field partition schematic diagram, a first pulse reflux temperature rise process is performed on each independent thermal control partition to obtain first real-time temperature data of each independent thermal control partition; Performing difference analysis on the first real-time temperature data of each independent thermal control zone and the theoretical temperature data of each independent thermal control zone to obtain temperature deviation data of each independent thermal control zone; Performing a second pulse reflux temperature rise treatment on each of the independent thermal control partitions, recording the temperature change rate of each of the independent thermal control partitions, and obtaining temperature response data of each of the independent thermal control partitions; Calculating the heat accumulation rate of each of the independent thermal control partitions according to the temperature deviation data and the temperature response data, and generating a dynamic compensation coefficient of each of the independent thermal control partitions; According to the dynamic compensation coefficient and the temperature response data, the pulse reflow parameters of each independent thermal control partition are dynamically adjusted.

[0008] Optionally, during the pulse reflow temperature rise process, real-time melting state detection is performed on each of the independent thermal control zones to obtain surface morphology data and temperature distribution data of each of the independent thermal control zones, and the melting state of the solder joints of each of the independent thermal control zones is judged according to the surface morphology data and the temperature distribution data, including: During the heating stage of the pulsed reflow heating treatment, performing a first solder joint surface inspection on each of the independent thermal control partitions to obtain initial surface topography data and initial surface reflectivity data; During the heat preservation stage of the pulse reflow heating treatment, a second solder joint surface inspection is performed on each of the independent thermal control zones to obtain instantaneous surface morphology data and instantaneous surface reflectivity data; Dynamically comparing the instantaneous surface topography data with the initial surface topography data, and performing a variation analysis on the instantaneous surface reflectivity data and the initial surface reflectivity data to generate surface topography data; Continuously collecting temperature change data during the pulsed reflux heating process, performing time series analysis on the temperature change data, and generating temperature distribution data; The melting state of the solder joints of each of the independent thermal control zones is determined according to the surface morphology data and the temperature distribution data.

[0009] Optionally, performing atmosphere switching processing on the independent thermal control partition in the cooling stage according to the melting state of the solder joints of each of the independent thermal control partitions to generate atmosphere switching partitions includes: Perform stress distribution detection on the independent thermal control partitions that have completed melting, obtain interfacial tension data and heat conduction rate data, and generate the partitions to be switched; According to the interfacial tension data and the heat conduction rate data, performing a first atmosphere switch on the partition to be switched, and switching the atmosphere environment of the partition to be switched to a first inert gas environment; Performing surface tension detection on the partition to be switched under the first inert gas environment to obtain surface stress distribution data; Performing a second atmosphere switching according to the surface stress distribution data, switching the first inert gas environment to a second inert gas environment; The second inert gas environment is locally pressure-regulated to match the curing rate with the temperature change of the adjacent independent thermal control partitions to generate atmosphere switching partitions.

[0010] Optionally, the micro-cycle steady-state detection is performed on the atmosphere switching partition to obtain the solder joint surface deformation velocity data and the heat distribution data of the atmosphere switching partition, and the atmosphere switching partition with abnormalities is locally reheated according to the solder joint surface deformation velocity data and the heat distribution data to generate the reheating partition, including: The atmosphere switching partition is subjected to stable temperature reduction control to control the temperature within the temperature range of the plastic state of the soldering point; Performing multi-cycle continuous detection on the atmosphere switching partition within the plastic state temperature range of the solder joint to obtain solder joint surface profile data and regional temperature data at multiple consecutive time points; Calculating the speed of the change of the solder joint surface profile data at a plurality of consecutive time points to generate solder joint surface deformation speed data; Performing distribution analysis on changes in the regional temperature data at multiple consecutive time points to generate thermal distribution data; According to the solder point surface deformation velocity data and the heat distribution data, determining a local area in the atmosphere switching partition where micro-collapse or bump retraction exists, and generating abnormal area data; The abnormal area is subjected to a fixed-point heating treatment to raise the temperature of the abnormal area to a temperature range of 90% to 95% of the melting point of the solder, thereby generating a reheating partition.

[0011] Optionally, the calculating the speed of the change of the solder joint surface profile data at a plurality of consecutive time points to generate solder joint surface deformation speed data includes: Extracting the edge contour coordinates of the weld points at the plurality of consecutive time points according to the weld point surface contour data, and generating weld point contour point cloud data; Performing time-series registration on the solder joint contour point cloud data to generate solder joint displacement vector diagrams at adjacent time points; According to the solder joint displacement vector diagram, the displacement distances of various areas on the solder joint surface at adjacent time points are calculated to generate displacement distribution data; Classifying the displacement distribution data into deformation modes, and dividing the displacement distribution data into micro-collapse deformation data, convex point retraction deformation data and steady-state deformation data; The solder joint surface deformation speed data is generated according to the variation of the micro-collapse deformation data, the bump retraction deformation data and the steady-state deformation data within a unit time.

[0012] Optionally, the dynamically cooling the reheating partition and regulating the cooling rate of the reheating partition according to the temperature history data of the reheating partition include: Performing time series analysis on the cooling process data of the reheating zone to generate temperature history data; Divide the reheating zone into temperature intervals according to the temperature history data, and divide the reheating zone into a high temperature interval, a medium temperature interval and a low temperature interval; Dynamically tracking the surface morphology of the reheating zone in the high temperature interval, the medium temperature interval and the low temperature interval to obtain deformation data of each temperature interval; Generate corresponding target cooling rate data according to the deformation amount data of the high temperature interval, the medium temperature interval and the low temperature interval; A first cooling rate is used for cooling control in the high temperature range, a second cooling rate which is 50% lower than the first cooling rate is used for cooling control in the medium temperature range, and a third cooling rate which is 20% lower than the first cooling rate is used for cooling control in the low temperature range; The first cooling rate, the second cooling rate and the third cooling rate are controlled in real time according to the temperature history data and the deformation data in each temperature interval.

[0013] Optionally, the cooling control is performed on the high temperature interval using a first cooling rate, the cooling control is performed on the medium temperature interval using a second cooling rate that is 50% lower than the first cooling rate, and the cooling control is performed on the low temperature interval using a third cooling rate that is 20% lower than the first cooling rate, including: In the high temperature interval, obtaining stress release data of the reheating zone, setting a reference cooling rate according to the stress release data, and using the reference cooling rate as the first cooling rate; Analyzing surface stress data of the reheating zone when it transitions from the high temperature interval to the medium temperature interval, and determining a variation period of a second cooling rate according to the surface stress data; According to the variation cycle of the second cooling rate, the first cooling rate is reduced by 50% to generate a second cooling rate in the medium temperature range; analyzing the solder point solidification data when the reheating zone transitions from the medium temperature interval to the low temperature interval, and determining a variation period of a third cooling rate according to the solder point solidification data; According to the variation cycle of the third cooling rate, the first cooling rate is reduced by 20% to generate a third cooling rate in the low temperature range; The reheating zones are subjected to zone cooling control, wherein the first cooling rate is adopted in the high temperature range, the second cooling rate is adopted in the medium temperature range, and the third cooling rate is adopted in the low temperature range.

[0014] A second aspect of the present invention provides a semiconductor chip welding device, the semiconductor chip welding device comprising: A thermal field scanning module is used to perform thermal field scanning on a multi-layer stacked package to obtain overall thermal imaging data, and generate a thermal field partition diagram containing multiple independent thermal control partitions according to the overall thermal imaging data; A pulse heating module, used to perform pulse reflux heating treatment on each independent thermal control partition according to the position information of each independent thermal control partition in the thermal field partition schematic diagram, obtain real-time temperature data of each independent thermal control partition, and dynamically adjust the pulse reflux parameters of each independent thermal control partition according to the real-time temperature data; A melting detection module, used to perform real-time melting state detection on each of the independent thermal control zones during the pulse reflow heating process, obtain surface morphology data and temperature distribution data of each of the independent thermal control zones, and determine the melting state of the solder joints of each of the independent thermal control zones according to the surface morphology data and the temperature distribution data; An atmosphere switching module, used to perform atmosphere switching processing on the independent thermal control partitions in the cooling stage according to the melting state of the solder joints of each of the independent thermal control partitions, and generate atmosphere switching partitions; A reheating detection module is used to perform micro-cycle steady-state detection on the atmosphere switching partition, obtain the solder joint surface deformation velocity data and thermal distribution data of the atmosphere switching partition, and perform local reheating treatment on the atmosphere switching partition with abnormalities according to the solder joint surface deformation velocity data and the thermal distribution data to generate a reheating partition; The dynamic cooling module is used to perform dynamic cooling processing on the reheating partition and regulate the cooling rate of the reheating partition according to the temperature history data of the reheating partition.

[0015] This application achieves accurate identification and targeted regulation of differentiated thermal capacity between chips in each layer by introducing refined thermal field scanning and regionalized pulse reflow strategies in the welding process of multi-layer stacked packages. Specifically, a full-scale thermal field scan is first performed on the overall package to distinguish several areas with obvious thermal hysteresis or insufficient thermal diffusion, and then based on the temperature distribution results obtained by the scan, independent pulse reflow parameters are used to ensure that each area can be fully melted within the expected temperature curve. This move can overcome the temperature gradient imbalance caused by a single heating method, thereby avoiding the "semi-molten" state of solder joints in some areas due to uneven heat distribution. Through this process, the potential micro-voids and stress concentration inside the solder joints are significantly reduced, and the pre-buried failure risks are also suppressed at an early stage.

[0016] On this basis, the present application also introduces a series of complementary links such as real-time molten state detection, atmosphere switching, and local reheating compensation to form a complete closed-loop control path. By collecting surface morphology and temperature change data in real time, the area that has not been completely melted or the convex point collapse has begun to appear can be found at the first time. Subsequently, the dynamic switching of atmosphere and pressure is used to alleviate the stress impact during the cooling process, and the detected local defective area is reheated at a fixed point when necessary, so that the solder joint is close to the ideal melting temperature again. Finally, combined with the cooling rate control of multiple temperature sections, all interlayer joints can be solidified at an appropriate speed in different temperature ranges. In this way, the consistency of the internal interconnection of the package can be improved, and the possible microcracks or deformation risks can be reduced to an extremely low level through zoning management, thereby achieving better long-term service reliability. With the above-mentioned interlocking technical design, the present application effectively solves the problem of insufficient local reflow that is difficult to eradicate in traditional processes, and achieves a comprehensive improvement in the welding quality of multi-layer stacked packages without changing the structure and the solder itself. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0018] Figure 1 A schematic diagram of an embodiment of a method for welding a semiconductor chip according to an embodiment of the present invention; Figure 2 FIG. 1 is a schematic diagram of an embodiment of a semiconductor chip welding device according to an embodiment of the present invention.

[0019] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0022] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, "and / or" in the full text includes three solutions. Taking A and / or B as an example, it includes technical solution A, technical solution B, and technical solution that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, which must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0023] An embodiment of the present application provides a method for welding a semiconductor chip. Figure 1A flow chart of a semiconductor chip welding method provided in an embodiment of the present application. In this embodiment, the method includes: See also Figure 1 , performing thermal field scanning on the multi-layer stacked package to obtain overall thermal imaging data, and generating a thermal field partition diagram including a plurality of independent thermal control partitions according to the overall thermal imaging data; In one embodiment of the present invention, the thermal field scanning of the multi-layer stacked package is performed to obtain overall thermal imaging data, and a thermal field partition diagram containing multiple independent thermal control partitions is generated based on the overall thermal imaging data, including: performing multi-angle thermal field scanning on the multi-layer stacked package to obtain top thermal imaging data, side thermal imaging data and bottom thermal imaging data of the multi-layer stacked package; comparing and analyzing the top thermal imaging data, the side thermal imaging data and the bottom thermal imaging data with a standard temperature curve to generate an initial temperature distribution diagram of the multi-layer stacked package; determining the thermal diffusion coefficient and the expected thermal hysteresis time of each region in the multi-layer stacked package based on the initial temperature distribution diagram to generate a thermal characteristic distribution diagram of the multi-layer stacked package; performing critical temperature analysis on the temperature gradient change region in the thermal characteristic distribution diagram to determine the maximum allowable temperature value and the minimum required temperature value of each region of the multi-layer stacked package to generate a temperature boundary distribution diagram and temperature threshold data of the multi-layer stacked package; generating a thermal field partition diagram containing multiple independent thermal control partitions based on the temperature boundary distribution diagram and the temperature threshold data.

[0024] Specifically, when performing multi-angle thermal field scanning on a multi-layer stacked package, an infrared thermal imager or a high-resolution optical thermal detection device can be installed on a movable platform to obtain thermal imaging data of the top, side and bottom respectively during rotation or tilting. In order to ensure the comparability of these thermal imaging data, it is necessary to maintain the same ambient temperature and scanning speed in the same scanning cycle, and the acquired image or temperature field data is uniformly saved in a file of the same format. In actual operation, the multi-layer stacked package can be fixed in the center of the platform, and the top, side and bottom shooting and temperature measurement can be completed by appropriate angle transformation. This method can cover the entire package and the chip connection parts in all directions to avoid missing temperature anomalies caused by package deformation or local shadows. When acquiring data, the infrared thermal imager will convert the infrared radiation information in a specific band into a temperature value. If there is a high reflectivity material or a strong curvature on the surface, the influence of optical reflection on the measurement can also be reduced by surface coating or coating correction before imaging, thereby obtaining a more accurate surface temperature distribution. In one example, if the shell material of the top chip is highly reflective, a uniform coating dose can be used to correct it before scanning, so that the top thermal imaging data is more in line with the needs of subsequent analysis.

[0025] After obtaining the top thermal imaging data, the side thermal imaging data, and the bottom thermal imaging data, these data need to be compared and analyzed with the standard temperature curve to generate an initial temperature distribution map. The source of the standard temperature curve can be the preset temperature curve of the reflow soldering process, or it can be a typical heating and insulation curve formed after multiple experiments or industry standards, which is used to determine when to start melting, when to enter the constant temperature zone, and when to cool down. When comparing the actual scanning results with this standard curve, the image recognition algorithm or the temperature data matching algorithm can be used to first filter out the heat accumulation values ​​corresponding to the same time point and the same heat treatment stage, and then project these results into a three-dimensional coordinate system or a numerical matrix to observe the temperature difference values ​​under each scanning angle. If the temperature of the top chip is significantly higher than the standard curve when it arrives at the reflow high temperature zone, while the side chip is slightly lower, it can be explained that the top part dissipates heat more slowly or the heat capacity difference is greater, which will eventually appear as different colors or area separations on the initial temperature distribution map. The initial temperature distribution diagram will mark the location of the area and the degree of heat deviation from the standard curve in the visual interface. If a certain area shows a large positive deviation, it means that there may be a risk of overheating. If it shows a large negative deviation, it means that there may be temperature lag during reflow.

[0026] Next, based on the above initial temperature distribution map, it is necessary to further calculate the thermal diffusion coefficient and expected thermal hysteresis time of each area of ​​the multilayer stacked package, and then convert this series of calculation results into a thermal characteristic distribution map. The thermal diffusion coefficient refers to the ability of a unit volume of material to conduct heat in a unit time, involving the chip material, pad geometry and connection medium characteristics inside the package. By building a material parameter table or finite element model in the software, the temperature data obtained by scanning can be further fitted to obtain the thermal diffusion coefficient value of each area under the reflow high temperature condition. The expected thermal hysteresis time is comprehensively estimated based on the thermal diffusion coefficient, the thickness of the packaging material and the temperature change rate obtained by scanning, representing the average delay required for a region to reach the peak temperature from the beginning of heat absorption. If the data analysis results of a certain interlayer area show that the thermal diffusion coefficient is relatively small and the expected thermal hysteresis time is too long, it means that this area needs to be heated more strongly or more concentratedly in the subsequent stage to avoid ending the reflow stage before the solder joint is completely melted. At the same time, if the thermal diffusion coefficient of another area is very high, it means that its heat dissipation speed is also fast, and it may cool down in advance in the later stage of reflow, so it is necessary to set a gentler temperature decay curve for this area. The thermal characteristic distribution diagram will intuitively present the thermal diffusion and hysteresis information of each area, and display the thermal conduction and thermal response characteristics of different parts through color scale, numerical grid or text annotation.

[0027] When conducting critical temperature analysis on the temperature gradient change area in the thermal characteristic distribution diagram, it is necessary to calculate the maximum allowable temperature value and the minimum required temperature value corresponding to each area based on the thermal diffusion coefficient and hysteresis time, combined with the maximum allowable temperature threshold of the reflow soldering itself and the material tolerance limit of the chip and the substrate. This process can be based on a data table driven analysis model, input the current maximum observed temperature of each area and the expected heat treatment time period, and automatically solve the upper and lower thresholds to ensure that the chip does not have the risk of overheating damage or insufficient melting. For example, in a high-computing power processing module, if the top metal package has a higher thermal critical point, but the polymer material of the adjacent layer has a lower temperature tolerance, it is necessary to set a lower maximum allowable temperature value in this area to prevent the internal polymer layer from cracking due to overheating. Correspondingly, if the bottom substrate has a large thermal capacity and is very sensitive to too low soldering temperature, it is necessary to set a higher minimum required temperature value to ensure that the solder joints in this area can be completely melted during reflow. After summarizing these upper and lower thresholds, the most critical temperature control link inside the package can be calibrated according to their overlap in spatial distribution, and a temperature boundary distribution map and temperature threshold data can be generated.

[0028] After obtaining the temperature boundary distribution map and temperature threshold data, the partitioning algorithm can be used to divide the entire multi-layer stacked package into several independent thermal control partitions to form the final thermal field partition diagram. This diagram is usually divided into regions by color or lines, and each region has its own corresponding maximum allowable temperature value and minimum required temperature value. Before the start of the welding process, the diagram is linked with the pulse reflow temperature rise control system. By setting different heating timings, heating intensities and insulation time for different partitions, the welding quality can be improved more targeted. If the temperature in the partition exceeds the maximum allowable temperature value during the reflow peak, the instruction to automatically reduce the heating power will be triggered to prevent stress problems caused by overheating of the solder joint. If the temperature is lower than the minimum required temperature value for a long time, the function of appropriately extending the insulation or re-pulsing heating will be automatically activated to keep the solder in a sufficiently molten state. When this thermal field partition diagram is implemented in actual production, it can be combined with multiple groups of independent temperature control modules in the reflow oven or heat treatment platform to manage different corners of the package and even multiple layers of chips in layers, thereby improving the overall consistency of welding quality. Through such a complete process of multi-angle scanning, comparative analysis, thermal characteristics calculation, critical temperature analysis and zoning, the temperature characteristics of each area can be fully understood in the preparation stage of the reflow process, making the subsequent pulse heating and dynamic regulation more targeted and significantly reducing the risk of local "incomplete reflow" or high-temperature damage.

[0029] Please continue reading Figure 1, according to the position information of each independent thermal control partition in the thermal field partition schematic diagram, pulse reflux heating treatment is performed on each independent thermal control partition to obtain real-time temperature data of each independent thermal control partition, and pulse reflux parameters of each independent thermal control partition are dynamically adjusted according to the real-time temperature data; In one embodiment of the present invention, the pulse reflow heating process is performed on each of the independent thermal control partitions according to the position information of each of the independent thermal control partitions in the thermal field partition schematic diagram to obtain real-time temperature data of each of the independent thermal control partitions, and the pulse reflow parameters of each of the independent thermal control partitions are dynamically adjusted according to the real-time temperature data, including: performing a first pulse reflow heating process on each of the independent thermal control partitions according to the position information of each of the independent thermal control partitions in the thermal field partition schematic diagram to obtain first real-time temperature data of each of the independent thermal control partitions; adjusting the pulse reflow parameters of each of the independent thermal control partitions according to the real-time temperature data ... The temperature deviation data of each independent thermal control partition is obtained by performing difference analysis with the theoretical temperature data of each independent thermal control partition; a second pulse reflux heating treatment is performed on each independent thermal control partition, and the temperature change rate of each independent thermal control partition is recorded to obtain the temperature response data of each independent thermal control partition; based on the temperature deviation data and the temperature response data, the heat accumulation rate of each independent thermal control partition is calculated to generate the dynamic compensation coefficient of each independent thermal control partition; based on the dynamic compensation coefficient and the temperature response data, the pulse reflux parameters of each independent thermal control partition are dynamically adjusted.

[0030] Specifically, when the first pulse reflow heating treatment is performed on multiple independent thermal control partitions, a multi-channel temperature controller or a multi-stage programmable reflow soldering machine can be used to make each partition receive a short and high-intensity heat pulse in the same time period. The heat pulse is usually completed by combining rapid heating and short-term heat preservation, the purpose of which is to make the area where the solder joint is located quickly approach the melting point temperature zone. During the implementation process, a separate heating unit can be configured for each partition inside the reflow soldering furnace and connected to the corresponding temperature sensor to ensure that each partition can operate independently according to the set pulse parameters. To illustrate this process, a multi-layer stacked package structure can be set, in which the top chip and the middle substrate position form an independent thermal control partition, and the lower substrate and its surrounding components together constitute another independent thermal control partition. When the device starts the pulse reflow heating, the top chip partition will quickly heat up to close to the melting point of the solder at a specific power, and the lower substrate partition will receive heat pulses of different waveforms in the same time period according to its position and material properties. Through this partitioned pulse method, the first real-time temperature data of each independent thermal control partition can be obtained after the first stage of heating is completed. The temperature data is usually collected by a temperature sensor inside the partition or an external infrared detector, and the result is recorded in real time in the data storage module for difference analysis with the theoretical temperature data.

[0031] When performing a difference analysis on the first real-time temperature data and the theoretical temperature data of each independent thermal control partition, it is necessary to first map the actually collected temperature values ​​to the same time coordinate or the same welding stage. The theoretical temperature data can be derived from the expected temperature curve targeted at the partition, or it can be based on previous simulations or existing welding process standards. By calculating the difference between the actual temperature and the theoretical temperature, it can be determined whether a partition is overheated or underheated during this pulse heating stage. If the temperature difference of a partition is positive and the amplitude is significant, it means that the heat obtained by the partition within the set period exceeds expectations, which will cause overheating risks to the local solder. If the difference is negative and the amplitude is large, it means that the partition has not been fully heated and may subsequently be insufficiently melted. In an example scenario, if the thermal capacity of the top chip partition is small, the theoretical temperature data is set to a heating range of 150 to 180°C, and the temperature detected after the short pulse actually exceeds 190°C, it can be judged that the partition has obvious overshoot. The difference analysis presents these differences in the form of numerical values ​​or charts through data processing software, generates temperature deviation data for each independent thermal control partition, and provides an objective basis for the next step of secondary pulse reflow heating.

[0032] During the second pulse reflow heating, each partition will appropriately adjust the pulse power or pulse duration based on the results of the differential analysis. When recording the temperature change rate of each independent thermal control partition, the temperature acquisition frequency can be increased to a higher level within the control system to ensure that the rapid change process of pulse heating can be captured. The temperature change rate can be calculated by dividing the temperature increment over a period of time by the corresponding time and compared between different partitions. For example, the middle substrate partition showed obvious underheating in the first pulse, so the pulse power was increased and the duration was slightly extended in the second pulse. The results showed that the temperature of the partition increased rapidly in a short period of time, and the temperature change rate increased significantly. By saving the real-time data of this process, the temperature response data can be further obtained, that is, the response of each independent thermal control partition to the heat energy input under the specified pulse conditions. The temperature response data is generally represented by a time-temperature curve or a time-temperature increment diagram. If it is found that the temperature response of a certain area is slow, it is necessary to take targeted compensation measures in the subsequent steps to ensure that the area can achieve full melting at the reflow peak stage.

[0033] When calculating the heat accumulation rate based on the temperature deviation data and the temperature response data, it is necessary to integrate or average the heat increment obtained by each partition during the two pulse stages, and determine the effective heat accumulated over a period of time in combination with the number of solder joints in the partition and their material properties. The heat accumulation rate refers to the effective heat energy actually transferred to the solder joint area per unit time. If the accumulated heat obtained by a partition under the same pulse conditions is too low, the solder joint may not reach the ideal melting temperature during the peak reflow period. In order to make the analysis process more intuitive, a curve can be constructed on the software interface, with the horizontal axis being time and the vertical axis being the heat accumulation value of each partition. The curve is then compared with the critical heat required for solder melting to determine whether there is still a lack of partition distance from the ideal state. When the heat accumulation rate is calculated, a dynamic compensation coefficient is generated to characterize the sensitivity of the partition to the thermal pulse. If the heat accumulation rate of a zone is lower than the set threshold, the dynamic compensation coefficient will be set to a value slightly greater than 1 so that the power can be increased or the pulse duration can be extended in subsequent pulses. For those partitions that heat up too quickly, the dynamic compensation coefficient can be set below 1 to reduce heat input.

[0034] Finally, when the pulse reflow parameters of each independent thermal control partition are dynamically adjusted according to the dynamic compensation coefficient and temperature response data, an adaptive algorithm will be implemented in the control software to include the real-time temperature data of each partition, the deviation and response of the two previous pulses, and automatically generate the temperature control instructions and time period settings for the next round of pulses. If a zone performs stably in the previous two pulses and the heat accumulation rate is close to the ideal value, the subsequent pulses can maintain the original power and cycle without generating additional heating or cooling burdens. For example, when the top chip partition can keep up with the theoretical temperature curve in both pulses, and the temperature deviation and temperature response values ​​are within a reasonable range, the dynamic compensation coefficient can be marked as 1 in the control system to indicate that no additional intervention is required. Through this iterative dynamic adjustment method, all partitions can maintain a high degree of consistency in different reflow stages, ultimately achieving a more uniform and controllable melting process and reducing the probability of failure caused by insufficient or overshooting of local temperature. The entire implementation process is continuously monitored and corrected through sensor temperature measurement, pulse control and data modeling to ensure that there will be no unexpected consequences of pulse reflow heating.

[0035] Please continue reading Figure 1 , during the pulse reflow temperature rise treatment, performing real-time melting state detection on each of the independent thermal control zones, obtaining surface morphology data and temperature distribution data of each of the independent thermal control zones, and judging the melting state of the solder joints of each of the independent thermal control zones according to the surface morphology data and the temperature distribution data; In one embodiment of the present invention, during the pulse reflow heating process, real-time melting state detection is performed on each of the independent thermal control zones to obtain surface morphology data and temperature distribution data of each of the independent thermal control zones, and the melting state of the solder joints of each of the independent thermal control zones is judged according to the surface morphology data and the temperature distribution data, including: during the heating stage of the pulse reflow heating process, a first solder joint surface detection is performed on each of the independent thermal control zones to obtain initial surface morphology data and initial surface reflectivity data; during the heat preservation stage of the pulse reflow heating process, the surface morphology data and the temperature distribution data of each of the independent thermal control zones are detected; The independent thermal control partition performs a second solder joint surface inspection to obtain instantaneous surface morphology data and instantaneous surface reflectivity data; dynamically compares the instantaneous surface morphology data with the initial surface morphology data, and performs a change analysis on the instantaneous surface reflectivity data and the initial surface reflectivity data to generate surface morphology data; continuously collects temperature change data during the pulse reflow heating process, performs a time series analysis on the temperature change data, and generates temperature distribution data; and determines the melting state of the solder joints of each of the independent thermal control partitions based on the surface morphology data and the temperature distribution data.

[0036] Specifically, during the heating stage of the pulsed reflow heating treatment, the first solder joint surface inspection can be performed on the solder joints in each independent thermal control partition through a high-speed visual inspection device or an observation window with precise optical measurement capabilities, so as to obtain the initial surface morphology data and initial surface reflectivity data. This inspection process is carried out immediately after the pulse heating of each partition is started, aiming to capture the initial state of the solder joint at a lower temperature for subsequent comparative analysis. If there are multiple welding areas inside a partition, they can be photographed or imaged synchronously in the same heating period of the partition, and the height difference, contour continuity, surface brightness and other information of the solder joint surface captured can be converted into morphology and reflectivity data. For the collection of morphology data, multi-angle structured light technology or laser line scanners can be used to make the light source form recognizable patterns or lines on the solder joint surface, and the surface height can be reconstructed by analyzing the line distortion in the image. For the measurement of reflectivity data, a special filter or photometer module can be installed in front of the lens of the high-speed imaging device to read the surface reflection parameters according to the intensity of the echo light signal. If the package surface is coated with a protective film or has a high-brightness metal area, a coating correction or compensation operation can be performed before detection to avoid measurement errors caused by the reflective properties of different materials. After completing the above operations, the initial surface morphology data and the initial surface reflectivity data will be stored in the central control system together with the time stamp and temperature record corresponding to the partition.

[0037] In the heat preservation stage of pulse reflow heating treatment, in order to grasp the dynamic evolution of the solder joint when it is close to the melting point of the solder, it is necessary to perform a second solder joint surface inspection on each independent thermal control partition to obtain instantaneous surface morphology data and instantaneous surface reflectivity data. The temperature in this stage is usually higher than that in the heating stage, and the internal structure and appearance of the solder joint are closer to the actual molten state. The inspection method is similar to the first surface inspection, still relying on high-speed visual instruments or structured light scanning methods, but at this time, attention needs to be paid to the deformation of the solder joint surface under heating conditions, the smoothness after melting, and the degree of change in reflectivity. If the surface glossiness is significantly improved in a certain partition and the solder joint contour tends to be hemispherical, it indicates that it has reached a relatively ideal molten state. After the inspection is completed, the instantaneous surface morphology data and the instantaneous surface reflectivity data will be dynamically compared with the initial surface morphology data and the initial surface reflectivity data obtained previously, and the change between the two time periods will be calculated by differential analysis or optical feature matching, so as to generate more intuitive surface morphology data. The change analysis can use pixel-level matching or reference point positioning algorithms, and present the specific deformation value and the increase or decrease in optical reflectivity on the output interface of the inspection software. If the reflectivity change of the solder joint in a certain partition increases significantly, but the morphology change does not show a complete molten shape, it may mean that there is partial melting or a lot of flux residue on the surface.

[0038] During the pulsed reflow heating process, in addition to the surface inspection at two key moments, temperature change data will be continuously collected and time series analysis will be performed to form temperature distribution data. This collection process is usually based on temperature sensors installed inside or near the partition, and can also rely on infrared scanning probes to capture the temperature field changes outside the package. In order to synchronize with the surface inspection as much as possible, a higher data acquisition frequency needs to be set so that the corresponding fluctuations of the temperature curve can be recorded in time every time the heating power changes. Subsequently, by performing segmented fitting or interpolation operations on these temperature data in the control system, a whole temperature time series from heating to insulation to the initial cooling in the partition can be generated. If the temperature curve is combined with the surface morphology data, the critical time point from solid to melting of the solder joint can be found, and it can be determined whether the melting temperature range required by the solder has been met. For example, if the temperature curve of a partition climbs above the melting point in a short period of time, but the surface inspection results still show that the overall contour of the solder joint has not undergone the expected smoothing change, it can be determined that a temporary hot spot or uneven heat distribution has occurred in the area, and the pulse parameters need to be further corrected later.

[0039] When the surface morphology data and temperature distribution data are updated synchronously, the melting state of the solder joints in each independent thermal control partition can be judged based on the comprehensive results of the two. The judgment logic can set several judgment thresholds in the system, including the melting temperature range, the critical value of the morphology smoothness, and the reflectivity increment range. If the temperature distribution data of a partition at this stage meets expectations, the heating time has exceeded the shortest melting maintenance time, and the surface morphology data has changed significantly compared with the initial value, then the partition is considered to have achieved full melting. If a partition shows a high temperature value but the morphology does not change enough accordingly, it means that the local heating is uneven or the surface oxide film hinders the melting of the solder, and subsequent heating or local reheating is required to eliminate the defect. Through this real-time dynamic detection method, not only can the surface evolution of the solder joint be accurately obtained at the critical moment of the reflow process, but also timely adjustments can be made based on the temperature timing analysis, so that each partition can meet the ideal melting conditions to a greater extent. If local spherical bulges or abnormal depressions are found during the morphology and reflectivity detection process, they can be quickly located and additional pulse corrections can be applied in subsequent steps. By linking with the temperature data, the problem area can be reheated to the appropriate temperature range to avoid the final unstable solder joint structure.

[0040] Please continue reading Figure 1 , performing atmosphere switching processing on the independent thermal control partition in the cooling stage according to the melting state of the solder joints of each of the independent thermal control partitions, to generate an atmosphere switching partition; In one embodiment of the present invention, according to the melting state of the solder joints of each of the independent thermal control partitions, the independent thermal control partitions in the cooling stage are subjected to atmosphere switching processing to generate atmosphere switching partitions, including: stress distribution detection of the independent thermal control partitions that have completed melting, obtaining interface tension data and heat conduction rate data, and generating partitions to be switched; according to the interface tension data and the heat conduction rate data, performing a first atmosphere switching on the partitions to be switched, switching the atmosphere environment of the partitions to be switched to a first inert gas environment; performing surface tension detection on the partitions to be switched under the first inert gas environment, and obtaining surface stress distribution data; performing a second atmosphere switching according to the surface stress distribution data, switching the first inert gas environment to a second inert gas environment; and locally adjusting the pressure of the second inert gas environment to match the curing rate with the temperature change of adjacent independent thermal control partitions to generate atmosphere switching partitions.

[0041] Specifically, when stress distribution detection is performed on the independent thermal control partitions that have completed melting, the mechanical equilibrium relationship between the solder joint and the chip substrate can be recorded by an optical stress monitoring device or a digital image correlation method after reflow treatment and while still at a relatively high temperature, thereby obtaining interfacial tension data and heat conduction rate data. The interfacial tension data can reflect the tension formed between the solder joint surface and the surrounding materials after melting, while the heat conduction rate data reflects the heat dissipation speed and heat loss efficiency inside the partition at this time. If the interfacial tension data value of a partition is significantly higher, it means that the solder joint still has a certain internal stress when it has just completed melting; if the heat conduction rate data of the partition is too low, it means that it is not easy for the molten area to quickly dissipate excess heat. By comprehensively analyzing these two types of data, the control system can determine which partitions need to further adjust the atmosphere and pressure in the future, and mark them as partitions to be switched.

[0042] When the first atmosphere switch is performed on the partition to be switched, the atmosphere environment can be changed from the initial reflux gas to the first inert gas environment, such as high-purity nitrogen. This high-purity nitrogen can provide relatively stable thermal convection conditions in the local area, and its chemical inertness helps to reduce the oxidation level on the surface or interface of the solder joint. When the first inert gas enters the partition, the stress state between the solder joint and the substrate will change accordingly, so the surface tension or interface stress needs to be measured again. In order to illustrate the specific process of this step, a reflux chamber with a multi-channel gas exchange system can be set up. This system injects high-purity nitrogen into the partition to be switched according to the stress monitoring results, so that the atmosphere in the local space quickly reaches the preset concentration and pressure level. If the detection finds that there is still a certain amount of excessive stress concentration or uneven heat dissipation on the surface of the solder joint, continue to fine-tune the local temperature and pressure in this environment to allow the solder joint to obtain a more stable cooling process in the early stage of solidification.

[0043] After completing the surface tension test and obtaining the surface stress distribution data for the partition to be switched under the first inert gas environment, if the results show that the internal mechanical condition has not yet reached the ideal distribution or there is still thermal stress retention, a second atmosphere switch can be performed to replace the original high-purity nitrogen environment with a second inert gas environment. The second inert gas environment can be argon or an argon-nitrogen mixed gas containing trace active components. Compared with the first inert gas environment, it has higher or lower thermal conductivity or physical properties that are more suitable for taking away local residual heat. By introducing inert gases of different compositions, the stress release process of the internal solder joints will enter a new equilibrium point, especially for the intersection of high-temperature and low-temperature partitions in multi-layer structures. Atmosphere switching can make the surface or interface stress distribution more balanced. If it is detected that the surface stress of the solder joint has been greatly reduced at this stage, it means that the second inert gas is more conducive to ensuring the mechanical stability of this partition.

[0044] In order to match the solidification rate with the temperature change of the adjacent independent thermal control zone, the zone will be locally pressure-regulated in the second inert gas environment, that is, the local gas pressure of the zone is continuously adjusted through the multi-way pressure regulating valve and the gas flow control system, so as to achieve a cooling rate coordinated with other zones on the cooling curve. If the temperature drop rate of the zone is too high, it may cause the solder joint to shrink or the stress to accumulate rapidly; on the contrary, if the drop rate is too low, the solidification process will be delayed and the exposure time of the solder joint to the high temperature state will be increased. By observing the real-time data of the temperature sensor and the stress sensor, the system can make pressure corrections within a time interval of microseconds or milliseconds, and gradually confirm the most suitable pressure range so that the zone can achieve a more ideal balance in the solidification stage. With the completion of this step, the state of the atmosphere switching zone is confirmed, that is, the zone has passed through the first inert environment of high-purity nitrogen from the previous initial molten state, and then passed through the second inert environment of argon or argon-nitrogen mixed gas, and assisted by local pressure regulation, so as to achieve a more balanced condition in terms of mechanical stress and heat dissipation rate. As a result, the subsequent heat dissipation and solidification process can continue in a relatively stable environment, and the partition is more synchronized with other areas in the overall cooling of the multi-layer stacked package, reducing hidden dangers such as cracks or delamination caused by uneven heat.

[0045] Please continue reading Figure 1 , performing micro-cycle steady-state detection on the atmosphere switching partition, obtaining the solder joint surface deformation velocity data and thermal distribution data of the atmosphere switching partition, and performing local reheating treatment on the atmosphere switching partition with abnormality according to the solder joint surface deformation velocity data and the thermal distribution data to generate a reheating partition; In one embodiment of the present invention, the atmosphere switching partition is subjected to micro-cycle steady-state detection to obtain the surface deformation velocity data and thermal distribution data of the solder joint of the atmosphere switching partition, and the atmosphere switching partition with abnormality is subjected to local reheating treatment according to the surface deformation velocity data and the thermal distribution data of the solder joint to generate a reheating partition, including: performing stable temperature reduction control on the atmosphere switching partition to control the temperature within the plastic state temperature range of the solder joint; performing multi-cycle continuous detection on the atmosphere switching partition within the plastic state temperature range of the solder joint to obtain the surface profile data of the solder joint at multiple consecutive time points. and regional temperature data; performing speed calculation on the change of the solder joint surface profile data at multiple consecutive time points to generate solder joint surface deformation speed data; performing distribution analysis on the change of the regional temperature data at multiple consecutive time points to generate thermal distribution data; determining the local area where micro-collapse or bump retraction exists in the atmosphere switching partition based on the solder joint surface deformation speed data and the thermal distribution data, and generating abnormal area data; performing fixed-point heating treatment on the abnormal area, raising the temperature of the abnormal area to a temperature range of 90% to 95% of the solder melting point temperature value, and generating a reheating partition.

[0046] Specifically, when the atmosphere switching partition is stably cooled, the ambient temperature of the partition needs to be adjusted to a temperature range that is suitable for the plastic state of the solder joint, so that the internal solder joint enters a slow cooling stage in a more flexible thermal environment. This temperature range is often slightly lower than the complete melting point of the solder, but it is sufficient to allow the surface and internal structure of the solder joint to maintain a certain plasticity, so that minor deformations can be observed and corrected. To illustrate this process, an independent temperature control module can be arranged in a local area of ​​the reflow oven, and the inert gas concentration and flow rate can be set according to the previous atmosphere switching results, so that the interior of the partition slowly drops to the target temperature within tens of seconds to minutes. At this point, the solder joint is in a critical temperature zone that is neither completely hardened nor melted again, so it is possible to more intuitively monitor whether there is an imbalance of local depression or bulging of bumps.

[0047] When performing multi-cycle continuous detection of the atmosphere switching partition within the temperature range of the plastic state of the solder joint, a high-speed visible light imaging or laser reflection device can be used to continuously capture the solder joint surface profile at several short time points, and the obtained image or laser reflection data can be combined with the corresponding regional temperature data to form a multi-dimensional record of the time series. If the partition contains multiple solder joints, the surface profile can be collected one by one by means of programmable moving lens or multi-camera synchronous shooting. For regional temperature data, it can be collected through thermocouples buried in the partition or its adjacent position, or through infrared thermal imaging equipment to ensure that it can correspond to the surface profile data in each detection cycle. The solder joint morphology and temperature information collected in these time periods are uniformly saved in the data management module, which can be used to further analyze the deformation speed and heat distribution law of the solder joint surface.

[0048] When calculating the speed of the change of the surface profile data of the solder joint at multiple consecutive time points, it is necessary to first coordinate the three-dimensional or two-dimensional profile of each time point, and then divide the profile difference of adjacent time points by the interval time to obtain the speed value of the local deformation. If a certain area has a micro-depression within a detection cycle, it will appear as a continuous decrease in the height value in the profile comparison. The deformation speed can be obtained by performing a differential operation on these values. If it is found that this speed exceeds a certain threshold, it means that the local solder joint is collapsing inward or convexing outward, which may be caused by internal stress or uneven cooling. Subsequently, by summarizing and visualizing these deformation speed data, the location of the most significant deformation area in the partition and its evolution trend can be accurately grasped.

[0049] When performing distribution analysis on the changes in regional temperature data at multiple consecutive time points, the temperature point cloud or temperature matrix within the same detection cycle will be interpolated or differentially processed, and combined with the known solder physical properties to determine whether each detection cycle is approaching the critical point of solidification. If the temperature drop rate at a certain location is significantly faster and the deformation rate data also exceeds the normal range, it means that there may be a risk of thermal stress concentration or uneven material shrinkage. By jointly analyzing the deformation rate and temperature distribution, abnormal evolution in the plastic state can be more objectively identified.

[0050] When it is determined that there is a local area of ​​micro-collapse or bump retraction in the atmosphere switching partition, the corresponding abnormal area data will be generated on the system interface to indicate the specific coordinates or chip positions that need to be corrected urgently. In order to prevent the abnormality from continuing to expand and causing irreversible gaps or cracks to form after the solder joints are fully hardened in the later stage, these abnormal areas need to be heated at a fixed point. This treatment process uses a local heating module or a laser irradiation device to raise the temperature of the abnormal area to between 90% and 95% of the melting point of the solder, so that the solder joint enters a semi-fluid state that is close to melting but retains a certain viscosity. In this temperature range, the micro-collapse or retraction that occurred previously has the opportunity to be repaired by the surface tension of the material itself and the local flow effect. If an example scenario is set, a certain area continues to show a negative deformation speed value in multiple tests, which may be caused by internal holes or incomplete bonding of the solder joints. After the temperature is brought to a near-molten state by local heating, the adjacent particles or metal media of the solder joint can be rearranged and the gaps can be filled, so that the position gradually restores a stable morphology.

[0051] When this reheating operation is completed, the local area becomes a reheating partition and continues to undergo a relatively slow solidification process in a specific thermal environment. The entire process is tested multiple times by the automated control software. Once it is observed that the deformation speed has returned to the ideal range and the temperature drop rate no longer surges, it can be said that the abnormality has been alleviated to a certain extent after fixed-point heating. With the help of the above series of steps, the risk of the atmosphere switching partition evolving into a permanent welding defect due to local stress or thermal unevenness during the plastic stage can be significantly reduced, and corrective measures can be taken in a timely manner when minor deformations are found, thereby enhancing the overall interconnection reliability of the multi-layer package.

[0052] In one embodiment of the present invention, the speed calculation of the change of the solder joint surface contour data at multiple consecutive time points to generate solder joint surface deformation speed data includes: extracting the solder joint edge contour coordinates at the multiple consecutive time points according to the solder joint surface contour data to generate solder joint contour point cloud data; performing time series registration on the solder joint contour point cloud data to generate a solder joint displacement vector map at adjacent time points; calculating the displacement distance of each area on the solder joint surface at adjacent time points according to the solder joint displacement vector map to generate displacement distribution data; classifying the deformation mode of the displacement distribution data to divide the displacement distribution data into micro-collapse deformation data, bump retraction deformation data and steady-state deformation data; generating solder joint surface deformation speed data according to the amount of change of the micro-collapse deformation data, the bump retraction deformation data and the steady-state deformation data per unit time.

[0053] Specifically, when extracting the edge contour coordinates of the weld at multiple consecutive time points based on the weld surface contour data, it is necessary to first convert the collected surface image or scanning result into a computable geometric form. The image processing algorithm or three-dimensional reconstruction algorithm can be called at each time point to match the edge pixels of the weld to the corresponding spatial coordinate system, and generate the weld contour point cloud data. If a laser line scanner or structured light imaging device is used, the weld surface can be divided into several lines or grids during detection, and then the height information and plane coordinate information are extracted line by line and block by block in the algorithm, and recorded in the database in chronological order. In this way, when comparing adjacent time points, the coordinates in each point cloud can be compared with the original reference, thereby reflecting the deformation of the weld in the local area. If the size of the weld in a certain partition is large, a denser scanning interval can be set in the equipment control software to avoid ignoring tiny edge depressions or protrusions. After the weld contour point cloud data is initially generated, these point clouds will be processed to remove obvious noise or erroneous data caused by abnormal surface reflection, so as to perform more accurate time sequence registration later.

[0054] When performing time series registration of solder joint contour point cloud data, a matching relationship will be established between adjacent time points, and the coordinates of the same spatial position or the same solder joint surface feature point at different time points will be accurately matched. In order to achieve this process, feature matching and minimum mean square error fitting algorithms can be used to gradually overlap the point clouds located in the same area, and external calibration information (such as height reference or specific position markers) can be used to ensure the accuracy of registration when possible. After completing the registration, the system will generate a solder joint displacement vector diagram of adjacent time points for the successfully registered point cloud. Each vector in the vector diagram indicates the displacement direction and displacement amplitude of a point on the solder joint surface at two adjacent time points. If a certain area is offset in the same direction multiple times in the time series, it will show a consistent trend in the vector diagram, which is convenient for observing the overall or local extrusion, subsidence or bulge of the solder joint.

[0055] Based on the solder joint displacement vector diagram, the displacement distance of each area on the solder joint surface at adjacent time points can be further calculated, and the displacement distribution data can be generated. The displacement distance is usually presented in the form of absolute or relative values, and can be visualized by color distribution or vector length. In an example scenario, if a vector is found to have an inward displacement between two adjacent time points in the center area of ​​the solder joint, and the total displacement exceeds a certain reference threshold, it means that the interior may be experiencing a slight collapse. By cross-comparing the vector diagram with the original height data of the solder joint surface, it can also be verified whether the collapse is accompanied by lateral expansion or contraction, so as to more objectively judge the mechanical state of the solder joint. If, on the contrary, it is observed that the vectors in the surrounding area are expanding outward, it is necessary to evaluate whether there is surface bulge caused by bump retraction or local stress release.

[0056] After obtaining the displacement distribution data, the deformation mode will be classified into micro-collapse deformation data, convex point retraction deformation data and steady-state deformation data. Micro-collapse deformation data refers to the displacement mode of inward contraction or sinking between adjacent time points, and the magnitude has not reached the level of complete failure; convex point retraction deformation data means that the surface may have protruded locally before, and then a rapid retraction phenomenon occurred; steady-state deformation data is used to indicate the area where the displacement is extremely small or evenly distributed in the time series comparison, and there is no significant height change. The significance of managing and analyzing different deformation modes separately is to quickly find the parts that are most prone to defects or residual stress concentration, and to better distinguish which parts show a stable and reliable state in the time series span. Through this classification, the detection system can output in the form of charts or lists which areas are in the process of micro-collapse, which areas have retraction characteristics, and which areas are basically stable.

[0057] According to the change in micro-collapse deformation data, bump retraction deformation data and steady-state deformation data per unit time, the system can generate solder joint surface deformation velocity data to numerically quantify the severity of deformation in a short period of time. Deformation velocity data is generally characterized by displacement or volume change per unit time. If a certain area undergoes significant displacement in a very short time, the velocity value will be high, and it is necessary to pay attention to whether there is stress concentration caused by cracks in the solder joint or sudden temperature change. If the micro-collapse deformation data continues to increase, it means that the center of the solder joint is shrinking inward and may endanger the final connection strength. If the bump retraction deformation data shows periodic fluctuations, it may mean that the packaging structure has repeated stress release under thermal cycling. Through these deformation velocity data, the comparison between different partitions and different time periods becomes intuitive. Once the deformation velocity of a partition increases sharply in a short period of time, it means that the area is at a high risk of potential failure. At this time, it can be treated in time with strategies such as local heating or atmosphere control. The entire process consists of multiple samplings, multiple time series alignments, and vector field-based calculations. It not only has a detailed spatial depiction, but also takes into account the judgment of deformation accumulation in the time series, so that the status of multi-layer stacked package solder joints can be more completely grasped and tracked.

[0058] Please continue reading Figure 1 , dynamically cool the reheating partition, and adjust the cooling rate of the reheating partition according to the temperature history data of the reheating partition.

[0059] In one embodiment of the present invention, the dynamic cooling process of the reheating partition is performed, and the cooling rate of the reheating partition is regulated according to the temperature history data of the reheating partition, including: performing time series analysis on the cooling process data of the reheating partition to generate temperature history data; dividing the reheating partition into temperature intervals according to the temperature history data, and dividing the reheating partition into a high temperature interval, a medium temperature interval and a low temperature interval; dynamically tracking the surface morphology of the reheating partition in the high temperature interval, the medium temperature interval and the low temperature interval to obtain deformation data of each temperature interval; generating corresponding target cooling rate data according to the deformation data of the high temperature interval, the medium temperature interval and the low temperature interval; cooling control is performed at a first cooling rate for the high temperature interval, cooling control is performed at a second cooling rate that is 50% lower than the first cooling rate for the medium temperature interval, and cooling control is performed at a third cooling rate that is 20% lower than the first cooling rate for the low temperature interval; and real-time regulation of the first cooling rate, the second cooling rate and the third cooling rate is performed according to the temperature history data and the deformation data of each temperature interval.

[0060] Specifically, when performing a time series analysis on the cooling process data of the reheating partition, after the local heating is completed, the temperature values ​​of the partition at different time points can be continuously collected through a temperature sensor array or an infrared detection device, and these data can be recorded in a one-dimensional or two-dimensional data table in chronological order. If the partition contains multiple solder joints or multiple sub-areas, it is necessary to generate an independent temperature change curve for each monitoring point to compare the cooling characteristics of different positions in subsequent analysis. In this scenario, a higher sampling frequency can be set, such as recording the temperature once per second or every several milliseconds, to ensure that the subtle fluctuations in the cooling process are captured. Subsequently, the system will interpolate or regress these discrete temperature points to form a relatively smooth temperature-time curve, which is convenient for observing the specific process of the partition temperature gradually evolving from a relatively high temperature range to a lower temperature range after the heating is completed. The curve can be presented in the data visualization interface in the form of a broken line or a color gradient, and the precise temperature values ​​of each time node are retained. After time series analysis, the system finally obtains temperature history data, including the overall temperature trend of the partition from the end of reheating to complete cooling and stabilization.

[0061] When dividing the reheating zone into temperature intervals based on the temperature history data, it is necessary to divide the cooling curve within the zone into several intervals according to the temperature level and downward rate of each stage. In one example, the part with a temperature higher than or close to the melting point of the solder can be defined as a high temperature interval, the area where the temperature is semi-solidified or the solder joint is still ductile can be defined as a medium temperature interval, and the area close to room temperature or significantly below the melting point can be defined as a low temperature interval. During the division process, the characteristics of the solder and the determination of the plastic range in the previous step can be combined to ensure that the divided intervals can truly reflect the process of the solder joint from flowing to semi-solidified and then to fully solidified. If multiple temperature sensors are detected in the zone, the data measured by different probes need to be corrected for consistency, so as to distinguish whether the zone is in the high temperature interval as a whole or has entered the medium temperature interval locally within the same period of time.

[0062] When dynamically tracking the surface morphology of the reheating partition in the high temperature range, medium temperature range and low temperature range, the height, curvature and surface light reflectivity of the solder joint can be measured and recorded multiple times in each temperature zone in combination with the optical imaging or laser scanning technology used in the previous step. The high temperature range usually corresponds to the state where the solder joint still has high fluidity or is close to the melting point, so a relatively soft surface can be observed during the inspection; the solder joint in the medium temperature range begins to gradually solidify but still retains a certain plasticity, and at this time, local bulges or slight indentations are prone to deformation behavior; the low temperature range is close to complete cooling and solidification. If the deformation rate still rises during this period, it means that the internal stress has not been completely released. During the measurement process, the morphological data of each cycle will be converted into a numerical height distribution map or a three-dimensional grid map, thereby forming the deformation data of each temperature range. If a large range of deformation is recorded in the high temperature range, it means that the temperature of the reheating partition has not yet stabilized; if local shrinkage or local bulging is detected in the medium temperature range, it is necessary to further reduce the cooling rate or heating channel to achieve a smoother solidification.

[0063] According to the deformation data of the high temperature interval, medium temperature interval and low temperature interval, the system will automatically generate the corresponding target cooling rate data. The generation process usually uses the model established by the previous experiment or simulation, takes the maximum deformation rate or stress increment allowed in each interval as the input parameter, and then outputs the optimal cooling rate of each interval to balance thermal shrinkage and stress release. If the deformation variable in the high temperature interval is large, the cooling rate can be increased to make the temperature pass through the over-softening stage quickly; if the medium temperature interval is prone to continuous collapse, the cooling rate can be moderately slowed down to allow the internal stress to be slowly released for a longer time. Therefore, the system will set the first cooling rate for the high temperature interval, the second cooling rate that is 50% lower than the first cooling rate for the medium temperature interval, and the third cooling rate that is 20% lower than the first cooling rate for the low temperature interval. Taking a numerical example, if the first cooling rate is set to cool down 6°C per second, the second cooling rate corresponding to the medium temperature interval is about 3°C / second, and the third cooling rate corresponding to the low temperature interval is about 1-2°C / second, making the overall transition more gradual.

[0064] While executing the above-mentioned multi-interval different cooling rates, it is also necessary to adjust the first cooling rate, the second cooling rate and the third cooling rate in real time according to the temperature history data and the deformation data of each temperature interval. If it is detected that a certain partition still has abnormal bump retraction or collapse speed continues to rise after entering the medium temperature interval, it means that the originally set second cooling rate is not enough to avoid stress concentration. The system will automatically reduce the cooling rate of the partition, or temporarily maintain a holding time higher than the minimum acceptable temperature to alleviate the problem of excessive internal stress release. If it is suddenly observed in the low temperature interval that the temperature curve drops slowly and the solder joint deformation stops accumulating, the system can consider slightly increasing the third cooling rate to speed up the overall process. Through this adaptive adjustment, the appropriate deformation and temperature matching relationship can be maintained throughout the cooling stage to avoid local overcooling or overheating at any time. In this way, each solder joint can complete solidification in a more stable manner in different temperature intervals, and obtain reliable mechanical properties under the condition of controllable deformation speed, thereby reducing the possibility of local cracks or voids.

[0065] In one embodiment of the present invention, the high temperature interval is cooled by a first cooling rate, the medium temperature interval is cooled by a second cooling rate that is 50% lower than the first cooling rate, and the low temperature interval is cooled by a third cooling rate that is 20% lower than the first cooling rate, including: in the high temperature interval, obtaining stress release data of the reheating partition, setting a reference cooling rate according to the stress release data, and using the reference cooling rate as the first cooling rate; analyzing the surface stress data of the reheating partition when it transitions from the high temperature interval to the medium temperature interval, and determining the variation period of the second cooling rate according to the surface stress data ; According to the variation cycle of the second cooling rate, the first cooling rate is decreased by 50% to generate the second cooling rate in the medium temperature range; the solder joint solidification data when the reheating partition transitions from the medium temperature range to the low temperature range is analyzed, and the variation cycle of the third cooling rate is determined according to the solder joint solidification data; according to the variation cycle of the third cooling rate, the first cooling rate is decreased by 20% to generate the third cooling rate in the low temperature range; the reheating partition is subjected to regional cooling control, and the first cooling rate is adopted in the high temperature range, the second cooling rate is adopted in the medium temperature range, and the third cooling rate is adopted in the low temperature range.

[0066] Specifically, when obtaining stress release data of the reheating partition in the high temperature range, the stress sensor or high-resolution optical stress detection device can be used to continuously monitor the solder joint and the package substrate when the partition has just completed the reheating stage and is still at a high temperature. If the measurement results show that the stress release rate is fast at this time, it means that there is still a certain difference in the temperature gradient inside and outside the solder joint, or the internal metal crystal phase is not yet completely stable. Based on these stress release data, the reasonable range of initial cooling of the partition can be determined, that is, by comparing the maximum stress change rate detected with the allowable range of the material, the reference cooling rate suitable for the solder joint to smoothly pass the high temperature cooling stage is calculated. Subsequently, the reference cooling rate is defined as the first cooling rate to indicate the upper limit of the speed that the partition should try to maintain when cooling in the high temperature range. In the example scenario, the reference cooling rate is set to cool down 5 degrees Celsius per second, and the device will automatically maintain or fine-tune this value in this range, and will not blindly cool down significantly to avoid extremely high internal stress or instantaneous deformation due to a sharp drop in temperature in a short period of time.

[0067] When the reheating zone transitions from the high temperature range to the medium temperature range, the analysis of the surface stress data can further reveal the local deformation of the solder joint within the critical temperature range. The medium temperature range is often the stage where the metal solder gradually changes from the semi-molten or high viscosity stage with plasticity to the stage close to solidification. If the surface stress data shows frequent fluctuations or local prominent values ​​during this period, it means that there is still a lot of energy being released inside, and it is necessary to appropriately extend the cooling cycle or reduce the cooling rate to make the stress environment of the solder joint more balanced. By fitting the time series distribution of these stress data, it can be determined in which time period or temperature range the second cooling rate should be adjusted. If the data shows that the surface stress changes drastically when the temperature just drops to a certain value, it can be inferred that special attention should be paid to the cooling rate reduction in a short period of time before and after this point. The resulting change cycle of the second cooling rate refers to the cumulative analysis result after multiple segmented sampling in this interval. After the first cooling rate is reduced by 50%, the second cooling rate in the medium temperature range is obtained. For example, if the first cooling rate is 5 degrees Celsius per second, a target rate of about 2.5 degrees Celsius per second may be obtained by decreasing the first cooling rate by 50%, so as to more gently handle the internal stress relief process.

[0068] When the partition continues to evolve from the medium temperature range to the low temperature range, it is necessary to combine the solder joint solidification data to determine the change cycle of the third cooling rate. The solder joint solidification data can be reflected by hardness detection or optical reflectivity monitoring: if it is detected that the hardness of the solder joint increases rapidly or the reflectivity increases significantly over time in a certain temperature range, it means that most areas have gradually converted to solid state; if the solidification speed is too fast, residual stress may be formed inside the material. Therefore, it is necessary to compare these solidification time points in the detection software to find the short-term window where deformation or stress unevenness is most likely to occur. If the analysis results show that the deformation is still large in the interval from the end of the medium temperature range to the beginning of the low temperature range, it can be determined that the third cooling rate should be further reduced to make the solidification process smooth and stable. Therefore, the first cooling rate is reduced by 20% to generate the third cooling rate in the low temperature range. Continuing the above example, if the first cooling rate is 5 degrees Celsius per second, it is reduced by 20% to about 4 degrees Celsius per second, but in this scheme, because the second cooling rate has been reduced to 2.5 degrees Celsius per second, the final third cooling rate will often be closer to the second cooling rate or lower. This allows the solder joints sufficient time to complete residual solidification and stress dissipation at near room temperature or low temperature environments.

[0069] In actual implementation, for solder joints in different parts of the reheating partition, this cooling scheme may be further refined into small grids or sub-partitions according to the temperature gradient distribution or local stress sensing data in the partition. In the high temperature range, the system follows the first cooling rate; after switching to the medium temperature range, the second cooling rate is automatically adopted; and when the temperature or time monitoring determines that it has entered the low temperature range, it switches to the third cooling rate again. The cooling control system can continuously compare the data of each sensor in this process, and if obvious stress anomalies or temperature lags are captured, the cooling rate can be fine-tuned within the partition to make each stage more consistent with the material properties and real-time measurement results. Through this regional cooling control, the high temperature period can quickly and safely remove excess heat, the medium temperature period can gradually stabilize the metal crystal phase, and the low temperature period tries to avoid compact cooling to cause surface cracking or plastic deformation accumulation. In this way, the reheating partition maintains reasonable stress and temperature changes throughout the cooling cycle, thereby improving the overall integrity and service life of the solder joint and package interface.

[0070] The above describes the semiconductor chip welding method according to the embodiment of the present invention. The following describes the semiconductor chip welding device according to the embodiment of the present invention. Figure 2 , an embodiment of a semiconductor chip welding device in an embodiment of the present invention comprises: The thermal field scanning module 101 is used to perform thermal field scanning on the multi-layer stacked package to obtain overall thermal imaging data, and generate a thermal field partition diagram including multiple independent thermal control partitions according to the overall thermal imaging data; The pulse heating module 102 is used to perform pulse reflux heating processing on each independent thermal control partition according to the position information of each independent thermal control partition in the thermal field partition diagram, obtain real-time temperature data of each independent thermal control partition, and dynamically adjust the pulse reflux parameters of each independent thermal control partition according to the real-time temperature data; The melting detection module 103 is used to perform real-time melting state detection on each of the independent thermal control zones during the pulse reflow heating process, obtain surface morphology data and temperature distribution data of each of the independent thermal control zones, and determine the melting state of the solder joints of each of the independent thermal control zones according to the surface morphology data and the temperature distribution data; The atmosphere switching module 104 is used to perform atmosphere switching processing on the independent thermal control partitions in the cooling stage according to the melting state of the solder points of each of the independent thermal control partitions, and generate atmosphere switching partitions; The reheating detection module 105 is used to perform micro-cycle steady-state detection on the atmosphere switching partition, obtain the solder joint surface deformation velocity data and thermal distribution data of the atmosphere switching partition, and perform local reheating processing on the atmosphere switching partition with abnormality according to the solder joint surface deformation velocity data and the thermal distribution data to generate a reheating partition; The dynamic cooling module 106 is used to perform dynamic cooling processing on the reheating partition and adjust the cooling rate of the reheating partition according to the temperature history data of the reheating partition.

[0071] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the inventive concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A method for welding a semiconductor chip, characterized in that: include: Performing thermal field scanning on the multi-layer stacked package to obtain overall thermal imaging data, and generating a thermal field partition schematic diagram including a plurality of independent thermal control partitions according to the overall thermal imaging data; According to the position information of each independent thermal control partition in the thermal field partition schematic diagram, pulse reflux heating treatment is performed on each independent thermal control partition to obtain real-time temperature data of each independent thermal control partition, and pulse reflux parameters of each independent thermal control partition are dynamically adjusted according to the real-time temperature data; During the pulse reflow temperature rise treatment, real-time melting state detection is performed on each of the independent thermal control zones to obtain surface morphology data and temperature distribution data of each of the independent thermal control zones, and the melting state of the solder joints of each of the independent thermal control zones is determined based on the surface morphology data and the temperature distribution data; According to the melting state of the solder joints of each of the independent thermal control partitions, an atmosphere switching process is performed on the independent thermal control partition in the cooling stage to generate an atmosphere switching partition; Performing micro-cycle steady-state detection on the atmosphere switching partition to obtain solder joint surface deformation velocity data and heat distribution data of the atmosphere switching partition, and performing local reheating treatment on the atmosphere switching partition with abnormality according to the solder joint surface deformation velocity data and the heat distribution data to generate a reheating partition; The reheating partition is subjected to dynamic cooling processing, and the cooling rate of the reheating partition is regulated according to the temperature history data of the reheating partition.

2. The method for welding a semiconductor chip according to claim 1, characterized in that: The thermal field scanning is performed on the multi-layer stacked package to obtain overall thermal imaging data, and a thermal field partition diagram including multiple independent thermal control partitions is generated according to the overall thermal imaging data, including: Performing multi-angle thermal field scanning on the multi-layer stack package to obtain top thermal imaging data, side thermal imaging data, and bottom thermal imaging data of the multi-layer stack package; Comparing and analyzing the top thermal imaging data, the side thermal imaging data, and the bottom thermal imaging data with a standard temperature curve to generate an initial temperature distribution diagram of the multi-layer stack package; Determine the thermal diffusion coefficient and the expected thermal hysteresis time of each region in the multi-layer stack package according to the initial temperature distribution map, and generate a thermal characteristic distribution map of the multi-layer stack package; Performing critical temperature analysis on the temperature gradient change area in the thermal characteristic distribution map, determining the maximum allowable temperature value and the minimum required temperature value of each area of ​​the multi-layer stack package, and generating a temperature boundary distribution map and temperature threshold data of the multi-layer stack package; A thermal field partition diagram including a plurality of independent thermal control partitions is generated according to the temperature boundary distribution diagram and the temperature threshold data.

3. The method for welding a semiconductor chip according to claim 1, characterized in that: According to the position information of each independent thermal control partition in the thermal field partition schematic diagram, pulse reflux heating treatment is performed on each independent thermal control partition to obtain real-time temperature data of each independent thermal control partition, and pulse reflux parameters of each independent thermal control partition are dynamically adjusted according to the real-time temperature data, including: According to the position information of each independent thermal control partition in the thermal field partition schematic diagram, a first pulse reflux temperature rise process is performed on each independent thermal control partition to obtain first real-time temperature data of each independent thermal control partition; Performing difference analysis on the first real-time temperature data of each independent thermal control zone and the theoretical temperature data of each independent thermal control zone to obtain temperature deviation data of each independent thermal control zone; Performing a second pulse reflux temperature rise treatment on each of the independent thermal control partitions, recording the temperature change rate of each of the independent thermal control partitions, and obtaining temperature response data of each of the independent thermal control partitions; Calculating the heat accumulation rate of each of the independent thermal control partitions according to the temperature deviation data and the temperature response data, and generating a dynamic compensation coefficient of each of the independent thermal control partitions; According to the dynamic compensation coefficient and the temperature response data, the pulse reflow parameters of each independent thermal control partition are dynamically adjusted.

4. The method for welding a semiconductor chip according to claim 1, characterized in that: During the pulse reflow temperature rise process, real-time melting state detection is performed on each of the independent thermal control zones to obtain surface morphology data and temperature distribution data of each of the independent thermal control zones, and the melting state of the solder joints of each of the independent thermal control zones is determined according to the surface morphology data and the temperature distribution data, including: During the heating stage of the pulsed reflow heating treatment, performing a first solder joint surface inspection on each of the independent thermal control partitions to obtain initial surface topography data and initial surface reflectivity data; During the heat preservation stage of the pulse reflow heating treatment, a second solder joint surface inspection is performed on each of the independent thermal control zones to obtain instantaneous surface morphology data and instantaneous surface reflectivity data; Dynamically comparing the instantaneous surface topography data with the initial surface topography data, and performing a variation analysis on the instantaneous surface reflectivity data and the initial surface reflectivity data to generate surface topography data; Continuously collecting temperature change data during the pulsed reflux heating process, performing time series analysis on the temperature change data, and generating temperature distribution data; The melting state of the solder joints of each of the independent thermal control zones is determined according to the surface morphology data and the temperature distribution data.

5. The method for welding a semiconductor chip according to claim 1, characterized in that: The step of performing atmosphere switching processing on the independent thermal control partition in the cooling stage according to the melting state of the solder joints of each independent thermal control partition to generate an atmosphere switching partition includes: Perform stress distribution detection on the independent thermal control partitions that have completed melting, obtain interfacial tension data and heat conduction rate data, and generate the partitions to be switched; According to the interfacial tension data and the heat conduction rate data, performing a first atmosphere switch on the partition to be switched, and switching the atmosphere environment of the partition to be switched to a first inert gas environment; Performing surface tension detection on the partition to be switched under the first inert gas environment to obtain surface stress distribution data; Performing a second atmosphere switching according to the surface stress distribution data, switching the first inert gas environment to a second inert gas environment; The second inert gas environment is locally pressure-regulated to match the curing rate with the temperature change of the adjacent independent thermal control partitions to generate atmosphere switching partitions.

6. The method for welding a semiconductor chip according to claim 1, characterized in that: The micro-cycle steady-state detection is performed on the atmosphere switching partition to obtain the solder joint surface deformation velocity data and the heat distribution data of the atmosphere switching partition, and the atmosphere switching partition with abnormalities is locally reheated according to the solder joint surface deformation velocity data and the heat distribution data to generate the reheating partition, including: The atmosphere switching partition is subjected to stable temperature reduction control to control the temperature within the temperature range of the plastic state of the soldering point; Performing multi-cycle continuous detection on the atmosphere switching partition within the plastic state temperature range of the solder joint to obtain solder joint surface profile data and regional temperature data at multiple consecutive time points; Calculating the speed of the change of the solder joint surface profile data at a plurality of consecutive time points to generate solder joint surface deformation speed data; Performing distribution analysis on changes in the regional temperature data at multiple consecutive time points to generate thermal distribution data; According to the solder point surface deformation velocity data and the heat distribution data, determining a local area in the atmosphere switching partition where micro-collapse or bump retraction exists, and generating abnormal area data; The abnormal area is subjected to a fixed-point heating treatment to raise the temperature of the abnormal area to a temperature range of 90% to 95% of the melting point of the solder, thereby generating a reheating partition.

7. The method for welding a semiconductor chip according to claim 6, characterized in that: The calculating the speed of the change of the solder joint surface profile data at a plurality of consecutive time points to generate solder joint surface deformation speed data includes: Extracting the edge contour coordinates of the solder joints at the plurality of consecutive time points according to the solder joint surface contour data, and generating solder joint contour point cloud data; Performing time-series registration on the solder joint contour point cloud data to generate solder joint displacement vector diagrams at adjacent time points; According to the solder joint displacement vector diagram, the displacement distances of various regions on the solder joint surface at adjacent time points are calculated to generate displacement distribution data; Classifying the displacement distribution data into deformation modes, and dividing the displacement distribution data into micro-collapse deformation data, convex point retraction deformation data and steady-state deformation data; The solder joint surface deformation speed data is generated according to the variation of the micro-collapse deformation data, the bump retraction deformation data and the steady-state deformation data within a unit time.

8. The method for welding a semiconductor chip according to claim 1, characterized in that: The dynamically cooling the reheating partition and regulating the cooling rate of the reheating partition according to the temperature history data of the reheating partition include: Performing time series analysis on the cooling process data of the reheating zone to generate temperature history data; Divide the reheating zone into temperature intervals according to the temperature history data, and divide the reheating zone into a high temperature interval, a medium temperature interval and a low temperature interval; Dynamically tracking the surface morphology of the reheating zone in the high temperature interval, the medium temperature interval and the low temperature interval to obtain deformation data of each temperature interval; Generate corresponding target cooling rate data according to the deformation amount data of the high temperature interval, the medium temperature interval and the low temperature interval; A first cooling rate is used for cooling control in the high temperature range, a second cooling rate which is 50% lower than the first cooling rate is used for cooling control in the medium temperature range, and a third cooling rate which is 20% lower than the first cooling rate is used for cooling control in the low temperature range; The first cooling rate, the second cooling rate and the third cooling rate are controlled in real time according to the temperature history data and the deformation amount data in each temperature interval.

9. The method for welding a semiconductor chip according to claim 8, characterized in that: The step of cooling the high temperature interval using a first cooling rate, cooling the medium temperature interval using a second cooling rate that is 50% lower than the first cooling rate, and cooling the low temperature interval using a third cooling rate that is 20% lower than the first cooling rate comprises: In the high temperature interval, obtaining stress release data of the reheating zone, setting a reference cooling rate according to the stress release data, and using the reference cooling rate as the first cooling rate; analyzing surface stress data of the reheating zone when it transitions from the high temperature interval to the medium temperature interval, and determining a variation period of a second cooling rate according to the surface stress data; According to the variation cycle of the second cooling rate, the first cooling rate is reduced by 50% to generate a second cooling rate in the medium temperature range; analyzing the solder point solidification data when the reheating zone transitions from the medium temperature interval to the low temperature interval, and determining a variation period of a third cooling rate according to the solder point solidification data; According to the variation cycle of the third cooling rate, the first cooling rate is reduced by 20% to generate a third cooling rate in the low temperature range; The reheating zones are subjected to zone cooling control, wherein the first cooling rate is adopted in the high temperature range, the second cooling rate is adopted in the medium temperature range, and the third cooling rate is adopted in the low temperature range.

10. A semiconductor chip welding device, characterized in that: The semiconductor chip welding device adopts the semiconductor chip welding method according to any one of claims 1 to 9, and the semiconductor chip welding device comprises: A thermal field scanning module is used to perform thermal field scanning on a multi-layer stacked package to obtain overall thermal imaging data, and generate a thermal field partition diagram containing multiple independent thermal control partitions according to the overall thermal imaging data; A pulse heating module, used to perform pulse reflux heating treatment on each independent thermal control partition according to the position information of each independent thermal control partition in the thermal field partition schematic diagram, obtain real-time temperature data of each independent thermal control partition, and dynamically adjust the pulse reflux parameters of each independent thermal control partition according to the real-time temperature data; A melting detection module, used to perform real-time melting state detection on each of the independent thermal control zones during the pulse reflow heating process, obtain surface morphology data and temperature distribution data of each of the independent thermal control zones, and determine the melting state of the solder joints of each of the independent thermal control zones according to the surface morphology data and the temperature distribution data; An atmosphere switching module, used to perform atmosphere switching processing on the independent thermal control partitions in the cooling stage according to the melting state of the solder points of each of the independent thermal control partitions, and generate atmosphere switching partitions; A reheating detection module is used to perform micro-cycle steady-state detection on the atmosphere switching partition, obtain the solder joint surface deformation velocity data and thermal distribution data of the atmosphere switching partition, and perform local reheating treatment on the atmosphere switching partition with abnormalities according to the solder joint surface deformation velocity data and the thermal distribution data to generate a reheating partition; The dynamic cooling module is used to perform dynamic cooling processing on the reheating partition and regulate the cooling rate of the reheating partition according to the temperature history data of the reheating partition.

Images