Chip underfill optimization method based on temperature detection

Abstract

The application relates to the technical field of electronic component packaging, in particular to a chip bottom filling optimization method based on temperature detection. The method collects a surface temperature distribution diagram of the bottom of a substrate in real time, and then fuses the surface temperature distribution diagram with a thermal conductivity distribution diagram to obtain a heat source-heat dissipation channel intensity diagram. According to the heat transfer relationship between the heat source point and the heat dissipation point, the material flow imbalance degree is determined, and the material flow risk index at a real-time moment can be obtained in combination with the real-time material flow speed. Through temperature changes at adjacent moments, local thermal shock is obtained, a thermal stress damage diagram is drawn, and finally, the process optimization direction can be determined through the real-time acquisition of the thermal stress damage diagram during the process. Through real-time information acquisition and in combination with prior information in a design file, the application can effectively warn the cavity risk in the filling process, and based on the extracted features, the process can be reasonably improved and optimized, and the process reliability is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electronic component packaging, in particular to a chip bottom filling optimization method based on temperature detection. BACKGROUND

[0002] Advanced packaging refers to a semiconductor packaging method that improves chip performance through new interconnection and integration technology. In advanced semiconductor packaging, chips with different functions are flip-chip bonded in high density on a substrate with a through-glass via (TGV), which is a key heterogeneous integration technology. The TGV structure provides vertical electrical interconnection and also serves as an efficient local heat dissipation channel. Epoxy resin is injected between the chip and the substrate for bottom filling (a necessary process to ensure the stability of the packaging structure, which aims to alleviate the thermal mechanical stress caused by the mismatch of the thermal expansion coefficients of the materials, and to protect the solder joints. This process usually relies on capillary action, and the filler flows on the heated substrate due to reduced viscosity, eventually filling the small gap between the chip and the substrate.

[0003] In the filling process, two main problems are faced. First, the power consumption of heterogeneous chips varies when they are working or preheating during the process, and the high thermal conductivity of the TGV itself forms a complex and non-uniform temperature field on the substrate. Due to the high correlation between the viscosity of the filler and the temperature, the non-uniform temperature field will directly cause a large difference in the local flow rate of the filler at different parts of the chip bottom, which will easily cause the mismatched convergence of the two fluid fronts and form air pocket defects. Second, the edge of the TGV is a double mutation point of material properties (such as glass, copper, filler) and thermal properties (such as thermal conductivity). Such discontinuous interfaces are high-risk areas for stress concentration after experiencing temperature cycles during solidification and operation, and are prone to micro-cracks. The existing technology lacks effective online observation means to predict and control the above flow defects and stress risks, and process optimization relies mainly on detection after filling and a large number of trial-and-error experiments, which is costly and time-consuming. SUMMARY

[0004] In order to solve the technical problem that the existing technology lacks effective online observation means to predict and control air pocket defects and stress risks, the purpose of the present application is to provide a chip bottom filling optimization method based on temperature detection, and the technical scheme adopted is as follows:

[0005] The present application proposes a chip bottom filling optimization method based on temperature detection, which comprises:

[0006] Obtain the thermal conductivity distribution map of the bottom of the packaging area according to the design file; during the bottom filling process, the surface temperature distribution map of the bottom is collected in real time;

[0007] combining the gradient of each point in the surface temperature distribution map with the thermal conductivity distribution map to obtain a heat source-heat dissipation channel intensity map; the heat source-heat dissipation channel intensity map includes heat source points and heat dissipation points, the heat source points and the heat dissipation points are matched with each other, the material flow imbalance degree is obtained according to the optimal matching result, and the flow risk index is obtained in combination with the material flow speed information collected in real time;

[0008] determining the process risk in real time according to the flow risk index, and adjusting the temperature of the corresponding position point through the optimal matching result;

[0009] For each position point at the bottom, the local thermal shock is obtained according to the gradient of the position point in the thermal conductivity distribution map and the temperature change at the adjacent moment, the position point damage condition is judged according to the local thermal shock and the thermal expansion coefficient of the material at the position point, the thermal force damage value of each position point at the real time moment is obtained, and a thermal stress damage map is formed; after the filling is completed, the process optimization direction is analyzed according to the thermal stress damage map.

[0010] Further, the method for obtaining the heat source-heat dissipation channel intensity map comprises:

[0011] For each position point at the bottom, the gradient of the position point in the surface temperature distribution map is multiplied by the opposite number of the thermal conductivity in the thermal conductivity distribution to obtain a heat flux vector field; the divergence of each position point in the heat flux vector field is calculated, and the divergence constitutes the heat source-heat dissipation channel intensity map.

[0012] Further, the positioning method of the heat source points and the heat dissipation points in the heat source-heat dissipation channel intensity map comprises:

[0013] The local maximum point in the heat source-heat dissipation channel intensity map is taken as the heat source point, and the local minimum point is taken as the heat dissipation point.

[0014] Further, the matching method of the heat source points and the heat dissipation points comprises:

[0015] Combining any one heat source point and any one heat dissipation point to obtain a thermal force interaction vector; multiplying the distance between the heat source point and the heat dissipation point by the element value of the heat source point in the heat source-heat dissipation channel intensity map to obtain the module length of the thermal force interaction vector, and the direction of the heat source point pointing to the heat dissipation point is the direction of the thermal force interaction vector;

[0016] The square of the module length of the thermal force interaction direction is taken as the matching cost, and the Hungarian algorithm is used for optimization to obtain the optimal matching result.

[0017] Further, the method for obtaining the material flow imbalance degree comprises:

[0018] The thermal force interaction vectors corresponding to the optimal matching results are vector summed, and the module length of the sum vector is taken as the material flow imbalance degree.

[0019] Further, the flow risk indicator acquisition method comprises:

[0020] The instantaneous velocity difference and the cumulative position difference of the filling material edges on both sides of the bottom gap region are monitored in real time, the instantaneous velocity difference and the cumulative position difference are weighted and summed to obtain a flow anomaly feature, and the flow anomaly feature is gained according to the material flow imbalance degree to obtain the flow risk indicator.

[0021] Further, the local thermal shock acquisition method comprises:

[0022] The gradient absolute value of the position point in the thermal conductivity distribution map is taken as the interface thermal conductivity mutation degree, and the temperature change amount at the adjacent time is gained according to the interface thermal conductivity mutation degree to obtain the local thermal shock.

[0023] Further, the damage condition of the position point is judged, comprising:

[0024] The damage bearing threshold of the position point is obtained according to the gradient absolute value of the thermal expansion coefficient at the position point, if the local thermal shock is greater than the damage bearing threshold, it is judged that there is damage, otherwise it is judged that no damage is caused.

[0025] Further, the thermal force damage value acquisition method comprises:

[0026] If it is judged that there is damage, the difference between the local thermal shock and the damage bearing threshold is taken as the damage increment at the adjacent time, if it is judged that no damage is caused, the damage increment is set to 0, the sum of the damage increment at the real-time time and the thermal force damage value at the previous time is taken as the thermal force damage value at the real-time time, and the thermal force damage value at the initial time is the corresponding damage increment.

[0027] Further, the process optimization direction is analyzed according to the thermal stress damage map construction process, comprising:

[0028] The high-risk point in the thermal stress damage map is determined by the threshold screening method, for the high-risk point, if the damage bearing threshold is less than a preset first threshold, the process optimization direction is to adjust the substrate structure, and if the local thermal shock is greater than a preset second threshold, the process optimization direction is to adjust the temperature control.

[0029] The present application has the following beneficial effects:

[0030] The present application collects the surface temperature distribution of the substrate bottom in real time, and then fuses it with the thermal conductivity distribution, so that the heat source-heat dissipation channel intensity diagram can reflect the heating characteristics and heat dissipation characteristics at each position point of the bottom, and thus the heat source points and heat dissipation points can be effectively identified, and then the material flow imbalance degree can be determined according to the heat transfer relationship between the heat source points and the heat dissipation points. The material flow imbalance degree is a static characteristic obtained based on the temperature and the heat conduction relationship, and thus the flow risk index of the material at the real time can be obtained by combining the real-time material flow speed. That is, the flow risk index can quantify the risk of generating a cavity defect based on static and dynamic characteristics, so that sensitive early warning can be realized in the filling process. After determining that the process risk will occur, the temperature control can be performed on the obviously abnormal heat transfer area by backtracking the matching relationship between the heat source points and the heat dissipation points, and then the flow characteristics of the material can be adjusted to avoid the formation of cavities. In order to further optimize the process, the local thermal shock is obtained through the temperature change at adjacent time points, and the local thermal shock represents the thermal stress risk generated by the position point in the filling process, so that the thermal stress damage diagram can be further drawn. Finally, the process can be analyzed by reviewing the thermal stress damage diagram obtained in real time after the process ends, so as to determine the process optimization direction and improve the reliability of the filling process. Through real-time information acquisition and combining the prior information in the design file, the present application can effectively early warn the cavity risk in the filling process, and based on the extracted features, the process can be reasonably improved and optimized to improve the process reliability. BRIEF DESCRIPTION OF DRAWINGS

[0031] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without any creative effort.

[0032] Figure 1 A flow chart of a chip bottom filling optimization method based on temperature detection provided by an embodiment of the present application. DETAILED DESCRIPTION

[0033] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following will combine the drawings and the preferred embodiments to specifically describe the specific implementation, structure, features and effects of the chip bottom filling optimization method based on temperature detection according to the present application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0034] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0035] The application provides a chip underfill optimization method based on temperature detection.

[0036] Referring to Figure 1 The application provides a chip underfill optimization method based on temperature detection.

[0037] Step S1: Obtain a thermal conductivity distribution map of the bottom of the packaging area according to a design file; in the underfill process, a surface temperature distribution map of the bottom is collected in real time.

[0038] Before the underfill process starts, the thermal conductivity distribution map of the bottom can be determined according to the design file (CAD design drawing and material list). The size of the thermal conductivity distribution map is the size of the packaging area. In the embodiment of the application, a two-dimensional grid is drawn according to the packaging area. The value of each node on the grid is the thermal conductivity value of the filling material at the corresponding position of the bottom area, that is, the thermal conductivity distribution map can also be regarded as a two-dimensional matrix. The thermal conductivity distribution map is static data determined according to the prior data in the design file. In the embodiment of the application, the surface temperature distribution map of the bottom is collected in real time during the underfill process. In the subsequent analysis process, the effective reference features are extracted by combining the dynamic characteristics collected in real time and the prior static data.

[0039] It should be noted that for the underfill process, other known parameters can be obtained based on the design file for analysis in the process, such as the thermal expansion coefficient of the filling material at each position of the bottom area. A thermal expansion coefficient distribution map can also be generated. According to the known relationship curve between the viscosity of the filling material and the temperature, the real-time viscosity distribution map can be generated according to the surface temperature distribution map of the bottom collected in real time. The construction of two-dimensional images of various parameters can intuitively and visually present the real-time characteristics of the underfill process, which is convenient for data processing and process control.

[0040] In the embodiment of the application, the surface temperature distribution map is collected once per second in real time. The infrared image acquisition device is used for collection. The collected infrared image is the surface temperature distribution map. The thermal value in the infrared image can be equivalent to the temperature at the corresponding position. The coordinate system of the surface temperature distribution map and the distribution map such as the thermal conductivity distribution map is the same. The data value of the position point is the temperature at the real-time moment.

[0041] Step S2: fusing the gradient of each point in the surface temperature distribution map with the thermal conductivity distribution map to obtain a heat source-heat dissipation channel intensity map; the heat source-heat dissipation channel intensity map includes heat source points and heat dissipation points, and the heat source points and the heat dissipation points are matched with each other to obtain a material flow imbalance degree according to an optimal matching result, and a flow risk index is obtained in combination with real-time collected material flow speed information.

[0042] The temperature at the bottom of the packaging area can control the viscosity of the filling material, and then control the flowability of the material. Therefore, in order to evaluate whether there will be a void defect in the filling process, it is necessary to quantify the heat conduction characteristics in the filling process. Because the gradient can represent the change direction and change rate of the data at the position point, the gradient of each point in the surface temperature distribution map is fused with the thermal conductivity distribution map to obtain a heat source-heat dissipation channel intensity map in the embodiment of the present application. The heat source-heat dissipation channel intensity map can reflect the heat transfer characteristics of each position point, which includes heat source points and heat dissipation points. By matching the heat source points and the heat dissipation points, the optimal matching result obtained is the heat conduction channel in the current bottom area. In the heat conduction channel, the longer the length and the greater the heat generation of the heat source point, the more unbalanced the heat conduction, and the more unbalanced the flow of the filling material. Therefore, the material flow imbalance degree can be obtained based on the optimal matching result. The material flow imbalance degree is a static characteristic obtained based on the relationship between temperature and heat conduction. Therefore, in combination with the real-time material flow speed, the flow risk index of the material at the real-time moment can be obtained. That is, the more unbalanced the material flow speed at the real-time moment, and the greater the material flow imbalance degree, the more unstable the flow of the material in the filling process, which is prone to cause the convergence of the fluid front on both sides without matching, thereby wrapping the gas to form a void defect. Therefore, the flow risk index obtained based on the two characteristics can quantify the risk of void defects generated at the real-time moment, and can effectively feedback a warning signal for reference by the staff.

[0043] Preferably, in the embodiment of the present application, the heat source-heat dissipation channel intensity map is obtained by the following method:

[0044] For each position point at the bottom, the gradient of the position point in the surface temperature distribution map is multiplied by the opposite number of the thermal conductivity in the thermal conductivity distribution to obtain a heat flux vector field. The vector field can depict the flow direction and intensity of heat in the filling process. Considering that the divergence can further characterize the divergence degree of the vector field at each point in space, the divergence of each position point in the heat flux vector field can be calculated, and the divergence constitutes the heat source-heat dissipation channel intensity map. That is, the data size of each position point in the heat source-heat dissipation channel intensity map can represent the intensity of heat divergence. If the position point is a heat source point, the greater the value, the greater the heat generation power of the heat source point. If the position point is a heat dissipation point, the greater the absolute value, the greater the heat dissipation efficiency of the heat dissipation point.

[0045] Preferably, in the embodiment of the present application, the positioning method of the heat source points and the heat dissipation points in the heat source-heat dissipation channel intensity map comprises:

[0046] The local maximum points in the heat source-heat dissipation channel intensity map are taken as the heat source points, and the local minimum points are taken as the heat dissipation points. It should be noted that the method for searching the local maximum points and the local minimum points in the two-dimensional matrix is a technical means known to those skilled in the art, and will not be described here.

[0047] It should be noted that in another embodiment of the present application, the geometric centers of the TGVs can also be considered as heat dissipation points, so the geometric centers of a preset number of TGVs can be extracted based on the design file as heat dissipation points. The specific extraction number can be set according to the size of the bottom region, and will not be described here.

[0048] Preferably, in the embodiment of the present application, the matching method of the heat source points and the heat dissipation points comprises:

[0049] Any one heat source point and any one heat dissipation point are combined to obtain a heat force interaction vector; the distance between the heat source point and the heat dissipation point is multiplied by the element value of the heat source point in the heat source-heat dissipation channel intensity map to obtain the module length of the heat force interaction vector, and the direction of the heat source point pointing to the heat dissipation point is the direction of the heat force interaction vector. That is, the longer the module length of the heat force interaction vector, the farther the distance between the heat source point and the heat dissipation point, and the higher the heat generation efficiency of the heat source point, and the greater the matching cost of the two points.

[0050] The square of the module length of the heat force interaction direction is taken as the matching cost, and the Hungarian algorithm is used for optimization to obtain the optimal matching result. In the optimal matching result, the heat source point and the heat dissipation point can obtain the optimal matching result, and then a plurality of heat force interaction vectors are obtained.

[0051] Further, based on the optimal matching result, the heat force interaction vectors corresponding to the optimal matching result are vector summed, and the module length of the sum vector is taken as the material flow imbalance degree. That is, the greater the module length of the sum vector, the higher the heat generation efficiency of the heat source point at the optimal heat dissipation path, and the longer the heat conduction channel, which indicates that the flowability of the filling material is more unbalanced in the dimension of heat conduction at the real time, and the material flow imbalance degree is greater.

[0052] Preferably, in the embodiment of the present application, the acquisition method of the flow risk index comprises:

[0053] Real-time monitor the instantaneous velocity difference and the cumulative position difference of the edge of the filling material on both sides of the bottom gap area; the instantaneous velocity difference and the cumulative position difference are weighted and summed to obtain the flow anomaly feature. In the embodiment of the present application, the edge of the material during the filling process can be collected by image processing means, and then the movement of the edge is determined by tracking algorithm to obtain the flow velocity of the material on each side of the gap area at the real-time moment and the flow distance relative to the initial position. The absolute value of the difference between the flow velocities of the materials on both sides is taken as the instantaneous velocity difference, and the absolute value of the difference between the flow distances on both sides is taken as the cumulative position difference. In the embodiment of the present application, in order to avoid the influence of dimension on operation, the instantaneous velocity difference and the cumulative position difference are normalized by range standardization in their respective dimensions, and the weight of the normalized instantaneous velocity difference and the cumulative position difference is set to 0.5, that is, the flow anomaly feature is obtained by averaging the two difference features.

[0054] According to the material flow imbalance degree, the flow anomaly feature is gained, and the flow risk indicator is obtained. In the embodiment of the present application, after the material flow imbalance degree is normalized by range standardization, the gain coefficient is obtained by adding 1 to the positive integer 1, and the product of the gain coefficient and the flow anomaly feature is taken as the flow risk indicator.

[0055] Step S3: Real-time judge the process risk according to the flow risk indicator, and adjust the temperature of the corresponding position point through the optimal matching result.

[0056] The greater the flow risk indicator is, the more likely the temperature distribution feature at the real-time moment will form a void defect, so the process risk can be judged in real time according to the flow risk indicator. And after it is determined that there is a process risk, instead of blindly adjusting the global temperature, the embodiment of the present application analyzes the core motivation of the current high flow risk indicator based on the optimal matching result, determines the position point to be adjusted by analyzing the heat conduction channel formed between the heat source point and the heat dissipation point in the optimal matching result, and then implements localized and targeted temperature regulation. For example, if the flow front below the heat source chip A flows too slowly because it is close to the heat dissipation TGV array B, the system will instruct the heating table to moderately increase the temperature of the region corresponding to the TGV array B to reduce the viscosity of the filling material in this region, thereby accelerating the flow; on the contrary, if the material flow below the heat source chip C is too fast because it is far away from the heat dissipation TGV array, the temperature of the region corresponding to the TGV array is reduced. This adjustment method can be regarded as dynamically changing the real-time viscosity distribution map to offset the generated process risk until the flow risk indicator returns to the safe interval, realizing closed-loop and real-time monitoring, and effectively preventing the generation of void defects in the filling process.

[0057] In the embodiment of the present application, the flow risk index threshold value can be preset, and if the flow risk index is greater than the preset flow risk index threshold value, it is judged that process risk will occur, and the temperature control strategy is executed. The flow risk index threshold value can be set according to the expected control accuracy in the actual implementation scene, which is not described here.

[0058] In the embodiment of the present application, for the thermal interaction vector under the optimal matching result, the obtained sum vector is subjected to cosine similarity acquisition with each thermal interaction vector, and if the cosine similarity is greater than 0.7, it is judged that the thermal interaction vector is the dominant heat conduction channel under the current optimal matching result, and then the temperature control is performed on the heat dissipation points in these dominant heat conduction channels. The instantaneous flow velocity of the material edge in each dominant heat conduction channel can be compared with the average instantaneous flow velocity in the non-dominant heat conduction channel to judge whether the current dominant heat conduction channel belongs to the case of too fast flow or too slow flow, and then the temperature is accurately controlled. In the specific temperature control, a step-by-step control method can be used, for example, the temperature is adjusted by one degree at a time until the obtained flow risk index is less than the flow risk index threshold value, and the temperature control is completed.

[0059] Step S4: for each position point at the bottom, the local thermal shock is obtained according to the gradient of the position point in the thermal conductivity distribution map and the temperature change at the adjacent moment; the position point damage condition is judged according to the local thermal shock and the thermal expansion coefficient of the material at the position point, the thermal stress damage value of each position point at the real time moment is obtained and constitutes a thermal stress damage map; after the filling is completed, the process optimization direction is analyzed according to the thermal stress damage map.

[0060] In order to optimize the whole filling process and improve the reliability of the process, the embodiment of the present application further obtains the thermal stress damage map. The conventional stress analysis method can only determine the high stress area, but cannot determine the cause of the formation of the high stress area. Compared with the conventional stress analysis method, the embodiment of the present application can not only identify the high risk point by constructing the thermal stress damage map, but also determine the process improvement direction by backtracking the process of the thermal stress damage map, so that the subsequent improvement work can be improved in a certain direction, which greatly improves the process reliability and improves the product development efficiency.

[0061] In the construction process of the thermal stress damage map, for each position point, the gradient in the thermal conductivity distribution map can quantify the sensitivity of the position point to amplify local temperature fluctuations in the dynamic thermal process, and then the local thermal impact at the position point can be determined according to the gradient characteristics and the temperature change at the adjacent time. That is, the greater the sensitivity to temperature fluctuations at the position point, and the greater the temperature change characteristics at the adjacent time, the greater the local thermal impact on the position point. Further, in combination with the thermal expansion coefficient of the filling material at the position point, the damage condition of the position point can be determined, and then the thermal stress damage value of each position point at the real time is obtained and the thermal stress damage map is constructed. That is, in the construction process of the thermal stress damage map, the main analysis includes the characteristics of the material itself and the temperature change characteristics, and based on the two directions, the process can be optimized.

[0062] Preferably, in the embodiment of the present application, the method for obtaining the local thermal impact comprises:

[0063] Taking the absolute value of the gradient of the position point in the thermal conductivity distribution map as the interface thermal conductivity mutation degree; gain the temperature change at the adjacent time according to the interface thermal conductivity mutation degree, and obtain the local thermal impact.

[0064] In the embodiment of the present application, the temperature change at the adjacent time is the absolute value of the temperature difference between the real time and the previous time at the position point; after the interface thermal conductivity mutation degree is normalized, 1 is added to obtain a gain coefficient, and the product of the gain coefficient and the temperature change is the local thermal impact.

[0065] Preferably, in the embodiment of the present application, considering that the position with a large difference in thermal expansion coefficient will produce permanent damage after bearing thermal impact, that is, the position is more fragile. Therefore, the damage bearing threshold of the position point can be obtained according to the gradient absolute value of the thermal expansion coefficient at the position point, that is, the greater the gradient absolute value, the smaller the damage bearing threshold corresponding to the position point. If the local thermal impact is greater than the damage bearing threshold, it is judged that there is damage; otherwise, it is judged that no damage is caused.

[0066] In the embodiment of the present application, the quantification method of the damage bearing threshold is: taking the basic threshold corresponding to the reference material as the denominator, and taking the sum of 1 and the normalized gradient absolute value of the thermal expansion coefficient as the denominator to obtain the damage bearing threshold. The basic threshold is the prior data of the reference material, which can be directly obtained by consulting materials, that is, the method for obtaining the damage bearing threshold is essentially to take the basic threshold as a reference, to reduce the basic threshold by the gradient absolute value of the thermal expansion coefficient, and then to obtain the damage bearing threshold at the position.

[0067] Further, the method for obtaining the thermal stress damage value comprises:

[0068] If it is judged that there is damage, the difference between the local thermal shock and the damage bearing threshold is taken as the damage increment at the adjacent time; if it is judged that there is no damage, the damage increment is set to 0; the sum of the damage increment at the real-time time and the thermal damage value at the previous time is taken as the thermal damage value at the real-time time; wherein because there is no migration time at the initial time, the thermal damage value at the initial time is the corresponding damage increment.

[0069] Further, according to the thermal stress damage map construction process, the process optimization direction is analyzed, including:

[0070] The threshold value screening method is used to determine the high-risk point in the thermal stress damage map. The threshold value setting in the high-risk point screening process can also be set based on the expected accuracy in the actual implementation process; similarly, it can also be obtained according to the adaptive threshold value obtaining method such as the Otsu threshold value algorithm in the prior art, which will not be repeated and limited here. For each high-risk point, its characteristics in the thermal stress damage map obtaining process are analyzed to determine whether it is a high risk caused by the substrate design stage or a high risk in the temperature control process.

[0071] For the high-risk point, if the damage bearing threshold is less than a preset first threshold value, it means that the damage bearing threshold is too low. According to the above content, it can be known that the damage bearing threshold is too low because the gradient of the thermal expansion coefficient at the high-risk point causes the risk of the high-risk point, which is caused by the difference between the TGV and the thermal expansion coefficient of the substrate material. Therefore, the process optimization direction is to adjust the substrate structure, and it is recommended that the design engineer adjusts the TGV geometry at this place, changes the relative distance of the TGV and the chip corner point and other key structures, or introduces a stress buffer structure, etc. to fundamentally improve the design and improve its inherent thermal shock resistance. Finally, a “design optimization report” can be generated to record the analyzed information and displayed to the relevant staff.

[0072] If the local thermal shock is greater than a preset second threshold value, it means that the risk at the high-risk point position is caused by the improper temperature curve in the filling process, and the process optimization direction is to adjust the temperature control. The actual temperature change curve of the high-risk point in the curing process can be displayed to the staff, and those sharp rising or falling times that cause large local thermal shock are specially marked. This will provide direct and quantitative evidence for process engineers to adjust the temperature rising and falling rate of the curing furnace or the holding platform setting to achieve a more gentle local temperature transition, thereby reducing the damage introduced by the process without changing the design. Finally, a “process optimization report” can be generated to record the analyzed information and displayed to the relevant staff.

[0073] The setting method of the first threshold value and the second threshold value can be set according to actual implementation scenes, and details are not described herein. In the embodiment of the application, the first threshold value is set as a data value at the position of the top 20% in ascending order after the damage bearing threshold value obtained by calculation; and the second threshold value is set as a data value at the position of the last 80% in ascending order after the local thermal shock obtained by calculation.

[0074] In summary, the embodiment of the application collects the surface temperature distribution map of the substrate bottom in real time, and then fuses it with the thermal conductivity distribution map to obtain a heat source-heat dissipation channel strength map. The material flow imbalance degree is determined according to the heat transfer relationship between the heat source point and the heat dissipation point, and the material flow risk index at the real-time moment can be obtained by combining the real-time material flow speed. The local thermal shock is obtained through the temperature change at adjacent moments, the thermal stress damage map is drawn, and finally the process optimization direction can be determined by reviewing and analyzing the thermal stress damage map obtained in real time after the process ends. The application can effectively warn the cavity risk in the filling process by real-time information acquisition combined with the prior information in the design file, and can reasonably improve and optimize the process based on the extracted features, thereby improving the process reliability.

[0075] It should be noted that the above-mentioned embodiment sequence of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or may be advantageous.

[0076] Each embodiment in the specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment focuses on the difference from other embodiments.

Claims

1. A temperature detection based chip underfill optimization method, comprising: The method comprises: obtaining a thermal conductivity distribution map of the bottom of the packaging area according to a design file; and collecting a surface temperature distribution map of the bottom in real time during the bottom filling process; combining the gradient of each point in the surface temperature distribution map with the thermal conductivity distribution map to obtain a heat source-heat dissipation channel intensity map; the heat source-heat dissipation channel intensity map comprises heat source points and heat dissipation points, the heat source points and the heat dissipation points are matched with each other, the material flow imbalance degree is obtained according to the optimal matching result, and the flow risk index is obtained in combination with the material flow velocity information collected in real time; judging the process risk in real time according to the flow risk index, and adjusting the temperature of the corresponding position point through the optimal matching result; for each position point of the bottom, obtaining a local thermal shock according to the gradient of the position point in the thermal conductivity distribution map and the temperature change at the adjacent time; judging the damage condition of the position point according to the local thermal shock and the thermal expansion coefficient of the material at the position point, obtaining the thermal force damage value of each position point at the real time and constructing a thermal stress damage map; and analyzing the process optimization direction according to the thermal stress damage map after the filling is completed.

2. The chip underfill optimization method based on temperature detection according to claim 1, wherein, The method for obtaining the heat source-heat dissipation channel intensity map comprises: for each position point of the bottom, multiplying the gradient of the position point in the surface temperature distribution map by the opposite number of the thermal conductivity in the thermal conductivity distribution to obtain a heat flux vector field; and calculating the divergence of each position point in the heat flux vector field, and the divergence constitutes the heat source-heat dissipation channel intensity map.

3. The method of claim 1, wherein the method further comprises: The positioning method of the heat source points and the heat dissipation points in the heat source-heat dissipation channel intensity map comprises: taking the local maximum value point in the heat source-heat dissipation channel intensity map as the heat source point and taking the local minimum value point as the heat dissipation point.

4. The chip underfill optimization method based on temperature detection of claim 1, wherein, The matching method of the heat source points and the heat dissipation points comprises: combining any one heat source point and any one heat dissipation point to obtain a thermal force interaction vector; multiplying the distance between the heat source point and the heat dissipation point by the element value of the heat source point in the heat source-heat dissipation channel intensity map to obtain the module length of the thermal force interaction vector, and the direction in which the heat source point points to the heat dissipation point is the direction of the thermal force interaction vector; taking the square of the module length of the thermal force interaction direction as the matching cost, and using the Hungarian algorithm to optimize to obtain the optimal matching result.

5. The chip underfill optimization method based on temperature detection according to claim 4, wherein, The method for obtaining the material flow imbalance degree comprises: performing vector summation on the thermal force interaction vector corresponding to the optimal matching result, and taking the module length of the sum vector as the material flow imbalance degree.

6. The method of claim 1, wherein the method further comprises: The method for obtaining the flow risk index comprises: monitoring the instantaneous velocity difference and the cumulative position difference of the filling material edges on both sides of the bottom gap area in real time; performing weighted summation on the instantaneous velocity difference and the cumulative position difference to obtain a flow abnormality feature; and performing gain on the flow abnormality feature according to the material flow imbalance degree to obtain the flow risk index.

7. The method of claim 1, wherein the method further comprises: The method for obtaining the local thermal shock comprises: taking the absolute value of the gradient of the position point in the thermal conductivity distribution map as an interface thermal conductivity mutation degree; and performing gain on the temperature change at the adjacent time according to the interface thermal conductivity mutation degree to obtain the local thermal shock.

8. The method of claim 1, wherein the method further comprises: The method for judging the damage condition of the position point comprises: The damage bearing threshold of the position point is obtained according to the gradient absolute value of the thermal expansion coefficient at the position point; if the local thermal shock is greater than the damage bearing threshold, it is judged that there is damage; otherwise, it is judged that no damage is caused.

9. The chip underfill optimization method based on temperature detection according to claim 8, wherein, The method for obtaining the thermal damage value comprises: If it is judged that there is damage, the difference between the local thermal shock and the damage bearing threshold is taken as the damage increment at the adjacent time; if it is judged that no damage is caused, the damage increment is set to 0; the sum of the damage increment at the real-time time and the thermal damage value at the previous time is taken as the thermal damage value at the real-time time; wherein the thermal damage value at the initial time is the corresponding damage increment.

10. The method of claim 9, wherein the temperature detection-based optimization of underfilling of a chip is based on a temperature of the chip. The process optimization direction is analyzed according to the thermal stress damage map construction process, which comprises: The high-risk points in the thermal stress damage map are determined by using a threshold screening method; for the high-risk points, if the damage bearing threshold is less than a preset first threshold, the process optimization direction is to adjust the substrate structure; if the local thermal shock is greater than a preset second threshold, the process optimization direction is to adjust the temperature control.

Images