Method for filling the interface of integrated circuit package silicon powder

Abstract

The application discloses an integrated circuit packaging silicon powder interface filling method and belongs to the field of integrated circuit packaging materials and packaging processes. In order to solve the problem that the interface of electronic-grade silicon powder as a packaging filler is rich, cavities and weak combination affect the packaging reliability, the application generates an interface risk field through packaging interface data set, configures a main body accumulation particle group and an interface filling particle group, adopts interface filling slurry for pulse pre-filling and combines gel window control solidification, and the technical effects of improving interface filling continuity and reducing thermal cycle failure risk are realized.

Description

Technical Field

[0001] This invention relates to the field of integrated circuit packaging materials and packaging processes, and more particularly to a method for filling the interface of silicon micropowder in integrated circuit packaging. Background Technology

[0002] In integrated circuit packaging, epoxy molding materials typically require the addition of electronic-grade silicon micropowder to adjust the coefficient of thermal expansion, flow properties, mechanical properties, and thermal conductivity. As chip size, lead frame structure, packaging substrate step structure, and local gap morphology become more complex, the particle size distribution, surface state, and flow migration behavior of silicon micropowder in the resin system have a more direct impact on the interface filling quality.

[0003] Current encapsulation filling processes primarily control the overall filler ratio, resin viscosity, or molding pressure. However, at chip edges, lead frame steps, and localized areas of the encapsulation substrate, mismatches may exist between the surface activity, particle size distribution, and resin wettability of the silicon micropowder, easily leading to localized enrichment, voids, and weak bonding interfaces. During curing and thermal cycling, these interface defects can cause stress concentration, warping, cracking, or delamination, and affect the heat dissipation stability of the encapsulation structure.

[0004] Therefore, there is a need for a method for filling silicon micropowder interfaces in integrated circuit packaging that can overcome the shortcomings of the existing technologies. Summary of the Invention

[0005] One objective of this invention is to propose a method for interface filling of silicon micropowder in integrated circuit packaging. Addressing the problems of localized enrichment, voids, and weak bonding caused by mismatches in the wettability, particle size distribution, and interface filling requirements of electronic-grade silicon micropowder with resin in existing technologies, this invention proposes a technical solution based on the configuration of the main packing particle group and the interface filling particle group using a packaging interface dataset and an interface risk field. It also employs an interface filling slurry for pulse pre-filling and gel window curing control. This invention has the technical effects of improving the continuity of interface particle distribution, enhancing resin wetting stability, and reducing the risk of thermal cycling failure.

[0006] This invention provides a method for filling the interface of silicon micropowder in integrated circuit packaging, comprising: S1, acquiring packaging geometry data, interface gap data, resin viscosity-temperature data, silicon micropowder particle size distribution data, silicon micropowder surface energy data, resin contact angle data, and historical defect data of the device to be packaged, and generating a packaging interface dataset; S2, calculating the interface risk field based on the packaging interface dataset, and dividing the area to be filled into an interface filling area and a main filling area according to the interface risk field; S3, determining the input of the particle packing fraction model and the surface energy matching index based on the packaging interface dataset and the interface risk field, and determining the proportion of the main packing particle group, the particle size proportion of the interface filling particle group, and the coupling treatment amount accordingly, and preparing the interface filling slurry and the main encapsulation material; S4, performing pre-filling of the interface filling area with the interface filling slurry by alternating vacuum holding and positive pressure pulse, and generating pulse correction parameters based on the pressure attenuation signal, dielectric curing signal, and acoustic void signal; S5, completing the filling of the main encapsulation material according to the pulse correction parameters, performing pre-geling pressure holding within the gel window corresponding to the dielectric curing signal, and performing post-curing after the pre-geling pressure holding is completed.

[0007] Optionally, S1 includes: The chip edge gap, lead frame step height and packaging substrate surface undulation are converted into a geometric gap matrix; Convert the viscosity values ​​of the resin at at least two temperature sampling points into a viscosity-temperature curve; The particle size distribution, specific surface area, surface hydroxyl density, and contact angle test values ​​of silicon micropowder are combined into a filler surface condition table; The locations of voids, delaminations, and cracks in historical defect data are mapped to the device coordinate system to obtain the packaging interface dataset.

[0008] Optionally, S2 includes: The encapsulation interface dataset is discretized into multiple grid cells, and the gap resistance component, thermal expansion mismatch component, thermal path weight component, and historical defect component of each grid cell are calculated. The interface risk value is obtained by summing the above components after normalizing them according to preset weights. Grid cells with interface risk values ​​not less than a preset risk threshold are defined as interface filling areas, and the remaining grid cells are defined as main body filling areas. The interface risk field consists of the interface risk value and area marker of each grid cell.

[0009] Optionally, S3 includes: Input the candidate particle size distribution of the main particle group into the particle packing fraction model, and select the particle size distribution that achieves the target filling volume fraction and the slurry viscosity does not exceed the set viscosity threshold as the proportion of the main packing particle group. The surface energy matching index is generated by weighting the difference between the surface energy of the resin and the surface energy of the coupled silicon micropowder, the contact angle, and the specific surface area of ​​the filler. The particle size ratio and coupling treatment amount of the interface filling particle group are determined by looking up the treatment amount mapping table based on the surface energy matching index. Furthermore, the silicon micropowder used for the interface filling particle group adopts a double-layer reactive surface structure, which includes a silane coupling inner layer that is condensed with hydroxyl groups on the surface of the silicon micropowder, and a resin compatible outer layer containing one of the reactive groups of epoxy, amino, acid anhydride compatibility group or mercapto group. The coupling treatment amount is determined based on the surface energy matching index, combined with the molar amount of inner reactive groups and the molar amount of outer reactive groups; Furthermore, the interface risk value of each grid cell in the interface filling region is extracted from the interface risk field, and the geometric gap matrix of the corresponding grid cell is extracted from the encapsulated interface dataset. The statistical value of the gap width of the corresponding grid cell is determined as the target gap width, and the upper limit of the particle size is determined by the product of the target gap width and the preset gap ratio coefficient, and the candidate set of particle sizes is screened. After sorting the interface risk values ​​of each grid cell in the interface filling area, divide the quantile intervals according to the preset quantile points; Based on the particle size ratio of the interface filling particle group determined by the processing volume mapping table, the mass ratio of fine particle gradation is corrected according to the particle size candidate set, the quantile interval and the surface energy matching index, so as to obtain the interface fine particle ratio, which is a component of the particle size ratio of the interface filling particle group.

[0010] Optionally, S4 includes: After applying a vacuum holding phase to the interface filling area, a positive pressure pulse phase is applied, and this process is repeated a preset number of cycles. The pressure attenuation signal is generated from the pressure drop between adjacent pulses, the dielectric solidification signal is generated from the dielectric loss factor curve, and the acoustic void signal is generated from the acoustic echo amplitude. Based on the above signals, pulse correction parameters are generated, including vacuum holding time, positive pressure pulse amplitude, positive pressure pulse count, and trigger flag for transitioning to body filling; Furthermore, the pressure attenuation signal, dielectric curing signal, and acoustic void signal are normalized into pressure attenuation component, curing process component, and void echo component, respectively. The interface void risk value is obtained by fusing the pressure attenuation component and the void echo component according to a preset weight, and the solidification process component is used as the solidification process value. When the interface void risk value is not less than the preset void threshold and the curing process value has not reached the preset gel trigger value, if the void echo component is not less than the preset echo component threshold, the number of positive pressure pulses will be increased; if the pressure attenuation component is not less than the preset pressure component threshold, the vacuum holding time will be increased. When the curing process value reaches the preset gel trigger value, the trigger flag for transitioning to the main body filling will be set to active.

[0011] Optionally, S5 includes: The main encapsulation material is injected into the main filling area according to the pulse correction parameters, so that the main deposited particles and the interface filling particles form a continuous transition at the boundary of the area. The time period during which the first slope of the dielectric loss factor curve enters the preset slope range but does not reach the gel endpoint threshold is defined as the gel window. Pre-geling and pressure holding are performed within the gel window, and post-curing is performed after the pre-geling and pressure holding is completed according to a preset heating rate and holding time. Furthermore, the mold cavity temperature and encapsulation pressure are collected during the pre-gelling and pressure holding period, and the particle freezing control conditions are determined based on the gel window, mold cavity temperature, and encapsulation pressure. The particle freezing control conditions include gelation endpoint criteria, pressure holding criteria, and temperature criteria; When the particle freezing control conditions are met, i.e. the dielectric curing signal reaches the gel endpoint threshold, the encapsulation pressure is within the preset holding pressure range, and the mold cavity temperature is within the preset curing temperature range, the particle migration control ends and the post-curing stage begins, so that the interface filling particle group maintains the distribution state formed by the interface filling slurry pre-filling in the interface filling area.

[0012] The beneficial effects of this invention are: 1. By calculating the interface risk field based on the packaging interface dataset, and dividing the interface filling area and the main filling area according to the interface risk field, the silicon micropowder configuration and filling sequence can correspond to the local interface risks near the chip, lead frame or packaging substrate, thereby reducing the disordered particle enrichment and filling blind zone in the interface area.

[0013] 2. By determining the proportion of the main packing particle group, the particle size ratio of the interface filling particle group, and the coupling treatment amount based on the particle packing fraction model and surface energy matching index, and by adopting the interface filling particle group with a double-layer reactive surface structure, it is beneficial to improve the wetting and reaction bonding stability between silicon micropowder and encapsulation resin.

[0014] 3. By using interface-filling slurry for pre-filling with alternating vacuum holding and positive pressure pulses, and generating pulse correction parameters based on pressure attenuation signals, dielectric curing signals, and acoustic void signals, the interface filling state can be improved before pre-gelling pressure holding and post-curing, reducing the risk of voids, cracks, and delamination. Attached Figure Description

[0015] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a method for filling the interface of silicon micropowder in integrated circuit packaging.

[0016] Figure 2 This is a flowchart of step S4 of the present invention, which performs pulse feedback correction based on pressure attenuation, dielectric curing, and acoustic void signals. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0018] refer to Figures 1-2 A method for filling silicon micropowder interfaces in integrated circuit packaging includes: S1, acquiring packaging geometry data, interface gap data, resin viscosity-temperature data, silicon micropowder particle size distribution data, silicon micropowder surface energy data, resin contact angle data, and historical defect data of the device to be packaged, and generating a packaging interface dataset; S2, calculating the interface risk field based on the packaging interface dataset, and dividing the area to be filled into an interface filling area and a main filling area according to the interface risk field; S3, determining the input of the particle packing fraction model and the surface energy matching index based on the packaging interface dataset and the interface risk field, and determining the proportion of the main packing particle group, the particle size proportion of the interface filling particle group, and the coupling treatment amount accordingly, and preparing the interface filling slurry and the main encapsulation material; S4, performing pre-filling of the interface filling area with the interface filling slurry using alternating vacuum holding and positive pressure pulses, and generating pulse correction parameters based on the pressure attenuation signal, dielectric curing signal, and acoustic void signal; S5, completing the filling of the main encapsulation material according to the pulse correction parameters, performing pre-geling pressure holding within the gel window corresponding to the dielectric curing signal, and performing post-curing after the pre-geling pressure holding is completed.

[0019] In this specific embodiment, S1 includes: Establish device coordinate system And take the geometric center of the mold cavity as the origin. Using the main plane of the packaging substrate as Plane and using the thickness direction as Along the axial direction, the three-dimensional topography of the chip edge, lead frame steps, and surface undulations of the packaging substrate is acquired using a combination of industrial CT and white light interferometry and unified into a coordinate system. Below, the voxel resolution of industrial CT is set to... The surface scanning step size of the white light interferometer is set to... Furthermore, the scanning area covers the interface neighborhood extending 2mm outward from the outer edge of the chip. After denoising, fitting to the reference plane, and registering the chip edge gap, lead frame step height, and surface undulations of the packaging substrate, respectively, the data is then processed according to... The grid resolution is discretized into a geometric gap matrix. , where matrix elements Indicates that the grid index is along Effective interface gap width in the axial direction and with The effective interface gap width is defined as the remaining flowable space after subtracting the lower surface profile height from the upper surface profile height at the grid location and deducting the height occupied by the leader frame step in the projected area. The viscosity values ​​of the encapsulating resin at two temperature sampling points were subjected to rheological testing and converted into viscosity-temperature profiles. The temperature sampling points were set as follows: and The steady-state shear viscosity at the two temperatures is denoted as follows: and And the shear rate is set to The viscosity-temperature curve is constructed using the Arrhenius form. , , ; in Indicates temperature as The resin viscosity at that time and the unit is Indicates resin temperature and the unit is Represents the natural logarithm function. This represents the intercept parameter of the viscosity-temperature curve and is dimensionless. This represents the temperature-sensitive parameter of the viscosity-temperature curve, and the unit is . and These represent the first and second temperature sampling points, respectively, with units of 1. and They represent in and The viscosity measured below and the unit is ; The particle size distribution data of silicon micropowder was measured using a laser particle size analyzer and recorded as a volume fraction distribution curve. The specific surface area was measured using the BET nitrogen adsorption method and... The surface hydroxyl density was recorded and measured by chemical titration. The contact angle between the silicon powder and the resin was measured using the same resin system on a pressed silicon powder sample via the seat drop method and recorded in degrees. Simultaneously, the surface energy data of the silicon powder was measured using reversed-phase gas chromatography at 303.15 K, and the dispersed and polar components were output and unified. The above particle size distribution, specific surface area, surface hydroxyl density and contact angle test values ​​are combined into a filler surface condition table and indexed by the same batch of silicon micro powder number; The locations of voids, delaminations, and cracks from historical defect data are extracted from failure analysis images and converted into three-dimensional coordinate point sets. Each defect point set is aligned with at least two alignment marks and coordinate systems on the device. After rigid body registration, it is mapped to the device coordinate system. This causes the defect coordinates to be aligned with the geometric gap matrix. They have the same spatial reference; Finally, the geometric gap matrix Viscosity-temperature curve parameters The surface condition table of filler, surface energy data of silicon micropowder, resin contact angle data, and the mapped historical defect point set are written into the same structured record according to device model and material batch to form a packaging interface dataset, which is used for subsequent interface risk field calculation and region division.

[0020] In this specific embodiment, S2 includes: Read the package interface dataset and use the device coordinate system With geometric gap matrix The mesh division discretizes the projected region of the device to be filled into a grid. Consistent set of grid cells ,in Indicates the first Line 1 The index of the column grid cell and the grid cell in the coordinate system This corresponds to a fixed two-dimensional position range; For each grid cell The gap resistance component, thermal expansion mismatch component, thermal path weight component, and historical defect component are calculated separately to form four component matrices. The gap resistance component adopts the construction method of "flow resistance in narrow gaps increases inversely to the cube of the gap width". Specifically, the filling temperature set in the packaging process is used. Substitute the viscosity-temperature curve obtained in step S1 Seek and with As a viscosity measure, The initial value of the gap resistance is used as the gap scale to construct each grid cell, so that... smaller and The larger the value, the larger the original value, thus indicating that the area is more likely to form a filling dead zone and retain gas; The thermal expansion mismatch component is determined based on the interface type of the mesh element, identifying the materials on both sides involved in the mismatch and taking their coefficients of linear expansion. The interface type is determined by the geometric gap matrix. The height of the superimposed lead frame step and the chip edge profile are in The positional relationship is determined, and the coefficient of linear expansion of the material is read from the packaging material library and fixed for the silicon chip. Copper lead frame Packaging substrate Cured encapsulant and with thermal cycling temperature difference As a load scale, the material on both sides of the interface corresponding to each mesh element and Multiplying them yields the original mismatch value, making The larger the mismatch component, the greater the likelihood of delamination and cracking at that location. The thermal path weight components were determined through steady-state thermal conductivity simulation. The simulation model directly used the package geometry data from step S1 and discretized it using the three-dimensional finite volume method. The mesh was refined to a finer density at the chip edge and lead frame step region. The thermal conductivity of the material is fixed as that of silicon chips. ,copper Encapsulation material Packaging substrate The boundary condition is set to apply uniform heat source power to the active area of ​​the chip. Surface heat transfer coefficient of the packaged exterior And ambient temperature The interface normal heat flux density output by the simulation is used in each The original value of the thermal path is obtained by averaging the area on the grid cells, so that the weight component is larger at the more concentrated the heat flow, so as to characterize the higher the sensitivity of the void or weak bond to heat dissipation and thermal gradient. Historical defect components are mapped from S1 to the coordinate system. The point set of void locations, layer locations, and crack locations is generated using a fixed radius. The neighborhood statistics method counts the number of defect points within the radius of the center point of each grid cell and assigns weights according to defect type. The weights for voids, delamination, and cracks are set to 1, 2, 3, and 4, respectively. This results in areas with previously high failure rates having a higher defect component in the current device; The above four types of original values ​​are respectively applied to all grid cells. The internal components are converted to normalized components using minimum-maximum normalization. subscript Represents the gap resistance component. Indicates the thermal expansion mismatch component. Represents the hot path weight components, The historical defect component, after normalization, has a fixed range of values. Furthermore, a higher value indicates a higher risk contribution; The interface risk value matrix is ​​obtained by summing the four normalized components according to preset weights. and based on risk threshold Each grid cell is marked with a region label, and the interface risk value is calculated using the formula. ; in Represents grid cells The interface risk value is dimensionless. Indicate the weight of the gap resistance component and take Indicate the weight of the thermal expansion mismatch component and take Indicate the weight of the hot path components and take the weight. Represent the weights of historical defect components and take... Represents grid cells The normalized component of the gap resistance, Represents grid cells The normalized component of thermal expansion mismatch Represents grid cells normalized components of hot path weights Represents grid cells The normalization component of historical defects and These represent the grid row index and column index, respectively, and are integers. when The grid cell is then marked as an interface filling region and written into the region marking matrix. ,when Mark the grid cell as the main fill region and write it. Thus, the interface risk value matrix is ​​obtained. With region label matrix The interface risk field, which together constitutes the risk field, is used as the region input for step S3.

[0021] In this specific embodiment, S3 includes: Read the region marking matrix based on the interface risk field And The set of grid cells is defined as the set of interface fill areas. ,Will The set of grid cells is defined as the set of main filling regions. Simultaneously extract from the encapsulated interface dataset Corresponding geometric gap matrix elements And calculate the statistical value of the target gap width in the interface filling area, and take the statistical value of the target gap width as... All The 10th percentile, based on which the upper limit of the allowable particle size of the interface filling particle group is determined, is the statistical value of the target gap width and the gap ratio coefficient. The product of the particles is then used to select all particle size distribution combinations whose maximum particle size is not greater than the upper limit of the particle size in the silicon micropowder particle size database to form a particle size candidate set, which is then used to construct the particle size ratio of the interface filling particle group. The proportion of the main packing particle group is determined by the particle packing fraction model. The particle packing fraction model is constructed using a discrete element random close-packed model, with particle size distribution, particle density, and interparticle friction as inputs and the limiting packing fraction as the output. In the model, the particle shape is set to spherical and the particle density is set to... The static friction coefficient between particles is set to The normal restoration coefficient is set to The computational domain uses cubic elements with periodic boundary conditions and a side length of 1.0 mm. The number of particles released in each simulation is set to [value missing]. The particle size is randomly assigned according to the input particle size distribution, and isotropic compression is used to gradually increase the volume fraction of the system while keeping the average kinetic energy of the system below a certain level. And the change in volume fraction over 1000 consecutive iterations is less than As a convergence criterion, the volume fraction at convergence is defined as the limiting packing fraction of the particle size distribution. The candidate particle size distribution of the main particle group is set to coarse particle size distribution. Medium particle size distribution Fine-grained distribution The three-peak combination and the enumeration of the mass ratio of the three are step by step. All combinations of input particle packing fraction models were used, and a main encapsulating material with a limiting packing fraction of not less than 0.74 and a total filler volume fraction of 0.72 was prepared using the same resin system. With shear rate The measured steady-state viscosity does not exceed the viscosity threshold. The combination of these elements was determined as the main particle size distribution and fixed as coarse particles. medium-sized particles fine particles ; The interface-filling particle group is constructed as a three-peak combination and fixed as ultrafine particles within the above-mentioned particle size candidate set. fine particles Medium and fine particles Its baseline quality ratio is determined by the index result of the processing volume mapping table and then finely corrected according to the interface risk quantile. The surface energy matching index is generated from the resin surface energy, the surface energy of the coupled silicon powder, the contact angle of the resin to the silicon powder, and the specific surface area of ​​the silicon powder. It is used to look up the treatment quantity mapping table. The surface energy matching index is defined by the following formula: ; in This indicates that the surface energy matching index is dimensionless. Indicates the weight of the surface energy difference and takes Represents the surface energy of the resin, with units of . And the resin surface energy data recorded in the encapsulation interface dataset can be read directly. This represents the surface energy of silicon micropowder after coupling treatment, and the unit is 1. And the surface energy data of silicon micropowder recorded in the data set of the packaging interface can be read directly. This represents the absolute difference between the surface energy of the resin and the surface energy of the coupled silica powder, expressed in units of... Indicate the contact angle weight and take This represents the static contact angle of the resin on the silicon micropowder tablet sample, in degrees, and is directly read from the resin contact angle data recorded in the encapsulation interface dataset. The angle constant used to represent the normalized contact angle, with the unit being degrees. Represent the specific surface area weight and take This represents the specific surface area of ​​silicon micropowder, with units of 1. And it can be directly read from the specific surface area data recorded in the encapsulated interface dataset. This represents the minimum specific surface area of ​​all candidate particle size pairs in the database for this batch of silicon micropowder, expressed in units of... This represents the maximum specific surface area of ​​all candidate particle size pairs in the database for this batch of silicon micropowder, expressed in units of [missing information]. ; Processing volume mapping table For indexing and outputting coupling processing volume Coupled processing volume Defined as the mass percentage of coupling agent to silicon micropowder, with units of . The mapping table is fixed as when hour ,when hour ,when hour Based on this, the reference particle size ratio of the interface filling particle group was simultaneously determined to be ultrafine particles. fine particles Medium and fine particles ; The silica micropowder used for interface filling particle groups adopts a double-layer reactive surface structure and is constructed by an inner silane coupling condensation and an outer resin-compatible reactive layer. The inner layer uses... - Aminopropyltriethoxysilane as a silane coupling agent and according to the coupling treatment amount Feeding is carried out, and the treatment process is set at a volume ratio of ethanol to deionized water. Prepare silane hydrolysate in solvent and then prepare the solution. Adjust to 4.5, add silicon micropowder to the hydrolysate and... Hydrolysis and adsorption were completed by mechanical stirring for 60 minutes, followed by... After heat treatment for 2 hours, the condensation with the hydroxyl groups on the surface of the silicon micropowder is completed to form a silane coupling inner layer; The outer layer uses an epoxy-terminated phenolic epoxy oligomer as the resin-compatible outer layer material and is grafted with silane-coupled inner layer exposed amino groups as reaction sites. The grafting process is set to dissolve the phenolic epoxy oligomer in acetone solvent to form a mass fraction The solution was prepared and the silicon micropowder that had undergone inner layer treatment was added to the solution. Stirring for 90 minutes allows some epoxy groups to undergo ring-opening reactions and form an outer layer. At the same time, the amount of outer layer material is controlled so that unreacted epoxy groups are retained as resin reaction groups after the outer layer reaction, thus obtaining a double-layer reactive surface silica powder with an inner silane coupling structure and an outer epoxy compatibility structure. Risk quantile correction is performed on the fine-grained gradation of the interface filling particle group, specifically by extracting risk quantiles from the interface risk field. Interface risk value of each grid cell Sort by numerical value from smallest to largest and then by quantile. Divide the data into four quantile intervals and calculate the corresponding geometric gap matrix within each quantile interval. The 10th percentile is calculated as the gap constraint scale for this quantile interval. Simultaneously, a monotonic correspondence is established between the ultrafine particle mass ratio and the interface risk quantile interval, starting from the baseline particle size ratio, and the correction rule is fixed as the ultrafine particle ratio of the first quantile interval. The proportion of ultrafine particles in the second quantile interval The proportion of ultrafine particles in the third quantile interval The proportion of ultrafine particles in the fourth quantile interval When increasing the proportion, the proportion of medium and fine particles is simultaneously decreased to maintain the total amount of interfacial filling particles. Furthermore, the upper limit constraint of the particle size candidate set is satisfied throughout the process, thereby obtaining the proportion of fine particles at the interface and using it as a component of the final particle size proportion of the interface filling particle group. Interface-filling slurry and main encapsulant were prepared separately. The main encapsulant was metered into the resin system at a ratio of the main particle size distribution to the total filler volume fraction of 0.72. The mixture was stirred at 1200 rpm for 8 minutes in a planetary vacuum mixer while maintaining a vacuum of -0.08 MPa to complete degassing and dispersion. The interface-filled slurry was metered and added to the same resin system according to the final interface-filled particle size ratio, with a total filler volume fraction of 0.60 to reduce flow resistance during interface pre-filling. The mixture was stirred under the same vacuum conditions for 10 minutes, and after discharge, it was... Metal mesh filtration removes agglomerates, resulting in an interface-filling slurry for S4 and a body encapsulant for S5.

[0022] In this specific embodiment, S4 includes: The interface filling slurry is collected along the interface filling area through the dispensing valve. The corresponding projected boundary forms a continuous adhesive line and covers the slurry. The thickness of the coating is controlled to 0.30mm. The mold is then closed, and the mold cavity is connected to a pressure control unit integrating vacuum and pressurization to execute a pre-filling cycle that alternates between a vacuum holding phase and a positive pressure pulse phase. The initial process parameters for the pre-filling cycle are set to the absolute pressure of the vacuum holding phase. Vacuum holding time Positive pressure pulse absolute pressure Positive pressure pulse duration Maximum number of loops Each cycle follows the sequence of "vacuum hold → positive pressure pulse → return to atmospheric pressure baseline". And maintain the 2s" sequence to promote the venting of the interface filling slurry in the narrow gap and enter the step and edge blind zone under pressure; A capacitive pressure sensor is installed inside the mold cavity to collect the mold cavity pressure curve. and with sampling frequency Record, define the first The pressure at the end of the baseline holding segment after the second cycle is: And the pressure decay signal is defined as ,in Indicates the cyclic index, ranging from 1 to... integers, Indicates the first The baseline pressure corresponding to the next cycle, and the unit is... It represents the pressure drop between adjacent cycle baselines and is expressed in kPa. The pressure decay signal is used to characterize the change in compressible volume fraction caused by the continuous release or leakage of gas in the interfacial micropores. Dielectric curing monitoring electrodes are embedded at the mold cavity interface, and dielectric loss factor curves are collected by a dielectric analyzer. And the sampling frequency is set to The dielectric curing signal is defined by the sampled value of the dielectric loss factor at the end of each cycle of vacuum holding. ,in Used to characterize the changes in polarization loss caused by resin reaction and as input for the curing process criterion; A piezoelectric ultrasonic transducer with a center frequency of 25MHz was fixed on the outer wall of the mold cavity. A single-pulse ultrasonic wave was emitted during the baseline holding segment of each cycle, and the echo envelope was acquired to define the acoustic cavity signal. S1 represents the maximum amplitude of the echo envelope within a threshold time window at the interface depth, expressed in units of V, where the start and end of the time window are determined by the geometric gap matrix of S1. The calculated interface depth range is converted to ensure that the amplitude corresponds to interface holes and layered reflections; The above three types of signals are normalized into pressure attenuation components. Curing process components With void echo component Normalization uses a fixed-range minimum-maximum normalization and sets... When a signal exceeds the upper limit of its corresponding range, its normalization result is truncated to 1; when it falls below the lower limit, it is truncated to 0 to ensure that all components are within the range. And possess a consistent scale of comparison; The interface void risk value is obtained by fusing the pressure attenuation component and the void echo component according to preset weights. And the curing process components As a value for the solidification process, the risk value of interface voids is calculated using the formula: ; in the formula Indicates the first The interface void risk value corresponding to the next cycle is dimensionless. Indicates the weight of the pressure attenuation component and takes Indicates the weight of the void echo component and takes Indicates the first The pressure decay component in the next cycle is dimensionless. Indicates the first The void echo component of the secondary cycle is dimensionless. Represents the cycle number and is an integer; Set a preset hole threshold Preset echo component threshold Preset pressure component threshold Preset gel trigger value ,when and When the pulse correction logic is entered, pulse correction parameters are generated accordingly. Then set the upper limit for the number of positive pressure pulses. Increase the current value by 2 and simultaneously increase the absolute pressure of the positive pressure pulse. Increase by 50 kPa to enhance the compaction and venting capacity in areas with strong interface reflection. Then the vacuum holding time Increase the current value by 10 seconds to enhance the pumping time for the continuous gas release zone, and the above two modifications can take effect simultaneously after the same cycle and be executed according to the updated parameters in the next cycle; when The trigger flag that will then transition to the main body filling. Set as valid and set ,in This indicates that the main body is filled with a trigger marker and is a dimensionless binary quantity, while freezing subsequent pulse corrections and using the current... Compared with the actual number of positive pressure pulses executed As the output pulse correction parameter, the pulse correction parameter includes the vacuum holding time. Positive pressure pulse amplitude Number of positive pressure pulses And the trigger flag for transitioning to the main body filling This information is then written into the process controller for the timing control of the S5 body package filling process.

[0023] In this specific embodiment, S5 includes: The process controller reads the pulse correction parameters and triggers the switch to the main body filling. The main body encapsulation material filling process is initiated, and the main body encapsulation material is heated and maintained at a constant temperature during the filling process. Then, the main body filling area is filled by injecting the transfer molding plunger. The displacement of the piston filling the main body is controlled by a closed loop and the displacement rate is set to... To ensure stable flow front advancement, the mold cavity pressure is controlled in a closed loop and based on the encapsulation pressure setpoint. Perform pressure holding in the filling section, combining the main filling area and the interface filling area. At the regional boundary, a continuous transition of particle populations is achieved through a "low-shear transition zone," which is defined as... Boundary direction Laterally extending 0.50 mm to form an annular band-shaped region and reducing the plunger displacement rate within this region. And maintain the mold cavity pressure not lower than This allows the main aggregated particle group to form a pushing connection with the interface-filling particle group at the flow front and suppress particle classification at the boundary. After the substrate filling is completed, continue to collect the dielectric loss factor curve corresponding to the dielectric curing signal. and with sampling frequency Record the encapsulation pressure within the mold cavity. With mold cavity temperature ,in This represents the time counted from the start of the main body filling, and the unit is 1. Indicates time as The dielectric loss factor at that time is dimensionless. Indicates time as The packaging pressure at that time and the unit is Indicates time as The temperature of the mold cavity at that time, and the unit is K; To determine the gel window, first... A smooth curve is obtained by using Savitzky-Golay filtering with a window length of 11 points and a polynomial order of 2. And calculate its first-order slope. ,in This represents the first-order slope of the dielectric loss factor as a function of time, with units of 1. This represents the dielectric loss factor after filtering and is dimensionless. Indicates time Differentiation operator and The unit is seconds (s). Will satisfy And simultaneously satisfy The continuous time period is defined as the gel window. ,in This represents the set of gel window time periods, with the unit being seconds. This indicates the lower limit of the gel window slope. This indicates the upper limit of the gel window slope. This represents the gel endpoint threshold and is dimensionless. At the gel window Perform pre-gel holding pressure and set the encapsulation pressure to [value]. This pressure is maintained until the end of the gelation window, thereby constraining volume shrinkage and gas compression at the interface and inhibiting particle redistribution before the resin reaches the gelation endpoint. During the pre-gelation pressure holding period, particle freezing control conditions are determined in real time. These conditions consist of a gelation endpoint criterion, a pressure holding criterion, and a temperature criterion. The gelation endpoint criterion is defined as follows: and The pressure holding criterion is defined as lasting 30 seconds. And last for 30 seconds and take Temperature criterion is defined as follows: And last for 30 seconds and take When all three criteria mentioned above are met simultaneously, particle migration control ends and the plunger displacement output is turned off, causing the mold cavity to enter a stable pressure-holding state, thereby allowing the interface-filling particle group to fill the interface filling area. The internal distribution remains formed by the pre-filling of the interface-filling slurry; Post-curing was then performed, employing a deterministic procedure of segmented heating and holding, while maintaining the mold cavity closed, the mold cavity temperature was increased at a certain rate. Rise to And keep warm Then, at a heating rate Rise to And keep warm After completion, release the pressure and open the mold to remove the packaged device to end step S5.

[0024] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0025] This invention utilizes a continuous technical chain between encapsulation interface datasets, interface risk fields, particle group ratio configuration, interface filling slurry prefilling, and gel window control to coordinate the spatial distribution of silicon micropowder fillers, the control of particle migration during resin impregnation and curing stages, thereby forming corresponding technical treatment methods for local interface defects near chips, lead frames, and packaging substrates.

[0026] This invention employs a double-layer reactive surface silicon micropowder and a pulse correction mechanism based on pressure, dielectric, and acoustic signal feedback. This allows the surface compatibility, fine particle size distribution, and filling pulse parameters of the interface filling particle group to be constrained and corrected according to interface risk and curing process, thereby better achieving the technical effect of reducing the risk of weak bonding zones and crack delamination after thermal cycling.

Claims

1. A method for filling the interface of silicon micropowder in integrated circuit packaging, characterized in that, include: S1. Obtain the packaging geometry data, interface gap data, resin viscosity-temperature data, silicon micropowder particle size distribution data, silicon micropowder surface energy data, resin contact angle data, and historical defect data of the device to be packaged, and generate the packaging interface dataset. S2. Calculate the interface risk field based on the encapsulated interface dataset, and divide the area to be filled into the interface filling area and the main body filling area according to the interface risk field. S3. Determine the input and surface energy matching index of the particle packing fraction model based on the encapsulation interface dataset and interface risk field, and accordingly determine the proportion of the bulk particle group, the particle size ratio of the interface filling particle group, and the coupling treatment amount, and prepare the interface filling slurry and the bulk encapsulation material; S4. Use the interface filling slurry to perform pre-filling of the interface filling area with alternating vacuum holding and positive pressure pulses, and generate pulse correction parameters based on the pressure attenuation signal, dielectric curing signal, and acoustic void signal; S5. Complete the filling of the bulk encapsulation material according to the pulse correction parameters, perform pre-gel holding pressure within the gel window corresponding to the dielectric curing signal, and perform post-curing after the pre-gel holding pressure is completed.

2. The method for filling the interface of silicon micropowder in integrated circuit packaging according to claim 1, characterized in that, S1 includes: converting chip edge gaps, lead frame step heights, and packaging substrate surface undulations into a geometric gap matrix; converting the viscosity values ​​of the resin at at least two temperature sampling points into viscosity-temperature curves; merging the particle size distribution, specific surface area, surface hydroxyl density, and contact angle test values ​​of silicon micropowder into a filler surface state table; and mapping the void locations, delamination locations, and crack locations in historical defect data to the device coordinate system to obtain the packaging interface dataset.

3. The method for filling the interface of silicon micropowder in integrated circuit packaging according to claim 1, characterized in that, S2 includes: The encapsulation interface dataset is discretized into multiple grid cells, and the gap resistance component, thermal expansion mismatch component, thermal path weight component, and historical defect component of each grid cell are calculated. The interface risk value is obtained by summing the above components after normalizing them according to preset weights. Grid cells with interface risk values ​​not less than a preset risk threshold are defined as interface filling areas, and the remaining grid cells are defined as main body filling areas. The interface risk field consists of the interface risk value and area marker of each grid cell.

4. The method for filling the interface of silicon micropowder in integrated circuit packaging according to claim 1, characterized in that, S3 includes: inputting the candidate particle size distribution of the main particle group into the particle packing fraction model, selecting the particle size distribution whose packing fraction output by the model reaches the target filling volume fraction and whose slurry viscosity does not exceed a set viscosity threshold as the proportion of the main packing particle group; generating a surface energy matching index by weighting the difference between the resin surface energy and the surface energy of the coupled silica powder, the contact angle, and the specific surface area of ​​the filler; and determining the particle size proportion and coupling treatment amount of the interface filling particle group by looking up the treatment amount mapping table according to the surface energy matching index.

5. The method for filling the interface of silicon micropowder in integrated circuit packaging according to claim 1, characterized in that, S4 includes: applying a positive pressure pulse stage after applying a vacuum holding stage in the interface filling region, and repeating the process a preset number of cycles; generating a pressure attenuation signal from the pressure drop between adjacent pulses, generating a dielectric curing signal from the dielectric loss factor curve, and generating an acoustic void signal from the acoustic echo amplitude; generating pulse correction parameters based on the above signals, wherein the pulse correction parameters include vacuum holding time, positive pressure pulse amplitude, positive pressure pulse count, and a trigger mark for transitioning to the main body filling.

6. The method for filling the interface of silicon micropowder in integrated circuit packaging according to claim 1, characterized in that, S5 includes: injecting the main encapsulant into the main filling area according to the pulse correction parameters, so that the main stacked particles and the interface filling particles form a continuous transition at the boundary of the area; defining the time period when the first slope of the dielectric loss factor curve enters the preset slope range and does not reach the gel endpoint threshold as the gel window; performing pre-geling and pressure holding within the gel window, and performing post-curing according to the preset heating rate and holding time after the pre-geling and pressure holding is completed.

7. The method for filling the interface of silicon micropowder in integrated circuit packaging according to claim 4, characterized in that, S3 includes: the silicon micropowder used for the interface filling particle group adopts a double-layer reactive surface structure, the double-layer reactive surface structure includes a silane coupling inner layer that is condensed with hydroxyl groups on the surface of the silicon micropowder, and a resin-compatible outer layer containing one of the reactive groups selected from epoxy, amino, acid anhydride, or thiol groups; the coupling treatment amount is determined based on the surface energy matching index and in combination with the molar amount of the inner layer reactive groups and the molar amount of the outer layer reactive groups.

8. The method for filling the interface of silicon micropowder in integrated circuit packaging according to claim 7, characterized in that, S3 includes: extracting the interface risk value of each grid cell in the interface filling region from the interface risk field, and extracting the geometric gap matrix of the corresponding grid cell from the encapsulation interface dataset; determining the gap width statistics of the corresponding grid cell as the target gap width, and determining the upper limit of the particle size according to the product of the target gap width and the preset gap ratio coefficient, and screening the particle size candidate set; sorting the interface risk values ​​of each grid cell in the interface filling region and dividing the quantile interval according to the preset quantile point; using the particle size ratio of the interface filling particle group determined by the processing volume mapping table as a benchmark, and correcting the mass ratio of fine particle gradation according to the particle size candidate set, the quantile interval and the surface energy matching index, to obtain the interface fine particle ratio as a component of the particle size ratio of the interface filling particle group.

9. The method for filling the interface of silicon micropowder in integrated circuit packaging according to claim 5, characterized in that, S4 includes: The pressure attenuation signal, dielectric curing signal and acoustic void signal are respectively normalized into pressure attenuation component, curing process component and void echo component. The interface void risk value is obtained by fusing the pressure attenuation component and the void echo component according to a preset weight, and the solidification process component is used as the solidification process value. When the interface void risk value is not less than the preset void threshold and the curing process value has not reached the preset gel trigger value, if the void echo component is not less than the preset echo component threshold, the number of positive pressure pulses will be increased; if the pressure attenuation component is not less than the preset pressure component threshold, the vacuum holding time will be increased. When the curing process value reaches the preset gel trigger value, the trigger flag for transitioning to the main body filling will be set to active.

10. The method for filling the interface of silicon micropowder in integrated circuit packaging according to claim 6, characterized in that, S5 includes: During the pre-gelation and pressure holding period, the mold cavity temperature and encapsulation pressure are collected, and the particle freezing control conditions are determined based on the gel window, mold cavity temperature, and encapsulation pressure. The particle freezing control conditions include gelation endpoint criteria, pressure holding criteria, and temperature criteria. When the particle freezing control conditions are met, i.e., the dielectric curing signal reaches the gelation endpoint threshold, the encapsulation pressure is within the preset pressure holding range, and the mold cavity temperature is within the preset curing temperature range, the particle migration control ends and the post-curing stage begins, so that the interface-filling particle group maintains the distribution state formed by the pre-filling of the interface-filling slurry within the interface-filling area.

Images