Process optimization method and system applied to liquid resin wafer plastic packaging

Abstract

The present application relates to wafer plastic packaging technical field, especially to a kind of process optimization method and system applied to liquid resin wafer plastic packaging, comprising: obtaining liquid resin initial characteristic, release film initial state and wafer size structure feature;Based on liquid resin initial characteristic and release film initial state, determine release film preheating control parameter, and preheating process is carried out in this way;After preheating process ends, feature recognition is carried out to release film target area, and release film surface feature and target area size feature are obtained;Release film surface feature, target area size feature and liquid resin initial characteristic are simulated, and according to flowability simulation result, determine extrusion parameter, and extrusion process is carried out in this way.It is filled by flowability simulation and mold control model in advance to judge the filling behavior of resin under different working conditions, solve the problem of uneven filling caused by fixed parameters.

Description

Technical Field

[0001] This invention relates to the technical field of wafer molding and encapsulation, and more particularly to a process optimization method and system for liquid resin wafer molding and encapsulation. Background Technology

[0002] Epoxy molding compound (EMC) is a core material for semiconductor chip packaging, accounting for over 97% of the total packaging materials. It plays an irreplaceable role in protecting chips from physical damage, isolating them from moisture and contaminants, and improving heat dissipation performance. As semiconductor packaging develops towards higher density and thinner profiles, liquid resins are increasingly widely used in wafer-level molding due to their advantages such as good fluidity and adaptability to complex structures.

[0003] In existing liquid resin wafer molding processes, most process parameters are pre-set and do not take into account the differences in the initial characteristics of the liquid resin, the state of the release film, and the wafer structure. This makes it difficult to adapt to the flow characteristics of the liquid resin, resulting in uneven filling during molding, forming surface unevenness and color difference defects. Moreover, there is a height difference of 2-3μm in the color difference area. This surface unevenness not only affects the accuracy of the chip step difference, but also causes a large number of abnormal images during optical inspection, leading to equipment jamming or missed detection risks, and seriously reducing production yield and efficiency. Summary of the Invention

[0004] This invention provides a process optimization method and system for liquid resin wafer molding by predicting the filling behavior of resin under different working conditions through flow simulation and mold control model, which can effectively solve the problems in the background art.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a process optimization method for liquid resin wafer encapsulation, comprising:

[0006] Acquire the initial properties of the liquid resin, the initial state of the release membrane, and the wafer-scale structural features;

[0007] Based on the initial characteristics of the liquid resin and the initial state of the release membrane, preheating control parameters for the release membrane are determined, and the preheating process is carried out accordingly.

[0008] After the preheating process is completed, feature recognition is performed on the target area of ​​the release membrane to obtain the surface features of the release membrane and the size features of the target area;

[0009] The surface features of the release membrane, the size features of the target area, and the initial characteristics of the liquid resin are simulated for flowability. Based on the flowability simulation results, the extrusion parameters are determined and the extrusion process is carried out accordingly.

[0010] After the extrusion process is completed, the real-time characteristics of the liquid resin are obtained and input into the molding control model along with the wafer size and structural features to obtain the molding process parameters, and the molding process is carried out accordingly.

[0011] In conjunction with the first aspect, in one possible design, the preheating control parameters of the release membrane include preheating temperature and preheating time;

[0012] The preheating temperature is determined based on the viscosity-temperature curve of the initial properties of the liquid resin and the thermal conductivity of the release membrane in its initial state.

[0013] The preheating time is determined based on the target area size characteristics of the release membrane and the expected flowability improvement time of the liquid resin at the preheating temperature, so that the target area of ​​the release membrane can reach a temperature state that is compatible with the initial characteristics of the liquid resin after preheating.

[0014] In conjunction with the first aspect, in one possible design, the formula for calculating the preheating temperature is:

[0015] ;

[0016] in, Indicates the preheating temperature; Indicates the resin viscous flow activation energy; Represents the gas constant; Indicates the target viscosity; This represents the pre-exponential factor, a benchmark value used to quantify how material properties change with temperature; Indicates ambient temperature.

[0017] In conjunction with the first aspect, in one possible design, the formula for calculating the preheating time is:

[0018] ;

[0019] in, Indicates the preheating time; This represents the density of the release membrane, determined from its initial state. This indicates the specific heat capacity of the release membrane, determined from the initial state of the release membrane. The target region's feature size is represented by the equivalent diameter of the target region, calculated from the target region's size feature. This represents the thermal conductivity coefficient of the release membrane, determined from the initial state of the release membrane; Indicates the initial temperature difference; This indicates the final temperature difference.

[0020] In conjunction with the first aspect, in one possible design, the surface characteristics of the release membrane include surface roughness, surface smoothness, and surface temperature.

[0021] In conjunction with the first aspect, in one possible design, the extrusion parameters include the spiral path, extrusion speed, and extrusion pressure.

[0022] In conjunction with the first aspect, in one possible design, the mold closing process parameters include mold closing speed and mold closing pressure.

[0023] In conjunction with the first aspect, in one possible design, the formula for the liquidity simulation is:

[0024] ;

[0025] Where U represents the radial diffusion uniformity of the resin during the annular extrusion process; k r This represents the path correction factor; P represents the extrusion pressure. Indicates the extrusion speed; d represents the maximum radius of the spiral trajectory; Indicates the dynamic viscosity of the resin; R a Indicates the surface roughness of the release membrane; represents the roughness influence coefficient; s represents the helical pitch; r represents the current radial position.

[0026] In conjunction with the first aspect, in one possible design, the mold-closing control model is constructed based on finite element analysis and includes a geometric modeling module, a physical parameter module, and a numerical solution module;

[0027] The geometric modeling module is used to generate a three-dimensional geometric model that is consistent with the actual mold cavity based on the wafer size and structural features;

[0028] The physical parameters module is used to store the physical properties of the resin and the mechanical parameters of the mold;

[0029] The numerical solution module incorporates a coupled fluid dynamics and curing dynamics algorithm to simulate the flow and curing behavior of resin during the mold closing process.

[0030] Secondly, the present invention also provides a process optimization system for liquid resin wafer encapsulation, comprising:

[0031] The feature acquisition module is used to acquire the initial properties of the liquid resin, the initial state of the release membrane, and the wafer size and structural features.

[0032] The preheating control module is used to determine the preheating control parameters of the release membrane based on the initial characteristics of the liquid resin and the initial state of the release membrane, and to control the execution of the preheating process accordingly.

[0033] The target area identification module is used to identify the features of the target area of ​​the release membrane after the preheating process, and to obtain the surface features of the release membrane and the size features of the target area.

[0034] The extrusion parameter determination module is used to perform flow simulation on the surface characteristics of the release membrane, the size characteristics of the target area, and the initial characteristics of the liquid resin, and determine the extrusion parameters based on the flow simulation results.

[0035] The extrusion execution module is used to execute the extrusion process according to the extrusion parameters;

[0036] The mold closing parameter determination module is used to obtain the real-time characteristics of the liquid resin after the extrusion process, and input them along with the wafer size and structural features into the molding and mold closing control model to obtain the mold closing process parameters.

[0037] The mold closing execution module is used to execute the mold closing process according to the mold closing process parameters.

[0038] The technical solution of this invention achieves the following technical effects: By acquiring the initial characteristics of liquid resin, the state of the release film, and the wafer structure features in real time, this invention enables precise matching of preheating, extrusion, and mold closing parameters with actual working conditions, fundamentally reducing the problem of uneven filling caused by rigid parameters; the preheating, extrusion, and mold closing processes are linked with process parameters to form a closed loop, with preheating parameters serving as a factor affecting the flowability of subsequent extrusion, extrusion parameters dynamically adjusted according to the state of the release film after preheating, and mold closing parameters optimized based on the real-time state of the resin after extrusion, avoiding cumulative errors caused by independent control of each process; through flowability simulation and mold closing control models, the filling behavior of resin under different working conditions can be predicted in advance, solving the problem of uneven filling caused by fixed parameters, thereby reducing surface unevenness and color difference defects. Attached Figure Description

[0039] Figure 1 This is a logic flowchart of the process optimization method applied to liquid resin wafer encapsulation in this invention;

[0040] Figure 2 This is a structural block diagram of the process optimization system applied to liquid resin wafer encapsulation in this invention. Detailed Implementation

[0041] This application will now be described with reference to the accompanying drawings.

[0042] like Figure 1 As shown, the process optimization method of the present invention for liquid resin wafer encapsulation specifically includes the following steps:

[0043] Step S100: Obtain the initial properties of the liquid resin, the initial state of the release membrane, and the wafer size and structural features;

[0044] Step S200: Based on the initial characteristics of the liquid resin and the initial state of the release membrane, determine the preheating control parameters of the release membrane, and perform the preheating process accordingly;

[0045] Step S300: After the preheating process is completed, feature recognition is performed on the target area of ​​the release membrane to obtain the surface features of the release membrane and the size features of the target area; the surface features of the release membrane include surface roughness, surface flatness and surface temperature; the target area is the liquid resin extrusion area;

[0046] Step S400: Perform flow simulation on the surface features of the release membrane, the size features of the target area, and the initial characteristics of the liquid resin, and determine the extrusion parameters based on the flow simulation results, and perform the extrusion process accordingly; the extrusion parameters include spiral path, extrusion speed, and extrusion pressure;

[0047] Step S500: After the extrusion process is completed, the real-time characteristics of the liquid resin are obtained and input into the molding control model along with the wafer size and structural features to obtain the molding process parameters, and the molding process is performed accordingly; the molding process parameters include molding speed and molding pressure.

[0048] In this embodiment, the present invention acquires the initial characteristics of the liquid resin, the state of the release film, and the wafer structure features in real time, enabling precise matching of preheating, extrusion, and mold closing parameters with actual working conditions. This fundamentally reduces the problem of uneven filling caused by rigid parameters. The preheating, extrusion, and mold closing processes are linked with process parameters to form a closed loop. The preheating parameters serve as a factor affecting the flowability of subsequent extrusion, the extrusion parameters are dynamically adjusted based on the state of the release film after preheating, and the mold closing parameters are optimized based on the real-time state of the resin after extrusion, avoiding the cumulative error caused by independent control of each process. Through flowability simulation and mold closing control models, the filling behavior of the resin under different working conditions can be predicted in advance, solving the problem of uneven filling caused by fixed parameters, thereby reducing surface unevenness and color difference defects.

[0049] In existing processes, parameter deviations in a single step can be amplified in subsequent steps. This method, however, uses the logic of providing an adaptation basis for subsequent steps with preceding parameters and adjusting subsequent parameters based on the real-time status of preceding parameters to achieve cross-step error compensation. For example, if the surface temperature of the release film is slightly lower than expected after preheating, the extrusion speed can be adjusted through flowability simulation during the extrusion stage, and the pressure can be optimized through real-time characteristics during the mold closing stage. The collaboration of multiple links can offset the small deviations in a single step, thereby improving the surface flatness control accuracy.

[0050] When dealing with different batches of liquid resin, release membranes with different aging levels, or irregularly shaped wafer structures, existing fixed parameters can lead to a sharp increase in defect rate due to fixed settings. However, this method can automatically adapt to extreme differences through full-process feature acquisition and model calculation. For example, for the combination of high-viscosity resin and rough release membrane, the preheating time will be increased, the extrusion speed will be reduced, and the mold closing speed will be reduced. The three factors work together to stabilize the product yield under complex working conditions.

[0051] Individual optimization of each process can only reduce a single defect. However, the overall logic of this method, through prediction, adjustment, and feedback loops, not only directly reduces the rework rate caused by surface defects, but also indirectly eliminates the chain costs caused by defects, such as the labor cost of re-judging abnormal images in AOI detection and the adaptation cost of subsequent packaging processes caused by step difference exceeding tolerance. This forms a positive cycle of process optimization, cost reduction, and efficiency improvement, and the overall benefits are better than the simple sum of individual process optimizations.

[0052] In some embodiments of the present invention, regarding the acquisition of the initial characteristics of the liquid resin, firstly, an appropriate sample is extracted from the batch of liquid resin to be used; for example, a clean sampling instrument is used to collect a certain volume of liquid resin to ensure that the sample is representative and can reflect the characteristics of the entire batch of resin; then, using a rotational viscometer, according to relevant industry standards, at specific temperatures and shear rates, the viscosity, viscous flow activation energy, power-law exponent, and other non-Newtonian fluid characteristic parameters of the liquid resin are detected; for example, the rotor speed and test time of the viscometer are set, the viscosity values ​​corresponding to different shear rates are recorded, and key parameters such as viscous flow activation energy are calculated through data fitting and other methods; then, using Fourier transform... Analytical techniques such as Fourier transform infrared spectroscopy and differential scanning calorimetry are employed. Fourier transform infrared spectroscopy can detect the types and contents of functional groups in the resin, determining whether its chemical structure meets process requirements. Differential scanning calorimetry can determine the glass transition temperature and heat of reaction of the resin, identifying the phase transition and reaction characteristics during the molding process. Additionally, the density and surface tension of the resin can be measured. Density can be measured using the specific gravity bottle method, where a known volume of resin sample is placed in a specific gravity bottle and weighed using a balance. Surface tension is tested using methods such as the pendant drop method and capillary rise method. Surface tension affects the spreading and filling behavior of the resin on the release membrane surface, serving as an input parameter for flowability simulation.

[0053] To obtain the initial state of the release membrane, firstly, the release membrane to be used is determined. If the release membrane has a protective film, it needs to be properly pretreated in a clean environment to remove the protective film and expose the effective surface in contact with the liquid resin. Then, the thermal conductivity of the release membrane is measured using a thermal conductivity meter, and its specific heat capacity and other parameters are tested using differential scanning calorimetry (DSC). For example, the release membrane sample is made into a specified shape, placed in the thermal conductivity meter's test chamber, and the thermal conductivity data is obtained according to the instrument's operating procedures. The change in specific heat capacity of the release membrane in different temperature ranges is then obtained through DSC testing. The initial temperature of the release membrane is measured using a high-precision temperature sensor to ensure uniform distribution of measurement points. Multiple measurements are taken and the average value is calculated to ensure accurate temperature data. The cleanliness of the release membrane surface is checked using a visual inspection system or a contact angle meter. Visual inspection can observe whether there are contaminants such as dust or oil on the surface. The contact angle meter indirectly judges the surface cleanliness and hydrophilicity / hydrophobicity by measuring the contact angle of liquid water droplets on the release membrane surface. If the contact angle is abnormal, it indicates that there may be contaminants on the surface or improper chemical modification, and the release membrane needs to be cleaned or replaced.

[0054] To obtain the initial state of the release film, firstly, the wafer to be encapsulated is placed on a high-precision support platform to ensure its level and stability. Vacuum adsorption or mechanical clamping can be used for fixation to prevent wafer movement during subsequent measurements from affecting data accuracy. Then, equipment such as an optical profilometer and laser rangefinder are used to measure the wafer's basic dimensions, such as diameter and thickness. The optical profilometer can quickly scan the wafer surface to obtain wafer thickness distribution data and determine if there are any thickness unevenness issues. The laser rangefinder can accurately measure the distance from the wafer edge to the center to determine if the wafer diameter meets process requirements. Wafer inspection equipment, such as a wafer defect detection system including optical imaging and electron beam detection modules, is used to identify the chip array layout, chip size, and spacing between chips (i.e., the target area size), as well as the step structure on the chip surface. For example, high-resolution optical imaging clearly presents the chip geometry and arrangement, and measures the gap size between chips. These chip structural features affect the liquid resin filling path and pressure distribution during mold closing.

[0055] In this embodiment, by standardizing and precisely acquiring initial characteristics, the initial conditions for molding each batch and each wafer are controllable. Regardless of whether it is a change of liquid resin between different batches, or a change of release membrane or wafer, subsequent process parameters can be adjusted based on accurate initial data, ensuring the consistency of the molding process in different production cycles and under different production environments, and reducing product quality differences caused by fluctuations in initial conditions. By mastering the initial state of liquid resin, release membrane, and wafer, risk factors that may affect molding quality can be identified in advance, such as abnormal resin viscosity, release membrane surface contamination, and wafer structural defects. For adjustable factors, such as resin viscosity which can be improved through pretreatment and release membrane surface contamination which can be cleaned, intervention can be carried out before the process begins. For non-adjustable factors, such as the inherent structure of the wafer, these are fully considered when adapting subsequent process parameters to prevent defects such as bubbles, incomplete filling, and surface unevenness from the source.

[0056] In some embodiments of the present invention, the preheating process of the release film has a significant impact on the subsequent extrusion and molding effects. The flowability and filling behavior of liquid resin are significantly affected by temperature changes, and the initial state of the release film also alters the resin's flow characteristics on its surface. Therefore, to ensure that the liquid resin can uniformly and smoothly fill the mold cavity during extrusion, it is necessary to accurately determine the preheating temperature and preheating time based on the initial characteristics of the liquid resin and the initial state of the release film. This helps the release film reach a temperature state compatible with the liquid resin, thereby improving the resin's flowability and filling performance, reducing filling defects, surface unevenness, and color differences caused by insufficient or excessive preheating, and improving wafer surface quality and the overall performance of the molded product.

[0057] Specifically, the method for determining the preheating temperature is as follows:

[0058] Step S211: Based on the viscosity-temperature curve of the initial properties of the liquid resin, determine the viscosity change law of the liquid resin at different temperatures; the viscosity-temperature curve reflects the trend of the viscosity of the liquid resin changing with temperature.

[0059] Step S212: Based on the thermal conductivity of the release membrane in its initial state, calculate the resistance and efficiency of the release membrane to heat transfer; release membranes with poor thermal conductivity require higher preheating temperatures or longer preheating times to reach the required temperature state.

[0060] Step S213: Determine the target viscosity range that the liquid resin should reach after preheating according to the target flowability requirements; a lower viscosity helps improve the flowability of the resin, but it should not be too low, so as not to affect the subsequent mold closing and curing effect.

[0061] Step S214: Taking into account the above factors, determine a suitable preheating temperature range through experimental data and simulation calculations to ensure that the release membrane and liquid resin achieve a good matching state in the subsequent extrusion process.

[0062] More specifically, the formula for calculating the preheating temperature is:

[0063] ;

[0064] in, Indicates the preheating temperature; It represents the resin viscosity activation energy, which is determined by the initial characteristics of the liquid resin. It is obtained by measuring the viscosity of the liquid resin at different temperatures and fitting it with the Arrhenius equation, reflecting the intrinsic characteristics of the viscosity of the liquid resin as a function of temperature. R represents the gas constant, R = 8.314 J / (mol·K); The target viscosity is indicated and adjusted according to flowability requirements. It is selected by referring to process experience and experimental data, and taking into account factors such as wafer structure complexity, to ensure that the liquid resin has good flowability and filling performance during the extrusion process. This represents the pre-exponential factor, a benchmark value used to quantify the change of material properties with temperature, obtained by measuring the viscosity of liquid resins and fitting the data. It indicates the ambient temperature and uses a high-precision temperature sensor to measure and record the real-time ambient temperature value near the preheating equipment or in the production workshop.

[0065] In this embodiment, the formula for calculating the preheating temperature is directly related to the resin's viscosity activation energy and pre-exponential factor, making the preheating temperature strongly correlated with the resin material properties. For example, when different batches or formulations of resin are used, the formula automatically calculates the new preheating temperature, avoiding the flowability problems caused by traditional fixed temperatures. Ambient temperature is introduced as a benchmark, allowing the preheating temperature to automatically adjust according to fluctuations in the workshop environment. For example, the preheating temperature calculated by the formula will change accordingly in high-temperature summer environments and low-temperature winter environments. Through a formulaic calculation method, complex material properties and process requirements are transformed into specific numerical parameters, making the preheating process more controllable and repeatable. When using different production batches, production lines, or changing different types of liquid resins and release membranes, the relevant parameters only need to be re-measured and substituted into the formula to accurately determine the preheating temperature, enhancing the stability and consistency of the process.

[0066] The method for determining the preheating time is as follows:

[0067] Step S221: Consider the size characteristics of the target area of ​​the release membrane, including the length, width, and shape of the target area; larger target areas require a longer preheating time to ensure that heat can be evenly transferred to the entire area.

[0068] Step S222: Analyze the expected flowability improvement time of liquid resin at the preheating temperature; different types of liquid resins have different response rates to temperature, and the time required to reach the expected flowability also varies.

[0069] Step S223: Based on the thermal inertia of the release membrane and the heating efficiency of the heating equipment, calculate the time required for heat to be transferred to the target area of ​​the release membrane during the preheating process;

[0070] Step S224: Based on the above factors, determine a reasonable preheating time range to ensure that the target area of ​​the release membrane can reach a temperature state that matches the initial characteristics of the liquid resin after preheating, while avoiding excessive preheating that could lead to energy waste or changes in material properties.

[0071] More specifically, the formula for calculating the preheating time is:

[0072] ;

[0073] in, Indicates the preheating time; The density of the release membrane is determined from its initial state using specialized material testing instruments such as a densitometer and a differential scanning calorimeter, reflecting the thermophysical properties of the release membrane material. This indicates the specific heat capacity of the release membrane, determined from the initial state of the release membrane. The characteristic size of the target area is usually taken as the equivalent diameter of the target area. Based on the shape and size of the target area of ​​the release membrane, it is measured and calculated by geometric calculation or image analysis software to characterize the size of the target area and the distance of heat transfer. The thermal conductivity coefficient of the release membrane is determined from the initial state of the release membrane and tested using a thermal conductivity meter according to relevant standards. It reflects the ability of the release membrane material to transfer heat. This represents the initial temperature difference, i.e. ; This represents the final temperature difference, i.e. , This refers to the temperature of the heating source.

[0074] In this embodiment, the formula for calculating the preheating time comprehensively considers the thermophysical properties of the release film, such as its density, specific heat capacity, and thermal conductivity, as well as the size characteristics and temperature difference changes of the target area. This allows for accurate calculation of the time required to meet the preheating requirements, ensuring that the release film reaches the expected temperature state after preheating. This avoids process problems caused by insufficient or excessive preheating time, improving the quality and efficiency of the preheating process. By accurately calculating the preheating time, energy waste caused by overheating is avoided, reducing energy consumption in the production process and helping enterprises save energy, reduce emissions, and lower production costs. Reasonably determining the preheating time can reduce unnecessary preheating waiting time, improving equipment utilization and production efficiency. Under the premise of ensuring preheating quality, shortening the preheating time as much as possible makes the entire molding process more compact and efficient, which is beneficial for meeting the needs of large-scale production and shortening the production cycle.

[0075] In some embodiments of the present invention, regarding step S300, the roughness and smoothness of the release membrane surface directly affect the flow and spreading behavior of the liquid resin during the extrusion process, while the surface temperature affects the viscosity and flowability of the resin. If the preset extrusion parameters are directly used, defects such as uneven filling and bubbles may occur due to the mismatch between the actual state of the release membrane and the resin characteristics. In addition, the size characteristics of the target area determine the size and shape of the area to be filled, directly affecting the flow path of the resin.

[0076] Specifically, the method for obtaining the surface features of the release membrane is as follows:

[0077] a. Surface roughness: The target area of ​​the release membrane is scanned using a laser confocal microscope or atomic force microscope, with the scanning range covering the entire extrusion area and a step size not exceeding 0.5 micrometers. The key parameters to be detected are the arithmetic mean roughness and the maximum profile height. The distribution density of microscopic protrusions or depressions on the surface is recorded, such as the number of protrusions greater than 1 micrometer per square millimeter. The purpose is that roughness directly affects the resin flow resistance. For example, excessive roughness can cause a sudden drop in local flow velocity, forming dead corners in the filling process. Microscopic resistance parameters are needed to provide flowability simulation.

[0078] b. Surface smoothness: A white light interferometer is used to scan the entire target area with a resolution of no more than 0.1 micrometers to generate a three-dimensional topography map. The calculation parameters include flatness error, i.e., peak and valley values, and local tilt angles, such as the height difference between the edge and center of the target area. The purpose is that flatness deviations can cause resin to climb or accumulate during flow, and the gravity field distribution in the flow simulation needs to be corrected by the test data.

[0079] c. Surface Temperature: High-precision temperature sensors, such as thermocouples or infrared thermometers, are used to measure the surface temperature of the target area of ​​the release membrane. Thermocouples can directly contact the surface of the release membrane for measurement, while infrared thermometers measure the temperature non-contactly by detecting the infrared radiation emitted by the surface of the release membrane. Temperature sensors are placed at multiple locations in the target area of ​​the release membrane to obtain comprehensive temperature distribution data. The measurement data are recorded, the average surface temperature is calculated, and the temperature uniformity is analyzed.

[0080] The method for obtaining the target region size feature is as follows:

[0081] Based on the size and structural characteristics of the wafer and the requirements of the extrusion process, the target area on the release film is clearly defined, that is, the area where the liquid resin will be extruded; the boundary of the target area is marked on the release film to facilitate subsequent measurement and identification.

[0082] Optical measuring equipment such as optical microscopes or image analyzers, or mechanical measuring tools such as vernier calipers or coordinate measuring machines, are used to measure the dimensional characteristics of the target area, including length, width, area, and shape. For areas with complex shapes, image analysis software is used to process and analyze the measurement images to extract the required dimensional information. The measurement data is recorded, and the characteristic dimensions of the target area, such as the equivalent diameter, are calculated.

[0083] In this embodiment, by real-time detection of the surface temperature distribution after preheating, local areas that have not reached the required temperature can be identified, such as edges where the temperature is too low due to rapid heat dissipation. The resin flow parameters can be adjusted accordingly in the flowability simulation, such as increasing the extrusion speed in that area, to avoid filling defects caused by uneven preheating. Surface roughness and smoothness data can be converted into wall shear force coefficients in the flowability simulation, making the simulation results closer to reality.

[0084] As a preferred embodiment of the above, in the liquid resin wafer encapsulation process, the annular extrusion method uses a spiral trajectory to move the extrusion nozzle, uniformly extruding the liquid resin into the mold cavity to achieve efficient wafer encapsulation. The annular extrusion method uses a mechanical device to drive the extrusion nozzle to move, forming a spiral trajectory of the resin. The extrusion nozzle moves along the set spiral path, continuously extruding the liquid resin into the mold cavity. As the extrusion nozzle moves and the resin is continuously extruded, the liquid resin gradually fills the mold cavity.

[0085] Specifically, in order to ensure that the liquid resin can uniformly fill the mold cavity and reduce defects such as uneven filling and air bubbles, it is necessary to perform flowability simulation analysis on the surface characteristics of the release membrane, the size characteristics of the target area, and the initial properties of the liquid resin, thereby obtaining extrusion parameters, which include:

[0086] Spiral path: Parameters such as the spacing, radius, and starting point of the spiral affect the filling effect and filling time of the liquid resin. Spiral path planning can ensure that the liquid resin fills the mold cavity evenly and reduce defects such as uneven filling and air bubbles. Based on the size characteristics of the target area, the spiral movement path of the gate is planned, including the starting point, the ending point, and parameters such as the spacing and radius of the spiral.

[0087] Extrusion speed: The extrusion speed determines the filling rate of liquid resin. Extrusion speed that is too fast may lead to uneven filling or air bubbles, while extrusion speed that is too slow will reduce production efficiency. By controlling parameters such as motor speed or hydraulic flow of the extrusion equipment, the extrusion speed can be precisely controlled to ensure that the liquid resin can fill the mold cavity at a suitable rate.

[0088] Extrusion pressure: Extrusion pressure is the driving force for the flow of liquid resin; appropriate extrusion pressure can ensure that liquid resin fills the mold cavity smoothly, but excessive pressure may cause overflow of resin in the mold cavity or damage the mold; adjust the extrusion pressure according to the viscosity of liquid resin and the resistance of the mold cavity so that liquid resin can be extruded and filled into the mold cavity smoothly.

[0089] Furthermore, the flow characteristics of liquid resins are complex and influenced by multiple coupled factors. The flow behavior of liquid resins is not only affected by their initial properties but also closely related to the surface characteristics of the release membrane and the size characteristics of the target area. In existing processes, extrusion parameters are mostly set based on experience, which cannot accurately adapt to the dynamic changes of the above-mentioned multiple factors, easily leading to defects such as uneven resin filling, residual air bubbles, or incomplete filling. Through flowability simulation, the flow state of the resin under complex boundary conditions can be predicted before actual extrusion, thereby optimizing the extrusion parameters based on the simulation results to ensure that the resin can be uniformly filled in the target area.

[0090] Specifically, the formula for simulating the flowability of liquid resin is shown below:

[0091] ;

[0092] Wherein, U represents the radial diffusion uniformity of resin during the ring extrusion process, which is used to characterize the radial diffusion uniformity of resin in the mold cavity. The closer the U value is to 1, the more uniform the filling is.

[0093] k r k represents the path correction coefficient, determined by the position of the spiral's starting point. When the starting point is aligned with the center of the target area, k... r =1, k at offset r <1, the larger the offset, the more k r The smaller the value, the better it is used to correct the effect of the deviation in the starting position of the spiral path on the uniformity of diffusion, ensuring that the rubber compound diffuses evenly from the center area.

[0094] P represents the extrusion pressure, which is the basic driving force for resin extrusion and directly determines the flow potential energy of the resin. It needs to be matched with the resin viscosity and mold cavity resistance. High viscosity resins require higher P to ensure flow.

[0095] This indicates the extrusion speed, which is the speed at which the outlet moves along the spiral trajectory. It is used to determine the rate and density of the extrusion of the adhesive. Too fast a speed can easily lead to local turbulence and the generation of bubbles, while too slow a speed will reduce efficiency and may cause defects due to premature resin curing.

[0096] d represents the maximum radius of the spiral trajectory, corresponding to the size characteristics of the target area. It needs to cover the area of ​​the wafer to be packaged. d must match the radius of the target area. If it is too large, it may exceed the mold cavity and cause glue overflow. If it is too small, it cannot completely cover the wafer.

[0097] The dynamic viscosity of the resin is the internal friction force of the resin flow at temperature T. It is affected by temperature T and is related to the preheating time of the release membrane. The longer the preheating time, the higher T becomes, and the lower the dynamic viscosity of the resin becomes. The dynamic viscosity of the resin under different preheating times is determined by experiment.

[0098] R a This represents the surface roughness of the release membrane, and its value is the arithmetic mean of the microscopic irregularities on the surface of the release membrane. It affects the resin flow resistance. If the roughness is too high, it will increase the frictional resistance of the resin flow, and its effect needs to be offset by adjusting the extrusion pressure or speed.

[0099] This represents the roughness influence coefficient, which quantifies the amplification effect of roughness on resistance and is obtained through experimental fitting.

[0100] s represents the helical spacing, which is the distance between adjacent helical trajectories. If the spacing is too small, it will easily lead to the accumulation of adhesive material. If the spacing is too large, it will easily form a filling dead corner. It needs to be matched with the resin viscosity. For low viscosity resin, s can be appropriately increased, while for high viscosity resin, s needs to be decreased.

[0101] r represents the current radial position, i.e., the distance from the center of the mold cavity to the current detection point, reflecting the diffusion state in different areas. Edge regions require higher diffusion dynamics; the formula decreases U by increasing r, indicating that extrusion parameters in edge regions need adjustment, such as increasing local pressure or speed.

[0102] In the above formula, the numerator represents the positive factors driving uniform resin diffusion, comprehensively reflecting the dynamic strength and coverage efficiency of spiral extrusion. The larger the value, the easier it is for the resin to diffuse uniformly. The denominator represents the negative factors hindering uniform resin diffusion, comprehensively reflecting the total resistance to resin flow. The larger the value, the more difficult it is to ensure uniform diffusion. By calculating the U-value at different radial positions, it is possible to determine whether the annular extrusion parameters are reasonable: if the U-value in a certain area is too large, the extrusion speed in that area needs to be increased or decreased; if the U-value in a certain area is too small, P needs to be increased or s decreased to ensure sufficient resin diffusion. Finally, by optimizing the parameters to make the U-value consistent across the entire region, uniform resin filling is achieved.

[0103] More specifically, the process of determining extrusion parameters based on simulation results aims to ensure uniform resin diffusion within the target area. According to the process objectives, the U-value deviation at each location within the target area must be controlled within a preset deviation threshold. A smaller deviation indicates more uniform filling. If the simulation shows abnormal local U-values, the extrusion parameters need to be adjusted accordingly, as follows:

[0104] Determining the spiral path parameters: including spacing, radius, and starting point; if the simulation shows that the U value of a certain radial area is too small, it indicates that the resin diffusion in that area is insufficient, which may lead to filling dead corners or color differences. It is necessary to check whether the spiral spacing is too large. The spiral spacing can be reduced to make the adjacent rubber tracks denser and enhance the resin coverage in that area; if the U value of the edge area is consistently too small, it may be that the spiral radius does not cover the edge of the target area. The spiral radius needs to be appropriately increased, but not beyond the mold cavity range; if the simulation shows that the U value of a certain area is too large, it indicates that the resin accumulates in that area, which may lead to height differences. It is necessary to check whether the spiral spacing is too small, causing excessive accumulation of rubber. The spiral spacing can be increased to reduce the amount of rubber in that area; if the U value of the central area is too large, it may be that the spiral starting point is off-center from the center of the target area. The starting point needs to be adjusted to the center position so that the diffusion proceeds evenly from the center to the edge;

[0105] Determining the extrusion speed: If the simulation shows that the overall U value is too small, it indicates that the resin diffusion is slow and the filling efficiency is low. This means that the extrusion speed is too slow, and the spiral movement speed at the outlet needs to be increased to accelerate the extrusion frequency of the adhesive and promote resin diffusion. However, it is necessary to consider the resin characteristics to avoid excessive speed causing local turbulence, which may lead to bubbles or surface ripples. If the simulation shows a sudden increase in the local U value, such as at the turning point of the spiral trajectory, it means that the extrusion speed at that location is not matched with the curvature of the path. That is, the speed is not adapted when turning, resulting in the accumulation of adhesive. The local speed in the turning area needs to be reduced. The spiral movement speed can be dynamically adjusted by mechanical devices to restore the U value to a stable state.

[0106] Determining the extrusion pressure: If the simulation shows that the U value is generally low and evenly distributed, it indicates that the resin has insufficient fluidity and the overall filling is slow. It is necessary to increase the extrusion pressure to enhance the resin's flow dynamics and reduce the resistance caused by viscosity. However, the pressure must be controlled within the mold's tolerance range. If the simulation shows that the U value in the edge area is significantly lower than that in the center, it may be that the mold cavity resistance in the edge area is greater. It is necessary to specifically increase the local extrusion pressure in the edge area to ensure that the resin can diffuse smoothly to the edge and reduce the height difference.

[0107] In this embodiment, by using flowability simulation and precise extrusion parameter optimization, the defects caused by existing processes relying on experience-based parameter settings are effectively solved. On the one hand, it can dynamically adapt to the coupled influence of multiple factors, comprehensively considering the dynamic changes in the initial characteristics of the liquid resin, the surface characteristics of the release membrane, and the size characteristics of the target area, avoiding the disconnect between parameters and actual working conditions, and reducing problems such as uneven filling, residual bubbles, surface height differences, and color differences from the root. On the other hand, the filling effect is quantified by radial diffusion uniformity measurement, and the influence of each parameter on diffusion is clarified by formulas. The spiral path, extrusion speed, and extrusion pressure can be adjusted in a targeted manner to ensure that the resin is uniformly filled in the target area, improve the flatness of the wafer surface, improve process stability and yield, reduce rework and AOI inspection re-judgment pressure caused by defects, and improve production efficiency.

[0108] In some embodiments of the present invention, the mold closing process is a crucial step after extrusion, where the resin is compacted and shaped. If the mold closing speed is too fast, it will violently impact the liquid resin within the mold cavity, causing turbulence or localized accumulation of the resin under sudden pressure drops, disrupting the uniform distribution formed during extrusion. If the mold closing pressure is inappropriate, excessive pressure will cause the resin to overflow from the mold cavity, while insufficient pressure will fail to eliminate air bubbles or ensure adequate adhesion between the resin and the wafer surface, ultimately exacerbating surface unevenness and color differences. Therefore, based on the real-time characteristics of the resin and the wafer structure, the mold closing control model needs to accurately determine the mold closing parameters to ensure uniform resin curing and achieve controllable wafer surface smoothness. The specific implementation is as follows:

[0109] Step S501: After the extrusion process is completed, real-time data such as viscosity, temperature and flow rate of liquid resin are collected using online viscometers, temperature sensors and other equipment; the above data can reflect the actual state changes of liquid resin during the extrusion process.

[0110] Step S502: Retrieve wafer size and structural feature data from the wafer inspection system or production database, including wafer diameter, thickness, chip layout, wiring layers, and other information; the above data has been acquired and stored during the wafer manufacturing process through optical measurement equipment or other inspection methods;

[0111] Step S503: Initialize the mold closing control model according to the wafer size and structural characteristics and the initial properties of the liquid resin, and define the geometry, boundary conditions and initial state of the mold cavity; for example, set the length, width and height of the mold cavity in the mold closing control model, define the position of the mold cavity inlet and outlet, and the viscosity, temperature and flow rate of the liquid resin at the initial moment of mold closing.

[0112] Step S504: The mold closing control model uses numerical simulation algorithms to simulate the flow front position, pressure change curve and filling time of liquid resin under different mold closing speeds and pressures based on the real-time characteristics of the input liquid resin and the wafer size and structure features. The optimal parameter combination that can meet the filling quality requirements is selected, namely the optimal mold closing speed and mold closing pressure, to ensure that the liquid resin can be evenly distributed, fully filled and without defects during the mold closing process.

[0113] Step S505: Based on the mold closing speed and mold closing pressure output by the mold closing control model, set the operating parameters of the mold closing equipment, such as the hydraulic pressure and mechanical transmission speed of the mold closing machine;

[0114] Step S506: Start the mold closing equipment and close the mold according to the set mold closing speed and pressure, so that the liquid resin is evenly distributed and solidified in the mold cavity; during the mold closing process, monitor the changes in mold closing pressure and speed in real time to ensure that the mold closing process is carried out according to the predetermined process parameters; for example, the pressure and speed during the mold closing process are monitored in real time by pressure sensors and displacement sensors, and feedback control and adjustment are performed in a timely manner if there is any deviation.

[0115] In this embodiment, by accurately calculating the mold closing speed and pressure, it is possible to ensure that the liquid resin is evenly distributed in the mold cavity, reducing defects such as internal stress concentration, surface unevenness, and residual bubbles caused by improper mold closing parameters. Based on the precise control of the real-time characteristics of the liquid resin and the wafer size and structure features, process fluctuations caused by material properties and size differences are reduced, improving the stability and consistency of the mold closing process. For example, when switching between different batches of wafers or different types of liquid resin, it is only necessary to re-collect relevant data and input it into the model to quickly and accurately determine the appropriate mold closing process parameters, ensuring the stable operation of the process.

[0116] In some embodiments of the present invention, the mold closing control model is a numerical simulation tool for accurately calculating the mold closing speed and mold closing pressure; based on the real-time characteristics of liquid resin and the wafer size and structure features, it simulates the flow state of resin under different mold closing parameters, and outputs the optimal mold closing speed and pressure that can ensure uniform curing of resin without defects, thereby realizing closed-loop control from parameter calculation, simulation verification to process execution.

[0117] The mold closing control model is built based on finite element analysis and includes a geometric modeling module, a physical parameter module, and a numerical solution module. The geometric modeling module generates a three-dimensional geometric model consistent with the actual mold cavity based on the wafer size and structural characteristics, accurately restoring the length, width, height, internal steps, wiring, and other details of the mold cavity. The physical parameter module stores the physical properties of the resin and the mechanical parameters of the mold. The numerical solution module incorporates a fluid dynamics and curing dynamics coupled algorithm to simulate the flow and curing behavior of the resin during the mold closing process.

[0118] Based on the mold cavity geometry parameters, data is retrieved from the wafer inspection system to define the mold cavity boundary and internal structure in the model; the initial state of the resin is set by inputting real-time viscosity, temperature, flow rate and other data collected after extrusion into the model as resin state parameters at the initial moment of mold closing; constraints are defined by setting the rigid boundary of the mold and the curing threshold of the resin.

[0119] The mold closing control model simulates resin flow behavior through the following logic: Flow simulation is based on the constitutive equation of non-Newtonian fluids to calculate the position of the resin flow front and pressure distribution under different mold closing speeds, and to determine whether turbulence or accumulation exists; Curing simulation combines the resin curing kinetic equation to calculate the resin curing rate under different mold closing pressures, avoiding premature curing due to excessively high pressure or incomplete curing due to excessively low pressure.

[0120] The mold closing control model simulates multiple sets of parameters and selects mold closing process parameters that meet the following conditions: the resin flow front advances uniformly without local stagnation; the pressure distribution in the mold cavity is stable without sudden rises or falls; and the degree of curing reaches the preset value at the end of mold closing.

[0121] In this embodiment, a multi-module collaborative system is constructed through finite element analysis. Combining the real-time characteristics of the resin with the features of the wafer structure, the resin flow and curing state under different mold closing parameters are accurately simulated. The optimal mold closing speed and pressure are selected, and closed-loop control from parameter calculation to process execution is achieved. This avoids problems such as turbulence, accumulation, overflow, and bubbles caused by excessively fast mold closing speed or improper pressure, ensuring uniform resin curing, effectively improving wafer surface unevenness and color difference, and enhancing product quality stability. At the same time, the model can be adapted to different resin batches and wafer types, reducing process fluctuations, facilitating repeated implementation, and helping to improve production efficiency and yield.

[0122] like Figure 2 As shown, the present invention also provides a process optimization system for liquid resin wafer encapsulation, specifically including the following modules;

[0123] The feature acquisition module is used to acquire the initial properties of the liquid resin, the initial state of the release membrane, and the wafer size and structural features.

[0124] The preheating control module is used to determine the preheating control parameters of the release membrane based on the initial characteristics of the liquid resin and the initial state of the release membrane, and to control the execution of the preheating process accordingly.

[0125] The target area identification module is used to identify the features of the target area of ​​the release membrane after the preheating process, and to obtain the surface features of the release membrane and the size features of the target area.

[0126] The extrusion parameter determination module is used to perform flow simulation on the surface characteristics of the release membrane, the size characteristics of the target area, and the initial characteristics of the liquid resin, and determine the extrusion parameters based on the flow simulation results.

[0127] The extrusion execution module is used to execute the extrusion process according to the extrusion parameters;

[0128] The mold closing parameter determination module is used to obtain the real-time characteristics of the liquid resin after the extrusion process, and input them along with the wafer size and structural features into the molding and mold closing control model to obtain the mold closing process parameters.

[0129] The mold closing execution module is used to execute the mold closing process according to the mold closing process parameters.

[0130] In this embodiment, a series logic of a feature acquisition module, a preheating control module, a target area identification module, an extrusion parameter determination module, and a mold closing parameter determination module enables seamless integration from initial feature acquisition to dynamic adjustment of parameters for each process. Unlike traditional fixed-parameter systems, this system can capture real-time differences in liquid resin characteristics, release film state, and wafer structure, and convert these differences into suitable process parameters, fundamentally avoiding uneven filling caused by applying parameters indiscriminately. The target area identification module specifically identifies the surface features of the preheated release film and the size of the extrusion area, reducing extrusion deviations caused by unknown surface conditions. The extrusion parameter determination module generates suitable parameters through flow simulation, solving problems such as bubbles and incomplete filling caused by the sensitivity of liquid resin to flow characteristics. The mold closing parameter determination module adjusts the mold closing speed and pressure based on the real-time characteristics of the resin after extrusion, specifically improving surface unevenness and color difference defects, indirectly improving AOI inspection yield.

[0131] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A process optimization method for liquid resin wafer encapsulation, characterized in that, include: Acquire the initial properties of the liquid resin, the initial state of the release membrane, and the wafer-scale structural features; Based on the initial characteristics of the liquid resin and the initial state of the release membrane, preheating control parameters for the release membrane are determined, and the preheating process is carried out accordingly. After the preheating process is completed, feature recognition is performed on the target area of ​​the release membrane to obtain the surface features of the release membrane and the size features of the target area; The surface features of the release membrane, the size features of the target area, and the initial properties of the liquid resin are simulated for flowability. Based on the flowability simulation results, the extrusion parameters are determined, and the extrusion process is carried out accordingly. The extrusion parameters include the spiral path, extrusion speed, and extrusion pressure. After the extrusion process is completed, the real-time characteristics of the liquid resin are acquired and input into the molding and mold-closing control model along with the wafer size and structural features to obtain the mold-closing process parameters, and the mold-closing process is performed accordingly; the mold-closing process parameters include the mold-closing speed and the mold-closing pressure; The preheating control parameters for the release membrane include preheating temperature and preheating time; The preheating temperature is determined based on the viscosity-temperature curve of the initial properties of the liquid resin and the thermal conductivity of the release membrane in its initial state. The preheating time is determined based on the target area size characteristics of the release membrane and the expected flowability improvement time of the liquid resin at the preheating temperature, so that the target area of ​​the release membrane can reach a temperature state that is compatible with the initial characteristics of the liquid resin after preheating. The formula for calculating the preheating temperature is: ； in, Indicates the preheating temperature; Indicates the resin viscous flow activation energy; Represents the gas constant; Indicates the target viscosity; This represents the pre-exponential factor, a benchmark value used to quantify how material properties change with temperature; Indicates ambient temperature.

2. The process optimization method for liquid resin wafer encapsulation according to claim 1, characterized in that, The formula for calculating the preheating time is: ； in, Indicates the preheating time; This represents the density of the release membrane, determined from its initial state. This indicates the specific heat capacity of the release membrane, determined from the initial state of the release membrane. The target region's feature size is represented by the equivalent diameter of the target region, calculated from the target region's size feature. This represents the thermal conductivity coefficient of the release membrane, determined from the initial state of the release membrane; Indicates the initial temperature difference; This indicates the final temperature difference.

3. The process optimization method for liquid resin wafer encapsulation according to claim 2, characterized in that, The surface characteristics of the release membrane include surface roughness, surface smoothness, and surface temperature.

4. The process optimization method for liquid resin wafer encapsulation according to claim 1, characterized in that, The formula for the liquidity simulation is: ； Where U represents the radial diffusion uniformity of the resin during the annular extrusion process; k r This represents the path correction factor; P represents the extrusion pressure. Indicates the extrusion speed; d represents the maximum radius of the spiral trajectory; Indicates the dynamic viscosity of the resin; R a Indicates the surface roughness of the release membrane; represents the roughness influence coefficient; s represents the helical pitch; r represents the current radial position.

5. The process optimization method for liquid resin wafer encapsulation according to claim 1, characterized in that, The mold control model is constructed based on finite element analysis and includes a geometric modeling module, a physical parameter module, and a numerical solution module. The geometric modeling module is used to generate a three-dimensional geometric model that is consistent with the actual mold cavity based on the wafer size and structural features; The physical parameters module is used to store the physical properties of the resin and the mechanical parameters of the mold; The numerical solution module incorporates a coupled fluid dynamics and curing dynamics algorithm to simulate the flow and curing behavior of resin during the mold closing process.

6. A process optimization system for liquid resin wafer encapsulation, wherein the system is applied to the process optimization method for liquid resin wafer encapsulation as described in claim 1, characterized in that, include: The feature acquisition module is used to acquire the initial properties of the liquid resin, the initial state of the release membrane, and the wafer size and structural features. The preheating control module is used to determine the preheating control parameters of the release membrane based on the initial characteristics of the liquid resin and the initial state of the release membrane, and to control the execution of the preheating process accordingly. The target area identification module is used to identify the features of the target area of ​​the release membrane after the preheating process, and to obtain the surface features of the release membrane and the size features of the target area. The extrusion parameter determination module is used to perform flow simulation on the surface characteristics of the release membrane, the size characteristics of the target area, and the initial characteristics of the liquid resin, and determine the extrusion parameters based on the flow simulation results. The extrusion execution module is used to execute the extrusion process according to the extrusion parameters; The mold closing parameter determination module is used to obtain the real-time characteristics of the liquid resin after the extrusion process, and input them along with the wafer size and structural features into the molding and mold closing control model to obtain the mold closing process parameters. The mold closing execution module is used to execute the mold closing process according to the mold closing process parameters.

Images