Full-lamination packaging process of COB (Chip On Board) integrated packaging LED (Light Emitting Diode) display screen and dispensing equipment thereof

Abstract

The invention relates to the technical field of display screen packaging, in particular to a full-lamination packaging process and dispensing equipment for a COB integrated packaging LED display screen, and the full-lamination packaging process comprises the steps: placing a COB substrate after die bonding into a cavity, and obtaining three-dimensional topological data and environmental parameters of a chip array; executing multi-stage gradient air exhaust and applying dual-frequency ultrasonic pretreatment, activating surface energy and removing adsorbed gas; in a vacuum state, dispensing is carried out based on three-dimensional data, a colloid volume expansion rate and a bubble nucleation critical pressure difference threshold value are calculated in real time by utilizing a dynamic rheological pressure coupling model, dispensing pressure and speed are dynamically and reversely compensated, and it is ensured that colloid tensile stress is lower than a critical value; the in-situ optical detection unit is used for scanning and recognizing micron-sized bubbles or missing in real time, and the Z-axis height and the glue output amount of a subsequent dispensing path are fed back and corrected; and finally, a nonlinear variable-voltage oscillation program is executed to induce micro-bubble resonance fracture and solidification, so that the bubble defect in the packaging process is effectively eliminated, and the product yield is improved.

Description

Technical Field

[0001] This invention relates to the field of display screen packaging technology, and in particular to the full lamination packaging process and dispensing equipment for COB integrated packaged LED displays. Background Technology

[0002] COB (Chip-on-Board) technology directly connects LED chips to the circuit via wire bonding or flip-chip technology, and finally encapsulates the entire chip with potting compound for protection. This integrated packaging method eliminates the need for separate packaging of individual LED chips, achieving true chip-on-board (COB) packaging.

[0003] With the development of display technology, COB (Chip-on-Board) integrated packaging technology has been widely used in the field of micro-pitch LED displays due to its advantages such as high protection, high heat dissipation efficiency, and grain-free display effect. However, in the COB packaging process, the dispensing step is a critical step that determines the product yield. Existing dispensing processes mainly have the following problems: When dispensing adhesive in a vacuum environment, the adhesive will expand in volume instantly when it enters the low-pressure chamber from the high-pressure nozzle. If not properly controlled, it can easily lead to the precipitation of dissolved gas, forming micron-sized bubbles, or residual gas in the gap between chips that cannot be discharged, affecting optical performance and reliability.

[0004] Because the chip array after die bonding has microscopic height deviations, traditional fixed-height dispensing paths can easily lead to insufficient or excessive adhesive in some areas, and lack a real-time feedback correction mechanism. Summary of the Invention

[0005] The purpose of this invention is to provide a fully laminated packaging process and dispensing equipment for COB integrated packaged LED displays, which can dynamically compensate dispensing parameters, effectively suppress bubble nucleation, and perform closed-loop correction of filling defects to improve packaging quality.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a fully laminated packaging process for a COB integrated packaged LED display, including placing a die-bonded COB substrate into a packaging chamber and obtaining three-dimensional topological data of the chip array and initial environmental parameters, wherein the initial environmental parameters include ambient temperature data and pressure data. A multi-stage gradient evacuation procedure is executed to reduce the chamber pressure to a first vacuum level. At the same time, ultrasonic pretreatment at a specific frequency is applied to activate the chip surface energy and remove the adsorbed gas layer. The specific frequency specifically includes a first frequency and a second frequency applied sequentially or alternately. The first frequency is used to excite the microstructure resonance of the chip surface and the dam sidewall, and the second frequency is used to induce molecular polarization on the colloidal material surface. Dispensing is performed based on three-dimensional topology data while maintaining a second vacuum level. The volume expansion rate of the colloid at the moment of exiting the nozzle is calculated in real time according to the dynamic rheological pressure coupling model. The dispensing pressure and dispensing speed are dynamically compensated in reverse, so that the actual tensile stress inside the colloid is always lower than the critical pressure difference threshold for bubble nucleation. The dynamic rheological pressure coupling model establishes the mapping relationship between colloid viscosity, surface tension, ambient temperature, pressure data and volume expansion rate, and calculates the critical pressure difference threshold for bubble nucleation based on the volume expansion rate. During the dispensing process, an in-situ optical detection unit integrated within the encapsulation chamber is used to scan the coated area in real time through a pressure-resistant optical window to identify micron-sized bubbles or missing fillers. If an abnormality is detected, dispensing is paused and the Z-axis height and dispensing volume of the subsequent dispensing path are corrected based on the feedback data. After dispensing, a nonlinear pressure oscillation program is executed, which involves high-frequency, small-amplitude pressure fluctuations between vacuum and atmospheric pressure to induce the resonance and rupture of residual microbubbles. Subsequently, rapid pressure is applied to compact the material and complete the curing process.

[0007] The specific steps for placing the die-bonded COB substrate into the packaging chamber and obtaining the three-dimensional topology data and initial environmental parameters of the chip array include: By using an online 3D topography measuring instrument integrated into the packaging chamber, the chip array is automatically scanned non-contactly after the COB substrate is placed into the chamber, obtaining the precise coordinates and height deviation data of each chip, and generating a 3D surface profile map including the chip gap depth.

[0008] The specific steps of executing a multi-stage gradient evacuation procedure to reduce the chamber pressure to a first vacuum level, while simultaneously applying ultrasonic pretreatment at a specific frequency to activate the chip surface energy and remove the adsorbed gas layer include: Start the vacuum pump to perform a rough pumping of the encapsulation chamber, reducing the chamber pressure from atmospheric pressure to an intermediate transition pressure at the first pumping rate. After reaching the intermediate transition pressure, multi-stage gradient pumping is performed according to the preset step pressure reduction curve. Each step is kept at a constant pressure for a certain time, so that the gas molecules in the chamber are gradually released until the chamber pressure drops to the first vacuum level. After reaching the first vacuum level, the ultrasonic generator is activated to apply ultrasonic waves of a specific frequency to the surface of the COB substrate. Under the continuous action of the ultrasonic waves, the gas molecules physically adsorbed on the chip surface and in the micro gaps are desorbed and released. At the same time, the vacuum pump is kept in operation to continuously extract the desorbed gas from the chamber.

[0009] The specific steps of establishing the mapping relationship between colloidal viscosity, surface tension, ambient temperature, absolute pressure, and volume expansion rate in the dynamic rheological pressure coupling model, and calculating the critical pressure difference threshold for bubble nucleation based on the volume expansion rate, include: The rheological curves of the encapsulating colloid at different temperatures and shear rates were pre-determined, and the viscosity of the colloid as a function of temperature η(T) was fitted. The surface tension coefficient of the colloid within the process temperature range was also determined, and the surface tension as a function of temperature γ(T) was established. Based on colloidal viscosity and surface tension, a dynamic coupling model of volume expansion rate and pressure difference is established for the instant when the colloid enters the vacuum chamber from the high-pressure zone of the nozzle. During the dispensing process, ambient temperature data is collected in real time to calculate the corresponding colloid viscosity and surface tension. Based on the pressure difference between ambient pressure data and nozzle pressure, the dynamic coupling model is substituted to calculate the actual volume expansion rate of the colloid at the moment it leaves the nozzle. Based on the calculated actual volume expansion rate and the dissolved gas content in the colloid, the critical pressure difference threshold for the formation of bubbles under the current operating conditions is calculated using the bubble nucleation theory formula.

[0010] The specific steps of dispensing adhesive based on three-dimensional topological data while maintaining a second vacuum level, calculating the volume expansion rate of the adhesive at the instant of exiting the nozzle in real time according to the dynamic rheological pressure coupling model, and dynamically compensating for dispensing pressure and dispensing speed to ensure that the actual tensile stress inside the adhesive is always lower than the critical pressure difference threshold for bubble nucleation include: Based on the three-dimensional topology data of the chip array, a variable height dispensing path is generated, and the basic dispensing pressure and basic dispensing speed are preset. As the dispensing head moves along the preset path, the theoretical volume expansion rate of the colloid at the instant it leaves the nozzle is collected and calculated in real time at a millisecond sampling frequency. The calculated theoretical volume expansion rate is compared with the preset maximum allowable volume expansion rate. If the theoretical volume expansion rate exceeds the maximum volume expansion rate, the compensated dispensing pressure and the compensated dispensing speed are calculated in real time according to the preset reverse compensation algorithm to reduce the pressure difference when the adhesive flows out of the nozzle. The compensated dispensing pressure and dispensing speed are output to the dispensing actuator in real time. At the same time, the actual tensile stress is substituted into the model again to verify whether it is lower than the critical tensile stress corresponding to the critical pressure difference threshold P for bubble nucleation. If the verification passes, the dispensing continues. If the verification fails, the compensation parameters are further adjusted.

[0011] During the dispensing process, an in-situ optical detection unit integrated within the encapsulation cavity is used to scan the coated area in real time through a pressure-resistant optical window to identify micron-sized bubbles or filler defects. If an anomaly is detected, dispensing is paused, and the Z-axis height and dispensing volume of subsequent dispensing paths are corrected based on feedback data. The specific steps include: While the dispensing head moves along a preset path to apply the coating, the in-situ optical detection unit integrated on the top of the encapsulation chamber is activated. The detection unit performs line scanning on the coated area through a pressure-resistant optical window at a speed synchronized with the movement of the dispensing head to acquire high-resolution digital images. Based on the high-resolution digital image, the image is analyzed in real time using a preset defect identification algorithm; the algorithm can distinguish and mark micron-sized bubbles, chip sidewall filling defects, and uneven adhesive layer thickness. When the size or number of any defect exceeds the preset process tolerance threshold, the system immediately generates a pause command, interrupts the movement of the dispensing head and the dispensing action, and records the precise coordinates of the current abnormal point and its corresponding defect type and severity. The system calls the 3D topology data of the area near the abnormal point. If it is determined that the Z-axis height is set too high, resulting in insufficient filling, the Z-axis height adjustment ΔZ is generated for the subsequent areas with the same or similar topology. If it is determined that the glue output is insufficient, resulting in air bubbles, the glue output increase coefficient K is generated for the corresponding area.

[0012] Secondly, the present invention provides a dispensing device for a COB integrated packaged LED display screen, comprising a packaging chamber, a three-dimensional topology data acquisition module, a vacuum module, a dispensing module, and an optical detection unit; the packaging chamber is used to place the COB substrate; the three-dimensional topology data acquisition module is used to acquire the three-dimensional topology data of the COB substrate; the vacuum module is used to evacuate the packaging chamber; the dispensing module is used to dispense adhesive based on the three-dimensional topology data; and the optical detection module is used to scan the coated area in real time through a pressure-resistant optical window to identify micron-sized bubbles or filler defects.

[0013] The encapsulation chamber includes a base, a chamber body, and a feeding structure. The chamber body has a feed inlet and is fixed on the base. The feeding structure is located on one side of the feed inlet.

[0014] The feeding structure includes a door panel, a support platform, a connecting rod, a sealing strip, a control cylinder, and a connecting block. The door panel is rotatably mounted on the chamber body, the support platform is slidably mounted inside the chamber body, the connecting rod is rotatably connected to the door panel and the support platform and is located between the door panel and the support platform, the sealing strip is fixed on the door panel and is located on one side close to the chamber body, the control cylinder is rotatably mounted on one side of the chamber body, and the connecting block is rotatably connected to the control cylinder and slidably connected to the door panel.

[0015] The dispensing module includes a lifting structure, a support plate, a moving screw, a moving block, a moving motor, a longitudinal moving structure, and a dispensing head. The moving block is slidably disposed on the top of the chamber body. The moving screw is threadedly connected to the moving block. The output end of the moving motor is connected to the moving screw. The longitudinal moving structure is disposed on the moving block. The lifting structure is disposed on the longitudinal moving structure. The dispensing head is disposed on the lifting structure.

[0016] The advantages of the COB integrated packaged LED display full lamination packaging process and its dispensing equipment of the present invention are as follows: By removing adsorbed gases through ultrasonic pretreatment and compensating dispensing parameters in real time using a dynamic rheological pressure coupling model, bubble nucleation conditions are eliminated from both the source and the process, significantly reducing the generation rate of micron-sized bubbles.

[0017] The variable height dispensing path based on 3D topology data, combined with the real-time feedback correction mechanism of in-situ optical detection, effectively solves the problem of uneven filling caused by chip height deviation.

[0018] It achieves fully automated closed-loop control from environmental perception and model calculation to execution compensation, reducing reliance on human experience and improving process stability. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0020] Figure 1 This is a flowchart of the full lamination packaging process of the COB integrated packaged LED display screen according to the first embodiment of the present invention.

[0021] Figure 2 This is a flowchart of the first embodiment of the present invention, which describes the execution of a multi-stage gradient pumping procedure to reduce the chamber pressure to a first vacuum level, while simultaneously applying ultrasonic pretreatment at a specific frequency to activate the chip surface energy and remove the adsorbed gas layer.

[0022] Figure 3 This is a flowchart of the first embodiment of the present invention, which establishes the mapping relationship between colloidal viscosity, surface tension, ambient temperature, absolute pressure and volume expansion rate using a dynamic rheological pressure coupling model, and calculates the critical pressure difference threshold for bubble nucleation based on the volume expansion rate.

[0023] Figure 4This is a flowchart of the first embodiment of the present invention, which describes dispensing based on three-dimensional topology data while maintaining a second vacuum level. The flow chart calculates the volume expansion rate of the colloid at the moment of exiting the nozzle in real time according to the dynamic rheological pressure coupling model, and dynamically compensates for the dispensing pressure and dispensing speed in reverse, so that the actual tensile stress inside the colloid is always lower than the critical pressure difference threshold for bubble nucleation.

[0024] Figure 5 The first embodiment of the present invention describes a process in which an in-situ optical detection unit integrated within the encapsulation chamber is used to scan the coated area in real time through a pressure-resistant optical window during the dispensing process to identify micron-sized bubbles or missing fillers; if an abnormality is detected, dispensing is paused and the Z-axis height and dispensing volume of the subsequent dispensing path are corrected based on feedback data.

[0025] Figure 6 This is a structural diagram of the dispensing equipment for a COB integrated packaged LED display screen according to the second embodiment of the present invention. Figure 7 This is a cross-sectional structural diagram of the dispensing equipment for a COB integrated packaged LED display screen according to the second embodiment of the present invention.

[0026] In the figure: encapsulation chamber 101, three-dimensional topology data acquisition module 102, vacuum module 103, optical detection unit 105, door panel 106, support platform 107, connecting rod 108, sealing strip 109, control cylinder 110, connecting block 111, lifting structure 112, support plate 113, moving screw 114, moving block 115, moving motor 116, longitudinal moving structure 117, dispensing head 118. Detailed Implementation

[0027] The embodiments of the present invention are described in detail below. Examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, but should not be construed as limiting the present invention.

[0028] First embodiment: Please see Figures 1-5 This invention provides a fully laminated packaging process for COB integrated packaged LED displays, comprising: S101 places the die-bonded COB substrate into the packaging chamber and obtains the three-dimensional topology data of the chip array and the initial environmental parameters, including ambient temperature data and pressure data. The specific steps include: By using an online 3D topography measuring instrument integrated into the packaging chamber, the chip array is automatically scanned non-contactly after the COB substrate is placed into the chamber, obtaining the precise coordinates and height deviation data of each chip, and generating a 3D surface profile map including the chip gap depth.

[0029] After the COB substrate is placed into the packaging chamber and physically positioned, multiple sets of high-sensitivity temperature and pressure sensors distributed inside the chamber are immediately activated to monitor the microenvironment within the chamber at a preset high sampling frequency. The system continuously acquires and records the current ambient temperature data (resolution up to ±0.1℃) and ambient pressure data (resolution up to ±10Pa). These two types of data not only serve as a benchmark for compensating for environmental fluctuations in subsequent processes, but also correct for minor errors in optical measurements caused by changes in the air refractive index, ensuring the traceability of the measurement data.

[0030] Once environmental parameters stabilize, an online 3D topography measuring instrument integrated into the top or sidewall of the packaging chamber is triggered. This instrument employs advanced non-contact optical measurement technology (such as high-resolution structured grating projection or precision laser triangulation) and performs fully automated, high-density, high-speed scanning of the chip array on the COB substrate based on a preset path planning algorithm. The scanning process precisely captures the spatial position information of each chip by projecting structured light spots or laser beams onto the chip surface and receiving reflected light signals. By using image recognition and centroid algorithms, the precise two-dimensional coordinate data (X, Y values) of each chip in the substrate coordinate system is analyzed, with an accuracy down to the micrometer level, to identify chip positional offsets. The relative height value of each chip surface is measured and compared with a preset substrate reference plane (or ideal plane) to accurately calculate the height deviation data (Z value deviation, i.e., coplanarity error) of each chip relative to the reference plane, thereby quantifying the degree of chip warpage or tilt.

[0031] The built-in software of the 3D topography measuring instrument or the connected host computer system processes the raw point cloud data acquired by scanning in real time, including filtering and noise reduction, point cloud registration, and surface fitting. Finally, a surface contour map containing the chip gap depth and 3D structure is generated. This digitized contour map not only visually displays the coplanar distribution of the chip array but also automatically extracts the following key feature parameters through algorithms: The chip gap depth matrix accurately depicts the depth and width of the gaps between chips and between the chip edge and the substrate, forming a "groove depth distribution map" used to calculate the amount of adhesive applied.

[0032] Clearly depict surface undulations caused by uneven die-attach thickness or unevenness of the substrate itself.

[0033] Ultimately, all the collected 3D topology data (chip coordinate matrix, height deviation dataset, gap depth field) and initial environmental parameters (temperature, pressure) are integrated into a complete digital substrate feature information package. This information package is transmitted in real time to the central process control system, serving as the core decision-making basis for accurately calculating adhesive volume compensation, planning the dispensing path, and adjusting the nozzle height in the subsequent step S102 (intelligent dispensing), thereby achieving true "on-demand allocation" and adaptive control of the full bonding process.

[0034] S102 executes a multi-stage gradient pumping procedure to reduce the chamber pressure to the first vacuum level, while applying ultrasonic pretreatment at a specific frequency to activate the chip surface energy and remove the adsorbed gas layer. The specific steps include: S201 starts the vacuum pump to perform rough pumping of the encapsulation chamber, reducing the chamber pressure from atmospheric pressure to intermediate transition pressure at the first pumping rate. The high-performance vacuum pump unit connected to the encapsulation chamber is started, and the system enters the rough pumping stage. The control unit rapidly vents the chamber at a set first pumping rate (i.e., a relatively high pumping speed), smoothly reducing the internal pressure from standard atmospheric pressure (approximately 101 kPa) to a preset intermediate transition pressure (e.g., the range of 1 kPa to 10 kPa). This stage aims to quickly remove most of the free air molecules within the chamber, establishing a stable directional gas flow field, laying the foundation for subsequent fine gradient depressurization.

[0035] After reaching the intermediate transition pressure, S202 performs multi-stage gradient pumping according to the preset step pressure reduction curve. Each step maintains constant pressure for a certain time, allowing gas molecules in the chamber to be released gradually until the chamber pressure drops to the first vacuum level. Once the chamber pressure reaches the intermediate transition pressure threshold, the control system automatically switches to precision depressurization mode. The system executes a multi-stage gradient evacuation procedure based on a built-in stepped depressurization curve algorithm. Specifically, the pressure does not decrease linearly, but rather decreases step-by-step according to a set sequence. At each pressure step, the system maintains a preset constant pressure for a set period by adjusting the opening of the vacuum valve or the operating mode of the pump unit. This intermittent constant pressure maintenance mechanism aims to utilize the pressure difference to drive the gas molecules adsorbed deep within the substrate's microstructure (such as chip gaps and dammed grooves) to gradually diffuse and escape into the chamber's free space, thereby avoiding the "lock-in effect" or localized turbulence caused by excessively rapid pressure reduction. After several stages of continuous evacuation, the chamber pressure is ultimately precisely controlled at the first vacuum level (e.g., within the range of 1 Pa to 10 Pa), achieving a gentle yet thorough degassing effect.

[0036] After reaching the first vacuum level, S203 activates the ultrasonic generator to apply ultrasonic waves of a specific frequency to the surface of the COB substrate. The specific frequency specifically includes a first frequency and a second frequency applied sequentially or alternately. The first frequency is used to excite the microstructure resonance of the chip surface and the dam sidewalls, and the second frequency f2 is used to induce molecular polarization on the surface of the colloidal material. Under the continuous action of the ultrasonic waves, the gas molecules physically adsorbed on the chip surface and in the micro gaps are desorbed and released. At the same time, the vacuum pump is kept in working state to continuously extract the desorbed gas from the chamber.

[0037] After the chamber pressure stabilizes and reaches the first vacuum level, the system activates the ultrasonic generator integrated with the substrate stage to directionally apply ultrasonic energy of a specific frequency to the COB substrate surface and the chip array area. This process does not use a single frequency, but rather applies two different frequencies of ultrasonic waves sequentially or alternately according to the optimized process formulation to achieve a synergistic effect. First, a first frequency f1 (e.g., in the low-frequency range) is applied. This frequency is precisely tuned to excite elastic resonance in the microstructures of the chip (such as electrode pad edges, passivation layer interfaces) and the microstructures of the dam sidewalls. This micro-amplitude mechanical vibration physically loosens gas molecule clusters that have been adsorbed on the surface and in micro-gaps for a long time, breaking their van der Waals bond with the solid surface and causing them to transition from an adsorbed state to a free state. Based on the first frequency, a second frequency f2 (e.g., in the mid-to-high-frequency range) is applied. This frequency mainly acts at the molecular level of the material surface, inducing polarization effects or surface charge rearrangement in the molecules of the chip surface material (such as gallium nitride, sapphire, or silicon) and the dam material through high-frequency oscillation, thereby significantly increasing the surface free energy. Higher surface energy means that the adhesive can achieve a smaller contact angle in subsequent steps, resulting in better wetting and spreading.

[0038] Throughout the ultrasonic pretreatment process, the vacuum pump unit operates continuously to maintain a low-pressure environment within the chamber. Gas molecules desorbed by physical vibration at the first frequency f1, along with the gas layer originally adsorbed on the surface, are rapidly extracted from the chamber to prevent re-adhesion. After this step, the COB substrate surface is not only in a highly clean vacuum state but also possesses highly active surface energy characteristics, providing optimal interface conditions for the subsequent uniform coating, deep filling, and bubble-free bonding of the bonding adhesive.

[0039] S103 performs dispensing based on three-dimensional topology data while maintaining a second vacuum level. It calculates the volume expansion rate of the colloid at the moment of exiting the nozzle in real time according to the dynamic rheological pressure coupling model, and dynamically compensates for the dispensing pressure and dispensing speed in reverse, so that the actual tensile stress inside the colloid is always lower than the critical pressure difference threshold for bubble nucleation. The dynamic rheological pressure coupling model establishes the mapping relationship between colloid viscosity, surface tension, ambient temperature, pressure data and volume expansion rate, and calculates the critical pressure difference threshold for bubble nucleation based on the volume expansion rate. The dynamic rheological pressure coupling model establishes a mapping relationship between colloidal viscosity, surface tension, ambient temperature, absolute pressure, and volume expansion rate. The specific steps for calculating the critical pressure difference threshold for bubble nucleation based on the volume expansion rate include: S301 pre-determines the rheological curves of the encapsulating colloid at different temperatures and shear rates, fits the functional relationship η(T) of the colloid viscosity with respect to temperature; and determines the surface tension coefficient of the colloid within the process temperature range, establishing the surface tension as a function of temperature γ(T). Through preliminary rheological experiments, the rheological curves of the selected encapsulating colloid were pre-determined at different temperatures and shear rates. Based on the experimental data, a mathematical fitting method was used to establish the functional relationship η(T) of the colloid viscosity with respect to temperature. This function can describe the viscosity variation with temperature within the process temperature range. Simultaneously, the surface tension coefficient of the colloid was determined using the pendant drop method or the Williams plate method within the process temperature range, establishing the functional relationship γ(T) of surface tension with respect to temperature. These two functions constitute the basic physical property inputs of the model.

[0040] Based on colloidal viscosity and surface tension, S302 establishes a dynamic coupling model of the volume expansion rate and pressure difference of the colloid as it enters the vacuum chamber from the high-pressure zone of the nozzle. Based on the principles of fluid mechanics and thermodynamics, and combining the viscosity characteristics η(T) and surface tension characteristics γ(T) of colloids, a mathematical model is established to describe the volume expansion behavior of colloids as they instantaneously enter the low-pressure zone (second vacuum environment) of the chamber from the high-pressure zone inside the nozzle (dispensing pressure environment). The core of this model is to describe the dynamic coupling relationship between the volume expansion rate ε caused by the internal and external pressure difference ΔP at the moment of dispensing and the rheological properties of the colloid. That is, the higher the viscosity and the greater the surface tension of the colloid, the stronger its resistance to instantaneous pressure difference and the smaller its volume expansion rate; conversely, the lower the viscosity and surface tension, the smaller the volume expansion rate. This coupled model provides a theoretical framework for subsequent real-time calculations.

[0041] S303 collects ambient temperature data in real time during the dispensing process to calculate the corresponding colloid viscosity and surface tension. Based on the pressure difference between ambient pressure data and nozzle pressure, it substitutes the data into the dynamic coupling model to calculate the actual volume expansion rate of the colloid at the moment it leaves the nozzle. During the dispensing process, the system collects real-time ambient temperature data T and ambient pressure data P2 (i.e., the second vacuum value) within the encapsulation chamber, and simultaneously acquires the real-time dispensing pressure P1 inside the nozzle. The collected temperature data T is substituted into the functional relationships η(T) and γ(T) established in sub-step S301 to calculate the real-time viscosity and surface tension of the colloid under the current operating conditions. Subsequently, these real-time physical property parameters, along with the pressure difference ΔP (ΔP=P1-P2), are substituted into the dynamic coupling model established in sub-step S302 to calculate the actual volume expansion rate ε experienced by the colloid at the instant it leaves the nozzle. real .

[0042] Based on the calculated actual volume expansion rate and the dissolved gas content in the colloid, S304 calculates the critical pressure difference threshold for the formation of bubbles under the current operating conditions using the bubble nucleation theory formula.

[0043] Based on the calculated actual volume expansion rate ε real By combining the inherent physicochemical properties of colloidal materials (such as dissolved gas concentration and saturated solubility), and introducing the bubble nucleation theory (including homogeneous and heterogeneous nucleation mechanisms), a critical pressure difference threshold ΔP for bubble nucleation is established. critical The calculation formula is used to determine the expansion rate ε at the current expansion rate. real And the minimum pressure difference required for the aggregation of molecular particles in a colloid to form stable bubble nuclei at temperature T. The calculated ΔP critical As a target benchmark value for real-time control.

[0044] Ultimately, the control system will use the actual volume expansion rate ε real The corresponding internal tensile stress, and the critical pressure difference threshold ΔP calculated in sub-step S304. critical Real-time comparison is performed. If the internal tensile stress tends to exceed the critical threshold, the system immediately performs dynamic reverse compensation on the dispensing pressure P1 and / or dispensing speed (such as reducing the dispensing pressure or adjusting the screw speed), thereby keeping the actual tensile stress inside the colloid below the critical pressure difference threshold for bubble nucleation, achieving bubble-free dispensing.

[0045] The specific steps of dispensing adhesive based on three-dimensional topology data while maintaining a second vacuum level, calculating the volume expansion rate of the adhesive at the instant of exiting the nozzle in real time according to the dynamic rheological pressure coupling model, and dynamically compensating for dispensing pressure and dispensing speed to ensure that the actual tensile stress inside the adhesive is always lower than the critical pressure difference threshold for bubble nucleation include: S401 generates a variable height dispensing path based on the three-dimensional topology data of the chip array, and presets the basic dispensing pressure and basic dispensing speed. The control system first reads the 3D surface contour map and the precise coordinates and height deviation dataset of the chip array generated in step S101. Based on this topology information, the dispensing path planning algorithm automatically generates a variable-height dispensing path: that is, during the movement of the dispensing head, its Z-axis height will be dynamically adjusted according to the actual height of each chip and the depth of the chip gap, ensuring that the nozzle and the deposition surface are always kept at the optimal working distance. At the same time, the system presets the basic dispensing pressure P according to the adhesive characteristics and the target filling amount. base With the basic glue dispensing speed v base This serves as the initial benchmark value for subsequent dynamic compensation.

[0046] As the dispensing head moves along the preset path, S402 collects and calculates the theoretical volume expansion rate of the colloid at the instant it leaves the nozzle in real time at a millisecond sampling frequency. During the continuous dispensing process as the dispensing head moves along a preset path, the control system collects real-time ambient temperature data T(t), ambient pressure data P2(t) (i.e., the second vacuum value), and transient dispensing pressure P1(t) inside the nozzle at a high sampling frequency of milliseconds (e.g., 1-10 milliseconds). This real-time data stream is dynamically input into the dynamic coupling model established in step S302, instantly calculating the theoretical volume expansion rate ε of the colloid at the moment it leaves the nozzle. calc(t) The calculation process is synchronized with the physical movement of the dispensing head, enabling real-time monitoring of the expansion behavior of each dispensing droplet.

[0047] S403 compares the calculated theoretical volume expansion rate with the preset maximum allowable volume expansion rate; if the theoretical volume expansion rate exceeds the maximum volume expansion rate, it calculates the compensated dispensing pressure and the compensated dispensing speed in real time according to the preset reverse compensation algorithm, so as to reduce the pressure difference when the adhesive flows out of the nozzle. The system will calculate the theoretical volume expansion rate ε in real time. calc(t) Compared with the preset maximum allowable volume expansion rate ε max (This threshold is derived from the critical pressure difference for bubble nucleation) Real-time comparison is performed: If ε calc(t) ≤ ε max This indicates that the current operating condition is safe, and the original parameters should be maintained and the operation should continue.

[0048] If ε calc(t) > ε max This indicates that the current pressure difference is too large, and the tensile stress inside the colloid has approached or exceeded the critical value for bubble nucleation, posing a risk of bubble formation. At this point, the system immediately triggers a preset reverse compensation algorithm. This algorithm, based on the inverse function of the dynamic coupling model, aims to reduce the pressure difference and calculates the required post-compensation dispensing pressure P in real time. comp(t) With the compensated glue dispensing speed v comp(t)The core logic of compensation is to appropriately reduce the glue supply pressure P1(t) to decrease the difference ΔP between it and the chamber pressure P2(t), thereby suppressing the volume expansion rate.

[0049] S404 outputs the compensated dispensing pressure and dispensing speed to the dispensing actuator in real time, and at the same time substitutes them into the model again to verify whether the actual tensile stress is lower than the critical tensile stress corresponding to the critical pressure difference threshold P for bubble nucleation; if the verification passes, the dispensing continues; if the verification fails, the compensation parameters are further adjusted.

[0050] The system will calculate the compensated dispensing pressure P comp(t) With the compensated glue dispensing speed v comp(t) As control commands, they are output in real time to the dispensing actuator (such as a piezoelectric or screw dispensing valve) via a high-speed communication bus to achieve instantaneous adjustment of parameters.

[0051] To ensure the accuracy and reliability of the compensation, the system implements a closed-loop verification mechanism: while outputting the compensation parameters, it also verifies the compensated parameters (P). comp(t) v comp(t) Substituting the current environmental parameters back into the dynamic rheological pressure coupling model and the formula for calculating the critical pressure difference for bubble nucleation, we can verify the actual tensile stress σ inside the colloid at this point. real(t) Has it truly decreased to the critical tensile stress σ corresponding to the critical pressure difference threshold P? critical the following: If the verification passes (σ) real(t) < σ critical Once the system confirms that the compensation is effective, it continues to dispense adhesive along the preset path.

[0052] If the verification fails (σ) real(t) ≥ σ critical If the single compensation is insufficient or there are other interfering factors, the system will further adjust the compensation parameters based on sub-step S403 according to the preset iterative compensation algorithm until the verification is passed, or trigger a process abnormality alarm and suspend the operation.

[0053] During the dispensing process, S1O4 utilizes an in-situ optical detection unit integrated within the encapsulation chamber to scan the coated area in real time through a pressure-resistant optical window, identifying micron-sized bubbles or missing fillers. If an anomaly is detected, dispensing is paused, and the Z-axis height and dispensing volume of subsequent dispensing paths are corrected based on feedback data. The specific steps include: While the dispensing head moves along a preset path to apply the coating, the S501 activates the in-situ optical detection unit integrated on the top of the encapsulation chamber. This detection unit performs line scanning on the coated area through a pressure-resistant optical window at a speed synchronized with the movement of the dispensing head to acquire a high-resolution digital image. While the dispensing head moves along the preset path generated in step S401 and applies the adhesive, the system simultaneously activates the in-situ optical detection unit integrated at the top of the encapsulation chamber. This detection unit consists of a high-resolution line scan camera, a custom optical system, and a high-frequency light source. Its scanning field of view is aligned with the area just coated behind the dispensing head through a pressure-resistant optical window. Through precise synchronization by the motion control system, the detection unit continuously and comprehensively acquires high-speed images of the coated area at a line scan frequency matching the moving speed of the dispensing head, obtaining a high-definition digital image stream with micron-level resolution.

[0054] S502 performs real-time analysis of the image based on the high-resolution digital image using a preset defect identification algorithm; the algorithm can distinguish and mark micron-sized bubbles, chip sidewall filling defects, and uneven adhesive layer thickness; The acquired high-resolution digital image stream is transmitted to the image processing system in real time. The system then calls a pre-defined defect identification algorithm (either a deep learning-based semantic segmentation model or a traditional image processing algorithm) to analyze the images in real time. This algorithm, trained on a large number of samples, can accurately distinguish and quantify different types of process defects, including: Micron-level bubbles: Identify circular or elliptical low-grayscale areas inside the adhesive layer or at the interface, and count their size, quantity, and distribution location.

[0055] Chip sidewall filling defects: Detect the wetting between the chip sidewall and the colloid meniscus, and mark the areas where the colloid has not fully climbed or filled.

[0056] Uneven adhesive layer thickness: By analyzing the reflective properties of the adhesive surface or using structured light reflection, the uniformity of the adhesive layer thickness in the coated area is evaluated, and areas that are too thin or too thick are marked.

[0057] When the size or number of any defect detected by S503 exceeds the preset process tolerance threshold, the system immediately generates a pause command, interrupts the movement of the dispensing head and the dispensing action, and records the precise coordinates of the current abnormal point and its corresponding defect type and severity. The system presets process tolerance thresholds for various defects (e.g., maximum allowable diameter of bubbles, maximum number of bubbles per unit area, maximum allowable width of missing fillers, etc.). When the size or number of any defect exceeds the corresponding tolerance threshold in the real-time analysis results of sub-step S502, the control system immediately generates a hardware interrupt-level pause command. This command synchronously triggers the dispensing driver and motion controller via a high-speed bus, interrupting the XYZ axis movement of the dispensing head and the dispensing action of the dispensing valve within milliseconds, achieving immediate defect containment. Simultaneously, the system automatically records the precise three-dimensional coordinates (X, Y, Z) of the current abnormal point, along with its corresponding defect type, size data, and severity score, generating an abnormal event log.

[0058] The S504 system calls the 3D topology data of the area near the abnormal point. If it is determined that the Z-axis height is set too high, resulting in insufficient filling, it generates a Z-axis height reduction amount ΔZ for subsequent areas with the same or similar topology. If it is determined that the glue output is insufficient, resulting in air bubbles, it generates a glue output increase coefficient K for the corresponding area.

[0059] The system retrieves 3D topological data of the area near the anomaly point stored in the database (from step S101), and performs root cause analysis and parameter calculation based on the defect type: If the defect type is determined to be insufficient filling (e.g., sidewall not climbing) and 3D topology analysis reveals that the excessive chip height in that area leads to an excessively large actual spacing between the dispensing heads, then an automatic Z-axis height reduction ΔZ will be generated for subsequent areas with the same or similar topology. This ΔZ value will be used to update the working height of the dispensing heads in the corresponding height areas of the remaining path.

[0060] If the defect type is determined to be microbubbles and the analysis indicates that it is caused by insufficient local adhesive, an adhesive dispensing increase coefficient K (K>1) is generated for the corresponding feature area (such as a deeper chip gap). The system will dynamically amplify the basic dispensing speed command in step S403 according to the coefficient K to achieve on-demand adhesive compensation in local areas.

[0061] After S105 dispensing is completed, a non-linear pressure oscillation program is executed, which involves high-frequency small-amplitude pressure fluctuations between vacuum and atmospheric pressure to induce the resonance and rupture of residual microbubbles. Subsequently, rapid pressure is applied to compact the material and complete the curing process.

[0062] During the window period after dispensing but before the colloid gels, the control system closes some valves of the vacuum pump. Simultaneously, it introduces the coordinated action of a fine-tuning pressure relief valve and a micro-inflation valve to execute a nonlinear pressure-variable oscillation program while maintaining a basic seal within the chamber. This program causes the gas pressure within the chamber to oscillate at high frequency (e.g., 20-100Hz), with small amplitude (e.g., ±100-500Pa), in a non-sinusoidal waveform (such as sawtooth or triangular waves) between the current second vacuum level and a transition pressure slightly above vacuum. This nonlinear pressure fluctuation can penetrate the colloid surface and reach the deep microstructure interface. When the oscillation frequency matches the inherent resonant frequency of the residual micron-sized bubbles, the bubbles are forced to undergo volumetric expansion and contraction resonance, disrupting the stability of their gas-liquid interface and causing the bubble walls to rupture. The ruptured gas dissolves in the surrounding colloid in molecular form or is instantly absorbed by adjacent vacuum regions, thus achieving in-situ elimination of deep microscopic bubbles.

[0063] After the microbubbles are eliminated, the system enters the graded pressure leveling stage. Following a pre-set gradient pressure curve, the chamber pressure is gradually and progressively increased from a vacuum state to an intermediate pressure value close to atmospheric pressure. Each pressure level is maintained for a certain time, and the gradually increasing external pressure is used as the driving force to propel the colloid into corners and dams on the chip sidewalls that are difficult to completely fill through capillary action, further penetrating and wetting them. Simultaneously, the pressure changes on the colloid surface promote microflow, achieving spontaneous thickness homogenization.

[0064] After the colloid has fully leveled and all micro-interfaces have been wetted, the system performs a rapid pressurization procedure. By rapidly filling the chamber with a high-purity inert gas (such as nitrogen), the chamber pressure is increased from the intermediate pressure value to the preset curing pressure (usually slightly higher than standard atmospheric pressure, such as 1.2-1.5 atmospheres) in a short time (e.g., within a few seconds).

[0065] Second Embodiment Please see Figures 6-7 The present invention also provides a dispensing device for a COB integrated packaged LED display screen, including a packaging chamber 101, a three-dimensional topology data acquisition module 102, a vacuum module 103, a dispensing module, and an optical detection unit 105; the packaging chamber 101 is used to place the COB substrate; the three-dimensional topology data acquisition module 102 is used to acquire the three-dimensional topology data of the COB substrate; the vacuum module 103 is used to evacuate the packaging chamber 101; the dispensing module is used to dispense adhesive based on the three-dimensional topology data; the optical detection module is used to scan the coated area in real time through a pressure-resistant optical window to identify micron-sized bubbles or filler defects.

[0066] In this embodiment, the encapsulation chamber 101, serving as the physical support and core of the process environment control, is made of high-rigidity metal material, possessing excellent sealing performance and capable of withstanding and stably maintaining a wide range of pressure environments from high vacuum to positive pressure. Precisely calibrated pressure-resistant optical windows are embedded in the top or side of the chamber, providing a high-transmittance, distortion-free observation channel for optical monitoring of the internal process. A precisely temperature-controlled substrate stage is provided within the chamber for placing and fixing the COB substrate, ensuring its positional stability during the process.

[0067] The 3D topology data acquisition module 102 is integrated inside the packaging chamber 101 or coupled through an optical window, employing non-contact optical measurement technology (including but not limited to structured light projection, laser triangulation, or confocal microscopy). Its function is to automatically perform a high-precision 3D scan of the chip array after the COB substrate is placed in the chamber and before the dispensing process begins, acquiring 3D surface contour data including the precise planar coordinates (X, Y) of each chip, height deviation (Z value), and chip gap depth, providing a digital basis for subsequent dispensing path planning and parameter compensation.

[0068] The vacuum module 103 consists of a primary vacuum pump (such as a dry screw pump), a high vacuum pump (such as a molecular pump), and a precision pressure sensor and proportional control valve. This module is sealed to the encapsulation chamber 101 and is used to execute a multi-stage gradient evacuation procedure, precisely reducing the gas pressure in the chamber from atmospheric pressure to a preset process vacuum level (including a second vacuum level for pretreatment and various pressure nodes for curing pressure variation), and achieving precise closed-loop pressure control throughout the dispensing and curing process.

[0069] The dispensing module is mounted on a high-precision motion platform inside the encapsulation chamber 101, possessing XYZ three-axis linkage and possible rotational degrees of freedom. The core of this module is a piezoelectric, pneumatic, or screw-type precision dispensing valve. Its nozzle can dispensing along a generated variable-height path based on data collected by module 102 using three-dimensional topology data. The dispensing module's control system interacts in real-time with a dynamic rheological pressure coupling model, enabling millisecond-level dynamic compensation of dispensing pressure and dispensing speed based on model calculations, ensuring bubble-free colloid coating in a vacuum environment.

[0070] The optical inspection unit 105 is mounted above the optical window outside the packaging chamber 101, or integrated beside the dispensing head 118 inside the chamber. It employs a high-resolution linear or area array camera with a dedicated optical lens and a high-frequency light source. Its function is to perform real-time line scanning of the coated area through the pressure-resistant optical window, synchronized with the movement of the dispensing head 118, to acquire high-resolution digital images and transmit them to the image processing system during the dispensing process. This unit can identify microscopic defects such as micron-sized bubbles, missing chip sidewall fillers, and uneven adhesive layer thickness in real time.

[0071] The central control system serves as the core of equipment decision-making and coordination, and is electrically connected to all the aforementioned modules. It is internally equipped with a dynamic rheological pressure coupling model, a three-dimensional path planning algorithm, and a machine vision-based defect recognition algorithm. The control system receives three-dimensional topology data and feedback data from the optical detection unit 105, calculates and outputs instructions in real time to coordinate the pressure control of the vacuum module 103, the movement and dispensing parameters of the dispensing module, and executes process pause and parameter self-correction when defects are detected, thereby achieving intelligent closed-loop production throughout the entire process.

[0072] The encapsulation chamber 101 includes a base, a chamber body, and a feeding structure. The chamber body has a feed inlet and is fixed on the base. The feeding structure is located on one side of the feed inlet.

[0073] The base, serving as the mounting foundation for the entire packaging chamber 101, is made of high-rigidity metal material, possessing excellent flatness and shock resistance. The base integrates a vibration damping mechanism and temperature control piping, providing stable support and a foundation for environmental heat exchange for the chamber body. The chamber body, fixedly mounted on the base, is made of high-strength stainless steel or aluminum alloy, forming a sealed space for accommodating the COB substrate and performing vacuum processing. A feed port for substrate loading and unloading is provided on the front or side of the chamber body. The chamber body also integrates an air extraction interface, a sensor interface, and the aforementioned pressure-resistant optical window. The feeding structure is positioned on one side of the feed port, used to drive the opening and closing of the door panel 106 and synchronously load and push the substrate.

[0074] The feeding structure includes a door panel 106, a support platform 107, a connecting rod 108, a sealing strip 109, a control cylinder 110, and a connecting block 111. The door panel 106 is rotatably mounted on the chamber body, and the support platform 107 is slidably mounted within the chamber body. The connecting rod 108 is rotatably connected to the door panel 106 and the support platform 107 and is located between the door panel 106 and the support platform 107. The sealing strip 109 is fixed to the door panel 106 and is located close to the side of the chamber body. The control cylinder 110 is rotatably mounted on one side of the chamber body, and the connecting block 111 is rotatably connected to the control cylinder 110 and slidably connected to the door panel 106.

[0075] The door panel 106 is matched with the feed inlet on the chamber body and is specifically used to close or open the feed inlet. A sealing strip 109 is fixedly installed on the side of the door panel 106 facing the chamber body. The sealing strip 109 is made of a vacuum-resistant and high / low temperature-resistant elastic material (such as fluororubber). When the door panel 106 is closed, the sealing strip 109 fits tightly against the edge of the feed inlet to ensure the airtightness of the chamber and meet the requirements of high vacuum processes.

[0076] The support platform 107 is slidably disposed on the inner bottom surface or guide rail of the chamber body, and is used to support the COB substrate to be processed. The position of the support platform 107 corresponds to the feed port and can move between the inside and outside.

[0077] The connecting rod 108 is rotatably connected to both the door panel 106 and the support platform 107, and is located between the door panel 106 and the support platform 107, forming a linkage mechanism. Its design allows the opening and closing motion of the door panel 106 to be synchronously converted into the sliding motion of the support platform 107.

[0078] The cylinder end of the control cylinder 110 is rotatably mounted on one side of the chamber body (usually located on the side or above the feed inlet) via a support or hinge, serving as the power source for the feeding structure.

[0079] One end of the connecting block 111 is rotatably connected to the piston rod end of the control cylinder 110, while the other end is slidably connected to the outer surface of the door panel 106 (e.g., via a slide rail or groove provided on the door panel 106). This design converts the linear motion of the cylinder into a compound motion of the door panel 106.

[0080] The dispensing module includes a lifting structure 112, a support plate 113, a moving screw 114, a moving block 115, a moving motor 116, a longitudinal moving structure 117, and a dispensing head 118. The moving block 115 is slidably disposed on the top of the chamber body. The moving screw 114 is threadedly connected to the moving block 115. The output end of the moving motor 116 is connected to the moving screw 114. The longitudinal moving structure 117 is disposed on the moving block 115. The lifting structure 112 is disposed on the longitudinal moving structure 117. The dispensing head 118 is disposed on the lifting structure 112.

[0081] The movable block 115 is slidably mounted on a linear guide rail on the top of the chamber body, serving as a support base for the lateral movement of the dispensing module. The movable screw 114 is mounted on the top of the chamber body via a bearing housing and is threadedly connected to the threaded sleeve inside the movable block 115, forming a precision ball screw transmission pair. The movable motor 116 (preferably a servo motor or stepper motor) is fixedly mounted on the top of the chamber body, and its output end is fixedly connected to one end of the movable screw 114 via a coupling. When the movable motor 116 rotates, it drives the movable screw 114 to rotate, which in turn translates into precise linear reciprocating motion of the movable block 115 along the guide rail, realizing position control of the dispensing head 118 in the X-axis direction (or Y-axis direction, depending on the coordinate system definition).

[0082] The longitudinal moving structure 117 is fixedly mounted on the moving block 115 and moves laterally together with the moving block 115. The longitudinal moving structure 117 can adopt a screw drive or linear motor drive design similar to the lateral drive mechanism to drive the load on it to move precisely in a direction perpendicular to the lateral direction (i.e., the Y-axis or X-axis direction), thereby forming an XY plane motion platform.

[0083] The lifting structure 112 is fixedly mounted on the movable part of the longitudinal moving structure 117 and can move in the plane with the longitudinal moving structure 117. The lifting structure 112 adopts a high-precision vertical feed mechanism (such as a vertical ball screw slider module or a piezoelectric ceramic micro-motion platform), controlled by an independent lifting motor or driver, to drive the dispensing head 118 to move up and down along the Z-axis direction perpendicular to the XY plane. Its designed stroke needs to cover the chip height difference and process gap requirements, and it can respond to three-dimensional topology data to achieve variable height dispensing.

[0084] The dispensing head 118 is fixedly installed at the bottom movable end of the lifting structure 112 and moves up and down with the lifting structure 112. The dispensing head 118 integrates a precision dispensing valve (such as a piezoelectric jet valve, a time-pressure controller, or a screw pump) and is connected to a glue cartridge feeding system. Its nozzle faces downwards towards the working substrate, and is used to precisely spray or extrude a quantitative amount of encapsulating adhesive under the command of the control system.

[0085] The above-disclosed embodiments are merely one or more preferred embodiments of this application and should not be construed as limiting the scope of this application. Those skilled in the art can understand that all or part of the processes for implementing the above embodiments and equivalent changes made in accordance with the claims of this application still fall within the scope of this application.

Claims

1. A fully laminated packaging process for COB integrated packaged LED displays, characterized in that, This includes placing the die-bonded COB substrate into a packaging chamber and acquiring three-dimensional topology data and initial environmental parameters of the chip array, wherein the initial environmental parameters include ambient temperature data and pressure data. A multi-stage gradient evacuation procedure is executed to reduce the chamber pressure to a first vacuum level. At the same time, ultrasonic pretreatment at a specific frequency is applied to activate the chip surface energy and remove the adsorbed gas layer. The specific frequency specifically includes a first frequency and a second frequency applied sequentially or alternately. The first frequency is used to excite the microstructure resonance of the chip surface and the dam sidewall, and the second frequency is used to induce molecular polarization on the colloidal material surface. Dispensing is performed based on three-dimensional topology data while maintaining a second vacuum level. The volume expansion rate of the colloid at the moment of exiting the nozzle is calculated in real time according to the dynamic rheological pressure coupling model, and the dispensing pressure and dispensing speed are dynamically compensated in reverse, so that the actual tensile stress inside the colloid is always lower than the critical pressure difference threshold for bubble nucleation. The dynamic rheological pressure coupling model establishes the mapping relationship between colloid viscosity, surface tension, ambient temperature, pressure data and volume expansion rate, and calculates the critical pressure difference threshold for bubble nucleation based on the volume expansion rate. During the dispensing process, an in-situ optical detection unit integrated within the encapsulation chamber is used to scan the coated area in real time through a pressure-resistant optical window to identify micron-sized bubbles or filler defects. If an anomaly is detected, dispensing is paused and the Z-axis height and dispensing volume of subsequent dispensing paths are adjusted based on the feedback data. After dispensing, a nonlinear pressure oscillation program is executed, which involves high-frequency, small-amplitude pressure fluctuations between vacuum and atmospheric pressure to induce the resonance and rupture of residual microbubbles. Subsequently, rapid pressure is applied to compact the material and complete the curing process.

2. The full lamination packaging process for COB integrated packaged LED displays as described in claim 1, characterized in that, The specific steps for placing the die-bonded COB substrate into the packaging chamber and obtaining the three-dimensional topology data and initial environmental parameters of the chip array include: By integrating an online 3D topography measuring instrument into the packaging chamber, the chip array is automatically scanned non-contactly after the COB substrate is placed into the chamber, obtaining the precise coordinates and height deviation data of each chip, and generating a 3D surface profile map including the chip gap depth.

3. The full lamination packaging process for COB integrated packaged LED displays as described in claim 2, characterized in that, The specific steps of executing a multi-stage gradient evacuation procedure to reduce the chamber pressure to a first vacuum level, while simultaneously applying ultrasonic pretreatment at a specific frequency to activate the chip surface energy and remove the adsorbed gas layer include: Start the vacuum pump to perform rough pumping of the encapsulation chamber, and reduce the chamber pressure from atmospheric pressure to intermediate transition pressure at the first pumping rate. After reaching the intermediate transition pressure, multi-stage gradient pumping is performed according to the preset step pressure reduction curve. Each step is kept at a constant pressure for a certain time, so that the gas molecules in the chamber are gradually released until the chamber pressure drops to the first vacuum level. After reaching the first vacuum level, the ultrasonic generator is activated to apply ultrasonic waves of a specific frequency to the surface of the COB substrate. Under the continuous action of the ultrasonic waves, the gas molecules physically adsorbed on the chip surface and in the micro gaps are desorbed and released. At the same time, the vacuum pump is kept in operation to continuously extract the desorbed gas from the chamber.

4. The full lamination packaging process for COB integrated packaged LED displays as described in claim 3, characterized in that, The dynamic rheological pressure coupling model establishes a mapping relationship between colloidal viscosity, surface tension, ambient temperature, absolute pressure, and volume expansion rate. The specific steps for calculating the critical pressure difference threshold for bubble nucleation based on the volume expansion rate include: The rheological curves of the encapsulating colloid at different temperatures and shear rates were pre-determined, and the viscosity of the colloid as a function of temperature η(T) was fitted. The surface tension coefficient of the colloid within the process temperature range was also determined, and the surface tension as a function of temperature γ(T) was established. Based on colloidal viscosity and surface tension, a dynamic coupling model of volume expansion rate and pressure difference is established for the instant when the colloid enters the vacuum chamber from the high-pressure zone of the nozzle. During the dispensing process, ambient temperature data is collected in real time to calculate the corresponding colloid viscosity and surface tension. Based on the pressure difference between ambient pressure data and nozzle pressure, the dynamic coupling model is substituted to calculate the actual volume expansion rate of the colloid at the moment it leaves the nozzle. Based on the calculated actual volume expansion rate and the dissolved gas content in the colloid, the critical pressure difference threshold for the formation of bubbles under the current operating conditions is calculated using the bubble nucleation theory formula.

5. The full lamination packaging process for COB integrated packaged LED displays as described in claim 4, characterized in that, The specific steps of dispensing adhesive based on three-dimensional topology data while maintaining a second vacuum level, calculating the volume expansion rate of the adhesive at the instant of exiting the nozzle in real time according to the dynamic rheological pressure coupling model, and dynamically compensating for dispensing pressure and dispensing speed to ensure that the actual tensile stress inside the adhesive is always lower than the critical pressure difference threshold for bubble nucleation include: Based on the three-dimensional topology data of the chip array, a variable height dispensing path is generated, and the basic dispensing pressure and basic dispensing speed are preset. As the dispensing head moves along the preset path, the theoretical volume expansion rate of the colloid at the instant it leaves the nozzle is collected and calculated in real time at a millisecond sampling frequency. The calculated theoretical volume expansion rate is compared with the preset maximum allowable volume expansion rate. If the theoretical volume expansion rate exceeds the maximum volume expansion rate, the compensated dispensing pressure and the compensated dispensing speed are calculated in real time according to the preset reverse compensation algorithm to reduce the pressure difference when the adhesive flows out of the nozzle. The compensated dispensing pressure and dispensing speed are output to the dispensing actuator in real time. At the same time, the actual tensile stress is substituted into the model again to verify whether it is lower than the critical tensile stress corresponding to the critical pressure difference threshold P for bubble nucleation. If the verification passes, the dispensing continues. If the verification fails, the compensation parameters are further adjusted.

6. The full lamination packaging process for COB integrated packaged LED displays as described in claim 5, characterized in that, During the dispensing process, an in-situ optical detection unit integrated within the encapsulation chamber is used to scan the coated area in real time through a pressure-resistant optical window to identify micron-sized bubbles or filler defects. If an anomaly is detected, the specific steps for pausing dispensing and correcting the Z-axis height and dispensing volume of subsequent dispensing paths based on feedback data include: While the dispensing head moves along a preset path to apply the coating, the in-situ optical detection unit integrated on the top of the encapsulation chamber is activated. The detection unit performs line scanning on the coated area through a pressure-resistant optical window at a speed synchronized with the movement of the dispensing head to acquire high-resolution digital images. Based on the high-resolution digital image, the image is analyzed in real time using a preset defect identification algorithm; the algorithm can distinguish and mark micron-sized bubbles, chip sidewall filling defects, and uneven adhesive layer thickness. When the size or number of any defect exceeds the preset process tolerance threshold, the system immediately generates a pause command, interrupts the movement of the dispensing head and the dispensing action, and records the precise coordinates of the current abnormal point and its corresponding defect type and severity. The system calls the 3D topology data of the area near the abnormal point. If it is determined that the Z-axis height is set too high, resulting in insufficient filling, the Z-axis height adjustment ΔZ is generated for the subsequent areas with the same or similar topology. If it is determined that the glue output is insufficient, resulting in air bubbles, the glue output increase coefficient K is generated for the corresponding area.

7. A dispensing device for a COB integrated packaged LED display screen, applied to the full lamination packaging process of the COB integrated packaged LED display screen according to any one of claims 1 to 6, characterized in that, It includes a packaging chamber, a three-dimensional topology data acquisition module, a vacuum module, a dispensing module, and an optical detection unit; the packaging chamber is used to place the COB substrate; the three-dimensional topology data acquisition module is used to acquire the three-dimensional topology data of the COB substrate; the vacuum module is used to evacuate the packaging chamber. The dispensing module is used to dispense adhesive based on three-dimensional topology data; The optical detection module is used to scan the coated area in real time through a pressure-resistant optical window to identify micron-sized bubbles or missing fillers.

8. The dispensing equipment for COB integrated packaged LED displays as described in claim 7, characterized in that, The packaging chamber includes a base, a chamber body, and a feeding structure. The chamber body has a feed inlet and is fixed on the base. The feeding structure is located on one side of the feed inlet.

9. The dispensing equipment for COB integrated packaged LED displays as described in claim 8, characterized in that, The feeding structure includes a door panel, a support platform, a connecting rod, a sealing strip, a control cylinder, and a connecting block. The door panel is rotatably mounted on the chamber body, the support platform is slidably mounted inside the chamber body, the connecting rod is rotatably connected to the door panel and the support platform and is located between the door panel and the support platform, the sealing strip is fixed to the door panel and is located on one side close to the chamber body, the control cylinder is rotatably mounted on one side of the chamber body, and the connecting block is rotatably connected to the control cylinder and slidably connected to the door panel.

10. The dispensing equipment for COB integrated packaged LED displays as described in claim 9, characterized in that, The dispensing module includes a lifting structure, a support plate, a moving screw, a moving block, a moving motor, a longitudinal moving structure, and a dispensing head. The moving block is slidably disposed on the top of the chamber body. The moving screw is threadedly connected to the moving block. The output end of the moving motor is connected to the moving screw. The longitudinal moving structure is disposed on the moving block. The lifting structure is disposed on the longitudinal moving structure. The dispensing head is disposed on the lifting structure.

Images