High-efficiency heat dissipation thermoelectric separation metal substrate manufacturing method based on insulation layer optimization

Abstract

This invention discloses a method for manufacturing a high-efficiency heat dissipation and thermoelectric separation metal substrate based on insulation layer optimization, relating to the field of heat dissipation substrate manufacturing technology. The method includes: cleaning the top surface of a copper substrate heat sink before lamination and applying a peelable protective layer; forming a periphery sealing structure around the substrate; and setting capillary fracture steps and adhesive storage tanks / cavities around the opening of the pores to pre-determine the destination of excess adhesive. Subsequently, the laminated blank is subjected to vacuum preheating and degassing, segmented pressurized filling, and curing to guide excess adhesive into the peripheral adhesive storage area. After curing, the protective layer is peeled off within a controlled peeling temperature window at a controlled peeling angle and speed. Machine vision is used to obtain the residual adhesive coverage rate and a threshold is determined. If the threshold is exceeded, low-damage cleaning / activation re-inspection is performed; if the threshold is still exceeded, the process is terminated. The determination is linked to vacuum degree, temperature, pressure curves, and batch data for feedback adjustment, reducing dependence on laser / grinding / strong acid adhesive removal and improving mass production stability and traceability, making interface contamination risks more controllable.

Description

Technical Field

[0001] This invention relates to the field of heat dissipation substrate manufacturing technology, specifically to a method for manufacturing a high-efficiency heat dissipation and thermoelectric separation metal substrate based on insulation layer optimization. Background Technology

[0002] Thermoelectrically separated metal substrates (thermoelectrically separated copper substrates, double-sided metal substrates) are commonly used in LED modules such as automotive headlights and daytime running lights. They need to balance heat dissipation and insulation: the heat-conducting and heat-dissipating area forms a heat channel with the metal substrate, the electrical connection area is isolated from the metal substrate by PP or insulating thermally conductive medium, the pads and circuits form a raised heat sink / thermal conductive area on the metal substrate, the copper foil layer and the PP layer (or insulating thermally conductive layer) are laminated with windows, and drilling, pattern etching and surface treatment are performed.

[0003] For example, Chinese patent document CN111246656A (202010025603.2, June 5, 2020) discloses a copper-based circuit board for thermoelectric separation of LEDs and its preparation method. The method involves creating windows in the copper foil layer and PP layer, etching a protruding heat sink in the substrate layer, stacking and pressing the substrate, PP film, and copper foil together, then using a laser to remove excess adhesive around the heat sink for separation, followed by circuit etching and other processes. A mechanical method for removing residual adhesive during pressing is disclosed in Chinese patent document CN108848615A (201810842864.6, November 20, 2018). This method involves forming a raised copper block on a copper substrate, then forming slots corresponding to the raised copper block in an FR4 substrate and PP sheet, stacking and pressing them together, and then grinding to remove residual adhesive at the raised copper block after copper reduction.

[0004] For example, Chinese patent document CN111954368A (202010771225.2, November 17, 2020) discloses a double-sided metal substrate for thermoelectric separation and filling electroplating and its preparation method. The exposed copper foil in the heat-conducting area is removed by etching and the exposed insulating thermally conductive adhesive is etched with a chemical solution, and then filled by electroplating to meet assembly requirements.

[0005] For the aforementioned applications, lamination involves softening and flowing PP or insulating thermally conductive adhesive under heat and pressure to achieve bonding and interface filling. Due to the gaps and height differences between the heat sink / copper bump and the window sidewalls, and the significant temperature variations in the resin system viscosity, material will accumulate upwards along the gaps during lamination, resulting in excess and residual adhesive on or near the top surface of the heat sink. For automotive lighting substrates, the top surface of the heat sink serves as the interface for device mounting or contact with heat dissipation components. After curing, the residual adhesive adheres to the metal surface, affecting solder wetting, assembly bonding, and the consistency of thermal interface contact. Furthermore, under thermal cycling, reflow soldering, or humid conditions, it amplifies appearance and reliability risks.

[0006] For existing technologies, using laser-based adhesive removal for isolation requires controlling positioning and removal depth, as well as adding specialized processes; grinding to remove residual adhesive is a mechanical contact process, usually accompanied by changes in surface morphology and dust / foreign matter control; chemical etching combined with electroplating requires multiple wet processes, with high requirements for chemical and waste management and process consistency. Therefore, under mass production conditions with high requirements for cycle time, yield, and consistency, how to stably control and reduce residual adhesive contamination on the top surface of the heat sink during the lamination and lamination manufacturing of thermoelectric separation metal substrates, and ensure the interface quality of subsequent welding and assembly, still needs further exploration. Summary of the Invention

[0007] (a) Technical problems to be solved To address the shortcomings of existing technologies, this invention provides a highly efficient heat dissipation and thermoelectric separation metal substrate manufacturing method based on insulation layer optimization. By performing vacuum preheating and degassing, segmented pressurized filling, and curing on the laminated blank, excess adhesive is directed into the peripheral adhesive storage area. After curing, the protective layer is peeled off within a controlled peeling temperature window at a controlled peeling angle and speed. Machine vision is used to obtain the residual adhesive coverage rate and a threshold is determined. If the threshold is exceeded, low-damage cleaning / activation re-inspection is performed; if the threshold is still exceeded, the process is terminated. The determination is linked to vacuum degree, temperature, pressure curves, and batch data for feedback adjustment, reducing dependence on laser / grinding / strong acid adhesive removal and improving mass production stability and traceability. This makes the risk of interface contamination more controllable and solves the technical problems described in the background art.

[0008] (II) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: A method for manufacturing a high-efficiency heat dissipation and thermoelectric separation metal substrate based on insulation layer optimization includes aligning and laminating a copper substrate, an insulating thermally conductive film or PP film with a circuit copper foil layer, and further includes cleaning the top surface of the copper substrate heat sink, applying a peelable protective layer and forming a periphery sealing structure; and forming a capillary fracture step and a glue storage tank or glue storage cavity at the periphery of the hole opening to obtain a laminated blank. The laminated blanks are sequentially vacuum preheated and degassed, then filled under pressure in sections and cured to obtain a cured press-fit plate. Within the preset peeling temperature window, peel the peelable protective layer at the set peeling angle and peeling speed. Use machine vision to obtain the residual adhesive coverage and determine the threshold. If the threshold is exceeded, clean and re-inspect. If the threshold is still exceeded, exit. The threshold determination is associated with the vacuum degree, temperature and pressure curve records and the batches of peelable protective layer, insulating thermal conductive film or PP film to generate release, rework or scrap records, and the pressing sequence or capillary fracture step, glue storage tank or glue storage cavity and peripheral sealing structure are adjusted according to continuous exceeding of threshold or peeling failure.

[0009] Furthermore, the peelable protective layer is formed by stacking a heat-resistant support layer, a seepage barrier layer, and a release layer. When applied, it covers the top surface of the heat sink and extends to the periphery of the opening. The heat-resistant support layer is attached to the top surface of the heat sink, and the release layer is located on the side away from the top surface of the heat sink. A peeling start end is reserved at the edge of the peelable protective layer.

[0010] Furthermore, the peripheral sealing structure is a continuous circumferential sealing line formed around the periphery of the peelable protective layer. The circumferential sealing is closed once before the laminated blank enters the pressing front, and the start and end sections overlap. After sealing, the circumferential sealing line is inspected in a circumferential manner and the results are recorded. If there is a sealing gap, the sealing is reworked and redone.

[0011] Furthermore, when forming the perimeter sealing structure, an annular shallow groove is processed on the perimeter of the top surface of the heat sink. The perimeter of the peelable protective layer is embedded into the annular shallow groove and compacted and fixed. Before the pre-pressing of the stacked blanks, the inner edge of the annular shallow groove is checked circumferentially to ensure there is no warping or breakage. If any abnormality is found, the groove is reworked and re-embedded.

[0012] Furthermore, the capillary fracture steps are continuously arranged along the periphery of the scooping hole and are located opposite to the side wall of the heat dissipation plate. The glue storage tank or glue storage cavity is arranged along the outer periphery of the capillary fracture steps and is connected to the outer periphery glue storage area. When the stacked blanks are aligned, the glue storage tank or glue storage cavity is located on the outer side of the top surface of the heat dissipation plate.

[0013] Furthermore, the vacuum preheating and exhaust process includes a low-temperature platform for vacuuming and pressure holding, a medium-temperature platform for heating and maintaining temperature, and a high-temperature platform before entering the segmented pressurization and filling stage. The segmented pressurization and filling stage includes a low-pressure pre-compression stage, a linear ramp pressurization stage, and a ramp end pause stage, and maintains a set pressure and set temperature in the curing and shaping stage.

[0014] Furthermore, the low-temperature platform temperature is 80°C to 120°C, the medium-temperature platform temperature is 120°C to 160°C, and the high-temperature platform temperature is 160°C to 200°C; the low-pressure pre-compression section pressure is 0.2 MPa to 0.6 MPa, and the linear ramp pressurization section increases from 0.1 MPa to 0.5 MPa per minute to 1 MPa to 2.5 MPa per minute.

[0015] Furthermore, the peelable protective layer is peeled off when the peeling temperature window is 50 degrees Celsius to 90 degrees Celsius. Before peeling, a peeling start end is formed at the edge of the peelable protective layer. During peeling, the peeling angle is 150 degrees to 180 degrees and the peeling speed is 20 mm to 150 mm per second. The peeling process is recorded with the plate number as an association.

[0016] Furthermore, under fixed light source illumination, machine vision acquires images of the top surface of the heat sink and generates an effective area mask. After performing flat field correction, edge detection, threshold segmentation, and morphological operations on the image, candidate areas for residual adhesive are obtained. The residual adhesive coverage rate is calculated and judged according to the threshold. If the value exceeds the threshold, plasma cleaning is performed and re-inspected.

[0017] Furthermore, the residual adhesive coverage threshold determination result is linked to the vacuum degree, temperature and pressure curve records and the batch of peelable protective layer, insulating thermal conductive film or PP film, and a cyclic redundancy check code is used for verification when writing the board number; when the threshold is exceeded or peeling fails continuously, the ramp pressure increase section of the segmented pressurization filling and the vacuum preheating and exhaust time are adjusted, and the machine vision execution cycle is calibrated.

[0018] (III) Beneficial Effects This invention provides a method for manufacturing a high-efficiency heat dissipation and thermoelectric separation metal substrate based on insulation layer optimization, which has the following beneficial effects: Before lamination, the top surface of the heat sink is cleaned and covered with a peelable protective layer. A periphery sealing structure is formed around the periphery of the peelable protective layer, blocking the resin from the interface below the protective layer. This shifts the cleanliness of the heat sink top surface from post-lamination to pre-lamination controllable cleanliness. Capillary fracture steps are set around the periphery of the borehole opening, and a resin storage tank / cavity is set on the outer side. The resin climbing path is blocked by the abrupt cross-sectional threshold and contained in the resin storage tank / cavity. Most of the overflow resin is located in the outer periphery storage area, eliminating the risk of contaminating the top surface of the heat sink. The laminated blanks are sequentially vacuum preheated and degassed, then filled and cured in stages. This allows volatile gases to be eliminated before pressurization and interface filling to be completed during the gradual increase in pressure. The lamination sequence, along with the capillary fracture steps, resin storage tank / cavity, and periphery sealing structure, works in synergy to achieve controlled resin flow and prevent abnormal gushing before curing.

[0019] Within the peeling temperature window, the peelable protective layer is peeled off according to the controlled peeling angle and peeling speed. The excess / residual adhesive of the protective layer is carried away with the migration of the protective layer, and the interface fracture becomes more stable. Without using carbon dioxide laser debonding, grinding or strong acid etching as the main process, a deliverable heat sink top surface can be obtained.

[0020] Machine vision is used to obtain the residual adhesive coverage and determine the threshold. If the threshold is exceeded, low-damage cleaning / activation is triggered and re-inspection is performed. If the re-inspection still exceeds the threshold, the process is terminated. The threshold determination and re-inspection results are linked to the board number and archived to restrict rework to an feasible process and prevent unqualified interfaces from entering subsequent processes.

[0021] The threshold determination results are linked to the records of vacuum degree, temperature, and pressure curves, as well as the batches of peelable protective layers, insulating thermal conductive films, or PP films. When the threshold is exceeded or peeling fails, feedback is provided to adjust the pressing sequence and structural control points, forming a closed loop for sustainable mass production and a closed loop for traceable verification of mass production. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the manufacturing process of the efficient heat dissipation and thermoelectric separation metal substrate of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Please see Figure 1 This invention provides a method for manufacturing a high-efficiency heat dissipation and thermoelectric separation metal substrate based on insulation layer optimization. The method includes the following steps: First, before entering the vacuum segmented lamination process, the passive problem of removing excess adhesive and residual adhesive after lamination is transformed into a controllable precondition for isolation and capture before lamination. This allows the laminated blank to have repeatable starting boundaries in terms of structure, material state, and alignment, thereby providing a stable input for subsequent curing, peeling, and inspection closed loop.

[0025] In the manufacturing of thermoelectrically separated copper substrates for automotive lighting, the top surface of the boss heat sink serves as both a thermal interface and a mounting or assembly contact surface. During lamination, the insulating and thermally conductive insulating sheet or PP insulating sheet softens and flows under temperature and pressure, easily climbing along the gaps in the windowed sidewalls and forming cured residue on the top surface of the heat sink. If only laser or grinding is used for adhesive removal after lamination, it often increases the requirements for alignment and depth control, and introduces changes in surface morphology and pressure on dust and waste liquid management.

[0026] Therefore, interface cleanliness and adhesive overflow destination are treated as the same causal chain: first, a cleanable interface is established on the top surface of the heat sink, then geometric boundaries are used to guide the adhesive overflow into non-critical areas, and finally, these states are solidified into verifiable pre-defined judgment quantities as release conditions for entering the next step.

[0027] Contamination on the top surface of the heat sink typically originates from oil films, fingerprints, microparticles, and debris brought in during the cutting and handling of laminated materials. If this contamination is not removed before lamination, localized floating points will appear between the protective layer and the top surface of the heat sink, preventing the lamination pressure from being evenly distributed. This, in turn, creates microchannels around the protective layer, providing a pathway for resin to penetrate beneath it. Therefore, cleaning is not just about making the surface cleaner, but also about establishing a continuous contact boundary for the subsequent sealing structure.

[0028] To bring the initial state of the heat sink top surface-protective layer interface to a repeatable range, solvent degreasing and particle removal are performed first, followed by surface activation to stabilize surface energy. Finally, the same set of acquisition and interpretation methods is used to confirm whether the cleanliness standard is met, avoiding batch drift caused by relying solely on visual judgment.

[0029] The process involves wiping the top surface of the heat sink with electronic-grade isopropanol (purity not less than 99.5%) to remove oil, rinsing with deionized water, and drying with a dust-free gas. After drying, the top surface of the heat sink is activated with low-power plasma and proceeds to the next process within a fixed time window to avoid re-adsorption. If the heat sink remains in an open environment for more than the preset time window after cleaning and activation, the top surface of the heat sink is prone to re-adsorbing organic matter and particles from the air, and the activation or cleaning process should be repeated before reapplying.

[0030] To ensure the cleanliness baseline can be used in subsequent processes, the cleanliness baseline image is used as the data input: a fixed-focus camera is used to acquire an image of the top surface of the heat sink under a fixed lighting angle. Particle candidate regions are obtained through grayscale thresholding, and isolated noise points are removed using morphological opening operations. If the candidate regions are continuous sheets, they are determined to require rework cleaning. In the general industrial vision controller, the threshold is set once by the calibration board during line change. For batches requiring verification, the contact angle correction threshold is checked but not a mandatory condition for each sheet. The process involves solvent degreasing, rinsing and drying, followed by plasma activation, and then transferring the sheet to the bonding station within a specified time. With the oil film and particles removed from the top surface of the heat sink, the interface floating point during protective layer bonding decreases, the surface energy state of the top surface of the heat sink converges, and the sealing structure more easily forms a continuous contact band. The cleanliness baseline image provides a fixed interpretation caliber, offering verifiable input for subsequent anomaly tracing.

[0031] During hot and cold pressing, single-layer films are prone to wrinkling or edge grooves, and resin can penetrate below the film along the grooves. Low molecular weight resin can diffuse to the metal interface, leaving a thin layer after the film is peeled off. The problem cannot be transferred from the residual adhesive to the residual thin layer. A multi-layer peelable protective layer can be used to seal the perimeter and change the film boundary from a free edge to a controlled edge.

[0032] To simultaneously achieve the goals of preventing displacement and resin migration during pressing and controllable peeling after curing, a heat-resistant support layer provides a rigid framework, a barrier layer inhibits resin penetration into the metal interface, and a release layer reduces resin adhesion to the outer surface of the protective layer. Subsequently, a sealing structure is formed around the protective layer to maintain a continuous compaction zone at the boundary during pre-pressing and pressing.

[0033] The heat-resistant support layer is made of polyimide film or polyester film with a thickness of 20μm to 60μm; the barrier layer is made of inorganic barrier coating or high-barrier resin coating containing inorganic flake filler, and the barrier layer forms a continuous coating surface on one side of the heat-resistant support layer; the release layer is made of silicone-based release coating or fluorine-containing release coating, and the release layer is located on the outer surface of the peelable protective layer and corresponds to the side that may come into contact with excess adhesive. The peelable protective layer is rolled onto the top surface of the heat sink by a pressure roller. The rolling direction of the pressure roller is parallel to the long side of the heat sink and maintains unidirectional air exhaust. The rolling is performed once or twice. After rolling, the surface of the peelable protective layer is inspected under side light and found to be free of wrinkles and continuous air bubbles.

[0034] The protective layer is cut to a contour that matches the top surface of the heat sink. After being placed with the help of a positioning fixture, it is slowly rolled in one direction with a rubber pressure roller to squeeze out the air at the interface from the same side and avoid the formation of closed air bubbles. A heat-resistant elastic pressure head is used to form a circumferential perimeter sealing structure along a circular trajectory, so that the edge of the protective layer and the outer perimeter of the heat sink form a continuous compaction zone. If the outer perimeter of the heat sink has a shallow groove, the edge of the protective layer is first embedded into the shallow groove before forming the perimeter sealing structure, so that the edge has both geometric locking and compaction sealing.

[0035] As an example: On the automotive LED board production line, the operator places the cleaned and activated copper substrate into a tray with positioning pins, takes a pre-cut peelable protective layer, attaches the barrier layer to the top surface of the heat sink, and gently places the protective layer into the positioning pin. Then, a pressure roller is used to roll it once from left to right to ensure there is no edge lifting. It is then sent to the sealing station, where a ring-shaped pressure head presses out a continuous circumferential sealing structure. After completion, the circumferential sealing structure is checked for breaks under side lighting. If a break is found, the protective layer is peeled off, reapplied, and sealed again. The visible result on-site is a continuous circumferential sealing structure and a wrinkle-free protective layer.

[0036] Preferably, the total thickness of the peelable protective layer is selected from 20μm to 80μm to ensure sufficient shape retention during roller pressing without creating obvious steps. The width of the compaction band of the circumferential sealing structure is selected from 0.2mm to 0.8mm to ensure continuous contact during hot pressing without encroaching on the effective area of ​​the opening. More preferably, when shallow grooves are set on the outer periphery of the heat sink, the depth of the grooves is selected from 10μm to 60μm to form a stable geometric lock after the edge of the protective layer is embedded, and then the compaction and sealing are completed by the circumferential sealing structure, thereby reducing the probability of edge lifting caused by handling vibration.

[0037] In use, first apply multiple peelable protective layers according to the fixture positioning and then use pressure rollers to release air. Next, form a perimeter seal using a peripheral sealing structure or shallow groove embedding. The protective layer remains flat under temperature and pressure, reducing wrinkles and grooves that provide channels for resin infiltration. The barrier layer inhibits resin diffusion to the metal interface and forms a cleanable interface after peeling.

[0038] The resin's ascent within the gap is influenced by both pressure and capillary action. When the window sidewall and the outer periphery of the heat sink form a continuous, smooth transition, the resin easily forms a thin layer along the transition surface and cures at the edge of the top surface of the heat sink. The capillary fracture step introduces abrupt changes in cross-section along the ascent path, forcing the thin layer to cross the step inflection point, thus raising the threshold and guiding the resin's position to the outside of the step. This step cooperates with the subsequent resin storage tank / cavity; the step is responsible for interrupting the ascent, and the resin storage tank / cavity is responsible for catching the guided overflow.

[0039] To provide robust constraints on material fluctuations using geometric boundaries, rather than leaving the problem to post-lamination processing, steps are preferentially formed around the openings in the insulating and thermally conductive sheet or PP insulating sheet. A height abrupt change from the upper to the lower plane is created through secondary die-cutting or localized embossing. If the insulating and thermally conductive sheet is a multi-layered structure, a pre-cured ring can be stacked around the openings as a step, ensuring a clear proportion between the step height and the thickness of the insulating and thermally conductive sheet.

[0040] The process involves first using a graded die to cut the full thickness to form a window, and then using a half-depth cut to form a stepped shoulder. A stylus-type profilometer is used to scan along the periphery of the window to obtain a height curve. First, cubic spline interpolation is used to resample the sampling points into an equally spaced sequence. Then, the first-order difference derivative is used to locate the inflection point position, confirming that the inflection point is continuous and has no reverse slope, thus avoiding the formation of a reverse slope that would cause the resin to climb again.

[0041] The input for this verification is a high-sampling sequence, and the output is the conclusion on the continuity of the step inflection point. If a gap exists, the insulating and thermally conductive insulating layer sheet should be replaced or the step forming should be redone, rather than leaving the gap for subsequent pressing to remedy, because the gap will become the preferred channel for glue overflow and amplify batch fluctuations.

[0042] In use, the periphery of the window is first die-cut in stages or locally imprinted to form capillary fracture steps. Then, contour scanning and interpolation differentiation are used to confirm the continuity of the step inflection points. The resin thin layer climbing path is interrupted by abrupt changes in cross-section, and the climbing termination position is more stably fixed on the outer side of the step. The consistency of the step inflection points is quantitatively verified, reducing the risk of glue overflow caused by local gaps.

[0043] Furthermore, even with capillary fracture steps, unavoidable excess adhesive volume may still occur during lamination. If this excess adhesive volume diffuses freely outside the step, it may still bypass the step and contact the edge of the top surface of the heat sink.

[0044] The adhesive reservoir / cavity provides a low-resistance volume, allowing excess adhesive to preferentially enter and remain in non-critical areas after curing. Meanwhile, alignment misalignment alters the gap distribution on the window sidewalls, leading to localized pressure concentration and pushing excess adhesive beyond the boundary. Therefore, the adhesive reservoir structure must be incorporated into the same decision-making logic as alignment pre-compression and sealing continuity.

[0045] To transform the adequacy of structural preparation into a calculable and verifiable threshold condition, the following steps are taken: First, the glue storage tank / cavity is formed and its relative position to the step is confirmed. Then, alignment and pre-compression are performed. Finally, the proportion of the sealing notch, alignment deviation, sidewall gap, step height, and glue storage volume are collected as key inputs to calculate the preset acceptance criteria, which are used to determine whether to proceed to the next step or rework.

[0046] An annular groove is set on the outer side of the step as a glue storage tank. The groove is formed by stamping or local imprinting. If a segmented glue storage chamber is used, the annular groove is divided into multiple chamber segments according to the direction of the heat sink array and the partition is retained to suppress glue overflow migration along the circumferential direction. After the lamination is aligned using positioning holes and rivet holes, pre-pressure is applied at room temperature to prevent the laminated blank from slipping before it is transported into the press. After pre-pressure, a camera is used to capture the concentricity image of the window and the heat sink. The contour is obtained by edge detection and the alignment deviation is calculated. The circumferential scanning of the peripheral sealing structure is used to count the length of the notch segment and convert it into the sealing notch ratio.

[0047] Among them, the dedicated derived judgment quantity of step one is introduced and the qualified judgment quantity is preset. This is used to compress multiple failure channels into a traceable threshold indicator: Where: Preset pass / fail judgment quantity : Characterizes the overall readiness of the laminated blank in four dimensions: sealing, alignment, steps, and adhesive storage volume; the value ranges from 0 to positive infinity, and the process uses a preset threshold for release or rework; sealing gap ratio Values ​​range from 0 to 1, representing the proportion of the notch length in the perimeter sealing structure to the total length of the perimeter sealing structure. The smaller the value, the more continuous the blockade; Alignment deviation The distance between the center of the window opening and the center of the heat sink is taken as 0 to the gap between the side walls. ; Sidewall gap The minimum distance between the window sidewall and the outer periphery of the heat sink, ranging from 20μm to 200μm; the effective volume of the glue storage tank. The glue storage tank / cavity is a geometric volume that can be used to collect excess glue, with a value of 0.1mm. 3 Up to 10mm 3 ; Expected volume of overflow The estimated volume of adhesive overflow per unit circumference of the heat sink under a given temperature and pressure curve, with a range of 0.05 mm. 3 up to 5mm 3 ; The volume of adhesive residue on the test piece of the same batch of materials was calculated by back-calculation. The expected volume of excess adhesive was obtained from test pieces made from the same batch of insulating sheet materials: After the test pieces were pressed and cured in step two, the height field of the excess adhesive area on the outer periphery was collected using a three-dimensional optical profilometer or confocal microscope. The volume of excess adhesive was calculated by integrating the pixel area with the height, and then normalized according to the circumferential length of the heat sink to obtain the volume corresponding to a single heat sink. ; Capillary fracture step height The height difference between the upper and lower planes of the step; the value ranges from 10μm to 200μm. The larger the relative sidewall gap, the lower the probability of the thin layer crossing the step; When the pre-set acceptable quantity is below the threshold, rework is performed based on the source of the defect: if the sealing gap ratio is too large, the bonding and pressing are redone; if the alignment deviation is too large, the alignment and pre-pressing are re-done; if there is a gap at the step inflection point, the step is replaced or redone; if the effective volume of the glue storage tank is insufficient, the insulating and thermally conductive insulating layer sheet with a larger glue storage structure is replaced. Through defect-rework path mapping, the main risks of the laminated blank can be contained before it enters the pressing machine. The glue storage tank / cavity is arranged around the perimeter of the window on the plane and is located outside the capillary fracture step; along the radial direction of the window, the distance from the inflection point of the capillary fracture step to the inner boundary of the glue storage tank / cavity is 0.1mm to 1.0mm, so as to ensure that the overflow glue can preferentially enter the glue storage tank / cavity before entering the top surface of the heat sink after crossing the step.

[0048] When using it, first form a ring or segmented glue storage tank and confirm its relative position with the step, then align and pre-press it and calculate the preset qualified judgment amount for release or rework.

[0049] Excess adhesive has a clearly defined storage volume, reducing the chance of it bypassing the steps and contacting the top surface of the heat sink. Alignment, sealing, steps, and adhesive accumulation are all unified into a single judgment quantity, making rework paths traceable and verifiable. The initial boundary of the laminated blank is converged, making it easier to keep the direction of excess adhesive under control in the next step of vacuum segmented pressing.

[0050] Step 2: Without disrupting the cleanable interface on the top surface of the heat sink and the controlled destination of the excess adhesive established in Step 1, the insulating thermally conductive insulating sheet or PP insulating sheet completes interface bonding and gap filling under controlled conditions through a continuous sequence of vacuum preheating and degassing, segmented pressurization and filling, and curing and shaping. The main volume of excess adhesive is stably guided into the adhesive storage tank / cavity, thereby outputting a cured and pressed board that can be peeled off within the temperature window and subjected to online closed-loop detection.

[0051] During lamination, the laminated blanks undergo three parallel processes: temperature increase, pressure increase, and resin cross-linking and curing. Deviation in any of these processes will amplify the risk of adhesive overflow: if venting is insufficient, entrained gas will form local jet channels during pressurization, and the overflowed adhesive will bypass the capillary fracture steps along the channels and contact the top edge of the heat sink; if the pressurization is too fast, the resin will be forced to flow laterally before a stable bonding interface is formed, and the adhesive reservoir / cavity will not be able to absorb all the overflowed adhesive, resulting in a breakthrough; if curing is too early, the resin viscosity will suddenly increase before the sidewall gaps are fully wetted, and unfilled defects will form at the interface, which are more likely to evolve into peeling or blistering during subsequent thermal cycling.

[0052] After the laminated blanks enter the vacuum press, the first challenge is whether the vacuum truly applies to the gaps between the windowed sidewalls and the perimeter sealing boundary. If the vacuum only exists within the cavity and fails to be effectively transmitted to the interior of the gaps through the microchannels between the layers, volatiles will remain trapped in localized areas during subsequent heating, eventually bursting through the resin layer in the form of bubbles during the pressurization stage. Therefore, vacuum sealing and exhaust channel management are considered the starting conditions for pressing, rather than assuming the evacuation action of the vacuum pump is inherently effective.

[0053] First, a sealing check is performed to confirm that there is no continuous leakage path between the vacuum chamber, the isolation membrane, and the pressed steel plate. Then, a three-stage process of isolation, observation, and re-vacuuming is used to determine whether there is slow leakage or localized air release around the laminated blank. Specifically, the laminated blank is placed between the pressed steel plates, the isolation membrane is laid on top, and the vacuum frame is sealed. The vacuum pump is started to bring the chamber to the preset vacuum level, and then the air inlet valve is closed, and the vacuum level is recorded over time. If the vacuum level recovers in a short time, the tightness of the frame sealing strip and whether foreign objects are trapped in the folded edge of the isolation membrane are checked first.

[0054] Furthermore, an exhaust channel area is reserved at the edge of the laminated blank to allow volatiles to be discharged along the channel area without passing through the perimeter sealing structure during vacuum pumping. For plate types with dense heat sink arrays, shallow grooves can be arranged on the pressed steel plate as auxiliary exhaust channels to make the exhaust direction consistent with the circumferential direction of the perimeter sealing structure, thereby reducing the risk of local edge lifting caused by exhaust flow passing through the perimeter sealing structure.

[0055] As an example, the operator places the laminated blank released in step one into the positioning block of the lower steel plate. After confirming that the alignment hole does not interfere with the block, an isolation film is applied, and the isolation film is pressed into the pressure strip of the vacuum frame around its perimeter. After starting the vacuum pump, the operator observes the vacuum gauge pointer drop and remain stable. Then, the inlet valve is closed and a period of time is waited. If the pointer rebounds, the pressure strip is lifted to check if the sealing strip is curled. After resealing, the vacuum is re-evacuated. After confirming that the pointer is stable, the evacuation direction is marked with a marker in the edge exhaust channel area as a basis for subsequent on-site investigation of abnormal glue overflow direction. Vacuum measurement can be performed using a capacitive vacuum gauge or a thermal conductivity vacuum gauge, or a measuring device with equivalent functions, as long as it can output a record of the vacuum degree changing over time.

[0056] In use, a vacuum seal is first established and leaks are screened through isolation observation. Then, the volatile emission path is managed through the edge exhaust channel area and shallow grooves in the steel plate. The vacuum effect is confirmed to be transmitted to the periphery and gap areas of the laminated blank, reducing the probability of bubbles breaking through the resin layer during the pressurization stage. The exhaust direction is constrained to the channel area, and the peripheral sealing structure is less likely to be lifted by the exhaust flow, thus maintaining the integrity of the protective layer boundary on the top surface of the heat sink.

[0057] During the heating process, the insulating and thermally conductive insulating sheet or PP insulating sheet undergoes a transformation from a glassy state to a highly elastic state and then to a flowable state, accompanied by the release of a small amount of volatiles. If the heating is too rapid, the volatiles cannot be discharged along the exhaust channel in time, and will form local pressure at the periphery of the protective layer or in the gap of the window sidewall, thus pushing the resin to the top edge of the heat sink during the subsequent pressurization stage. If the heating is too slow, the resin remains in a viscoelastic state for a long time, and will undergo slow creep under the action of gravity and micro-displacement between the plates, causing the resin reservoir / cavity to be filled with a thin layer of resin in advance, reducing its capacity to be contained during the pressurization stage. Therefore, a segmented heating method is adopted with an executable preheating termination criterion to allow the resin to enter the window that can be filled but is not forced to flow out.

[0058] First, the overall temperature is uniformly reached in the low-temperature section to release the adsorbed water and low-boiling-point components. Then, the resin viscosity is reduced in the medium-temperature section to begin wetting the gaps between the sidewalls. Finally, the resin is briefly stopped in the high-temperature section near the flow start point to ensure continuous wetting of the interface, while vacuum pumping is maintained at all times to remove volatiles.

[0059] The preheating process is divided into three continuous platforms: a low-temperature platform for temperature homogenization and initial venting, a medium-temperature platform for viscosity reduction and wetting initiation, and a high-temperature platform for wetting completion without entering strong flow. Preferably, the temperatures of the three platforms are set sequentially to 80°C to 120°C, 120°C to 160°C, and 160°C to 200°C, and maintained on each platform until the vacuum level stabilizes.

[0060] Furthermore, the vacuum pump's extraction valve is briefly closed at the end of the high-temperature platform, and the vacuum level is observed to see if it recovers rapidly. If the vacuum level remains stable within a specified time and there is no visible condensate in the exhaust channel area, the release of volatiles is considered essentially complete, and the process proceeds to the next step. If the vacuum level recovers or continuous condensation bands appear in the exhaust channel area, the process returns to the medium-temperature platform to extend the holding time. If necessary, the heating rate is reduced to avoid the instantaneous release of volatiles and the resulting shock. After the vacuum seal is established, the pressure holding and observation phase begins. If the vacuum level recovers beyond the preset recovery threshold within a fixed observation time, a leak is considered to exist. The recovery threshold is determined by the resolution of the press vacuum gauge and the reference stable section data, preferably 5 to 20 times the resolution.

[0061] To reduce the subjectivity of the termination criterion, the vacuum level over time can be recorded using a simple differential operation: the difference in vacuum level between adjacent time points is taken and compared with a preset threshold. If the difference is consistently lower than the threshold multiple times, the system is considered stable. This differential operation does not rely on a complex solver and can be implemented on the press's built-in controller or an external data acquisition device. If an external data acquisition device is used, time synchronization should be completed before the data enters the press to avoid curve alignment errors affecting the determination.

[0062] During use, preheating is performed in stages at three levels: low temperature, medium temperature, and high temperature, with continuous vacuuming. Vacuum stability and the condensation state in the exhaust channel area are used as criteria for terminating preheating. Volatile substances are fully released before pressurization and discharged along the exhaust channel, reducing the risk of jetting during pressurization. Resin wetting is completed within a controlled window, and the resin storage tank / cavity maintains an effective volume before pressurization, thus preserving subsequent storage capacity.

[0063] After preheating, the resin has wettable ability but has not yet been forced to flow at high speed. At this point, the pressurization method determines whether the overflow can be effectively controlled by the capillary fracture step and the resin storage tank / cavity from step one. If high pressure is applied directly, the resin will form a shear flow in the gap between the window sidewalls and diffuse circumferentially, and may even cross the step inflection point to enter the top edge of the heat sink. If the pressure rise rate is too low, the resin will remain in a viscoelastic state for too long, and the peripheral sealing structure will soften due to stress relaxation and produce slight warping. Therefore, segmented pressurization combined with ramp pressurization distributes the resin flow driving force in the order of interface first and then overflow. First, a low-pressure section is applied to allow the resin to complete the final wetting in the gap between the sidewalls and expel residual gas. Then, a medium-pressure ramp section is applied to promote interface filling and allow the overflow to slowly enter the resin storage tank / cavity. Finally, a short pause is set before reaching the target pressure to allow the resin to complete self-leveling in the resin storage tank / cavity without generating secondary impact.

[0064] In this process, while maintaining a vacuum, a low pressure is applied to transition the laminated blank from a pre-compressed state to a compressed state; preferably, the low pressure is 0.2 MPa to 0.6 MPa, and is maintained until the displacement of the press tends to slow down, indicating that the wetting in the gap is nearing completion.

[0065] The pressure is increased to 1.0 MPa to 2.5 MPa in a linear ramp manner, with a short pause at the end of the ramp. A linear ramp means increasing the pressure in fixed increments within adjacent control cycles, so that the pressure changes approximately linearly with time, thereby avoiding instantaneous resin gushing caused by step loading. For plate types with small sidewall gaps or preset acceptance values ​​close to the threshold, a gentler ramp and longer pauses should be used to allow the overflowing resin time to enter the resin storage tank / cavity instead of forming a thin layer of backflow on the outside of the step.

[0066] As an example, after the press reaches the preheating termination criterion, the operator selects the pressing formula with low-pressure start-ramp-pause-target pressure on the control panel, presses the start button, and observes the pressure display rise steadily from the low-pressure section. During the pause at the end of the ramp section, no continuous bubbles can be seen escaping from the edge exhaust channel area through the observation window, indicating that the gas in the gap has been basically exhausted. If continuous bubbles are observed in the channel area, manual pressure holding to prevent the current pressure from rising further is immediately executed, and the vacuuming time is extended until the bubbles disappear before resuming ramp-up. The press can be a hydraulic vacuum press or a servo press, or equivalent equipment, as long as it can execute segmented pressure formulas and record the pressure-time curve.

[0067] In use, the final wetting and venting within the gap are first completed under vacuum at a low pressure. Then, the pressure is increased linearly and paused at the end, allowing the overflow resin to slowly enter the resin storage tank / cavity and be locked in its destination. The resin filling sequence is limited to wetting first and then overflowing. The capillary fracture step is more likely to maintain the fracture effect and reduce the thin layer of overflow resin crossing the inflection point. The resin storage tank / cavity obtains continuous absorption time in the slope section, and the main volume of overflow resin is more stably contained in the resin storage area rather than at the top edge of the heat sink.

[0068] After segmented pressurization and filling, the resin needs to undergo curing and setting at the target pressure and temperature to transform it from a flowable state to a cross-linked cured state and lock in the interface morphology. If the pressure fluctuates significantly during the curing stage, it can cause micro-displacement of the resin near the gel point, potentially resulting in micro-cracks in the periphery sealing structure. If the pressure is released too early, the resin may rebound and create interfacial micro-voids before it has developed sufficient strength. If cooling is too rapid, the difference in thermal expansion between the metal substrate and the insulating thermally conductive sheet will generate shear stress at the periphery of the protective layer, increasing the risk of subsequent peeling failure.

[0069] Therefore, curing, cooling and depressurization are defined as a continuous time sequence, and the process curves are archived to generate a dedicated derived decision quantity for step two, which is used to transfer the stability of the pressing process to the peeling and detection strategy in step three.

[0070] First, the resin is cured and held under the target pressure to complete cross-linking and form a stable bonding interface. Then, while keeping the pressure constant, the temperature is lowered to an intermediate temperature range to increase the resin modulus before gradually depressurizing. Finally, after depressurization, the temperature is lowered to near room temperature and the vacuum is released to release the lamination stress in a controlled sequence.

[0071] Maintain the target temperature of 160°C to 200°C under a target pressure of 1.5 MPa to 3.5 MPa. The holding time depends on the curing system of the insulating thermally conductive sheet or PP insulating sheet. For insulating thermally conductive sheets containing more fillers, the holding time should be appropriately extended to avoid uneven curing caused by internal temperature lag.

[0072] First, the temperature is lowered to 120 to 160 degrees Celsius before the pressure is released in stages. The pressure is released in multiple small steps, allowing the interface to rebound and release at a higher modulus stage. After the pressure is released, the temperature is lowered to below 100 degrees Celsius before the vacuum is released and the cover is opened to remove the plate. This prevents the perimeter sealing structure from being damaged by sudden airflow and displacement impacts in a low-temperature, high-modulus state.

[0073] Furthermore, define the compression stability judgment quantity. The three-level judgment value is used to mark the peeling and detection strength of the pressed plate when it enters step three: When the vacuum degree does not abnormally rebound after passing the preheating termination criterion, the pressure curve does not exhibit excessive jumps in the ramp and curing holding sections, the temperature curve does not show obvious overshoot in the curing holding section, and no continuous overflow or breakthrough marks appear in the outer periphery glue accumulation area after pressing, the pressing stability judgment value is recorded as stable; when any of these conditions occur once but are eliminated by extending preheating or extending the pause and the final curve returns to smoothness, the pressing stability judgment value is recorded as questionable; when there is a continuous rebound, continuous jump, or systematic breakthrough marks in the glue accumulation area, the pressing stability judgment value is recorded as unstable. Specifically, regarding the pressing stability judgment value... When no abnormalities of vacuum, pressure, and temperature occur, and there are no continuous overflow marks on the outside of the glue storage tank / cavity after the board is removed, the pressing stability is determined to be stable. When any of the above abnormalities occurs once but is eliminated by extending the preheating or extending the stop at the end of the ramp, and there are no continuous marks after the board is removed, the pressing stability is determined to be suspicious. When any of the above abnormalities persists or there are continuous marks, the pressing stability is determined to be unstable.

[0074] The temperature profile is based on the target temperature of the curing and holding section. If a peak value exceeding the target temperature occurs after entering the curing and holding section, and the peak value exceeds the preset multiple of the stability accuracy of the temperature control system, it is determined to be a temperature overshoot.

[0075] The criteria for determining stability, suspicion, and instability are based on process records of vacuum level, pressure, and temperature, as well as visual inspection records after plate removal. These records, linked to the plate number, serve as inputs for setting the sampling frequency and tightening strategy for the peeling temperature window in step three. If the pressing stability is determined to be unstable, the pressed plate directly enters the enhanced testing path in step three, prioritizing a low-damage cleaning and re-inspection fault-tolerant strategy. If necessary, it will be deemed unqualified and exit the test in step three.

[0076] During use, first cure and maintain at the target pressure and temperature, then remove the plate in the order of cooling first, then depressurizing, and then releasing the vacuum. Based on the process curve and appearance record, generate a pressing stability judgment value. .

[0077] The sequence of curing, cooling, and depressurization is fixed, the interface morphology is locked in the high modulus stage, and the peripheral sealing structure is less prone to cracking during depressurization. The process curves and the appearance of the peripheral rubber accumulation zone are archived and transformed. This provides traceable input for the stripping and inspection in step three, reducing the variability caused by experience-based adjustments.

[0078] Step 3: Without introducing laser adhesive removal, grinding adhesive removal, or strong acid etching as the main process, the contaminant is transferred from the top surface of the heat sink to the peelable protective layer through controlled temperature window peeling. Then, machine vision is used to convert the absence of residue into an inspectable threshold judgment. When the threshold is exceeded, low-damage cleaning / activation and re-inspection are used to achieve closed-loop exit.

[0079] The top surface of the heat sink of the cured laminate is still covered by a peelable protective layer at the end of the pressing process, and there is excess cured adhesive in the outer adhesive reservoir / cavity. If the peelable protective layer is removed directly at room temperature, brittle tearing can easily occur between the protective layer and the excess cured adhesive, leaving discontinuous edges on the top surface of the heat sink. If it is removed at excessively high temperatures, the release interface of the peelable protective layer will soften and stick together, making it easier for the contaminant layer to remain on the top surface of the heat sink.

[0080] Therefore, the peeling action is limited to an operable temperature window, and the peeling angle, peeling speed, and the morphology of the initial tear are incorporated into the same action chain. Then, the top surface of the peeled heat sink is captured by a camera with a fixed optical path to obtain the residual adhesive coverage and compare it with the release threshold. When the limit is exceeded, mechanical grinding or high-energy removal is not used. Instead, low-damage cleaning / activation is used to restore the surface state and re-inspect, thereby forming a traceable closed loop within the mass production cycle.

[0081] After the cured laminate is removed from the press, a temperature gradient still exists within the board. The difference in thermal conductivity between the heat sink and the surrounding insulating thermally conductive sheets will cause the top surface of the heat sink to cool down either first or last. If peeling begins directly, the interface fracture mode of the protective layer will be inconsistent in different areas, resulting in localized residue or tearing of the protective layer. Therefore, the cured laminate needs to be reheated before peeling to ensure that the top surface of the heat sink is within a controllable temperature window during peeling.

[0082] First, the reheating method is determined based on the pressing stability judgment value: when the pressing stability judgment value is stable, conventional reheating is used; when the pressing stability judgment value is questionable or unstable, the reheating is extended and additional surface temperature verification points are added. Subsequently, during the reheating process, the temperature of the top surface of the heat sink and the temperature of the back of the board are collected simultaneously to avoid inferring solely from the cavity temperature.

[0083] The cured and pressed plate is placed on a constant temperature platform or a hot air circulation box. The constant temperature platform uses an aluminum alloy heating plate and an isolation pad is laid under the plate to prevent indentation. The hot air circulation box uses an air supply structure with a baffle to avoid local direct blowing. Temperature measurement is performed by attaching a contact thermocouple to the non-functional area on the outer side of the top surface of the heat sink, or by using an infrared thermometer to measure the center and four corners of the top surface of the heat sink. Either temperature measurement method outputs a recordable temperature sequence.

[0084] The continuous time that the temperature of the top surface of the heat sink remains between its upper and lower limits is used as the trigger condition for entering the peeling station, and the completion of temperature equalization is indicated by the temperature difference between the top and back surfaces not exceeding a preset value. When the temperature difference exceeds the limit, the temperature continues to recover until the temperature difference converges, to avoid inconsistent interface fracture caused by entering the peeling process when the temperature difference is large. The peeling temperature window is preferably 50℃ to 90℃, more preferably 60℃ to 80℃. Below the lower limit, it is easier for the surface to become brittle, and above the upper limit, it is easier for the surface to stick together. The temperature difference between the top and back surfaces is preferably no more than 10℃.

[0085] To reduce the arbitrariness of window settings, the side wall gap in step one can be adjusted. Capillary fracture step height The ratio of the effective volume of the adhesive reservoir to the expected volume of overflow adhesive is used as an empirical reference: the smaller the sidewall gap and the larger the capillary fracture step height, the more concentrated the cured adhesive overflow is in the adhesive reservoir / cavity, and the peel window can be selected lower to avoid adhesion; when the sidewall gap is larger or the peripheral sealing structure gap is more likely to appear, the peel window can be selected higher to reduce the risk of brittle fracture. The above reference does not change the physical definition of the window, but is only used to shorten the test piece exploration cycle when changing production lines.

[0086] During use, the laminated board is first reheated and cured using a constant temperature platform or hot air circulation box, and the temperature sequence of the top surface of the heat sink is collected. Then, the temperature is set at the window and the temperature difference between the top and back surfaces converges as the peeling trigger condition. The temperature gradient on the top surface of the heat sink is converged, and the fracture mode of the protective layer interface tends to be consistent within the board, reducing the number of local tear initiation points. The peeling entry condition is triggered by the recordable temperature sequence and temperature difference judgment, avoiding batch drift caused by relying solely on feel or appearance.

[0087] Furthermore, the peelable protective layer forms a continuous cover on the top surface of the heat sink, but the peripheral sealing structure compacts the edges of the protective layer. If the edges are pried off directly with fingers, gaps are easily created at the peripheral sealing structure. Once a gap is created, it will expand along the fiber direction and tear the protective layer. Therefore, it is necessary to first construct a controllable initial tear, and then use a clamp to maintain a stable angle and speed to complete the peeling. First, a short cut is made on the non-functional area on the outer side of the top surface of the heat sink using a cutting tool, so that the protective layer has graspable tongues; then the tongues are clamped in the peeling clamp and pulled at a fixed angle, so that the interface fracture is pushed along the top surface of the heat sink; during the pushing process, it is observed at the same time whether there are continuous adhesive lines and particle accumulation on the back of the protective layer to confirm the contamination migration path; if the protective layer breaks, the pulling is stopped immediately and the tongues are reconstructed according to the fracture location to prevent the broken fragments from falling onto the top surface of the heat sink.

[0088] The cutting tool uses a rounded-tip scriber, with a cutting depth that only penetrates the protective layer without damaging the surface of the heat sink. The cutting length is controlled within the range where the tongue can be clamped. For plates using shallow groove embedding for sealing, the cutting position is selected on the outside of the shallow groove to avoid disrupting the sealing continuity. Specifically, when forming the circumferential sealing line, the sealing fixture is a ring-shaped elastic pressure head with a contact surface width of 0.2mm to 0.8mm. The sealing action closes around the outer circumference of the heat sink, with the start and end points overlapping to form a continuous compacted band. After sealing, the sealing line is scanned circumferentially using side or oblique light. If any uncompacted sections or breaks are found, the sealing is deemed incomplete, and the sealing is reworked or the peelable protective layer is reapplied.

[0089] The peeling fixture uses a pneumatic fixture or a servo fixture. The end of the fixture is equipped with a soft pad to prevent damage to the tongue. The traction angle is kept in a large angle range so that the interface breaks and is pushed forward in the peeling form. The traction speed is kept stable and decelerated at the corner to avoid sudden speed changes that could cause brittle fracture. The traction angle is preferably 150° to 180° and the traction speed is preferably 20 mm / s to 150 mm / s.

[0090] The peeling direction is preferably opposite to the opening direction of the glue storage tank / cavity, so that the cured glue line in the glue storage area is more easily carried away rather than pushed towards the top surface of the heat sink.

[0091] As an example, the operator places the cured and pressed board, which has undergone reheating, into the positioning frame of the peeling station. The positioning frame fixes the long side of the board in the same direction. The operator marks a tongue on the outer side of the top surface of the heat sink, clamps the tongue with a clamp, and presses the traction button. The clamp moves along the guide rail and maintains a fixed angle. The protective layer is continuously pulled up from the top surface of the heat sink. A ring of cured adhesive is visible on the back of the protective layer, and the top surface of the heat sink has a continuous metallic luster. If the protective layer is partially torn at the end of the guide rail, the operator stops traction and marks a new tongue at the torn edge to continue peeling until the protective layer is completely removed. The removed protective layer is then affixed to a collection card with a board number for traceability.

[0092] In operation, a cutting tool is first used to construct the initial tearing tongue in the non-functional area. Then, a clamp is used to maintain the angle and speed for continuous peeling, stopping the reconstruction of the tongue upon breakage and continuing the process. The initial tearing moves the peeling starting point from the edge of the peripheral sealing structure to a controllable position, reducing the probability of damage to the peripheral sealing structure. A stable angle and speed allow the contaminant layer to migrate with the protective layer and reduce fragment fallout, facilitating subsequent machine vision identification of residual adhesive against a fixed background.

[0093] The top surface of the heat sink is a reflective metal surface, and residual adhesive often appears as a thin film or strip. If the optical path and camera position are not fixed, reflected highlights may be misjudged as adhesive residue or missed altogether. Therefore, a ring light source with a fixed incident angle is needed in conjunction with a polarizer to suppress specular reflection. Then, the top surface of the heat sink is imaged within a fixed field of view, and the adhesive coverage is obtained through a verifiable image processing workflow.

[0094] First, the camera is geometrically calibrated to ensure that the top surface of the heat sink is fixed in the field of view. Then, the original image is acquired and flat field correction is performed to eliminate uneven light source brightness. Next, edge detection is used to obtain an effective area mask of the top surface of the heat sink, and color or grayscale threshold segmentation is performed within the mask to obtain candidate regions for residual adhesive. Finally, connected component filtering is performed on the candidate regions to remove isolated noise points and output the residual adhesive coverage rate.

[0095] The camera is fixed on the inspection table after the stripping station. The ring light source and the normal of the top surface of the heat sink maintain a fixed angle. The polarizer is placed in front of the light source and the lens. The calibration plate is used to determine the ratio of pixels to actual length and lock the center position of the top surface of the heat sink in the image.

[0096] Flat field correction involves acquiring background images without a board and normalizing the original images pixel by pixel; edge detection uses a first-order gradient operator to generate an edge intensity map, and then uses a closing operation to fill in edge breakpoints to form an effective region mask; threshold segmentation uses an adaptive threshold to compress the surface brightness variations of different batches; connected component filtering uses the lower limit of area to filter dust points and the upper limit of aspect ratio to filter scratch images, outputs residual adhesive coverage, and binds the image to the board number for archiving.

[0097] The effective area mask is (Obtained through edge detection and closing operation). The candidate mask for residual adhesive is... (Obtained through adaptive thresholding, opening operation, and connected component filtering). Therefore: Protective layer integrity rate: After peeling, the peelable protective layer is attached to a collection card with a plate number, its image is acquired, and a protective layer mask is obtained. The expected protective layer mask is (Generated from a cut pattern template or a template image from the same batch that does not use a protective layer). Then: symbol (Mask of the effective area on the top surface of the heat sink) (Residual adhesive candidate mask) (Number of pixels in the effective area) (Measured protective layer mask) (Expected protective layer mask).

[0098] When the pressing stability determination is doubtful or unstable, the testing station adds a side supplementary light acquisition in the same field of view, and uses the union of the two segmentation results as the candidate area for residual glue to avoid accidental omissions by a single optical path; correspondingly, the release threshold is tightened and the board number is marked as a key backtracking set that needs to enter step four.

[0099] Define the criteria for achieving the decontamination and stripping standard. This is used to summarize residual adhesive coverage, protective layer integrity, and temperature window deviation into a single judgment metric, which facilitates correlation with release thresholds and rework thresholds. Where: the amount required to meet the decontamination and stripping standards : Characterizes the overall compliance level of temperature window stripping and decontamination results, with a value ranging from 0 to 1; residual adhesive coverage. : The ratio of the area of ​​the residual adhesive candidate region to the effective area of ​​the top surface of the heat sink, with a value of 0 to 1; the residual adhesive coverage rate is the ratio of the sum of the areas of the residual adhesive candidate regions at the pixel level to the pixel area of ​​the effective area of ​​the top surface of the heat sink. The residual adhesive candidate regions are obtained by threshold segmentation, morphological opening operation for noise reduction, and connected component filtering. Protective layer integrity The measured area of ​​the peelable protective layer after removal is the ratio of its expected area, ranging from 0 to 1. The peelable protective layer after removal is attached to a collection card with a board number. The collection card is placed under a fixed light source and a fixed camera to acquire images. The outline of the protective layer is obtained through edge detection, and the pixel area within the outline is calculated as the measured area. The expected area is calculated from the template outline of the same batch that does not use the protective layer or from the protective layer cut pattern. The protective layer integrity rate is the ratio of the measured area to the expected area. Temperature window deviation : The normalized deviation of the heat sink top surface temperature from the preset peeling temperature window during peeling; the value ranges from 0 to 2; when the temperature falls within the window, it is 0, and when the temperature exceeds the upper or lower limit of the window, it is 0. The excess amount is divided by the window width and capped at 2. During peeling, the temperature of the heat sink top surface is obtained by a thermocouple or infrared thermometer at a fixed position, with the measurement point located at the center of the heat sink top surface. The measurement time is the last sampled value before the peeling fixture begins traction, serving as the input for calculating the temperature window deviation. Under the above definition, is a bounded attenuation term used to compress the influence of temperature deviation on the peeling result to a range within which comparability between different plate types is achieved.

[0100] In use, a verifiable image is first obtained using a fixed optical path and polarization imaging to generate the residual adhesive coverage rate. Then, based on the residual adhesive coverage rate, the protective layer integrity rate, and the temperature window deviation, the peeling and decontamination compliance judgment quantity is calculated and archived. The residual adhesive coverage rate is generated by a fixed optical path and a verifiable processing flow, and specular reflection is suppressed to reduce release deviations caused by misjudgments.

[0101] When the amount of material removed to meet the decontamination standard is below the threshold, the cause may be residual debris from a torn protective layer or film residue due to temperature window deviation. If removed directly by grinding or strong acid, the morphology and chemical state of the heat sink's top surface will be significantly altered, making it difficult to maintain consistency in subsequent welding wetting and assembly.

[0102] Therefore, by adopting a low-damage cleaning / activation and re-inspection diversion, the processing actions are limited to the removal of films and particles, and the number of cleaning cycles is limited to a controllable upper limit.

[0103] First, locate the cause based on the components of the stripping and decontamination compliance determination quantity: when the protective layer integrity rate is low, prioritize removing fragments and check the edges of the surrounding sealing structure; when the temperature window deviation is large, prioritize activating to disrupt the continuity of the film and then wiping it again; when the residual adhesive coverage rate is high and the protective layer integrity rate is normal, it is determined to be an overflow of adhesive or traces of overflow from the adhesive storage tank / cavity, and enters the enhanced sampling inspection and exits if necessary.

[0104] The cleaning process involves two methods: First, a standard-pressure plasma cleaning gun is used to sweep the top surface of the heat sink. The sweeping path follows parallel lines while maintaining a fixed gun spacing. After cleaning, a lint-free cloth dampened with electronic-grade isopropyl alcohol is used to gently wipe in the same direction, and finally, the surface is dried with clean air. Alternatively, a low-power radio frequency plasma chamber can be used to treat the entire board, with immediate cleaning after exiting the chamber. In either method, mechanical abrasion contact is avoided. The nozzle distance of the standard-pressure plasma cleaning gun is 5mm to 20mm, the sweeping speed is 50mm / s to 200mm / s, and the number of sweeps is one or two. After cleaning, the surface is gently wiped with electronic-grade isopropyl alcohol and then dried.

[0105] After cleaning / activation, return to the camera station to re-inspect the residual adhesive coverage and recalculate the peeling and decontamination compliance judgment quantity; if the re-inspection is still below the threshold, it is judged as unqualified and exited, and the board number is marked as peeling rework failure. At the same time, the peeling temperature record, peeling direction record, and pressing stability judgment quantity are written into the traceability card; if the re-inspection reaches the threshold, it is released and the cleaning path and number of times are recorded for step four to correlate and analyze the rework ratio and incoming material batch.

[0106] To establish a verifiable test path, controls can be set up during the development phase: for the same plate type, peeling inside and outside the temperature window can be performed respectively, and the changing trends of the protective layer integrity rate and residual adhesive coverage rate can be recorded; for the same plate type, cleaning with atmospheric pressure plasma and re-inspection without cleaning can be performed respectively, and the changing trends of the peeling and decontamination compliance criteria can be recorded.

[0107] During use, first perform low-damage cleaning / activation with plasma cleaning and gentle wiping with isopropyl alcohol. Then, recalculate the peeling and decontamination compliance threshold according to the camera's re-inspection and release or exit based on the threshold. Cleaning / activation is limited to the removal of thin films and particles, maintaining a stable morphology on the top surface of the heat sink to avoid introducing grinding grooves and dust residue. The re-inspection brings the rework results back to the same judgment standard, with clear exit conditions to prevent the same board number from circulating and lingering on the production line.

[0108] Step 4: Convert the clean interface and test record of the heat sink top surface generated in Step 3 into executable release / isolation / exit procedures, and back-map the procedure results to the preset boundary in Step 1 and the pressing curve in Step 2 to form a closed-loop control for sustainable operation.

[0109] In step three, the peelable protective layer has already been removed and the residual adhesive coverage has been determined. Whether the surface condition is ready for delivery is also related to the stability of the lamination process, fluctuations in the flowability of incoming batches, and deviations in the peeling temperature window. Without the closed-loop treatment in step four, the production line often can only be stopped after the defect has spread, and it is difficult to distinguish which of the following chain of causes is causing adhesive leakage due to the sealing gap, gushing due to the lamination curve, or tearing due to the peeling conditions, resulting in repeated trial and error.

[0110] After the stripped and inspected boards exit from step three, the next process is usually circuit etching and surface treatment. Therefore, a clear release or isolation decision needs to be made before they enter the site, and this decision, along with the board number, incoming batch, lamination curve, and inspection images, must be archived. Using the same board number throughout the entire process is a constraint to avoid mixing qualified and unqualified boards on the turnover rack, which could break the traceability chain.

[0111] First, at the exit position, read the board number and retrieve the residual adhesive coverage, protective layer integrity rate, and temperature window deviation records archived in step three. Simultaneously, read the pressing stability judgment quantity and vacuum / temperature / pressure curve index archived in step two. Then, read the peelable protective layer batch number, the thermally conductive insulating layer sheet or PP insulating layer sheet batch number, and the sealing notch ratio and alignment deviation records archived in step one. Finally, write these fields into the same record entry and generate an entity identifier consistent with the treatment result. The record entry also includes the exit timestamp, operator number, and testing station number, and uses the test image file name and pressing curve file name as index fields, enabling direct location of the original data from the board number later.

[0112] If barcode scanning verification fails, no record entry is generated and a rescan is prompted; if consecutive failures occur, the board is forcibly isolated and temporarily stored to prevent unidentified boards from entering the next process. Entity identification uses paper labels or laser marking information for on-site sorting and subsequent sampling and positioning. Board number reading and record entry generation: The board number is read using a barcode scanner, and after verification of a certain length, it is written into a record entry. A cyclic redundancy check (CRC) code is used to delete misreads. The CRC code typically uses CRC-16 or CRC-32. The check code is generated from the board number string using a specific polynomial and stored along with the board number. During reading, it is calculated, compared, and misreads are deleted.

[0113] For physical sorting and preventing mixed placement: release boards are labeled with release tags and placed into release turnover racks, isolation boards are labeled with isolation tags and placed into isolation turnover racks, and exit boards are labeled with exit tags and placed into exit turnover racks; the three types of turnover racks are physically separated and placed in separate areas. Before putting items on the racks, operators should double-check that the label color matches the turnover rack color.

[0114] As an example, the operator places the stripping detection board into the scanning station. The scanner reads the board number and the printer outputs a disposal label. When the label indicates release, the board is placed into the release rack. When the label indicates isolation, the board is placed into the isolation rack and an isolation list is attached. When the label indicates exit, the board is placed into the exit rack and an exit list is registered.

[0115] The results visible on-site are as follows: the edge labels of the inner panels of the three types of turnover racks are of the same color, and the panel number can be traced back to the detection image and the pressing curve index in the record entry.

[0116] In use, the board number is first read and a record entry with a checksum is generated. Then, a physical label is generated according to the disposal status, and sorting and secondary confirmation are completed in the physical partition turnover rack. Release, isolation, and exit are implemented as visible sorting actions, avoiding traceability chain breaks caused by mixed storage during turnover. The board number, inspection image, pressing curve index, and incoming batch field are written into the same record entry, and the cause can be traced back level by level from the board number.

[0117] Judging release solely based on the residual adhesive coverage of a single piece can easily lead to situations where individual pieces are acceptable while the entire batch is drifting. Drifting often manifests gradually as an increase in the isolation ratio and the number of rework and re-inspections. To trigger review early and pinpoint drift to actionable steps, a closed-loop handling level is defined, and control limit trigger paths are provided. This ensures that the handling level determines not only the sorting destination but also whether a line stoppage for review is necessary and the priority of the backtracking path. This is specifically designed for the closed-loop handling level. When step three determines that the product is to be released and no cleaning rework occurs, the closed-loop handling level is the release level; when step three determines that the product is to be released but a cleaning rework occurs and the product only meets the standard after re-inspection, the closed-loop handling level is the isolation level; when the control limit triggering condition is met, the closed-loop handling level is the stop-line level; when step three determines that the product is to be withdrawn or multiple peeling and breakages occur for the same board number, the closed-loop handling level is the withdrawal level.

[0118] First, based on the release / rework / exit judgment and the component of the stripping and decontamination compliance judgment in step three, the single piece treatment level is given: release without cleaning and rework is recorded as release level, release but only meets the standard after cleaning and rework re-inspection is recorded as isolation level, and exit or multiple peeling and breakage of the same plate number is recorded as exit level.

[0119] Subsequently, within the same incoming batch window, the handling level sequence of consecutive board numbers is counted. If the isolation level or exit level reaches the preset control limit, a line stop review is triggered, and the batch window is marked as a drift window. The counting uses a combination of continuous counting and sliding counting: continuous counting captures sudden anomalies, and sliding counting captures gradual drift. The sliding counting window is automatically maintained by the manufacturing execution system in ascending order of board number. The closed-loop handling level of each board within the window comes from the recorded entries. For early identification of gradual drift, an exponentially weighted moving average can be optionally applied to the isolation level occurrence sequence, giving higher weight to recent isolation events, thus indicating risks before continuous counting is triggered. The weight coefficient decreases with board number and is written into the control limit version number. The sliding counting window length is preferably 20 to 100 consecutive board numbers, and the window length is determined according to the production line cycle time and incoming batch segmentation. The isolation level ratio threshold is determined by the upper limit of the isolation level ratio of historically stable batches, preferably 1.2 to 2.0 times this upper limit as the trigger threshold. The exponentially weighted moving average uses the closed-loop handling level sequence as input and adopts a recursive form. The weighting coefficients, ranging from 0.1 to 0.5, represent the closed-loop processing level corresponding to the th plate. As a drift warning quantity, it is used to provide early warning of risks.

[0120] Among them, the definition and recording of closed-loop treatment levels are as follows: the closed-loop treatment level takes discrete values. Where 0 represents the release level, 1 represents the isolation level, 2 represents the stop level, and 3 represents the exit level; when the single piece is handled at the isolation level or the exit level, the trigger reason field is written into the record entry. The trigger reason field is limited to one of the following four categories: excessive residual glue coverage, insufficient protective layer integrity, excessive temperature window deviation, and traces of glue accumulation area breakthrough.

[0121] To control the triggering of limits and the drift window marking: when the continuous isolation count reaches 3 to 5 or the proportion of isolation levels within the sliding window reaches a preset proportion, the closed-loop handling level is set to the stop-line level and a stop-line verification instruction is triggered. The stop-line verification instruction includes four actions: press machine seal inspection, press curve playback verification, peeling station temperature window verification, camera optical path cleaning, and calibration board re-photographing. These four actions are executed in a fixed order. The stop-line verification instruction is issued in the form of an electronic work order. The work order items include the board number range, incoming material batch number, trigger reason field statistics, and check code. After completing each action, the executor signs for each item in the work order and uploads on-site photos or curve screenshots as evidence.

[0122] In practice, single-piece decisions are first mapped to discrete closed-loop handling levels and written into the restricted cause field. Then, continuous counting and sliding counting are performed within the batch window to trigger stop-line review and mark the drift window. The handling level unifies release sorting and stop-line review into a single standard, and drift identification relies on recordable count triggers rather than empirical observation. The cause field is limited to four types of executable root causes, and feedback adjustments can be traced back along a fixed causal chain without being diluted by irrelevant descriptions.

[0123] If a production line stoppage or isolation occurs, simply recording the event without adjustment will cause the drift to recur after production resumes. Blindly adjusting the process may disrupt the synergistic boundary between steps one and two, reducing the ability to control the flow of excess adhesive. Therefore, the backtracking attribution is limited to a three-segment chain: structural boundary – pressing sequence – peeling condition. Priority backtracking paths are selected based on the trigger cause field, ultimately outputting a process adjustment work order that can be reproduced on-site.

[0124] First, determine the backtracking entry point based on the trigger cause field: If the trigger cause field indicates insufficient protective layer integrity, prioritize backtracking the continuity of the perimeter sealing structure in step one and the peeling angle / speed record in step three; if the trigger cause field indicates excessive temperature window deviation, prioritize backtracking the reheating equipment and temperature measurement points; if the trigger cause field indicates excessive residual adhesive coverage but normal protective layer integrity, prioritize backtracking the segmented pressurization ramp and preheating termination criteria in step two; if the trigger cause field indicates a breakthrough mark in the adhesive storage area, prioritize backtracking the capillary fracture step height. Effective volume and alignment deviation of the glue storage tank After the backtracking is completed, a four-tuple consisting of the action requiring adjustment, the adjustment range, the verification method, and the recovery conditions is generated and written into the process adjustment work order. Before the process adjustment work order is issued, the process engineer must preview it on the press machine recipe interface to confirm that the adjusted pressure, temperature, and time will not exceed the equipment's rated range, and record the recipe version number after execution so as to associate it with the subsequent treatment level reduction.

[0125] Among them, the proportion of sealing gaps is extracted from the record entries. Alignment deviation Sidewall gaps Capillary fracture step height The ratio of the effective volume of the glue storage tank to the expected overflow volume, and the start and end timestamps of the slope segment of the pressure curve in step two; the pressure curve is interpolated using cubic spline interpolation as in step one to generate an equally spaced sequence, and then a first-order difference is taken to identify whether there are abrupt changes in the slope segment. The pressure curve is based on the set value of the pressing formula. If the actual pressure in the slope segment exceeds the step increment within the controller sampling period, and the increment exceeds the preset multiple of the target increment of the slope segment, it is determined to be a pressure jump; if the actual pressure continues to exceed the upper or lower limit of the formula for a preset duration, it is determined to be a pressure over-limit.

[0126] For work order output and review: When tracing back to the pressing sequence, the work order requires breaking down the ramp section into more smaller segments and adding a pause at the end of the ramp, while reviewing the termination criteria for the vacuum preheating and exhaust section; when tracing back to the structural boundary, the work order requires replacing the insulating and thermally conductive sheet with one that has a larger effective volume of the adhesive storage tank or adjusting the depth of the step forming die to improve... The work order requires verification of the pre-set qualification threshold in step one. When the proportion of the sealing gap approaches the threshold, it also requires tightening the sealing continuity check in step one and cleaning the clamping fixture of the peripheral sealing structure to avoid local discontinuities caused by foreign objects. When tracing back to the peeling condition, the work order requires changing the fixture padding material and locking the traction angle within the predetermined range, while verifying the temperature measurement point. After the work order is executed, the recovery condition must be the image of residual adhesive coverage taken again along the same optical path and the treatment level dropping. The stop level cannot be lifted if the recovery condition is not met.

[0127] When in use, the structural boundary and pressing timing evidence are first extracted according to the trigger cause field to form a retrospective conclusion. Then, a process adjustment work order containing the action, amplitude, verification method and recovery conditions is output and the recovery is confirmed with the same detection caliber. Retrospective attribution is limited to three causal chains to avoid the destruction of the collaborative boundary due to adjustment without evidence. Interpolation and difference transform the continuity of the pressing ramp section into verifiable evidence. Process adjustment is based on curve characteristics rather than verbal description.

[0128] If the detection system drifts, the closed loop will misinterpret the detection drift as process drift and trigger unnecessary line stoppages. Detection drift is commonly caused by lens contamination, light source aging, changes in polarizer angle, and calibration plate wear. Meanwhile, control limits that are not updated for a long time can easily become either too strict or too lenient. To address this, executable calibration and control limit fixing actions are added, and the conditions for resuming production and exiting are written as verifiable clauses. First, the detection system calibration is triggered at the line stop level. If the calibration passes, process backtracking begins; if the calibration fails, the detection system is restored first, and then the inner panel of the isolation turnover rack is re-inspected.

[0129] After calibration, the control limits are updated based on the handling level and trigger reason fields within the drift window. The control limit update only allows changes to the continuous isolation count threshold and the sliding window length. Finally, when the exit conditions are met, the stop level is lifted and production is resumed, and the control limit version number is written to the record entry.

[0130] The calibration route and calibration evidence include: re-photographing the calibration board, which contains fixed patterns of high-reflection and low-reflection areas. After re-photographing, the edge position and grayscale distribution are obtained according to the algorithm. If the edge position offset exceeds the set pixel threshold or the grayscale distribution produces bimodal distortion, the calibration is deemed to have failed, and lens cleaning, light source replacement, or polarizer reset is performed. The calibration time, operator, and calibration board number are recorded in the record entries.

[0131] To control the solidification and exit conditions: the solidification of control limits adopts a temporary-then-solidification sequence. First, the continuous isolation count threshold is temporarily lowered to verify whether drift is suppressed. Then, when a number of consecutive board numbers are in the release level and the trigger reason field is empty, the new threshold is solidified. When the stop level is not triggered, lens wiping and calibration board re-shooting are performed according to a preset cycle, preferably within hours, to intercept detection drift in the early stage. The exit conditions are limited to three conditions being met simultaneously: all re-inspected boards on the isolation turnover rack are re-judged to the release or exit level and sorted; the sealing inspection of the press machine and the verification and recording of the press curve playback are completed; and the detection system calibration is passed and calibration evidence is generated. After the exit conditions are met, the stop level is lifted. After lifting, the first batch of board numbers enters the stricter sampling inspection until the sampling inspection is passed, and then the regular sampling inspection rhythm is resumed.

[0132] In use, the detection system is first calibrated and archived using a calibration board and lens / light source / polarizer adjustments. Then, after verification against a temporary threshold, the control limits are solidified, and production is resumed after a stoppage is lifted using three exit conditions. Detection system drift is independently identified and eliminated beforehand to prevent detection errors from being mistaken for process anomalies and triggering ineffective adjustments. Control limit updates are limited to two parameters, adapting to batch fluctuations while maintaining continuous judgment criteria, facilitating long-term trend traceability.

[0133] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0134] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0135] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0136] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0137] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for manufacturing a high-efficiency heat dissipation and thermoelectric separation metal substrate based on insulation layer optimization, comprising aligning and laminating a copper substrate, an insulating thermally conductive film or PP film, and a circuit copper foil layer, characterized in that: It also includes, Clean the top surface of the copper substrate heat sink, apply a peelable protective layer and form a perimeter sealing structure; form capillary fracture steps and glue storage tanks or glue storage cavities around the opening of the hole to obtain the laminated blank. The laminated blanks are sequentially vacuum preheated and degassed, then filled under pressure in sections and cured to obtain a cured press-fit plate. Within the preset peeling temperature window, peel the peelable protective layer at the set peeling angle and peeling speed. Use machine vision to obtain the residual adhesive coverage and determine the threshold. If the threshold is exceeded, clean and re-inspect. If the threshold is still exceeded, exit. The threshold determination is associated with the vacuum degree, temperature and pressure curve records and the batches of peelable protective layer, insulating thermal conductive film or PP film to generate release, rework or scrap records, and the pressing sequence or capillary fracture step, glue tank or glue storage cavity and peripheral sealing structure are adjusted according to continuous exceeding of threshold or peeling failure. The capillary fracture steps are continuously arranged along the periphery of the scooping hole and are located opposite to the side wall of the heat sink. The glue storage tank or glue storage cavity is arranged circumferentially along the outer side of the capillary fracture steps and is connected to the outer peripheral glue storage area. When the laminated blanks are aligned, the glue storage tank or glue storage cavity is located outside the top surface of the heat sink. The capillary fracture steps introduce abrupt changes in cross-section on the climbing path, so that the resin thin layer must cross the step inflection point to climb, thereby raising the crossing threshold and guiding the resin to the outside of the step. The steps cooperate with the subsequent glue storage tank or glue storage cavity. The steps are responsible for disconnecting the climbing, and the glue storage tank or glue storage cavity is responsible for catching the guided overflow glue.

2. The method for manufacturing a high-efficiency heat dissipation thermoelectric separation metal substrate according to claim 1, characterized in that: The peelable protective layer is formed by stacking a heat-resistant support layer, a seepage barrier layer and a release layer. When applied, it covers the top surface of the heat sink and extends to the perimeter of the opening of the vent. The heat-resistant support layer is attached to the top surface of the heat sink, and the release layer is located on the side away from the top surface of the heat sink. A peeling start end is reserved at the edge of the peelable protective layer.

3. The method for manufacturing a high-efficiency heat dissipation thermoelectric separation metal substrate according to claim 2, characterized in that: The perimeter sealing structure is a continuous circumferential sealing line formed around the perimeter of the peelable protective layer. The circumferential sealing is closed once before the laminated blank enters the pressing front, and the start and end sections overlap. After sealing, the circumferential sealing line is inspected and the results are recorded. If there is a sealing gap, the sealing is reworked and redone.

4. The method for manufacturing a high-efficiency heat dissipation thermoelectric separation metal substrate according to claim 3, characterized in that: When forming the perimeter sealing structure, a shallow annular groove is machined on the perimeter of the top surface of the heat sink. The perimeter of the peelable protective layer is embedded into the shallow annular groove and compacted and fixed. Before the pre-pressing of the stacked blanks, the inner edge of the shallow annular groove is checked circumferentially to ensure there is no warping or breakage. If any abnormality is found, the groove is reworked and re-embedded.

5. The method for manufacturing a high-efficiency heat dissipation thermoelectric separation metal substrate according to claim 4, characterized in that: The vacuum preheating and degassing process includes a low-temperature platform for vacuuming and pressure holding, a medium-temperature platform for heating and maintaining temperature, and a high-temperature platform before entering the segmented pressurization and filling stage. The segmented pressurization and filling stage includes a low-pressure pre-compression stage, a linear ramp pressurization stage, and a ramp end pause stage, and maintains a set pressure and set temperature in the curing and shaping stage.

6. The method for manufacturing a high-efficiency heat dissipation thermoelectric separation metal substrate according to claim 5, characterized in that: The low-temperature platform temperature is 80°C to 120°C, the medium-temperature platform temperature is 120°C to 160°C, and the high-temperature platform temperature is 160°C to 200°C; the low-pressure pre-compression section pressure is 0.2 MPa to 0.6 MPa, and the linear ramp pressurization section increases the pressure from 0.1 MPa to 0.5 MPa per minute to 1 MPa to 2.5 MPa per minute.

7. The method for manufacturing a high-efficiency heat dissipation thermoelectric separation metal substrate according to claim 6, characterized in that: The peelable protective layer is peeled off when the peeling temperature window is 50 degrees Celsius to 90 degrees Celsius. Before peeling, a peeling start end is formed at the edge of the peelable protective layer. During peeling, the peeling angle is 150 degrees to 180 degrees and the peeling speed is 20 mm to 150 mm per second. The peeling process is recorded with the plate number as an association.

8. The method for manufacturing a high-efficiency heat dissipation thermoelectric separation metal substrate according to claim 7, characterized in that: Machine vision acquires images of the top surface of the heat sink under fixed light source illumination and generates an effective area mask. After performing flat field correction, edge detection, threshold segmentation and morphological operations on the image, candidate areas for residual adhesive are obtained. The residual adhesive coverage rate is calculated and judged according to the threshold. If the value exceeds the threshold, plasma cleaning is performed and re-inspection is carried out.

9. The method for manufacturing a high-efficiency heat dissipation thermoelectric separation metal substrate according to claim 8, characterized in that: The residual adhesive coverage threshold judgment result is bound to the vacuum degree, temperature and pressure curve records and the batch of peelable protective layer, insulating thermal conductive film or PP film, and a cyclic redundancy check code is used for verification when writing the board number. When consecutive threshold exceedances or stripping failures occur, adjust the ramp pressurization section of the segmented pressurization filling and the vacuum preheating and exhaust time, and calibrate the machine vision execution cycle.

Images