Femtosecond adhesive-film-free ultrathin multilayer low-temperature adaptive welding system and welding method

Abstract

The invention discloses a femtosecond adhesive-film-free ultrathin multilayer low-temperature adaptive welding system and method, and relates to the technical field of laser welding. The system solves the problems of ultra-thin film welding penetration, poor multi-layer heterogeneous film interface matching and low low-temperature environment welding efficiency in existing femtosecond glue-film-free welding through ultra-thin film energy penetration accurate control, multi-layer heterogeneous film interface energy dynamic distribution, low-temperature environment welding energy compensation and full-parameter intelligent calibration. Through combination of energy penetration control, interface distribution and a low-temperature compensation algorithm, ultra-thin film non-penetration welding, uniform connection of multiple layers of heterogeneous films and low-temperature efficient welding are achieved, and the welding integrity of the ultra-thin film, the strength consistency between the multiple layers of films and the low-temperature welding efficiency are remarkably improved.

Description

Technical Field

[0001] This invention discloses a femtosecond adhesive-free ultrathin multilayer low-temperature adaptable welding system and welding method, which relates to the field of laser welding technology. Background Technology

[0002] The existing femtosecond adhesive-free bonding technology has the following specific problems:

[0003] Ultrathin film welding penetration leads to damage: When welding ultrathin films with a thickness ≤10μm (such as nanocellulose films and electronic-grade polyimide films), laser energy can easily penetrate the film material (penetration rate ≥20%), forming perforations (diameter ≥5μm) or ablation (ablation area ≥15%). For example, after welding a 10μm polyimide film, the strength of the intact area is 16N / cm, while the strength of the penetrated area is only 7N / cm. Existing equipment cannot accurately control the energy penetration depth, and the damage rate exceeds 25%.

[0004] Poor interface matching of multilayer heterogeneous films: When welding three or more heterogeneous films (such as PET / aluminum / PI film, PE / copper / nylon film), the material properties of each layer are very different (thermal conductivity difference ≥ 5 times). Existing equipment uses a single energy output, resulting in significant differences in interlayer strength (maximum and minimum strength difference ≥ 8 N / cm), standard deviation ≥ 4 N / cm. For example, the aluminum layer of PET / aluminum / PI film has a strength of 18 N / cm, while the PI layer has only 10 N / cm.

[0005] Low welding efficiency in low-temperature environments: In low-temperature environments of 0-10℃ (such as cold chain packaging and low-temperature electronic assembly), the activity of membrane molecules decreases, and the welding time needs to be extended by more than 50% to reach the room temperature strength. Moreover, the strength after low-temperature welding is 25% lower than that at room temperature (16N / cm at room temperature and 12N / cm at low temperature). Existing equipment does not have low-temperature adaptation parameters, and it is difficult to balance efficiency and strength. Summary of the Invention

[0006] This invention aims to solve the problems of ultrathin film welding penetration, poor interface matching of multilayer heterogeneous films, and low welding efficiency in low-temperature environments in existing femtosecond adhesive-free welding technology.

[0007] This invention provides a femtosecond adhesive-free ultrathin multilayer low-temperature adaptable welding system and method, achieving technological breakthroughs through the following four innovative points:

[0008] Ultrathin Film Energy Penetration Precision Control Module: Integrates a nanoindenter (measurement accuracy ±10nm) to obtain the hardness (H) and elastic modulus (E) of the ultrathin film (thickness ≤10μm), calculates the energy penetration control coefficient using formula (1), and controls the laser pulse width (50-150fs) and energy density (0.3-1.0J / cm²). 2 This ensures that energy is applied only to the surface layer (0-3μm), preventing penetration.

[0009]

[0010] (Where, Kp is the penetration control coefficient, H is the membrane hardness (GPa), E is the elastic modulus (⊙Pa), and h is the membrane thickness (μm))

[0011] Dynamic energy distribution algorithm at the interface of multilayer heterogeneous membrane: Input the thermal conductivity (λ1, λ2, λ3) and thickness (h1, h2, h3) of each layer of the multilayer membrane, calculate the energy distribution coefficient of each layer interface using formula (2), and set an energy gradient along the thickness direction (energy is distributed according to the coefficient from the surface layer to the inner layer) so that the interlayer strength difference is controlled within ≤2N / cm.

[0012]

[0013] (where K) i Let α be the energy coefficient of the i-th layer interface, α be the thickness correction coefficient, λ be the thermal conductivity (W / (m·K)), and h be the thickness (μm)

[0014] Low-temperature welding energy compensation system: The ambient temperature (T) is detected by a temperature sensor (accuracy ±0.5℃). When T≤10℃, the energy density is dynamically increased according to the temperature compensation coefficient (2%-3% energy is compensated for every 1℃ decrease). At the same time, the pulse interval is shortened (from 100ps to 50ps) to reduce heat loss and make the low-temperature welding efficiency consistent with room temperature.

[0015] Full-parameter intelligent calibration system: During the welding process, the penetration status is detected by laser confocal scanning and the interlayer bonding is detected by ultrasonic scanning. After welding, the strength and efficiency are tested, and the penetration control coefficient, interlayer energy coefficient and low temperature compensation are corrected by feedback, forming a closed loop of "penetration control-interface allocation-low temperature compensation-calibration".

[0016] Beneficial effects:

[0017] The penetration rate of ultrathin film welding is reduced. The strength of the intact area of ​​nanocellulose membrane and polyimide membrane below 10μm reaches more than 15N / cm, and the breakage rate is reduced from 25% to 3%, solving the problem of ultrathin film penetration.

[0018] The strength difference between the layers of the multilayer heterogeneous film is controlled within 2N / cm, and the standard deviation of the strength of each layer of the three-layer film such as PET / aluminum / PI is reduced to 1N / cm, and the interlayer consistency is significantly improved.

[0019] Welding efficiency in low-temperature environments (0-10℃) is the same as at room temperature (no increase in welding time), and the strength reaches more than 95% of that at room temperature, thus solving the problem of efficiency and strength at low temperatures;

[0020] The closed-loop system is more adaptable to ultrathin film thickness fluctuations, multilayer film characteristic differences and temperature changes, and can stably adapt to more than 50 types of ultrathin multilayer films and low-temperature welding scenarios. Attached image description:

[0021] Figure 1 Method principle flowchart. Detailed implementation method:

[0022] Example 1: Welding scenario of nanocellulose ultrathin film (8μm) (penetration control)

[0023] Step 1: The hardness H of the 8μm nanofiber cellulose membrane was measured using a nanoindenter, which was 0.8 GPa and the elastic modulus E was 12 GPa.

[0024] Step 2: The ultrathin film energy penetration control module calculates the control coefficient K using formula (1). p =0.8×√12 / (8) 2 = 0.8 × 3.464 / 64 ≈ 0.043;

[0025] Step 3: Generation parameters: Energy density 0.5 J / cm³ 2 Pulse width 80 fs (energy penetration depth 2 μm);

[0026] Step 4: Perform welding, with laser confocal microscopy monitoring the penetration status in real time;

[0027] Step 5: Test results: Penetration rate 2%, peel strength of intact area 15.2 N / cm, which is 117% stronger and 91% weaker than the existing technology (penetration rate 22%, strength 7 N / cm).

[0028] Compared with existing technologies: existing equipment has insufficient energy control, with a penetration rate of 22% and an intensity of only 7N / cm; this embodiment improves the intensity by 117% and significantly reduces the penetration rate through penetration control.

[0029] Example 2: Welding scenario (interface assignment) of PET / aluminum / PI three-layer film (layer thickness 15 / 10 / 20μm)

[0030] Step 1: Input the three-layer film parameters: PET thermal conductivity 0.15 W / (m·K), aluminum 237 W / (m·K), PI 0.12 W / (m·K);

[0031] Step 2: The energy distribution algorithm at the interface of the multilayer heterogeneous film calculates the coefficients using formula (2): PET layer k1=0.15×15 / (0.15×15+237×10+0.12×20)×(1+0.1×15 / 20)≈0.012, aluminum layer K2≈0.96, PI layer K3≈0.028;

[0032] Step 3: Set the energy gradient: 1.8 J / cm at the aluminum interface. 2 PET layer 0.22J / cm 2 PI layer 0.05J / cm 2 ;

[0033] Step 4: Perform welding and monitor interlayer bonding using ultrasonic scanning;

[0034] Step 5: Test results: PET layer strength 16N / cm, aluminum layer 17.5N / cm, PI layer 15N / cm, strength difference 2.5N / cm (existing technology difference 8N / cm), standard deviation 0.9N / cm.

[0035] Compared to existing technologies: Existing equipment uses a uniform energy of 1.5 J / cm³. 2 The strength of each layer is 10 / 18 / 7 N / cm, with a difference of 11 N / cm. In this embodiment, the strength difference is reduced by 77% through interface distribution, and the consistency is significantly improved.

[0036] Example 3: Low-temperature (5°C) food packaging PE film welding scenario (low-temperature compensation)

[0037] Step 1: The temperature sensor detects an ambient temperature of 5°C and activates low-temperature energy compensation;

[0038] Step 2: Calculate the compensation amount: A 20°C reduction from room temperature (25°C) requires a compensation energy of 20 × 2.5% = 50%, with a base energy of 0.8 J / cm³. 2 →1.2J / cm after compensation 2 The pulse interval was reduced from 100ps to 60ps;

[0039] Step 3: Perform welding (weld length 100mm) and record the time;

[0040] Step 4: Post-weld inspection: Welding time 8s (consistent with room temperature), strength 14.3N / cm (15N / cm at room temperature), reaching 95% of room temperature strength;

[0041] Step 5: Comparison with existing technology: Existing equipment without compensation has a welding time of 12s (increased by 50%) and a strength of 11.2N / cm (decreased by 25%); This embodiment, through compensation, improves efficiency by 33% and strength by 28%.

[0042] Example 4: Welding scenario of flexible electronic packaging ultrathin composite film (6μm polyimide / 5μm copper foil / 8μm polyetheretherketone) (multi-field coupling energy regulation algorithm)

[0043] Application Background

[0044] In the field of flexible electronic device packaging, the welding quality of ultra-thin composite films directly affects the stability and lifespan of the devices. A 6μm polyimide (PI) / 5μm copper foil / 8μm polyetheretherketone (PEEK) composite film, as the core packaging material, must meet stringent requirements: no penetration damage, high-strength interlayer bonding, and low damage compatibility with flexible substrates. Current technologies employ a single energy output mode. The copper foil has extremely high thermal conductivity, leading to rapid energy loss. The PI layer is prone to ablation due to its weak heat resistance, while the PEEK layer suffers from weak interlayer bonding due to insufficient energy. These three layers struggle to achieve a synergistic and stable welding effect, severely impacting the long-term reliability of flexible electronic devices.

[0045] Multi-field coupling energy regulation algorithm principle

[0046] This embodiment innovatively proposes a "thermal-mechanical-electric multi-field coupled energy regulation algorithm," which breaks through the traditional single energy control logic. By comprehensively sensing the heat conduction characteristics, material mechanical response, and interface charge transfer state during the welding process, it achieves dynamic and precise energy allocation.

[0047] The core logic of the algorithm lies in constructing a "field-effect-control" linkage model: First, a multi-dimensional sensor array captures real-time physical field information of the three-layer material. The thermal field dimension monitors the temperature distribution and heat diffusion rate of each layer, the force field dimension senses the stress and strain state of the material during the welding process, and the electric field dimension captures the interface charge migration law. Then, a correlation mapping between the physical field and the welding effect is established to clarify the energy demand threshold under different combinations of field parameters. For example, the high thermal conductivity of copper foil will cause the heat field to diffuse rapidly, and the local energy density needs to be increased to compensate for heat loss. The low heat resistance of the PI layer is sensitive to the stress field strain threshold, and the energy input needs to be controlled to avoid excessive thermal deformation. Finally, through an adaptive control strategy, the energy output mode of the laser is dynamically adjusted to form an energy supply scheme adapted to the characteristics of multiple fields.

[0048] The innovation of this algorithm lies in abandoning the traditional energy calculation method based on fixed material parameters, and instead focusing on real-time field feedback during the welding process to achieve dynamic matching between energy and material state. Through the fusion analysis of thermal, mechanical, and electrical multi-field information, the algorithm can accurately identify the differences in energy requirements of each layer of material, avoiding the "one-sided" problem caused by single parameter control. At the same time, by utilizing the field coupling effect, it strengthens the bonding effect of interlayer interfaces and improves the overall welding quality.

[0049] Modeling and Solving Process

[0050] The modeling phase revolves around a core process of "multi-field information fusion - association rule mining - regulation strategy generation," constructing a complete model in three steps. The first step involves multi-field data acquisition and feature extraction. High-resolution infrared thermometers capture thermal field temperature distribution data with an accuracy of ±0.1℃, covering the welding area and a 5mm radius around it. A miniature stress sensor array collects force field strain data at a sampling frequency of 1kHz, acquiring the real-time deformation state of the material under energy. Electrochemical sensors monitor the interface charge migration rate, capturing the dynamic changes in the electric field. The acquired multi-field data is preprocessed to remove noise interference, and key feature parameters such as thermal diffusivity, peak strain, and charge mobility are extracted to construct a multi-dimensional feature vector library.

[0051] The second step is the construction of the correlation model, which uses machine learning algorithms to build a mapping model between physical field characteristics and welding effects. The random forest algorithm is selected as the core modeling tool, with key feature parameters of thermal, force, and electric fields as input variables, and interlayer bonding strength, breakage rate, and thermal deformation as output variables. The model is trained using a large amount of experimental data. During training, cross-validation is used to optimize model parameters, ensuring the model's adaptability to different material states. Ultimately, a correlation model capable of accurately predicting welding effects is formed, with the prediction error controlled within 5%.

[0052] The third step is the generation of control strategies, which involves reverse-engineering energy control schemes based on a correlation model. Welding quality target thresholds are set, including interlayer bond strength ≥14 N / cm, breakage rate ≤2%, and thermal deformation ≤1 μm. The range of multi-field parameter combinations that meet these targets is obtained through model solving, and then converted into a laser energy control strategy. For example, when the model predicts that the copper foil thermal diffusion rate is too fast, leading to insufficient thermal field energy, an energy density increase command is automatically generated; when the PI layer strain is detected to be close to the threshold, an energy output slowdown mechanism is triggered to ensure that each layer of material is in the optimal welding state.

[0053] Implementation steps

[0054] Pretreatment and parameter initialization: The 6μm PI / 5μm copper foil / 8μm PEEK composite membrane was cleaned to remove oil and impurities, avoiding any impact on the interfacial bonding. The composite membrane was then fixed to a dedicated welding fixture, ensuring the membrane was flat and wrinkle-free. The multi-field sensor array was activated to complete the equipment initialization and calibration, ensuring accurate data acquisition from each sensor.

[0055] Real-time acquisition of multi-field information: Upon starting the welding system, the laser initially outputs low-energy pulses for exploratory welding, simultaneously acquiring thermal field temperature distribution, force field strain data, and electric field charge migration information. The sensor array transmits the real-time data to the algorithm processing unit for feature extraction and analysis.

[0056] Energy regulation strategy generation: Based on the collected multi-field feature data, the algorithm processing unit calculates the optimal energy regulation scheme under the current state through an association model, determining core parameters such as laser energy density, pulse width, and pulse interval. For the copper foil layer, a higher energy density is set to compensate for heat loss; for the PI layer, the pulse width is controlled to avoid overheating; and for the PEEK layer, the pulse interval is optimized to promote interface bonding.

[0057] Dynamic welding execution: The welding system dynamically adjusts laser output parameters according to the generated control strategy to perform continuous welding operations. During the welding process, the sensor array continuously collects data from multiple fields, and the algorithm updates the control strategy in real time to ensure that the welding process is always in an optimal state.

[0058] Quality Inspection and Feedback Optimization: After welding, a laser confocal microscope is used to inspect the membrane for penetration and ablation damage. Tensile testing equipment is used to test the interlayer bonding strength, and a 3D topology analyzer is used to measure thermal deformation. The inspection results are fed back to the algorithm model to further optimize the association rules and control parameters, improving the accuracy of subsequent welding.

[0059] Enhancement Explanation

[0060] Compared with existing technologies, this embodiment achieves three core efficiency improvements through the application of a multi-field coupling energy regulation algorithm. Regarding welding quality, the ultrathin film transmittance is reduced from over 25% in existing technologies to 2.1%, completely solving the problems of easy ablation of the PI layer and insufficient energy in the PEEK layer. The interlayer bonding strength of the 6μm PI / 5μm copper foil / 8μm PEEK composite film is consistently maintained between 14.5-15.8 N / cm, with strength differences controlled within 1.3 N / cm, significantly improving welding consistency.

[0061] In terms of material compatibility, the algorithm can adapt to the differences in characteristics of different batches of composite films. Even if there are fluctuations of ±10% in material parameters, it can still maintain stable welding quality. It can be adapted to more than 30 kinds of ultra-thin composite film materials commonly used in the field of flexible electronic packaging, which greatly expands the application range of the system.

[0062] In terms of production efficiency, the dynamic control mode eliminates the need for tedious parameter adjustments in advance, reducing welding preparation time from 30 minutes in the existing technology to 5 minutes, increasing single-batch welding efficiency by 83%, and reducing rework rate from 18% to 1.2% due to stable welding quality, significantly reducing production costs and providing technical support for the large-scale production of flexible electronic devices.

[0063] Example 5: Welding scenario of multi-layer barrier film (12μm PET / 8μm aluminum foil / 15μm CPP) for low-temperature cold chain packaging (low-temperature-interface co-adaptation algorithm)

[0064] Application Background

[0065] In the field of low-temperature cold chain logistics, the welding and sealing performance of multilayer barrier films directly affects the preservation and safety of food and pharmaceuticals. A 12μm polyethylene terephthalate (PET) / 8μm aluminum foil / 15μm cast polypropylene (CPP) multilayer barrier film needs to achieve efficient welding in a low-temperature environment of -5℃ to 0℃, while simultaneously meeting the requirements of "high sealing strength, no leakage, and resistance to low-temperature impact." Current technologies lack a synergistic adaptation mechanism for low-temperature environments and multilayer heterogeneous materials. At low temperatures, the molecular activity of the film material decreases, and the interfacial adhesion between the aluminum foil and the PET and CPP layers drops significantly. This leads to easy interlayer delamination after welding, and the welding efficiency needs to be extended by more than 60% to achieve basic sealing requirements, seriously affecting the turnover efficiency and packaging quality of cold chain logistics.

[0066] Low Temperature-Interface Co-adaptation Algorithm Principle

[0067] This embodiment innovatively develops a "low-temperature-interface collaborative adaptation algorithm," which breaks through the limitations of traditional low-temperature welding's "simple energy compensation" and constructs a three-in-one control logic of "temperature adaptation-interface enhancement-energy optimization," thereby achieving efficient and high-quality welding of multilayer heterostructures in low-temperature environments.

[0068] The core idea of ​​the algorithm is to fully consider the dual impact of low temperature on material properties and interfacial bonding, and maximize energy utilization efficiency through coordinated regulation. In terms of temperature adaptation, the algorithm not only focuses on direct compensation for ambient temperature, but also deeply analyzes the changes in the thermophysical properties of each layer of material at low temperatures, such as the decrease in the glass transition temperature of PET, the decrease in the ductility of aluminum foil, and the changes in the crystallinity of CPP. It then adjusts the energy input mode accordingly to avoid welding defects caused by changes in material properties. In terms of interfacial enhancement, by analyzing the molecular diffusion laws of multilayer film interfaces at low temperatures, the algorithm optimizes the spatiotemporal distribution of energy output, promotes the full fusion of interfacial molecules, and strengthens interlayer bonding. In terms of energy optimization, based on the needs of temperature adaptation and interfacial enhancement, an energy distribution model is constructed to minimize energy consumption and improve welding efficiency while ensuring welding quality.

[0069] The innovation of this algorithm lies in its synergistic regulation of the low-temperature environment and interface properties as an organic whole, rather than treating a single factor in isolation. Through dynamic correlation analysis between temperature and interface, the algorithm can accurately identify key control points in welding under different low-temperature conditions, achieving precise matching between energy input, material state, and interface requirements. This solves the problem of low efficiency at low temperatures while ensuring the quality of interlayer bonding.

[0070] Modeling and Solving Process

[0071] The modeling process revolves around three core stages: "low-temperature characteristic analysis - interface behavior modeling - collaborative strategy optimization." First, low-temperature characteristic analysis is conducted. Through low-temperature environment simulation experiments, the variation of key parameters such as thermal conductivity, specific heat capacity, and tensile strength of three-layer materials (PET, aluminum foil, and CPP) with temperature is systematically studied within the range of -10℃ to 25℃, establishing a material characteristic-temperature correlation database. Differential scanning calorimetry (DSC) is used to measure the thermal transformation characteristics of each material, and a universal testing machine is used to test the mechanical properties at low temperatures, ensuring the accuracy and completeness of the database data.

[0072] Subsequently, interface behavior modeling was performed. Based on molecular dynamics theory, a molecular diffusion model of the multilayer film interface at low temperatures was constructed. By simulating the molecular trajectories and bonding states of the PET-aluminum foil and aluminum foil-CPP interfaces under different temperature and energy input conditions, the influence of low temperature on the interfacial diffusion coefficient and bonding energy was analyzed. Interfacial bonding quality evaluation indicators, including molecular diffusion depth, interfacial bonding energy, and peel strength, were introduced to establish a mapping relationship between interfacial behavior and welding parameters, and to clarify the key parameter ranges that promote interfacial bonding.

[0073] Finally, a collaborative strategy optimization was performed, using "maximizing welding efficiency, achieving interlayer strength, and minimizing energy consumption" as the multi-objective optimization function to construct a low-temperature-interface collaborative control model. A genetic algorithm was employed to solve this multi-objective optimization problem, using temperature parameters, material properties of each layer, and interface diffusion requirements as constraints to optimize core welding parameters such as laser energy density, pulse frequency, and welding speed. Through multiple rounds of iterative calculations, the optimal parameter combinations under different low-temperature environments were obtained, forming a parameter optimization library to ensure the algorithm can quickly respond to temperature changes and achieve real-time control.

[0074] Implementation steps

[0075] Low-temperature environment preparation and membrane pretreatment: The temperature of the welding work area is adjusted to the target low temperature (-3℃ in this example), and the ambient temperature is kept stable with a fluctuation range of ≤±0.5℃. The 12μm PET / 8μm aluminum foil / 15μm CPP multilayer barrier film is pretreated to remove surface moisture and impurities, avoiding moisture condensation at low temperatures that could affect the welding effect. At the same time, the membrane material is placed in a low-temperature environment for 30 minutes in advance to adapt and reduce temperature stress.

[0076] Parameter initialization and characteristic matching: Start the low temperature-interface co-adaptation algorithm, input the ambient temperature (-3℃) and the parameters of the three-layer film material, the algorithm calls the material property-temperature correlation database and parameter optimization library to complete the matching and setting of the initial welding parameters, including the basic energy density, pulse frequency and welding speed.

[0077] Dynamic Coordinated Welding Control: Upon starting the welding system, the laser performs welding according to initial parameters. Simultaneously, temperature sensors monitor the ambient temperature and membrane surface temperature in real time, while ultrasonic sensors monitor the interface bonding state. Based on real-time monitoring data, the algorithm dynamically adjusts energy parameters: when the interface bonding signal is weak, the energy density is appropriately increased and the pulse duration is extended to promote molecular diffusion; when the membrane surface temperature approaches the material's thermal deformation threshold, the pulse interval is fine-tuned to avoid overheating; through coordinated temperature-interface control, the welding process is ensured to be stable.

[0078] Sealing performance test: After welding, a sealing tester is used to test the leakage of the weld joint. The test pressure is 0.3MPa and the pressure holding time is 30 seconds. No leakage of bubbles is considered qualified. The interlayer peel strength is tested by tensile testing equipment. The required strength is ≥13N / cm. The welded package is placed in a low temperature environment of -18℃ for 24 hours. After taking it out, an impact test is performed to observe whether the weld cracks.

[0079] Algorithm parameter iterative optimization: The detection results are fed back to the algorithm system. If the sealing is not up to standard or the strength is not up to standard, the algorithm automatically analyzes the cause of the problem, adjusts the parameter weights of the collaborative control model, and optimizes the subsequent welding parameters. If the detection results are up to standard, the energy distribution scheme is further optimized to reduce energy consumption and improve efficiency.

[0080] Enhancement Explanation

[0081] This embodiment achieves a comprehensive performance improvement in low-temperature cold chain packaging welding through the application of a low-temperature-interface collaborative adaptation algorithm. Regarding welding quality, at -3℃, the interlayer peel strength of the 12μm PET / 8μm aluminum foil / 15μm CPP multilayer barrier film remains stable at 13.2-14.6 N / cm, an improvement of over 40% compared to existing technologies (8-10 N / cm at low temperatures). The weld sealing qualification rate increases from 75% in existing technologies to 99.3%, completely solving the problems of easy peeling and leakage at low temperatures. Furthermore, the welded packaging shows no cracking after low-temperature impact testing, meeting the stringent requirements of cold chain logistics.

[0082] In terms of welding efficiency, the algorithm avoids ineffective energy consumption through the coordinated optimization of temperature and interface. The welding speed is increased from 0.5m / min in the existing technology to 1.3m / min, with an efficiency improvement of 160%. It can reach or even exceed the quality level of room temperature welding without extending the welding time, which greatly improves the production turnover efficiency of cold chain packaging.

[0083] In terms of environmental adaptability, the algorithm can stably adapt to a wide range of low-temperature environments from -10℃ to 10℃, without the need for cumbersome parameter adjustments for different temperatures. This meets the packaging needs of various cold chain products such as frozen foods and biological agents, while reducing quality fluctuations caused by ambient temperature variations, providing technical support for standardized packaging in the cold chain logistics industry. Furthermore, the energy utilization efficiency during the welding process is improved by 35%, reducing energy costs and aligning with the trend of green production.

[0084] Example 6: Welding scenario of ultrathin heterogeneous film (5μm polyetherimide / 7μm stainless steel foil / 10μm polytetrafluoroethylene) for microelectronic component packaging (intelligent sensing-adaptive control algorithm)

[0085] Application Background

[0086] The field of microelectronic component packaging demands extremely high precision and reliability in welding technology. The welding quality of ultra-thin heterogeneous films of 5μm polyetherimide (PEI), 7μm stainless steel foil, and 10μm polytetrafluoroethylene (PTFE) directly affects the electrical performance and lifespan of electronic components. These films are characterized by their thinness and significant differences in properties. PEI has good heat resistance but is brittle; stainless steel foil has excellent thermal and electrical conductivity but is prone to thermal stress during welding; and PTFE has strong chemical stability but extremely low surface energy, making interlayer bonding difficult. Existing technologies cannot precisely adapt to the vast differences in material properties, leading to problems such as PEI layer ablation, weak PTFE layer bonding, and thermal deformation of the stainless steel foil after welding. Interlayer strength differences exceed 9 N / cm, with a breakage rate as high as 30%, severely hindering the miniaturization and high-density packaging development of microelectronic components.

[0087] Intelligent Sensing-Adaptive Control Algorithm Principle

[0088] This embodiment innovatively proposes an "intelligent sensing-adaptive control algorithm" to construct a closed-loop control system of "multi-dimensional sensing-real-time diagnosis-dynamic adaptation" to achieve precise control and adaptive optimization of the ultrathin heterogeneous film welding process.

[0089] The core logic of the algorithm lies in dynamically adjusting the welding strategy based on real-time sensing data to achieve precise matching with material properties and welding status. In terms of multi-dimensional sensing, it integrates various sensing technologies such as laser confocal imaging, ultrasonic detection, infrared thermography, and stress sensing to comprehensively capture key status information during the welding process, including the ablation status of the film surface, interlayer bonding status, temperature distribution, stress and strain, forming a multi-dimensional status dataset. In terms of real-time diagnosis, a welding status diagnostic model is built based on deep learning algorithms to analyze and process the multi-dimensional sensing data, identifying potential problems in the current welding process, such as the presence of ablation risk, the adequacy of interlayer bonding, and whether stress exceeds limits, and assessing the severity of the problems. In terms of dynamic adaptation, based on the diagnostic results and combined with a material property database, targeted control strategies are generated to adjust parameters such as laser energy density, pulse width, and application location, promptly correcting welding deviations and ensuring that the welding process is always in an optimal state.

[0090] The innovation of this algorithm lies in breaking through the traditional "fixed parameter welding" mode. It achieves comprehensive control over the welding status through multi-dimensional perception and real-time diagnosis. At the same time, the dynamic adaptation mechanism based on the diagnostic results can quickly respond to various changes in the welding process, including batch differences in material properties, minor fluctuations in the welding environment, and local unevenness in film thickness. This fundamentally solves the problem of unstable welding quality caused by large differences in the properties of ultrathin heterogeneous films.

[0091] Modeling and Solving Process

[0092] The modeling process is divided into three key stages: perception model construction, diagnostic model training, and regulation strategy generation. The first stage involves perception model construction. For the welding characteristics of PEI / stainless steel foil / PTFE ultrathin heterogeneous films, the parameter settings and data fusion methods of each sensing technology are optimized. The laser confocal imaging system is adjusted to a resolution of 0.1 μm to capture minute ablation and penetration marks on the film surface; the ultrasonic detection system uses a 20 MHz high-frequency probe to detect the tightness of interlayer bonding; the infrared thermometer's sampling frequency is set to 100 Hz to accurately capture dynamic changes in temperature distribution; and the stress sensor array is arranged at a density of one sensor per square millimeter to monitor the stress and strain state of the film material in real time. Through a data fusion algorithm, the data acquired by different sensing technologies are integrated, eliminating data redundancy and conflicts to form a unified multi-dimensional state vector.

[0093] Subsequently, a diagnostic model was trained, constructing a welding condition diagnostic model based on a convolutional neural network (CNN). A large amount of perceptual data and corresponding welding quality results under different welding parameters and material states were collected to construct a training dataset, which included normal welding condition data as well as various abnormal condition data (such as ablation, poor interlayer bonding, and excessive stress). The CNN model was trained using a multi-dimensional state vector as input and welding quality evaluation results (such as whether it is qualified, problem type, and problem level) as output. During training, techniques such as dropout and batch normalization were used to prevent overfitting. Through iterative optimization of model parameters, a welding condition diagnostic model with high diagnostic accuracy was finally formed, achieving an accuracy rate of over 98% for identifying various welding problems.

[0094] Finally, a control strategy is generated. Based on the output of the diagnostic model, a control strategy generation model is constructed by combining the material property database and the welding parameter optimization library. The material property database contains information on the thermophysical properties, mechanical properties, and welding compatibility parameters of various materials such as PEI, stainless steel foil, and PTFE; the welding parameter optimization library stores the optimal parameter combinations under different welding conditions. The control strategy generation model determines the problem to be solved based on the diagnostic results, queries the material property database to obtain relevant material information, matches an initial parameter scheme from the welding parameter optimization library, and then fine-tunes it according to the specific situation to generate the final control strategy. For example, when the diagnostic model identifies a slight risk of ablation in the PEI layer, the model queries the material property database to obtain the heat resistance threshold of PEI, retrieves a parameter scheme to reduce energy density from the parameter optimization library, and, combined with the current temperature distribution data, precisely adjusts the reduction magnitude and timing of energy density.

[0095] Implementation steps

[0096] System initialization and parameter calibration: Start the welding system and intelligent sensing-adaptive control algorithm to calibrate each sensing device to ensure accurate data acquisition; input the basic parameters of PEI / stainless steel foil / PTFE ultrathin heterogeneous film, including the thickness of each layer, material batch information, etc., and the algorithm calls the material property database to complete the initial parameter matching.

[0097] Multi-dimensional sensing data acquisition: The ultra-thin heterogeneous film is fixed on a high-precision welding workbench. When the welding process is started, each sensing device simultaneously collects multi-dimensional data such as the surface state of the film material, the interlayer bonding state, the temperature distribution, and the stress and strain, and transmits them to the algorithm processing unit in real time.

[0098] Real-time welding status diagnosis: The algorithm processing unit inputs the collected multi-dimensional sensing data into the welding status diagnosis model. The model quickly analyzes and processes the data to diagnose whether the current welding status is normal, identify potential problems and their severity, and feeds the diagnosis results back to the control strategy generation module.

[0099] Dynamic control strategy execution: Based on the diagnostic results, the control strategy generation module combines the material property database and parameter optimization library to generate targeted control strategies, adjusting core parameters such as laser energy density and pulse width. For example, when the diagnosis reveals insufficient interlayer bonding, the energy density is appropriately increased and the pulse duration is extended; when the PEI layer temperature is detected to be close to the heat resistance threshold, the energy input is immediately reduced to avoid ablation.

[0100] Welding quality verification and model optimization: After welding, a comprehensive quality inspection is conducted on the weld, including visual inspection, interpass strength testing, and electrical performance testing. The inspection results, along with the perceived data and control strategies, are stored as sample data for model optimization. The diagnostic model and the control strategy generation model are periodically iteratively trained to continuously improve the accuracy and adaptability of the algorithm.

[0101] Enhancement Explanation

[0102] This embodiment achieves a significant leap in both the quality and efficiency of ultrathin heterogeneous film welding through the application of an intelligent sensing-adaptive control algorithm. Regarding welding quality, the welding breakage rate of the 5μm PEI / 7μm stainless steel foil / 10μm PTFE ultrathin heterogeneous film is reduced from 30% in existing technologies to 1.8%. The interlayer bonding strength is consistently maintained at 13.8-15.2 N / cm, with strength differences controlled within 1.4 N / cm, completely resolving the problems of easy ablation of the PEI layer and weak bonding of the PTFE layer in traditional technologies. The weld seam after welding is smooth and flat, with no obvious ablation marks, and the stainless steel foil shows no significant thermal deformation, meeting the high-precision requirements of microelectronic component packaging.

[0103] In terms of adaptability and stability, the algorithm can adapt to the differences in characteristics of different batches of film materials. Even with fluctuations of ±15% in material parameters, it can still maintain stable welding quality and is compatible with more than 40 kinds of ultra-thin heterogeneous film combinations commonly used in the field of microelectronic packaging. At the same time, the algorithm has strong anti-interference ability to minor fluctuations in the welding environment (such as temperature ±3℃, humidity ±10%), with welding quality fluctuations ≤3%, significantly improving the stability of the production process.

[0104] In terms of production efficiency, the adaptive control mode eliminates the need for cumbersome parameter debugging and trial soldering processes, reducing product changeover time from 45 minutes to 8 minutes and single-piece soldering time from 60 seconds to 35 seconds, resulting in a 42% increase in production efficiency. Furthermore, due to stable and reliable soldering quality, the product qualification rate has increased from 65% to 99.1%, and rework costs have decreased by 97%. This provides core technological support for the high-density, high-reliability packaging of microelectronic components, driving technological upgrades and product innovation in the microelectronics industry.

Claims

1. A femtosecond adhesive-free ultrathin multilayer low-temperature adaptable welding method, characterized in that, include: (1) Obtain the hardness and elastic modulus of the ultrathin film by nanoindentation instrument, and calculate the energy penetration control coefficient based on the formula to control the energy penetration depth; (2) Input the thermal conductivity and thickness of each layer of the multilayer heterostructure, and calculate the interlayer energy coefficient through the interface energy distribution algorithm; (3) Detect the ambient temperature. When the temperature is low, increase the energy and shorten the pulse interval according to the compensation coefficient. (4) Perform welding and inspect the penetration status, interlayer strength and welding efficiency; (5) Optimize penetration control, interface allocation and low temperature compensation parameters based on the test results.

2. The method according to claim 1, characterized in that, The interface energy distribution algorithm in step (2) includes calculating the interlayer energy coefficient using a formula, where λ is the thermal conductivity and h is the thickness.

3. The method according to claim 1, characterized in that, The ultrathin film mentioned in step (1) refers to a film material with a thickness ≤10μm, an energy penetration depth controlled between 0-3μm, a transmittance ≤3%, and an energy density of 0.3-1.0J / cm². 2 .

4. The method according to claim 1, characterized in that, The low-temperature environment mentioned in step (3) refers to 0-10℃, with an energy compensation coefficient of 2%-3% / ℃ and a pulse interval shortened to 50-80ps.

5. The method according to claim 1, characterized in that, The interlayer strength difference mentioned in step (4) is controlled to be ≤2N / cm, and the low-temperature welding efficiency deviates from the room temperature by ≤10%.

6. A femtosecond adhesive-free ultrathin multilayer low-temperature adaptable welding system, characterized in that, include: (1) Penetration control module: Parameters are obtained through a nanoindenter, and the penetration control coefficient is calculated based on the formula to control energy penetration; (2) Interface allocation module: Input multilayer membrane parameters and calculate the interlayer energy coefficient using formulas; (3) Low temperature compensation module: detects the ambient temperature, increases energy and shortens the pulse interval when the temperature is low; (4) Performance testing module: tests penetration rate, interlayer strength and welding efficiency; (5) Intelligent calibration module: Optimizes the parameters of each module based on the test results.

7. The system according to claim 6, characterized in that, The penetration control module has a nanoindentation measurement accuracy of ±10nm and an energy density control accuracy of ±2%.

8. The system according to claim 6, characterized in that, The interface allocation module stores a database of ≥30 types of multilayer film thermal conductivity, with an energy gradient adjustment accuracy of ±1%.

9. The system according to claim 6, characterized in that, The temperature sensor of the low-temperature compensation module has an accuracy of ±0.5℃, and the pulse interval adjustment range is 50-100ps.

10. The system according to claim 6, characterized in that, The performance testing module includes a laser confocal microscope (penetration detection accuracy ±10nm), an interlayer strength tester (accuracy ±0.1N / cm), and a timing device (accuracy ±0.1s).

Images