Solar cell, laser sintering preparation method and device of electrode thereof

Abstract

This invention relates to the field of photovoltaic cell electrode manufacturing technology, specifically disclosing a laser sintering preparation method and apparatus for solar cells and their electrodes. The method includes: placing the solar cell electrode in a laser sintering preparation apparatus; activating a MEMS gas sensor array to detect the evaporation rate of the silver paste organic carrier in real time and constructing an evaporation rate distribution model; based on the evaporation rate model, using FPGA parallel processing and adaptive PWM to adjust the laser duty cycle to reduce heat accumulation, suppress nanoscale bulging at the electrode edge, and generate a heat input spectrum. This invention constructs a evaporation rate distribution model in real time using a MEMS gas sensor array, combined with FPGA parallel processing and adaptive PWM duty cycle adjustment, dynamically compensating for uneven heat input and effectively suppressing nanoscale bulging at the electrode edge. Simultaneously, the heat input spectrum is mapped to a thickness field, and the scanning speed is optimized through Kalman filtering and PID control to achieve thickness homogenization closed-loop control, significantly improving the electrode conductivity uniformity.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic cell electrode manufacturing technology, and in particular to a method and apparatus for laser sintering preparation of a solar cell and its electrodes. Background Technology

[0002] As a crucial component of renewable energy technology, solar cells directly impact the widespread adoption and efficiency of photovoltaic power generation due to their performance and cost. Electrode formation is a key process in solar cell manufacturing, directly affecting the cell's conductivity, photoelectric conversion efficiency, and long-term stability. Laser sintering, an emerging technology, uses high-energy lasers to locally heat printed or coated metal nanoparticles, rapidly fusing them to form a continuous conductive path. Laser sintering offers advantages such as localized heating, non-contact processing, high controllability, and high energy utilization. It can significantly improve the uniformity of the electrode's microstructure and conductivity while reducing the impact of overall thermal stress on the substrate.

[0003] During the laser sintering process of solar cell electrodes, uneven volatilization rates of the organic carrier in the silver paste can easily lead to bulging at the electrode edges, which directly affects the contact resistance. Furthermore, the imbalance of surface tension in the molten silver paste can cause fluctuations in electrode thickness, further deteriorating the uniformity of conductivity. At the same time, silver particles splashed in the laser path contaminate the focusing lens, causing the sintering linewidth to become out of control and drift, ultimately compromising the electrode morphology accuracy. Summary of the Invention

[0004] In order to overcome the shortcomings of the prior art, the present invention provides a method and apparatus for laser sintering preparation of solar cells and their electrodes, which can effectively solve the problems involved in the prior art.

[0005] The objective of this invention can be achieved through the following technical solution: Firstly, this invention provides a method for laser sintering to prepare a solar cell and its electrodes, comprising the following steps: Step 1: Place the solar cell electrode in the laser sintering fabrication device, start the MEMS gas sensor array, detect the volatilization rate of the silver paste organic carrier in real time, construct a volatilization rate distribution model, and quantify the spatial distribution of the volatilization rate in real time to provide a precise basis for subsequent heat input compensation. Step 2: Based on the evaporation rate model, the FPGA performs parallel processing and adopts adaptive PWM to adjust the laser duty cycle, reduce heat accumulation, suppress nanoscale bulging at the electrode edge, and generate a thermal input spectrum. Step 3: Map the thermal input spectrum to the thickness field, predict the electrode thickness deviation through Kalman filtering, calculate the scanning speed correction coefficient, dynamically adjust the spot dwell time to balance the surface tension of the molten silver paste, and control the electrode thickness fluctuation. Step 4: The laser sintering preparation device adjusts the scanning speed in real time according to the scanning speed correction coefficient to match the residence time with the heat input, thereby achieving electrode thickness homogenization, controlling conductivity uniformity, completing the main sintering, and significantly improving conductivity uniformity. Step 5: Calculate the contamination risk index by combining the laser duty cycle and the scanning speed correction coefficient. If it exceeds the preset risk threshold, it is determined that the splashed silver particles pose a risk of contamination to the focusing lens, triggering cleaning preparation, identifying splash risks, and avoiding sudden contamination of the focusing lens. Step 6: Drive the optical path cleaner installed in the laser sintering preparation device to dynamically remove splash particles, prevent sintering linewidth drift, ensure electrode morphology accuracy, complete the preparation of solar cell electrodes, stabilize the sintering linewidth, and ensure electrode morphology accuracy.

[0006] Preferably, step 1 specifically includes: Silver paste is uniformly coated onto the solar cell substrate to form a wet film electrode. The solar cell electrode is then placed in a laser sintering fabrication device. Multiple sensing units sensitive to the volatile components of the organic carrier are then deployed in a MEMS gas sensor array. These units are arranged in a grid pattern along the length and width of the wet film electrode. The instantaneous volatile concentration signal of each grid point is collected in a spatially resolved manner, effectively capturing the real-time concentration distribution of the volatile components. The instantaneous volatile concentration signals output by each sensing unit are sent to the FPGA after analog-to-digital conversion. A sliding window mean filter is used to suppress environmental noise, extract the temporal variation characteristics of the volatile rate of each grid point, form a discrete rate matrix, effectively filter out environmental interference, accurately extract the volatile rate characteristics and quantify its spatial distribution unevenness. Based on the discrete rate matrix, a radial basis function interpolation algorithm is used to estimate the rate in areas not directly sampled, generating a evaporation rate distribution model covering the entire wet film electrode surface. This achieves high-precision estimation of the evaporation rate across the entire surface, providing fine input for adaptive laser control.

[0007] Preferably, step 2 specifically includes: The generated evaporation rate distribution model is input into the parallel computing array of the FPGA in the form of a pixelated heat map. Each computing unit corresponds to a local area of ​​an electrode and independently calculates the deviation ratio between the evaporation rate of that area and the reference rate, thereby realizing pixel-level parallel computing of evaporation rate deviation and improving compensation response speed. Each calculation unit looks up the corresponding initial correction amount of PWM duty cycle in parallel according to the deviation ratio. The larger the deviation, the larger the correction amount. At the same time, the correction amount of adjacent areas is checked for spatial consistency to avoid local over-adjustment, eliminate the thickness change caused by local abnormal correction, and ensure sintering uniformity. The correction values ​​for all regions are reduced in parallel and then output to the laser drive module to update the laser duty cycle of each sintering position pixel by pixel. This enables adaptive real-time compensation for uneven evaporation rates, suppresses heat input fluctuations caused by differences in evaporation rates, and prevents electrode edge bulging.

[0008] Preferably, step 2 further includes: In the FPGA, a higher evaporation rate sensitivity coefficient is preset for the electrode edge region than for the center region. When the evaporation rate deviation ratio of the edge region exceeds the threshold, a duty cycle attenuation factor dedicated to the edge is triggered to further reduce the duty cycle of the edge laser, effectively suppressing nanoscale bulging at the electrode edge and improving the uniformity of contact resistance. The duty cycle after edge attenuation and the duty cycle of the central region are input together into the preset thermal accumulation prediction model. The thermal accumulation prediction model estimates the cumulative thermal equivalent of the current sintering point based on the historical values ​​of the duty cycle of multiple previous scanning cycles, determines whether it is close to the bulging critical value, realizes early warning of thermal accumulation, and avoids deformation caused by local overheating. If the heat accumulation prediction model determines that the accumulated heat equivalent exceeds the critical value, the duty cycle of the current sintering point will be further reduced, and the adjusted duty cycle distribution will be used as the output heat input spectrum to dynamically balance the surface tension of the molten silver paste and stabilize the electrode thickness.

[0009] Preferably, step 3 specifically includes: An empirical mapping function is established between the laser energy density of each pixel in the thermal input spectrum and the steady-state thickness of the molten silver paste. The product of duty cycle and residence time is transformed into a predicted thickness value, generating an initial thickness field distribution, effectively establishing a quantitative correlation between energy input and thickness response. The Kalman filter algorithm is used to take the initial thickness field as the state prior estimate. Combined with the measured thickness feedback value of the previous scanning cycle, the thickness state estimate of the current cycle is iteratively updated, and the thickness prediction deviation at each position is calculated to achieve dynamic correction and noise suppression of thickness prediction. Based on the sign and magnitude of the thickness prediction deviation, the PID controller calculates the scanning speed correction coefficient. When the deviation is positive, the scanning speed is increased to shorten the dwell time; when the deviation is negative, the scanning speed is decreased to extend the dwell time, thus achieving real-time reverse adjustment of the thickness deviation.

[0010] Preferably, step 4 specifically includes: The control system of the laser sintering preparation device receives the calculated scanning speed correction coefficient in real time, multiplies it with the preset reference scanning speed, and generates the instantaneous target scanning speed for each sintering position, ensuring that each pixel receives an accurate speed command before sintering and avoiding speed lag. A speed feedforward compensation algorithm is adopted to dynamically adjust the scanning motor drive voltage of the scanning mirror according to the rate of change of the target scanning speed, so that the actual scanning speed changes continuously in space, avoiding the introduction of new thickness inhomogeneity due to speed abrupt changes, suppressing speed abrupt changes, and preventing uneven spreading of molten silver paste. During continuous scanning, the control system feeds back the actual dwell time at each position to the Kalman filter in a closed loop, which serves as the measured thickness feedback value for the next cycle. This completes the full closed-loop control of the main sintering process, achieving full closed-loop control of thickness homogenization and continuously converging the thickness deviation.

[0011] Preferably, step 5 specifically includes: The actual laser duty cycle and scanning speed correction coefficient applied at each sintering position are collected in real time. The cumulative laser energy per unit length is obtained by multiplying the two and integrating them over time. This serves as the basic characteristic quantity for the probability of generating sputtered silver particles, providing a direct and reliable input basis for risk quantification. The basic characteristic quantities are input into a pre-calibrated nonlinear risk mapping model, and the fluctuation range of duty cycle change rate and scanning speed correction coefficient are considered at the same time. The pollution risk index is output to realize a multi-factor comprehensive assessment of splash probability and improve the accuracy of early warning. The pollution risk index is compared with a preset risk threshold. If the pollution risk index exceeds the threshold for a period of time longer than the set judgment window, the optical path cleaning preparation signal is triggered, and the position of exceeding the threshold is recorded to assist in cleaning, avoid accidental triggering due to instantaneous noise, and provide accurate spatial guidance for subsequent cleaning.

[0012] Secondly, the present invention also provides a laser sintering preparation apparatus for solar cells and their electrodes, used to implement the laser sintering preparation method for solar cells and their electrodes described above. The apparatus includes a chassis, a preparation working chamber fixedly installed on the top of the chassis to provide a sealed environment for laser sintering and prevent external airflow interference, a power supply detachably installed on the top of the preparation working chamber to provide stable energy input for laser sintering, an electrode laser sintering device installed at the output end of the power supply, and an optical path cleaner connected to the output end of the electrode laser sintering device to the top of the inner cavity of the preparation working chamber for dynamically removing splash particles, preventing sintering linewidth drift, and ensuring electrode morphology accuracy. The electrode laser sintering device includes a laser, the input end of which is connected to the output end of a power supply, and the laser is fixedly installed at the top of the inner cavity of the preparation working chamber. A scanning mirror is fixedly installed at the output end of the laser. The scanning mirror receives the laser beam and adjusts the output direction according to the control system command to achieve pixel-by-pixel scanning sintering. A focusing mirror is movably installed at the output end of the scanning mirror. The focusing mirror focuses the laser beam onto the electrode surface to form a tiny spot for precise sintering. The laser beam emitted by the scanning mirror is directed onto the electrode through the focusing mirror to achieve the laser sintering operation. The optical path cleaner includes an upper fixed shell fixedly installed at the top of the inner cavity of the preparation working chamber. The upper fixed shell provides stable support for the optical path cleaner and isolates external contaminants. A lower isolation plate is detachably installed at the bottom of the upper fixed shell. The lower isolation plate and the upper fixed shell together form an isolation cavity to prevent splash particles from entering the upstream of the optical path. An auxiliary cleaning component is installed at the bottom of the lower isolation plate. The auxiliary cleaning component is used to provide air knife cleaning when the focusing lens is contaminated to restore the cleanliness of the optical path. A focusing lens conversion component for rotating and replacing the focusing lens is installed on the side of the scanning lens. When the focusing lens is contaminated, the focusing lens conversion component automatically rotates to insert a spare focusing lens to ensure continuous sintering production.

[0013] Preferably, the focusing lens conversion component includes a drive motor fixedly mounted on the top of the upper fixed housing. The drive motor receives a contamination risk trigger signal and initiates a rotational action. The output end of the drive motor passes through the upper fixed housing and rotates with the inner wall of the upper fixed housing. A connecting rod is fixedly mounted on the output end of the drive motor, which transmits the rotational torque of the drive motor to the focusing lens. The end of the connecting rod is fixed to the outer edge surface of the focusing lens, and the top and bottom of the connecting rod, the focusing lens, and the lower isolation plate slide against the bottom and top of the upper fixed housing, respectively, to ensure that the focusing lens remains stable during rotation and replacement, and to avoid optical path deviation. A barrier strip is fixedly mounted on the bottom surface of the upper fixed housing. The bottom surface of the barrier strip contacts the top of the lower isolation plate. The barrier strip dynamically seals the gap when the focusing lens rotates, preventing splash particles from entering the optical path from the gap.

[0014] Preferably, the auxiliary cleaning component includes a cleaning chamber fixedly installed on the side of the lower isolation plate. The cleaning chamber collects the blown-off splash particles to prevent them from re-contaminating the focusing lens. A pulsed airflow pipe is fixedly installed at the bottom of the cleaning chamber. The pulsed airflow pipe introduces periodic pulsed airflow to enhance the removal effect on the attached particles. One end of the pulsed airflow pipe is externally connected to an air supply device, and the other end of the pulsed airflow pipe extends into the interior of the cleaning chamber and is fixedly connected to an auxiliary air knife ring. The auxiliary air knife ring evenly distributes the airflow onto the surface of the focusing lens, forming an air knife sweeping across the lens surface. A slag discharge pipe is fixedly connected to the side opening of the cleaning chamber. The end extends to the outside of the preparation working chamber and is connected to an external waste gas treatment device. The slag discharge pipe discharges the waste gas containing splash particles from the preparation working chamber to maintain the cleanliness of the chamber. The auxiliary air knife ring is inclinedly set below the focusing lens to be cleaned. The inclined setting causes the airflow to impact the surface of the focusing lens at an angle, effectively shearing and peeling off the splash particles. A conical air-gathering nozzle is fixedly installed at the top center of the auxiliary air knife ring. The air-gathering nozzle gathers the airflow and increases the local impact force to remove stubborn deposits. Several slit nozzles are opened at the top of the auxiliary air knife ring. The slit nozzles form a flat air knife that evenly covers the entire optical area of ​​the focusing lens, ensuring that there are no dead corners in the cleaning.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. A method and apparatus for laser sintering fabrication of solar cells and their electrodes, which constructs a evaporation rate distribution model in real time through a MEMS gas sensor array, combines FPGA parallel processing and adaptive PWM duty cycle adjustment to dynamically compensate for uneven heat input, effectively suppressing nanoscale bulges at the electrode edges. At the same time, the heat input spectrum is mapped to the thickness field, and the scanning speed is optimized through Kalman filtering and PID adjustment to achieve thickness homogenization closed-loop control, which significantly improves the conductivity uniformity of the electrodes.

[0016] 2. The laser sintering preparation method and apparatus for a solar cell and its electrodes integrates the laser duty cycle and scanning speed correction coefficient to calculate the contamination risk index. When the threshold is exceeded, the optical path cleaning preparation is triggered to identify the risk of contamination by splashed silver particles. The splashed particles are dynamically removed by driving the optical path cleaner to avoid sudden contamination of the focusing lens, prevent sintering linewidth drift, stabilize the electrode morphology accuracy, and improve process reliability.

[0017] 3. The method and apparatus for laser sintering preparation of solar cells and their electrodes adopts a gridded MEMS sensor array to collect the volatile concentration signal on the surface of the wet film electrode, and generates a full-surface volatile rate distribution model through sliding window filtering and radial basis function interpolation. The FPGA calculates the rate deviation of each region in parallel and corrects the laser duty cycle by looking up a table, realizing pixel-level adaptive real-time compensation and eliminating heat input fluctuations caused by uneven volatile rates.

[0018] 4. The laser sintering preparation method and apparatus for a solar cell and its electrodes includes a focusing lens conversion component in the optical path cleaner. The focusing lens is rotated and slid by a drive motor and connecting rod, which allows for quick replacement and position switching of the focusing lens. Combined with the closed cavity formed by the upper fixed shell and the lower isolation plate, it effectively isolates the diffusion of pollutants and significantly improves the ease of maintenance of the laser sintering apparatus.

[0019] 5. The laser sintering preparation method and apparatus for a solar cell and its electrodes include a pulsed airflow pipe, an auxiliary air knife ring, and a conical gas-gathering nozzle in the auxiliary cleaning component. A high-speed pulsed air curtain is generated by the inclined slit nozzle to dynamically purge the area below the focusing lens. At the same time, in conjunction with the slag discharge pipe connected to an external waste gas treatment device, it can actively and efficiently remove the splashed silver particles on the focusing lens, avoid long-term accumulation of particles that will cause contamination of the optical path, and ensure the stability of laser focusing and the consistency of electrode sintering. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the workflow of a laser sintering method for preparing a solar cell and its electrodes according to the present invention. Figure 2 This is a schematic diagram of the structure of a laser sintering fabrication apparatus for a solar cell and its electrodes according to the present invention; Figure 3 This is a partial structural schematic diagram of a laser sintering fabrication apparatus for a solar cell and its electrodes according to the present invention. Figure 4 This is a schematic diagram of the power supply, electrode laser sintering device, and optical path cleaner of a solar cell and its electrode laser sintering preparation apparatus according to the present invention. Figure 5 This is a schematic diagram of the structure of the electrode laser sintering device and the optical path cleaner of the laser sintering preparation apparatus for a solar cell and its electrodes according to the present invention; Figure 6 This is a schematic diagram of the focusing lens and optical path cleaner of a laser sintering fabrication apparatus for a solar cell and its electrodes according to the present invention. Figure 7 This is a partial cross-sectional schematic diagram of the focusing lens and optical path cleaner of a laser sintering fabrication apparatus for a solar cell and its electrodes according to the present invention. Figure 8 This is a partial structural schematic diagram of a focusing lens and an optical path cleaner of a laser sintering fabrication apparatus for a solar cell and its electrodes according to the present invention. Figure 9 This is an exploded structural diagram of the focusing lens and optical path cleaner of a laser sintering fabrication apparatus for a solar cell and its electrodes according to the present invention. Figure 10 This is a partial cross-sectional schematic diagram of the auxiliary cleaning component of a laser sintering fabrication apparatus for a solar cell and its electrodes, according to the present invention. Figure 1 ; Figure 11 This is a partial cross-sectional schematic diagram of the auxiliary cleaning component of a laser sintering fabrication apparatus for a solar cell and its electrodes, according to the present invention. Figure 1 ; Figure 12 This is a schematic diagram of the auxiliary air knife ring, gas-gathering nozzle, and slit nozzle of a laser sintering preparation apparatus for a solar cell and its electrodes according to the present invention.

[0021] In the diagram: 1. Chassis; 2. Preparation chamber; 3. Power supply; 4. Electrode laser sintering machine; 41. Laser; 42. Scanning mirror; 43. Focusing mirror; 5. Optical path cleaner; 51. Upper fixed shell; 52. Lower isolation plate; 53. Focusing mirror conversion component; 531. Drive motor; 532. Connecting rod; 54. Auxiliary cleaning component; 541. Cleaning chamber; 542. Pulsating airflow pipe; 543. Slag discharge pipe; 544. Auxiliary air knife ring; 545. Concentrating nozzle; 546. Slit nozzle; 55. Isolation strip. Detailed Implementation

[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0023] Example 1, please refer to Figure 1 This invention provides a technical solution: a laser sintering method for preparing a solar cell and its electrodes, comprising the following steps: Step 1: The solar cell electrode is placed in a laser sintering fabrication device. Silver paste is uniformly coated onto the solar cell substrate to form a wet film electrode. A MEMS gas sensor array is activated to monitor the evaporation rate of the organic carrier in real time, constructing an evaporation rate distribution model and quantifying the spatial distribution of the evaporation rate in real time. This provides a precise basis for subsequent heat input compensation. The silver paste is uniformly coated onto the solar cell substrate to form a wet film electrode, which is then placed in the laser sintering fabrication device. Multiple sensing units sensitive to the characteristic volatile components of the organic carrier are deployed in the MEMS gas sensor array, arranged in a grid pattern along the length and width of the wet film electrode, to collect data in a spatially resolved manner. The instantaneous volatile concentration signals of each grid point are collected to effectively capture the real-time concentration distribution of volatile components. The instantaneous volatile concentration signals output by each sensing unit are sent to the FPGA after analog-to-digital conversion. A sliding window mean filter is used to suppress environmental noise, and the temporal variation characteristics of the volatile rate of each grid point are extracted to form a discrete rate matrix. This effectively filters out environmental interference, accurately extracts the volatile rate characteristics, and quantifies their spatial distribution unevenness. Based on the discrete rate matrix, a radial basis function interpolation algorithm is used to estimate the rate of areas that are not directly sampled, generating a volatile rate distribution model covering the entire wet film electrode surface. This achieves high-precision estimation of the volatile rate across the entire surface, providing fine input for adaptive laser control. It should be noted that silver paste is uniformly coated onto the surface of the solar cell substrate using a screen printing process, with the wet film thickness controlled to be 10–15 μm, forming a wet film electrode. The substrate supporting the wet film electrode is placed on the vacuum adsorption platform of the laser sintering fabrication device. A MEMS gas sensor array is integrated inside the device. This array contains 32×32 sensing units, arranged in a grid pattern with equal spacing along the X-axis (length) and Y-axis (width) of the wet film electrode. Adjacent sensing units are equally spaced. Each sensing unit is a conductivity sensor sensitive to the main volatile components in the silver paste organic carrier—terpineol and ethyl cellulose pyrolysis products. The metal-oxide-semiconductor sensor operates at temperatures between 180 and 220°C, with a response time of no more than 0.5 seconds. It spatially resolves and acquires instantaneous volatile concentration signals at each grid point in parallel, with a sampling frequency set to 200 Hz. The analog concentration signals output by each sensing unit are converted to digital values ​​by a 16-bit analog-to-digital converter and then sent to the FPGA in real-time via a high-speed serial interface. Within the FPGA, each sensing unit is independently configured with a sliding window filter, with a window length set to 10 sampling points (50 ms), to suppress noise caused by hot airflow disturbances and environmental vibrations during laser sintering. The filtered signal is processed by differential operation to extract the first derivative of the volatile concentration with respect to time at each grid point, obtaining the temporal characteristic value of the volatile rate. After continuously acquiring rate data for three complete laser scanning cycles, each cycle lasting approximately 2 seconds, the median rate value of the same grid point is taken to form a 32×32 discrete rate matrix representing the typical volatile rate of each grid point under the current process conditions. This matrix quantifies the spatial unevenness of the volatile rate distribution on the wet film electrode surface caused by differences in coating thickness or edge thermal effects. The discrete rate matrix is ​​then input into the interpolation calculation module of the FPGA, and a radial basis function interpolation algorithm is used to interpolate the areas not directly sampled. For line rate estimation, a Gaussian kernel is selected as the basis function, and its shape parameter is set to about half of the spacing between sensing units to ensure the spatial smoothness of the interpolation result and to avoid introducing spurious oscillations. The interpolation calculation is performed in the X-axis and Y-axis directions with a step size of 0.1 mm. The original 32×32 grid is expanded into a fine grid to cover the entire wet film electrode surface. After the interpolation is completed, the estimated evaporation rate value is output for each fine grid point, and cubic spline filtering is used to eliminate possible interpolation overshoot at the boundary. The final generated evaporation rate distribution model is stored in the FPGA block random access memory in the form of a two-dimensional floating-point array. Step 2: Based on the evaporation rate model, the FPGA processes the data in parallel and uses adaptive PWM to adjust the laser duty cycle, reducing heat accumulation, suppressing nanoscale bulging at the electrode edges, and generating a thermal input spectrum. The generated evaporation rate distribution model is input into the FPGA's parallel computing array in the form of a pixelated thermal map. Each computing unit corresponds to a local region of the electrode and independently calculates the deviation ratio between the evaporation rate of that region and the reference rate, realizing pixel-level parallel calculation of evaporation rate deviation and improving the compensation response speed. Each computing unit obtains the corresponding initial correction amount of the PWM duty cycle by looking up a table in parallel according to the deviation ratio. The larger the deviation, the larger the correction amount. At the same time, the correction amounts of adjacent regions are spatially consistent to avoid local over-adjustment, eliminate thickness abrupt changes caused by local abnormal correction, and ensure sintering uniformity. The correction amounts of all regions are output to the laser drive module after parallel reduction, updating the laser duty cycle of each sintering position in a pixel-by-pixel manner, realizing adaptive real-time compensation for uneven evaporation rates, suppressing thermal input fluctuations caused by differences in evaporation rates, and preventing electrode edge bulging. It should be noted that the FPGA's parallel computing array contains 320×320 computing units, each corresponding to a fine grid point on the electrode surface. Each unit independently processes the evaporation rate data for a corresponding local region of 0.1mm×0.1mm. The FPGA reads a reference rate value from the block random access memory. This reference rate is pre-determined through evaporation rate calibration experiments on defect-free silver paste samples under standard process conditions and is set to 0.85 μg·mm. -2 ·s -1Each computing unit performs subtraction and division operations in parallel to calculate the deviation ratio between the evaporation rate of its region and the reference rate. The calculation cycle is 5μs. A deviation ratio greater than 1.2 is considered an area of ​​excessively fast evaporation, and a deviation ratio less than 0.8 is considered an area of ​​excessively slow evaporation. A deviation ratio between 0.8 and 1.2 is considered normal. All 320×320 deviation ratio calculation results are temporarily stored in the FPGA's register file. Each computing unit accesses the FPGA's internal preset correction lookup table in parallel according to the deviation ratio. This correction lookup table contains 256 entries, with the deviation ratio starting from 0.01 in increments of 0.01. Values ​​from 70 to 1.30 are mapped to the corresponding initial PWM duty cycle corrections. When the deviation ratio is 1.0, the correction is 0%; when the deviation ratio is 1.2, the correction is -12%; when the deviation ratio is 1.3, the correction is -18%; and when the deviation ratio is below 0.8, the correction is +5% to +10%. After the table lookup is complete, each calculation unit reads the correction values ​​from the eight adjacent directions for spatial consistency verification. The absolute difference between the unit's correction value and the average of the neighboring correction values ​​is calculated. If this difference exceeds 5% and the deviation ratio of the unit from the neighboring average exceeds 0.15, it is determined to be a local overshoot, and the unit is adjusted accordingly. The correction value is replaced with the neighborhood mean. After verification, all correction values ​​are aggregated by a parallel reduction tree with a depth of 10 levels and a total delay of 2μs, thus outputting the global correction value statistics. The laser drive module receives the parallel reduced correction value data and updates the laser duty cycle of each sintering position pixel by pixel. The laser reference duty cycle is set to 45%, and the actual duty cycle of each pixel is calculated by adding the correction value to the reference duty cycle. The effective range of the correction value is -18% to +10%, and the corresponding actual duty cycle adjustment range is 27% to 55%. The duty cycle update cycle is related to the pixel time of the FPGA. Clock synchronization with a clock frequency of 10MHz ensures that the duty cycle calculation and loading are completed before sintering each 0.1mm×0.1mm pixel. For extreme areas where the evaporation rate deviation ratio exceeds 1.3, the FPGA simultaneously triggers the overspeed flag, which, in conjunction with the laser drive module, limits the duty cycle change rate to within ±3% per pixel, preventing sudden changes in duty cycle from causing surface tension imbalance in the molten silver paste. The adaptive real-time compensation mechanism is completed without interrupting the laser scanning process, and the total delay from the evaporation rate model input to the duty cycle update is controlled within 15μs, meeting the real-time requirements of the laser sintering process. Furthermore, step 2 also includes: setting a higher evaporation rate sensitivity coefficient for the electrode edge region than for the center region in the FPGA; when the evaporation rate deviation ratio of the edge region exceeds the threshold, triggering a duty cycle attenuation factor dedicated to the edge, further reducing the duty cycle of the edge laser, effectively suppressing nanoscale bulging at the electrode edge, improving contact resistance uniformity, and inputting the duty cycle after edge attenuation and the duty cycle of the center region into a preset thermal accumulation prediction model. This thermal accumulation prediction model estimates the cumulative thermal equivalent of the current sintering point based on the historical values ​​of the duty cycle of multiple previous scan cycles, determines whether it is close to the bulging critical value, realizes early warning of thermal accumulation, and avoids deformation caused by local overheating. If the thermal accumulation prediction model determines that the cumulative thermal equivalent exceeds the critical value, the duty cycle of the current sintering point is further reduced, and the adjusted duty cycle distribution is output as a thermal input spectrum to dynamically balance the surface tension of the molten silver paste and stabilize the electrode thickness. It should be noted that within the FPGA, edge regions are first identified based on the fine mesh coordinates of the electrodes: all pixels less than 0.5 mm from the electrode's geometric boundary are marked as edge regions, and the rest as center regions. A preset evaporation rate sensitivity coefficient of 1.8 is applied to edge regions, meaning that when the actual evaporation rate of a pixel in an edge region differs from the reference rate by 0.85 μg·mm², the sensitivity coefficient is determined. -2 ·s -1 When the deviation ratio reaches 1.2, the equivalent deviation ratio of the region is amplified by 1.8 times and used for table lookup, making the correction calculation more aggressive. Simultaneously, the FPGA independently configures a dedicated duty cycle attenuation factor for the edge region. This factor adds a fixed attenuation of -6% to the initial correction obtained from the table lookup. When the original deviation ratio of the edge region exceeds the threshold of 1.2, the FPGA immediately triggers this attenuation factor, further reducing the edge laser duty cycle by 6 percentage points on top of the initial correction, thereby suppressing nanoscale bulges caused by faster heat accumulation at the edge than in the center region. The heat accumulation prediction model uses a sliding register stack with a depth of 5. The historical duty cycle values ​​of the same spatial position in the five consecutive scanning cycles before the current scanning point are stored. The weight coefficients corresponding to each cycle are set to 0.35, 0.25, 0.20, 0.12 and 0.08 respectively. The weight decreases with the cycle to reflect the dominant role of recent heat input on current heat accumulation. The heat accumulation prediction model multiplies each historical duty cycle by the corresponding weight and sums them, and then multiplies them by the laser peak power of 50W and the dwell time reference value of 0.5ms to calculate the cumulative heat equivalent of the current sintering point. This cumulative heat equivalent is compared with the critical threshold of 9.5mJ calibrated in advance through silver paste bulging experiment to determine whether it is close to the bulging critical state. The formula for calculating the cumulative heat equivalent at the current sintering point is as follows: ; In the formula: Q acc(t) represents the cumulative thermal equivalent at the current sintering point at time t. This value characterizes the total energy input by the laser at the current position and in the previous multiple scanning cycles, used to determine whether it is close to the critical state of silver paste bulging; t is the current time, corresponding to the system time of the laser scanning to the current sintering point; k is the cycle index number, ranging from 1 to 5, where k=1 represents the previous scanning cycle, k=2 represents the previous two scanning cycles, and so on up to k=5 representing the previous five scanning cycles; w k is the weighting coefficient for the k-th historical scan cycle; D(tk·T) is the laser duty cycle of the current sintering point coordinates at the same spatial location in the k-th historical scan cycle; T is the laser scan cycle, i.e., the time required for the laser to complete one full scan; P peak τ represents the peak laser power, which is the instantaneous output power of the laser when the pulse is on. base This is the baseline value for dwell time, which is the reference dwell time of the laser spot on a single pixel. When the heat accumulation prediction model determines that the accumulated heat equivalent at the current sintering point exceeds the critical threshold of 9.5mJ, the FPGA initiates a secondary duty cycle reduction mechanism. This mechanism dynamically calculates the reduction range based on the excess amount: for every 0.5mJ increase in the excess amount, the duty cycle decreases by an additional 2%, with a maximum additional decrease of no more than 10%. After completing the final duty cycle calculation, the FPGA organizes the adjusted duty cycles of all pixels into a 320×320 two-dimensional floating-point array according to spatial coordinates, and adds a timestamp and heat accumulation status flag to each array element to form a heat input map that can be directly called by subsequent processes. This map is output to the laser drive module in a pixel-by-pixel manner. Step 3: Map the thermal input spectrum to the thickness field, predict the electrode thickness deviation through Kalman filtering, calculate the scanning speed correction coefficient, dynamically adjust the spot dwell time to balance the surface tension of the molten silver paste, and control the electrode thickness fluctuation. Step 4: The laser sintering preparation device adjusts the scanning speed in real time according to the scanning speed correction coefficient to match the residence time with the heat input, thereby achieving electrode thickness homogenization, controlling conductivity uniformity, completing the main sintering, and significantly improving conductivity uniformity. Step 5: Calculate the contamination risk index by combining the laser duty cycle and the scanning speed correction coefficient. If it exceeds the preset risk threshold, it is determined that the splashed silver particles pose a risk of contamination to the focusing lens, triggering cleaning preparation, identifying splash risks, and avoiding sudden contamination of the focusing lens. Step 6: Drive the optical path cleaner installed in the laser sintering preparation device to dynamically remove splash particles, prevent sintering linewidth drift, ensure electrode morphology accuracy, complete the preparation of solar cell electrodes, stabilize the sintering linewidth, and ensure electrode morphology accuracy.

[0024] Example 2, as Figure 1As shown, based on Example 1, the present invention provides a technical solution: Step 3 specifically includes: establishing an empirical mapping function between the laser energy density of each pixel in the thermal input spectrum and the steady-state thickness of the molten silver paste, converting the product of duty cycle and dwell time into a predicted thickness value, generating an initial thickness field distribution, effectively establishing a quantitative correlation between energy input and thickness response, using a Kalman filter algorithm, taking the initial thickness field as a state prior estimate, combining the measured thickness feedback value of the previous scanning cycle, iteratively updating the thickness state estimate of the current cycle, and calculating the thickness prediction deviation at each position to achieve dynamic correction and noise suppression of thickness prediction. Based on the sign and magnitude of the thickness prediction deviation, the scanning speed correction coefficient is calculated by a PID controller. When the deviation is positive (thickness is too high), the scanning speed is increased to shorten the dwell time; when the deviation is negative, the scanning speed is decreased to extend the dwell time, thereby achieving real-time reverse adjustment of the thickness deviation. It should be noted that, within the FPGA, an empirical mapping function between the laser energy density of each pixel in the thermal input spectrum and the steady-state thickness of the molten silver paste was established based on offline calibration experiments. The calibration experiments used silver paste material and substrate identical to those used in the process. Under standard conditions of a laser peak power of 50W, a reference duty cycle of 45%, and a dwell time of 0.5ms, the duty cycle and scanning speed were varied using a controlled variable method. A line laser displacement sensor was used to measure the steady-state thickness at various positions of the electrode after sintering. The product of the duty cycle and the dwell time was defined as the dimensionless energy input factor. Steady-state thickness data corresponding to the energy input factor in the range of 0.20 to 0.55 were collected in steps of 0.01. A continuous mapping curve was generated using cubic spline interpolation. The mapping function is stored in the FPGA's block random access memory in the form of a lookup table. The input is the product of the duty cycle and dwell time of the current pixel, and the output is the predicted thickness value. For each 320×320 pixel in the thermal input map, the FPGA calls this mapping function in parallel to convert the product of duty cycle and dwell time into the predicted thickness value point by point, generating an initial thickness field distribution covering the entire electrode surface. A Kalman filter algorithm is used to dynamically optimize and estimate the initial thickness field. The state variable of the filter is the actual thickness value of each pixel in the current scanning cycle, and the observation variable is the thickness feedback value measured by the line laser displacement sensor in the previous scanning cycle. The process noise covariance matrix of the Kalman filter is set to 0.25μm based on the statistical characteristics of the silver paste coating thickness fluctuation. 2 The measurement noise covariance is set to 0.10 μm based on the sensor accuracy. 2At the beginning of each scanning cycle, the initial thickness field is used as the prior estimate of the state. The corresponding position values ​​are extracted from the posterior thickness state estimate stored in the previous cycle, and the state is updated by combining the Kalman gain to obtain the posterior thickness state estimate of the current cycle. After the state update is completed, the thickness prediction deviation of each pixel is calculated, which is the difference between the posterior and prior estimates. The sign of this deviation indicates whether the thickness is too thick or too thin, and the magnitude reflects the degree of deviation. The thickness prediction deviation and its corresponding spatial coordinates are packaged and stored in the FPGA register file for the PID controller to read in real time. The deviation update cycle is synchronized with the laser scanning pixel clock to ensure that the current thickness deviation information is obtained before each pixel is sintered. The PID controller adopts a positional digital PID algorithm with a proportional coefficient set to 0.85, an integral coefficient set to 0.12, and a derivative coefficient set to 0.06. The control cycle is consistent with the pixel clock at 10MHz. When the deviation value is positive (i.e., the thickness is too thick), the PID controller outputs a correction coefficient greater than 1.0 to make the target scanning speed higher than the reference speed, thereby shortening the dwell time and reducing the laser energy input. When the deviation value is negative (i.e., the thickness is too thin), the PID controller outputs a scanning speed correction coefficient less than 1.0 to make the target scanning speed lower than the reference speed, thereby extending the dwell time and increasing the laser energy input. The effective range of the scanning speed correction coefficient is limited to 0.70 to 1.30, corresponding to a scanning speed adjustment range of 70% to 130% of the reference speed. This prevents over-adjustment from causing new thickness non-uniformity. A first-order low-pass filter is connected in series at the output of the PID controller, with a cutoff frequency set to 100Hz, to suppress high-frequency fluctuations in the correction coefficient introduced by thickness measurement noise. The calculated scanning speed correction coefficient is output to the control system of the laser sintering preparation device in the form of a 320×320 two-dimensional array to drive the scanning mirror to perform speed regulation. Step 4 specifically includes: The control system of the laser sintering preparation device receives the calculated scanning speed correction coefficient in real time, multiplies it with the preset reference scanning speed, and generates the instantaneous target scanning speed for each sintering position. This ensures that each pixel receives an accurate speed command before sintering, avoiding speed lag. A speed feedforward compensation algorithm is used to dynamically adjust the scanning motor drive voltage of the scanning mirror according to the rate of change of the target scanning speed, so that the actual scanning speed changes continuously in space. This avoids the introduction of new thickness non-uniformity due to speed abrupt changes, suppresses speed abrupt changes, and prevents uneven spreading of molten silver paste. During continuous scanning, the control system feeds back the actual residence time of each position to the Kalman filter in a closed loop as the measured thickness feedback value for the next cycle, completing the full closed-loop control of the main sintering, realizing full closed-loop control of thickness homogenization, and continuously converging the thickness deviation. It should be noted that the control system receives a 320×320 two-dimensional scanning speed correction coefficient array calculated and output by the FPGA in real time. Each coefficient corresponds to a 0.1mm×0.1mm pixel on the electrode surface. The correction coefficient of each pixel is multiplied by the preset reference scanning speed to generate the instantaneous target scanning speed at each sintering position. The reference scanning speed is determined by the thermophysical parameters of the silver paste material. The control system adopts a distributed clock synchronization mechanism, writing the target scanning speed point by point into the speed register of the scanning mirror controller at a 10MHz pixel clock cycle, ensuring that the speed parameter is loaded before sintering for each pixel. The effective range of the target scanning speed corresponds to the correction coefficients from 0.70 to 1.30. A speed feedforward compensation algorithm is used to suppress the impact of sudden changes in scanning speed on thickness uniformity. The control system monitors the rate of change of the target scanning speed between adjacent pixels in real time, i.e., the derivative of speed with respect to displacement. When the absolute value of the rate of change exceeds a set value of 5, the feedforward compensation mechanism is triggered. The additional component of the scanning motor drive voltage of the scanning mirror is dynamically adjusted according to the magnitude and direction of the rate of change. The feedforward compensation coefficient is obtained by offline identification of the transfer function of the scanning mirror electromechanical system and is set to 0.32. For scanning speeds from high speed... In the region transitioning to low speed, feedforward compensation increases the braking voltage component in advance to prevent speed overshoot; in the region transitioning from low speed to high speed, feedforward compensation increases the acceleration voltage component in advance to prevent speed lag. The compensated actual scanning speed exhibits continuous spatial variation, with the speed jump amplitude between adjacent pixels controlled within ±2mm / s to avoid uneven spreading of molten silver paste due to sudden speed changes. The control system feeds back the actual scanning speed at each position in a closed loop to the data acquisition module, which calculates the actual dwell time, i.e., the spot diameter of 30μm divided by the actual scanning speed. The dwell time, as an indirect representation of the measured thickness feedback value, is fused with the thickness data measured by the line laser displacement sensor and input into the observation variable port of the Kalman filter as the basis for updating the thickness status in the next scanning cycle. The actual scanning speed is obtained in real time by the photoelectric encoder built into the scanning mirror at a sampling rate of 1MHz. After being timestamped with the laser emission pulse, the deviation record between the actual dwell time and the predicted dwell time of each pixel is formed, enabling the Kalman filter to continuously correct the estimation deviation of the process noise covariance matrix and complete the full closed-loop control of thickness homogenization in the main sintering process. Step 5 specifically includes: real-time acquisition of the actual laser duty cycle and scanning speed correction coefficient applied at each sintering position, multiplying the two and integrating them over time to obtain the cumulative laser energy per unit length, which serves as the basic characteristic quantity for the probability of generating sputtered silver particles, providing a direct and reliable input basis for risk quantification. The basic characteristic quantity is input into a pre-calibrated nonlinear risk mapping model, while considering the fluctuation range of the duty cycle change rate and the scanning speed correction coefficient, outputting a contamination risk index to achieve a multi-factor comprehensive assessment of the sputtering probability, improving the accuracy of early warning. The contamination risk index is compared with a preset risk threshold. If the contamination risk index exceeds the threshold for a period of time exceeding the set judgment window, an optical path cleaning preparation signal is triggered, and the position exceeding the threshold is recorded to assist in cleaning, avoiding false triggering due to instantaneous noise, and providing accurate spatial guidance for subsequent cleaning. It should be noted that during the laser sintering process, the actual laser duty cycle and scanning speed correction coefficient applied at each sintering position are collected in real time. The product of these two values ​​is then integrated over time to obtain the cumulative laser energy per unit length, which serves as a fundamental characteristic for the probability of generating sputtered silver particles. Specific acquisition parameters are as follows: the duty cycle sampling frequency is synchronized with the pixel clock at 10MHz; the scanning speed correction coefficient is provided by a 320×320 two-dimensional array; the integration step size corresponds to the dwell time of each 0.1mm×0.1mm pixel; and the formula for calculating the cumulative laser energy per unit length is... Divide by the scan length, where E unit The cumulative laser energy per unit length, D is the actual applied laser duty cycle, and k is the total laser energy. v P is the scan speed correction factor. peak Let dt be the peak laser power, dt be the integral time element corresponding to the continuous integral over the actual dwell time, and L be the scanning length. The integral result is the energy value per millimeter, which is directly related to the probability of generating sputtered silver particles; the higher the energy density, the greater the sputtering tendency. The calculated cumulative laser energy per unit length is used as the basic characteristic and input into a pre-calibrated nonlinear risk mapping model, which adopts the Sigmoid function form. The parameters calibrated through offline splash experiments are: slope coefficient k = 0.85 mm / mJ, threshold energy E0 = 2.5 mJ / mm, and the model simultaneously inputs the duty cycle change rate dD / dt and the fluctuation amplitude Δk of the scan speed correction coefficient. v These two indicators reflect the severity of laser energy fluctuations and the stability of speed control, respectively. When the duty cycle change rate exceeds 0.15 / ms or the fluctuation range of the scan speed correction coefficient exceeds 0.12, an increment of 0.15 is added to the basic risk index, resulting in the final pollution risk index R. risk It reflects in real time the probability of silver microparticles contaminating the focusing lens under the current sintering conditions; The calculation expression is as follows: ; In the formula: R risk The final contamination risk index comprehensively reflects the probability of silver microparticles sputtering and contaminating the focusing lens under current sintering conditions. A value closer to 1 indicates a higher contamination risk, while a value closer to 0 indicates a lower contamination risk. R base Basic pollution risk index; Δ dD / dt Add an increment to the duty cycle rate of change; An increment is added to the fluctuation range of the scan rate correction coefficient; α is the slope coefficient of the Sigmoid function; dD / dt is the duty cycle change rate; Δk v The fluctuation range of the scanning speed correction coefficient is determined. The contamination risk index is compared with the preset risk threshold. A sliding time window judgment mechanism is adopted, with the window duration set to 50ms, corresponding to the sintering time of approximately 5 consecutive pixels. If the contamination risk index continuously exceeds the risk threshold within the window, it is determined that the sputtered silver particles pose an actual contamination risk to the focusing lens, and the optical path cleaning preparation signal is immediately triggered. At the same time, the spatial coordinates of the position exceeding the threshold are recorded to form a high-risk area distribution map, which is stored in the block random access memory of the FPGA for subsequent optical path cleaner cleaning strategy optimization.

[0025] Example 3, as Figures 2 to 12 As shown, based on Embodiments 1 and 2, the present invention also provides a laser sintering preparation apparatus for solar cells and their electrodes, used to realize a laser sintering preparation method for solar cells and their electrodes. The apparatus includes a housing 1, a preparation working chamber 2 fixedly installed on the top of the housing 1, the preparation working chamber 2 providing a sealed environment for laser sintering to prevent external dust interference, a power supply 3 detachably installed on the top of the preparation working chamber 2, the power supply 3 providing stable energy output for laser sintering, an electrode laser sintering device 4 installed at the output end of the power supply 3, and an optical path cleaner 5 connected to the top of the inner cavity of the preparation working chamber 2 at the output end of the electrode laser sintering device 4. The optical path cleaner 5 is used to dynamically remove splashed silver particles during the sintering process to avoid contamination of the focusing lens. The electrode laser sintering unit 4 includes a laser 41. The input end of the laser 41 is connected to the output end of the power supply unit 3. The laser 41 is fixedly installed on the top of the inner cavity of the preparation working cavity 2. The laser 41 emits pulsed laser according to the duty cycle adjusted by adaptive PWM. A scanning mirror 42 is fixedly installed on the output end of the laser 41. A focusing mirror 43 is movably installed on the output end of the scanning mirror 42. The focusing mirror 43 focuses the laser beam onto the electrode surface to achieve fine sintering. The laser beam emitted by the scanning mirror 42 is projected onto the electrode through the focusing mirror 43 to achieve the laser sintering operation. The optical path cleaner 5 includes an upper fixed shell 51 fixedly installed at the top of the inner cavity of the preparation working chamber 2. The upper fixed shell 51 provides structural support for the optical path cleaner. A lower isolation plate 52 is detachably installed at the bottom of the upper fixed shell 51. The lower isolation plate 52 partially isolates the cleaning area from the sintering area to reduce the diffusion of splash particles. An auxiliary cleaning component 54 is installed at the bottom of the lower isolation plate 52. The auxiliary cleaning component 54 is activated after triggering the cleaning preparation signal to remove splash particles attached to the focusing lens. A focusing lens conversion component 53 for rotating and replacing the focusing lens 43 is installed on the side of the scanning lens 42. The focusing lens conversion component 53 includes a drive motor 531 fixedly installed at the top of the upper fixed shell 51. The drive motor 531 receives a contamination risk index exceeding a threshold. The cleaning preparation signal is triggered and the drive motor 531 is activated. The output end of the drive motor 531 passes through the upper fixed housing 51 and rotates with the inner wall of the upper fixed housing 51. A connecting rod 532 is fixedly installed at the output end of the drive motor 531. The connecting rod 532 transmits the rotational motion of the drive motor 531 to the focusing lens 43. The end of the connecting rod 532 is fixed to the outer edge surface of the focusing lens 43. The focusing lens 43 rotates under the drive of the connecting rod 532, realizing the position switch between the spare focusing lens and the contaminated focusing lens. The top and bottom of the connecting rod 532 and the focusing lens 43 slide with the bottom and top of the upper fixed housing 51 and the lower isolation plate 52, respectively, to ensure a smooth and jam-free rotation replacement process. An isolation strip 55 is fixedly installed on the bottom surface of the upper fixed housing 51. The isolation strip 55... The focusing lens maintains contact with the lower isolation plate during rotation to prevent splash particles from entering the optical path through gaps. The bottom surface of the insulating strip 55 contacts the top of the lower isolation plate 52. The auxiliary cleaning component 54 includes a cleaning cavity 541 fixedly installed on the side of the lower isolation plate 52. The cleaning cavity 541 provides a closed space for airflow cleaning to prevent secondary diffusion of splash particles. A pulsed airflow pipe 542 is fixedly installed at the bottom of the cleaning cavity 541. The pulsed airflow pipe 542 introduces pulsed airflow to enhance the stripping effect on the surface of the focusing lens. One end of the pulsed airflow pipe 542 is externally connected to an air supply device, and the other end of the pulsed airflow pipe 542 extends into the interior of the cleaning cavity 541 and is fixedly connected to an auxiliary air knife ring 544. The auxiliary air knife ring 544 distributes the airflow evenly. A slag discharge pipe 543 is fixedly connected to the side opening of the cleaning chamber 541 to the surface of the focusing lens. The slag discharge pipe 543 discharges the exhaust gas carrying splashed particles to the external treatment device. The end of the slag discharge pipe 543 extends to the outside of the preparation working chamber 2 and is connected to an external exhaust gas treatment device. The auxiliary air knife ring 544 is inclinedly set below the focusing lens 43 to be cleaned. The inclined setting allows the airflow to impact the surface of the focusing lens at a certain angle, improving the cleaning efficiency. A conical air-gathering nozzle 545 is fixedly installed at the top center of the auxiliary air knife ring 544. The air-gathering nozzle 545 concentrates the airflow and enhances the local purging force. Several slit nozzles 546 are opened at the top of the auxiliary air knife ring 544. The slit nozzles 546 form an air curtain, covering the entire area of ​​the focusing lens to be cleaned.

[0026] It should be noted that after startup, the power supply 3 provides energy to the laser 41. The laser 41 is fixedly installed at the top of the inner cavity of the preparation working chamber 2, and a scanning mirror 42 is fixedly installed at its output end. The laser beam emitted by the scanning mirror 42 is directed to the surface of the solar cell substrate coated with silver paste through the movable focusing mirror 43 to perform laser sintering operation. During the sintering process, the optical path cleaner 5 works in conjunction with the electrode laser sinterer 4. The optical path cleaner 5 includes an upper fixed shell 51 fixed to the top of the inner cavity of the preparation working chamber 2, with a lower isolation plate 52 detachably installed at its bottom. A focusing lens conversion component 53 is installed on the side of the scanning mirror 42. When silver particles are splashed onto the surface of the focusing lens 43, the drive motor 531 installed on the top of the upper fixed shell 51 is activated. Its output end passes through the upper fixed shell 51 and drives the connecting rod 532 to rotate. The end of the connecting rod 532 is fixed to the outer edge of the focusing lens 43, thereby driving the focusing lens 43 to slide between the upper fixed shell 51 and the lower isolation plate 52, realizing the rotational replacement of the focusing lens 43. The insulating soft strip 55 on the bottom surface of the upper fixed shell 51 contacts the top of the lower isolation plate 52 to maintain the cavity seal. The replaced contaminated focusing lens 43 moves to the position of the auxiliary cleaning component 54, which includes a fixed... The cleaning chamber 541 on the side of the lower isolation plate 52 has a pulsed airflow pipe 542 installed at its bottom. One end of the pulsed airflow pipe 542 is connected to an external air supply device, and the other end extends into the cleaning chamber 541 and is connected to an auxiliary air knife ring 544. The auxiliary air knife ring 544 is inclined and set below the focusing lens 43 to be cleaned. Pulsed airflow is sprayed out through the conical air-gathering nozzle 545 at the center of its top and multiple slit nozzles 546 to dynamically clean the surface of the focusing lens 43. The splash particles and waste gas generated during cleaning are discharged through the slag discharge pipe 543 at the side opening of the cleaning chamber 541. The end of the slag discharge pipe 543 extends to the outside of the preparation working chamber 2 and is connected to a waste gas treatment device. After cleaning is completed, the drive motor 531 reverses its action to move the cleaned focusing lens 43 back into the optical path and continue to perform laser sintering.

[0027] The following describes in detail the laser sintering preparation method and apparatus for this type of solar cell and its electrodes.

[0028] First, a solar cell substrate uniformly coated with silver paste is placed in the fabrication working chamber 2 to form a wet film electrode. The MEMS gas sensor array is activated, and the organic carrier volatile concentration signal is collected in real time through the sensor units arranged in a grid. After FPGA sliding window filtering and interpolation calculation, a volatile rate distribution model covering the entire electrode surface is generated. Based on this model, the FPGA parallel computing array calculates the deviation ratio between the volatile rate and the reference rate pixel by pixel. The initial correction amount of PWM duty cycle is obtained by looking up a table. After spatial consistency verification, the laser duty cycle is updated point by point. For the electrode edge area, a higher volatile rate sensitivity coefficient is preset and a duty cycle attenuation factor is added to suppress nanoscale bulges. At the same time, the historical value of the duty cycle of the previous scan cycle is used to input the heat accumulation prediction model to estimate the accumulated heat equivalent. If the critical value is exceeded, the duty cycle is further reduced. Finally, the heat input spectrum is output. Subsequently, an empirical mapping function between laser energy density and molten silver paste thickness is established, and the thermal input spectrum is transformed into an initial thickness field distribution. Using a Kalman filter algorithm, the current thickness state is iteratively estimated by combining the measured thickness feedback value of the previous cycle, and the thickness prediction deviation is calculated. The PID controller calculates the scanning speed correction coefficient based on the sign and magnitude of the deviation. When the deviation is too thick, the scanning speed is increased to shorten the dwell time, and when it is too thin, the scanning speed is decreased to prolong the dwell time. The control system of the laser sintering preparation device multiplies the scanning speed correction coefficient with the reference speed to generate the instantaneous target scanning speed of each pixel. Using a speed feedforward compensation algorithm, the scanning motor drive voltage is dynamically adjusted according to the speed change rate to avoid speed abrupt changes. The actual scanning speed is fed back to the Kalman filter as the observation input for the next cycle, realizing full closed-loop control of thickness homogenization. Simultaneously, the actual laser duty cycle and scanning speed correction coefficient are collected in real time, the cumulative laser energy per unit length is calculated, and the data is input into the nonlinear risk mapping model. Combining the duty cycle change rate and speed fluctuation amplitude, a contamination risk index is output. If the index continuously exceeds the risk threshold within the set window, the optical path cleaning preparation signal is triggered and the position exceeding the threshold is recorded. Finally, after the cleaning preparation is triggered, the optical path cleaner 5 is driven to work. The drive motor 531 of the focusing lens conversion component 53 rotates through the connecting rod 532 to replace the contaminated focusing lens 43. At the same time, the auxiliary cleaning component 54 is started. The external air supply device supplies air to the auxiliary air knife ring 544 through the pulsating airflow pipe 542. The inclined air knife is generated through the gas-gathering nozzle 545 and the slit nozzle 546 to remove splash particles. The exhaust gas is discharged through the slag discharge pipe 543, thereby preventing sintering linewidth drift and completing the preparation of the solar cell electrode.

[0029] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A solar cell, and the laser sintering preparation of its electrodes, characterized in that, Includes the following steps: Step 1: Place the solar cell electrode in the laser sintering fabrication device, start the MEMS gas sensor array, detect the volatilization rate of the silver paste organic carrier in real time, and construct a volatilization rate distribution model; Step 2: Based on the evaporation rate model, the FPGA performs parallel processing and adopts adaptive PWM to adjust the laser duty cycle, reduce heat accumulation, suppress nanoscale bulging at the electrode edge, and generate a thermal input spectrum. Step 3: Map the thermal input spectrum to the thickness field, predict the electrode thickness deviation through Kalman filtering, calculate the scanning speed correction coefficient, and dynamically adjust the spot dwell time to balance the surface tension of the molten silver paste. Step 4: The laser sintering preparation device adjusts the scanning speed in real time according to the scanning speed correction coefficient to match the residence time with the heat input, thereby achieving electrode thickness homogenization and completing the main sintering. Step 5: Calculate the contamination risk index by combining the laser duty cycle and the scanning speed correction coefficient. If it exceeds the preset risk threshold, it is determined that the sputtered silver particles pose a risk of contamination to the focusing lens, and cleaning preparation is triggered. Step 6: Drive the optical path cleaner mounted on the laser sintering preparation device to dynamically remove splash particles, prevent sintering linewidth drift, and complete the preparation of solar cell electrodes.

2. The laser sintering fabrication of a solar cell and its electrodes according to claim 1, characterized in that: Step 1 specifically includes: Silver paste is uniformly coated onto the solar cell substrate to form a wet film electrode. The solar cell electrode is then placed in a laser sintering fabrication device. Multiple sensing units sensitive to the volatile components of the organic carrier are then deployed in a MEMS gas sensor array and arranged in a grid pattern along the length and width of the wet film electrode. The instantaneous volatile concentration signal of each grid point is collected in a spatially resolved manner. The instantaneous volatile concentration signals output by each sensing unit are converted from analog to digital and then sent to the FPGA. A sliding window mean filter is used to suppress environmental noise and extract the temporal variation characteristics of the volatile rate of each grid point to form a discrete rate matrix. Based on the discrete rate matrix, a radial basis function interpolation algorithm is used to estimate the rate of areas not directly sampled, generating a volatile rate distribution model covering the entire wet film electrode surface.

3. The laser sintering fabrication of a solar cell and its electrodes according to claim 1, characterized in that: Step 2 specifically includes: The generated evaporation rate distribution model is input into the parallel computing array of the FPGA in the form of a pixelated heat map. Each computing unit corresponds to a local area of ​​an electrode and independently calculates the deviation ratio between the evaporation rate of that area and the reference rate. Each calculation unit looks up the corresponding initial correction amount of PWM duty cycle in parallel according to the deviation ratio. The larger the deviation, the larger the correction amount. At the same time, the correction amount of adjacent regions is checked for spatial consistency. The correction values ​​for all regions are reduced in parallel and then output to the laser drive module to update the laser duty cycle of each sintering position pixel by pixel.

4. The laser sintering fabrication of a solar cell and its electrodes according to claim 3, characterized in that: Step 2 also includes: In the FPGA, a higher evaporation rate sensitivity coefficient is preset for the electrode edge region than for the center region. When the evaporation rate deviation ratio of the edge region exceeds the threshold, a duty cycle attenuation factor dedicated to the edge is triggered to further reduce the duty cycle of the edge laser. The duty cycle after edge attenuation and the duty cycle of the central region are input together into the preset thermal accumulation prediction model. The thermal accumulation prediction model estimates the cumulative thermal equivalent of the current sintering point based on the historical values ​​of the duty cycle of multiple previous scan cycles, and determines whether it is close to the bulging critical value. If the heat accumulation prediction model determines that the accumulated heat equivalent exceeds the critical value, the duty cycle of the current sintering point will be further reduced, and the adjusted duty cycle distribution will be output as the heat input spectrum.

5. The laser sintering fabrication of a solar cell and its electrodes according to claim 1, characterized in that: Step 3 specifically includes: An empirical mapping function is established between the laser energy density of each pixel in the thermal input spectrum and the steady-state thickness of the molten silver paste. The product of duty cycle and residence time is converted into a predicted thickness value to generate an initial thickness field distribution. The Kalman filter algorithm is used to take the initial thickness field as the state prior estimate. Combined with the measured thickness feedback value of the previous scanning cycle, the thickness state estimate of the current cycle is iteratively updated, and the thickness prediction deviation at each position is calculated to achieve dynamic correction and noise suppression of thickness prediction. Based on the sign and magnitude of the thickness prediction deviation, the PID controller calculates the scanning speed correction coefficient. When the deviation is positive, the scanning speed is increased to shorten the dwell time, and when the deviation is negative, the scanning speed is decreased to prolong the dwell time.

6. The laser sintering fabrication of a solar cell and its electrodes according to claim 1, characterized in that: Step 4 specifically includes: The control system of the laser sintering preparation device receives the calculated scanning speed correction coefficient in real time, multiplies it with the preset reference scanning speed, and generates the instantaneous target scanning speed for each sintering position. A speed feedforward compensation algorithm is adopted to dynamically adjust the scanning motor drive voltage of the scanning mirror according to the rate of change of the target scanning speed, so that the actual scanning speed changes continuously in space. During continuous scanning, the control system feeds back the actual dwell time at each position to the Kalman filter in a closed loop, which serves as the measured thickness feedback value for the next cycle, thus completing the full closed-loop control of the main sintering.

7. The laser sintering fabrication of a solar cell and its electrodes according to claim 1, characterized in that: Step 5 specifically includes: The actual laser duty cycle and scanning speed correction coefficient applied at each sintering position are collected in real time. The cumulative laser energy per unit length is obtained by multiplying the two and integrating them over time, which serves as the basic characteristic quantity for the probability of generating sputtered silver particles. The basic characteristic quantities are input into a pre-calibrated nonlinear risk mapping model, and the fluctuation range of the duty cycle change rate and the scanning speed correction coefficient are considered to output the pollution risk index. The pollution risk index is compared with a preset risk threshold. If the pollution risk index exceeds the threshold for a period of time longer than the set judgment window, the optical path cleaning preparation signal is triggered, and the position of exceeding the threshold is recorded to assist in cleaning.

8. A laser sintering preparation apparatus for a solar cell and its electrodes, used to implement the laser sintering preparation method for a solar cell and its electrodes as described in any one of claims 1-7, comprising a chassis (1), characterized in that: The top of the chassis (1) is fixedly installed with a preparation working chamber (2), and the top of the preparation working chamber (2) is detachably installed with a power supply (3). The output end of the power supply (3) is installed with an electrode laser sintering device (4), and the output end of the electrode laser sintering device (4) is connected to an optical path cleaner (5) connected to the top of the inner cavity of the preparation working chamber (2). The electrode laser sintering machine (4) includes a laser (41), the input end of which is connected to the output end of the power supply (3), and the laser (41) is fixedly installed on the top of the inner cavity of the preparation working cavity (2). A scanning mirror (42) is fixedly installed on the output end of the laser (41), and a focusing mirror (43) is movably installed on the output end of the scanning mirror (42). The laser beam emitted by the scanning mirror (42) is directed onto the electrode through the focusing mirror (43) to realize the laser sintering operation. The optical path cleaner (5) includes an upper fixed shell (51) fixedly installed on the top of the inner cavity of the preparation working chamber (2), a lower isolation plate (52) is detachably installed on the bottom of the upper fixed shell (51), an auxiliary cleaning component (54) is installed on the bottom of the lower isolation plate (52), and a focusing lens conversion component (53) for rotating and replacing the focusing lens (43) is installed on the side of the scanning lens (42).

9. The laser sintering fabrication apparatus for a solar cell and its electrodes according to claim 8, characterized in that: The focusing lens conversion component (53) includes a drive motor (531) fixedly installed on the top of the upper fixed shell (51). The output end of the drive motor (531) passes through the upper fixed shell (51) and rotates with the inner wall of the upper fixed shell (51). A connecting rod (532) is fixedly installed on the output end of the drive motor (531). The end of the connecting rod (532) is fixed to the outer edge surface of the focusing lens (43). The top and bottom of the connecting rod (532), the focusing lens (43), and the bottom and top of the upper fixed shell (51) and the lower isolation plate (52) slide against each other, respectively. An insulating soft strip (55) is fixedly installed on the bottom surface of the upper fixed shell (51). The bottom surface of the insulating soft strip (55) contacts the top of the lower isolation plate (52).

10. The apparatus for laser sintering fabrication of a solar cell and its electrodes according to claim 9, characterized in that: The auxiliary cleaning component (54) includes a cleaning chamber (541) fixedly installed on the side of the lower isolation plate (52). A pulsating airflow pipe (542) is fixedly installed at the bottom of the cleaning chamber (541). One end of the pulsating airflow pipe (542) is connected to an air supply device. The other end of the pulsating airflow pipe (542) extends into the interior of the cleaning chamber (541) and is fixedly connected to an auxiliary air knife ring (544). A slag discharge pipe (543) is fixedly connected to the side opening of the cleaning chamber (541). The end of the slag discharge pipe (543) extends into the exterior of the preparation working chamber (2) and is connected to an exhaust gas treatment device. The auxiliary air knife ring (544) is inclinedly arranged below the focusing lens (43) to be cleaned. A conical gas-gathering nozzle (545) is fixedly installed at the top center of the auxiliary air knife ring (544). Several slit nozzles (546) are opened at the top of the auxiliary air knife ring (544).

Images