Bismuth telluride-based thermoelectric material wafer surface treatment method

Abstract

The invention relates to the technical field of thermoelectric material surface treatment, and provides a bismuth telluride-based thermoelectric material wafer surface treatment method which comprises the following steps: acquiring multi-dimensional data such as current density distribution, temperature field distribution, solution concentration gradient and plating thickness increment in real time through a sensor array; performing iterative calculation on the data based on a thermoelectric performance parameter optimization model to generate an optimal parameter combination of chemical oil removal, activation, nickel plating and tin and bismuth plating processes; dynamically adjusting the anode plate area ratio, the current density threshold value and the solution temperature fluctuation range of the electroplating equipment; a self-adaptive feedback mechanism is used for correcting parameter deviation, so that a uniform thermoelectric function layer structure is formed on the surface of the wafer. According to the method, the uniformity and the thermoelectric performance stability of the surface coating of the bismuth telluride thermoelectric wafer can be improved, the defect rate is reduced, and the production efficiency and the product consistency are improved.

Description

Technical Field

[0001] This invention relates to the field of thermoelectric material surface treatment technology, and more specifically, to a method for surface treatment of bismuth telluride-based thermoelectric material wafers. Background Technology

[0002] In the field of thermoelectric materials applications today, bismuth telluride-based thermoelectric materials have attracted much attention due to their excellent thermoelectric properties. This material has broad application prospects in thermoelectric power generation and refrigeration. However, to fully realize its performance, the wafer surface treatment process is crucial. Currently, traditional surface treatment methods for bismuth telluride-based thermoelectric material wafers mainly rely on fixed-parameter processes such as chemical degreasing, activation, nickel plating, and tin-bismuth plating. The parameter settings for these processes are usually based on experience, lacking real-time data acquisition and dynamic adjustment mechanisms. For example, in the chemical degreasing process, the coupling relationship between sodium hydroxide concentration and temperature is difficult to control precisely, leading to unstable degreasing results; in the nickel plating process, pH fluctuation compensation lacks accurate predictive models, affecting the plating quality; and in the tin-bismuth plating process, the matching relationship between cathode movement speed and current density is also difficult to adjust precisely, easily leading to uneven plating.

[0003] While existing surface treatment technologies can meet basic process requirements to some extent, they have significant limitations. Firstly, traditional methods cannot monitor multi-dimensional data during wafer surface treatment in real time, such as current density distribution, temperature field distribution, solution concentration gradient, and coating thickness increment. This results in a lack of effective data support for optimizing process parameters. Secondly, the lack of effective adaptive feedback mechanisms prevents dynamic adjustment of process parameters based on real-time monitoring data, leading to uneven thermoelectric functional layer structure on the treated wafer surface and affecting thermoelectric performance. Furthermore, existing technologies largely rely on experience for process parameter optimization, lacking scientific optimization models and algorithms, making it difficult to achieve globally optimal parameter combinations.

[0004] In the process of implementing the embodiments of the present invention, the inventors found that the prior art has at least the following problems or defects: it is impossible to collect multi-dimensional data in the wafer surface processing process in real time, it lacks a dynamic parameter adjustment mechanism based on real-time data, and it lacks a scientific thermoelectric performance parameter optimization model and algorithm, which leads to uneven thermoelectric functional layer structure on the surface of the processed wafer and the thermoelectric performance cannot reach the optimal level. Summary of the Invention

[0005] This invention provides a surface treatment method for bismuth telluride-based thermoelectric material wafers, comprising: real-time acquisition of multi-dimensional data during the wafer surface treatment process using a sensor array, the data including current density distribution, temperature field distribution, solution concentration gradient, and coating thickness increment; iterative calculation of the acquired data based on a thermoelectric performance parameter optimization model to generate the optimal parameter combination for chemical degreasing, activation, nickel plating, and tin-bismuth plating processes; dynamic adjustment of the anode plate area ratio, current density threshold, and solution temperature fluctuation range of the electroplating equipment according to the parameter combination; and correction of parameter deviations based on an adaptive feedback mechanism to form a uniform thermoelectric functional layer structure on the wafer surface.

[0006] Furthermore, the sensor array includes a four-dimensional current probe matrix, an infrared thermal imaging unit, an electrochemical impedance spectroscopy analysis module, and an X-ray fluorescence thickness gauge; the temporal resolution of the real-time acquired data reaches a preset time accuracy threshold, and the spatial resolution reaches a preset spatial accuracy threshold; the parameter combination includes the fluctuation range of hydrochloric acid concentration in the activation solution, the dynamic correction coefficient of the nickel plating pH value, and the matching factor of the tin-bismuth cathode movement speed.

[0007] Furthermore, the objective function expression of the thermoelectric performance parameter optimization model is: In the formula, Let be the objective function. This is the weighting factor for temperature field uniformity. γ is the coating density optimization coefficient, and γ is the thermal stress suppression factor. This represents the measured value at the i-th temperature monitoring point. The standard temperature threshold for the process. The ion mobility in the solution, This is the critical value for lattice distortion. This is the upper limit of thermal stress. The starting time of the integration. This is the end time for integration.

[0008] Furthermore, the iterative calculation includes the following steps: generating multiple sets of process parameter combinations in the initial parameter space; calculating the predicted coating bonding strength corresponding to each set of parameters through finite element simulation; using a genetic algorithm to select a subset of parameters whose bonding strength is greater than a preset strength threshold; performing secondary optimization within the parameter subset based on a particle swarm optimization algorithm, and finally outputting the global optimal solution that satisfies the dual constraints of thermal conductivity and electrical conductivity.

[0009] Furthermore, the dynamic adjustment includes an anode plate area ratio correction equation: In the formula, This is the adjusted anode plate area ratio. This is the initial area ratio setting value. This is the concentration compensation coefficient for the activating solution. The current density response factor, This refers to the real-time measured concentration of hydrochloric acid. For standard process concentration, Given the current current density, The reference current density is used.

[0010] Furthermore, the adaptive feedback mechanism is implemented through a coating defect detection subsystem, which includes: a laser confocal microscope for real-time scanning of the coating surface morphology; an image processing unit for calculating porosity, crack propagation index, and grain boundary segregation; and a parameter re-optimization module for regenerating the corrected set of process parameters when the porosity exceeds a preset porosity threshold or the grain boundary segregation exceeds a preset grain boundary segregation threshold.

[0011] Furthermore, the parameter calculation model for the chemical degreasing process includes a coupling equation between sodium hydroxide concentration and temperature: In the formula, Where A is the optimal concentration of sodium hydroxide and A is the surface energy correction factor. denoted as the activation energy threshold, R as the gas constant, T as the actual solution temperature, B as the time response factor, and t as the degreasing treatment duration.

[0012] Furthermore, the nickel plating process employs a dual-layer neural network to predict pH fluctuation compensation. The input layer includes real-time current efficiency, nickel ion consumption rate, and boric acid buffer capacity; the hidden layer contains neurons and uses the ReLU activation function; and the output layer generates instructions for the amount of pH adjuster to be added, achieving a control precision that reaches a preset pH adjustment precision threshold.

[0013] Furthermore, the cathode moving speed matching algorithm for the tin-bismuth plating process includes: establishing a mapping relationship between the cathode moving frequency f and the current density D. In the formula, As the reference moving frequency, The current density sensitivity coefficient is... For temperature compensation factor, This is the standard temperature value for the process. Given the current current density, As a reference current density, This represents the actual temperature of the solution.

[0014] Furthermore, the parameter calculation system is deployed on an edge computing terminal, including: a multi-channel data acquisition card connecting various sensors; an FPGA chip executing optimization algorithms in real time; an industrial control module outputting PWM power adjustment signals to the electroplating power supply and temperature control equipment; and a human-machine interface displaying a three-dimensional coating quality cloud map and parameter correction suggestions.

[0015] According to the above embodiments of the present invention, at least the following beneficial effects are achieved: 1. Improved coating uniformity and thermoelectric performance stability: By collecting multi-dimensional data such as current density, temperature field, solution concentration and coating thickness in real time through a sensor array, and combining iterative calculations of the thermoelectric performance parameter optimization model, the anode plate area ratio, current density threshold and solution temperature fluctuation range are dynamically adjusted to ensure optimal matching of process parameters such as chemical degreasing, activation, nickel plating, and tin-bismuth plating, thereby forming a uniform and dense thermoelectric functional layer on the wafer surface, significantly improving thermoelectric conversion efficiency and material reliability.

[0016] 2. Reduce coating defect rate and optimize process tolerance: Based on the adaptive feedback mechanism, the coating porosity, crack propagation and grain boundary segregation are detected in real time by laser confocal microscope. When the parameters exceed the standard, the parameter re-optimization module is triggered to automatically correct the process deviation, reduce coating defects caused by solution concentration fluctuations, uneven current distribution or temperature changes, and improve product yield and batch consistency.

[0017] 3. Achieve intelligent and efficient production control: The parameter optimization system is deployed using edge computing terminals. Combined with FPGA real-time computing and industrial control module output PWM power adjustment signals, it can quickly respond to process changes. At the same time, through a dual-layer neural network to predict the nickel plating pH value compensation amount and cathode moving speed matching algorithm, the coating bonding strength and structural density are optimized, reducing manual intervention and improving production efficiency and process repeatability. Attached Figure Description

[0018] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of the invention are illustrated in the drawings by way of example and not limitation, wherein: Figure 1 This is a schematic flowchart of a surface treatment method for bismuth telluride-based thermoelectric material wafers according to an embodiment of the present invention. Detailed Implementation

[0019] The principles and spirit of the invention will now be described with reference to several exemplary embodiments. It should be understood that these embodiments are provided merely to enable those skilled in the art to better understand and implement the invention, and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided to make the invention more thorough and complete, and to fully convey the scope of the invention to those skilled in the art.

[0020] Those skilled in the art will recognize that embodiments of the present invention can be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0021] It should be noted that the number of any elements in the accompanying drawings is for illustrative purposes only and not as a limitation, and any naming is for distinction only and has no limiting meaning.

[0022] The following is for reference. Figure 1 , Figure 1 This is a schematic flowchart illustrating a surface treatment method for a bismuth telluride-based thermoelectric material wafer according to an embodiment of the present invention. Figure 1 As shown, a method for surface treatment of bismuth telluride-based thermoelectric material wafers includes: S1, real-time acquisition of multi-dimensional data during the wafer surface treatment process using a sensor array, wherein the data includes current density distribution, temperature field distribution, solution concentration gradient, and coating thickness increment.

[0023] S2. Based on the thermoelectric performance parameter optimization model, iterative calculations are performed on the collected data to generate the optimal parameter combination for the chemical degreasing, activation, nickel plating, and tin-bismuth plating processes.

[0024] S3. Dynamically adjust the anode plate area ratio, current density threshold, and solution temperature fluctuation range of the electroplating equipment according to the parameter combination.

[0025] S4. Based on the adaptive feedback mechanism, the parameter deviation is corrected to form a uniform thermoelectric functional layer structure on the wafer surface.

[0026] It should be noted that this invention proposes a surface treatment method for bismuth telluride-based thermoelectric material wafers. The core of this method lies in the real-time acquisition of multi-dimensional data during the wafer surface treatment process using a sensor array. This multi-dimensional data includes current density distribution, temperature field distribution, solution concentration gradient, and coating thickness increment, which comprehensively reflect the physical and chemical changes during the wafer surface treatment process. Based on a thermoelectric performance parameter optimization model, the acquired data is iteratively calculated to generate the optimal parameter combination for chemical degreasing, activation, nickel plating, and tin-bismuth plating processes. This process involves complex physicochemical reactions; the optimization model ensures that the parameters of each process reach their optimal state, thereby improving the thermoelectric performance of the wafer. The anode plate area ratio, current density threshold, and solution temperature fluctuation range of the electroplating equipment are dynamically adjusted according to the parameter combination. This step ensures precise parameter control during the electroplating process to adapt to the optimal process requirements under different conditions. Finally, an adaptive feedback mechanism corrects the parameter deviation, resulting in a uniform thermoelectric functional layer structure on the wafer surface. This mechanism can monitor and adjust process parameters in real time, ensuring the quality of the final product.

[0027] Specifically, the sensor array includes a four-dimensional current probe matrix, an infrared thermal imaging unit, an electrochemical impedance spectroscopy (EIS) module, and an X-ray fluorescence thickness gauge. The four-dimensional current probe matrix is ​​used to accurately measure current density distribution, detecting minute current changes and providing data support for current control during the electroplating process. The infrared thermal imaging unit can monitor the temperature field distribution in real time, allowing for a direct view of temperature changes on the wafer surface, which is crucial for controlling temperature uniformity during electroplating. The EIS module analyzes solution concentration gradients, accurately measuring the concentration changes of various ions in the solution based on electrochemical principles, thus providing a basis for solution formulation and adjustment. The X-ray fluorescence thickness gauge measures the coating thickness increment, monitoring coating growth in real time to ensure the coating reaches the designed thickness. The real-time data acquisition achieves a preset temporal accuracy threshold and a preset spatial accuracy threshold, meaning the data acquisition system can acquire accurate data within a very short time and a very small spatial range, providing high-quality data support for subsequent optimization calculations. The parameter combination includes the fluctuation range of hydrochloric acid concentration in the activation solution, the dynamic correction coefficient of the nickel plating pH value, and the matching factor of the cathode moving speed in tin plating. These parameters are key factors affecting the surface treatment quality of the wafer. Optimizing these parameters can significantly improve the thermoelectric performance of the wafer.

[0028] Preferably, the objective function expression of the thermoelectric performance parameter optimization model is a complex function that comprehensively considers multiple factors. This can be understood as the model comprehensively optimizing factors such as temperature field uniformity, coating density, and thermal stress through parameters such as weighting factors, optimization coefficients, and suppression factors. The iterative calculation process includes generating multiple sets of process parameter combinations within the initial parameter space, calculating the predicted coating bonding strength corresponding to each set of parameters through finite element simulation, using a genetic algorithm to select a subset of parameters with bonding strength greater than a preset strength threshold, and then performing secondary optimization within the parameter subset based on a particle swarm optimization algorithm. Finally, it outputs the globally optimal solution that satisfies both thermal conductivity and electrical conductivity constraints. This process involves multiple advanced algorithms and models, which ensure that the optimal parameter combination is found under complex process conditions. The anode plate area ratio correction equation in the dynamic adjustment process is a mathematical model that dynamically adjusts the anode plate area ratio based on real-time detection data. Although the specific equation is not given, it can be understood that the model dynamically adjusts the anode plate area ratio by considering factors such as activation solution concentration and current density to adapt to different process conditions. The adaptive feedback mechanism is implemented through a coating defect detection subsystem. This subsystem includes a laser confocal microscope that scans the coating surface morphology in real time, and an image processing unit that calculates porosity, crack propagation index, and grain boundary segregation. When these parameters exceed preset thresholds, a parameter re-optimization module is triggered to regenerate a corrected set of process parameters. This mechanism can monitor the coating quality in real time and adjust process parameters promptly when problems are detected, ensuring the quality of the final product.

[0029] In some embodiments, the sensor array includes a four-dimensional current probe matrix, an infrared thermal imaging unit, an electrochemical impedance spectroscopy analysis module, and an X-ray fluorescence thickness gauge; the temporal resolution of the real-time acquired data reaches a preset time accuracy threshold, and the spatial resolution reaches a preset spatial accuracy threshold; the parameter combination includes the fluctuation range of hydrochloric acid concentration in the activation solution, the dynamic correction coefficient of the nickel plating pH value, and the matching factor of the tin-bismuth cathode movement speed.

[0030] The sensor array comprises a four-dimensional current probe matrix, an infrared thermal imaging unit, an electrochemical impedance spectroscopy (EIS) module, and an X-ray fluorescence thickness gauge. These devices can acquire multi-dimensional data in real time during the wafer surface treatment process, such as current density distribution, temperature field distribution, solution concentration gradient, and coating thickness increment. Data acquisition requires not only high temporal resolution to capture transient changes but also high spatial resolution to accurately reflect the state of different locations on the wafer surface. The fluctuation range of the activation solution hydrochloric acid concentration, the dynamic correction coefficient for nickel plating pH, and the matching factor for the cathode movement speed in tin-bismuth plating are crucial parameters affecting the surface treatment effect. Optimizing these parameters can improve the quality and efficiency of wafer surface treatment.

[0031] Specifically, the four-dimensional current probe matrix can simultaneously measure current density distribution in four dimensions, including current magnitude, direction, location, and temporal variation. The infrared thermal imaging unit generates temperature distribution images by detecting infrared radiation from the object's surface, enabling real-time monitoring of temperature changes on the wafer surface. The electrochemical impedance spectroscopy module is an electrochemical measurement technique that analyzes the electrochemical properties of a solution, including concentration gradients, by applying a small-amplitude AC signal and measuring its response. The X-ray fluorescence thickness gauge utilizes the principle of X-ray fluorescence to determine the coating thickness by measuring the intensity of X-rays. The collaborative operation of these devices provides comprehensive and accurate data support. The hydrochloric acid concentration fluctuation range of the activation solution refers to the allowable range of hydrochloric acid concentration variation during activation; this parameter directly affects the activation effect on the wafer surface. The dynamic pH correction coefficient for nickel plating is a parameter used to adjust the pH value of the solution during nickel plating. It can be dynamically adjusted based on real-time monitoring of pH changes to ensure stable nickel plating. The cathode moving speed matching factor for tin-bismuth plating is a parameter used to control the cathode moving speed during tin-bismuth plating; it needs to be matched with factors such as current density to ensure coating uniformity.

[0032] Preferably, the four-dimensional current probe matrix in the sensor array can achieve high-precision current density measurement through sensor layout and processing technology. The infrared thermal imaging unit can be set with different temperature ranges and resolutions to adapt to different measurement needs. The electrochemical impedance spectroscopy analysis module can optimize measurement accuracy by selecting appropriate frequency ranges and signal amplitudes. The X-ray fluorescence thickness gauge can improve the accuracy of thickness measurement by adjusting the X-ray energy and measurement time. In terms of data processing, advanced algorithms can be used to filter, calibrate, and analyze the acquired data to improve its reliability and usability. For example, regarding the fluctuation range of hydrochloric acid concentration in the activation solution, the optimal concentration range can be determined experimentally, and the amount of activation solution added can be dynamically adjusted based on real-time monitoring data. For the dynamic correction coefficient of nickel plating pH, a feedback control model based on real-time pH and a preset target pH can be established, and the amount of pH adjuster added can be adjusted by calculating the deviation. For the cathode moving speed matching factor in tin-bismuth plating, the cathode moving speed can be dynamically adjusted according to changes in current density to ensure the uniformity and quality of the plating layer. These optimization measures can further improve the accuracy and efficiency of wafer surface treatment.

[0033] In some embodiments, the objective function expression of the thermoelectric performance parameter optimization model is: In the formula, Let be the objective function. This is the weighting factor for temperature field uniformity. γ is the coating density optimization coefficient, and γ is the thermal stress suppression factor. This represents the measured value at the i-th temperature monitoring point. The standard temperature threshold for the process. The ion mobility in the solution, This is the critical value for lattice distortion. This is the upper limit of thermal stress. The starting time of the integration. This is the end time for integration.

[0034] The objective function comprehensively considers various factors affecting thermoelectric performance, such as temperature field uniformity, coating density, and thermal stress. By assigning different weighting factors to these factors, the optimization model can seek a balance among multiple objectives, thereby generating the optimal combination of process parameters. This process involves complex physicochemical principles and requires precise model construction and parameter settings.

[0035] Specifically, the temperature field uniformity weighting factor is a parameter used to measure the importance of temperature field uniformity in thermoelectric performance optimization. It reflects the degree of influence of temperature field uniformity on overall thermoelectric performance; a larger value indicates a greater contribution of temperature field uniformity to the optimization objective. The coating density optimization coefficient measures the importance of coating density in the optimization process, directly affecting the quality and performance of the coating. The thermal stress suppression factor is a parameter used to control the impact of thermal stress on thermoelectric performance. Excessive thermal stress may lead to a decrease in wafer performance, therefore it needs to be suppressed through the optimization model.

[0036] Preferably, the objective function construction process of the thermoelectric performance parameter optimization model can be further refined. First, the specific values ​​of each weighting factor need to be determined, which can be obtained through a combination of experimental data and theoretical analysis. For example, by experimentally measuring the thermoelectric performance under different temperature field uniformities, the reasonable range of the temperature field uniformity weighting factor α can be determined. Similarly, by analyzing the influence of coating density on thermoelectric performance, the specific value of the coating density optimization coefficient β can be determined. The thermal stress suppression factor γ can be determined by simulating the influence of thermal stress on wafer performance. In terms of data processing, the parameter calculations in the objective function can be dynamically updated using real-time data. For example, the measured values ​​at temperature monitoring points can be acquired in real time using an infrared thermal imaging unit, the solution ion mobility can be measured using an electrochemical impedance spectroscopy analysis module, and the critical value of lattice distortion can be indirectly estimated using equipment such as an X-ray fluorescence thickness gauge. Through the input of these real-time data, the optimization model can dynamically adjust the parameters to ensure optimal thermoelectric performance under different process conditions.

[0037] In some embodiments, the iterative calculation includes the following steps: generating multiple sets of process parameter combinations in the initial parameter space; calculating the predicted coating bonding strength corresponding to each set of parameters through finite element simulation; using a genetic algorithm to select a subset of parameters whose bonding strength is greater than a preset strength threshold; performing secondary optimization within the parameter subset based on a particle swarm optimization algorithm, and finally outputting the global optimal solution that satisfies the dual constraints of thermal conductivity and electrical conductivity.

[0038] It should be noted that the iterative calculation process mentioned in this invention is a key step in optimizing wafer surface treatment parameters. This process initializes multiple sets of process parameter combinations within the parameter space, uses finite element simulation to calculate the predicted coating bonding strength for each set of parameters, then uses a genetic algorithm to select a subset of parameters with bonding strength greater than a preset strength threshold, and finally performs secondary optimization within the parameter subset based on a particle swarm optimization algorithm to output a globally optimal solution that satisfies both thermal conductivity and electrical conductivity constraints. This process involves various advanced algorithms and models, ensuring that the optimal parameter combination is found under complex process conditions, thereby improving the quality and efficiency of wafer surface treatment.

[0039] Specifically, initializing multiple sets of process parameter combinations within the parameter space refers to generating a series of possible parameter combinations based on process experience and theoretical analysis before the optimization algorithm begins. These parameter combinations cover various factors that may affect the wafer surface treatment effect, such as current density, solution concentration, and temperature. Finite element simulation is a numerical analysis method that predicts the performance of the entire system by discretizing a complex physical problem into a finite number of elements and calculating the response of each element.

[0040] More specifically, in this invention, finite element simulation is used to calculate the coating bonding strength under different parameter combinations, providing a basis for subsequent parameter selection. Genetic algorithms are optimization algorithms that simulate biological evolution, selecting a subset of parameters with better performance from the initial parameter combinations through selection, crossover, and mutation operations. Particle swarm optimization is an optimization algorithm based on swarm intelligence, further optimizing the parameter subset by simulating the foraging behavior of bird flocks, ultimately finding the globally optimal solution that satisfies the dual constraints of thermal conductivity and electrical conductivity. The combined use of these algorithms can effectively improve the efficiency and accuracy of parameter optimization.

[0041] Preferably, the iterative calculation process can be further refined. When initializing the parameter space, the reasonable range and distribution of parameters can be determined by combining experimental data and theoretical analysis. For example, the current density can be initially set based on the material properties and process requirements of the wafer, the solution concentration can be adjusted based on the equilibrium conditions of the chemical reaction, and the temperature can be optimized based on the thermodynamic principles of the electroplating process. The construction of the finite element simulation model requires inputting detailed process parameters and material properties, such as the wafer's size, shape, material conductivity, and thermal conductivity, while also setting appropriate boundary and initial conditions. In the genetic algorithm, the selection operation can use roulette wheel selection based on the predicted value of the coating bonding strength, the crossover operation can use single-point or multi-point crossover, and the mutation operation can use uniform mutation or Gaussian mutation to increase the diversity of parameter combinations. In the particle swarm optimization algorithm, the position and velocity update rules for each particle can be adjusted according to the optimization objective. For example, by introducing inertia weights and learning factors, the global search capability and local search capability can be balanced. In terms of data processing, the results of the finite element simulation and algorithm calculations can be filtered and calibrated to improve the reliability and accuracy of the data. These refined steps can further improve the efficiency of iterative calculations and the quality of optimization results.

[0042] In some embodiments, the dynamic adjustment includes an anode plate area ratio correction equation: In the formula, This is the adjusted anode plate area ratio. This is the initial area ratio setting value. This is the concentration compensation coefficient for the activating solution. The current density response factor, This refers to the real-time measured concentration of hydrochloric acid. For standard process concentration, Given the current current density, The reference current density is used.

[0043] By using an anode plate area ratio correction equation, the anode plate area ratio is dynamically adjusted based on real-time monitored parameters such as hydrochloric acid concentration and current density. This dynamic adjustment ensures precise parameter control during electroplating to adapt to optimal process requirements under different conditions, thereby improving the quality and efficiency of wafer surface treatment.

[0044] Specifically, the anode plate area ratio correction equation is a mathematical model that dynamically adjusts the anode plate area ratio based on real-time detection data. The initial area ratio setting is a pre-set value for the anode plate area ratio based on process experience and theoretical analysis. The activator concentration compensation coefficient is a parameter used to compensate for the impact of changes in activator concentration on the anode plate area ratio; it reflects the degree of influence of changes in activator concentration on the electroplating process. The current density response factor is a parameter used to adjust the impact of changes in current density on the anode plate area ratio; it reflects the degree of influence of changes in current density on the electroplating process. The real-time detected hydrochloric acid concentration value is the actual concentration of hydrochloric acid in the activator solution measured in real-time by a sensor. The standard process concentration is the standard value of hydrochloric acid concentration set according to process requirements. The current current density is the current density value measured in real-time during the electroplating process, and the reference current density is the reference value of current density set according to process requirements. These parameters are comprehensively calculated through the anode plate area ratio correction equation to dynamically adjust the anode plate area ratio, ensuring the stability and uniformity of the electroplating process.

[0045] Preferably, the process of constructing the anode plate area ratio correction equation can be further refined. When constructing the equation, it is first necessary to determine the reasonable range of the initial area ratio setting through experiments. This can be done by conducting electroplating experiments at different initial area ratios and measuring the coating quality. The activator concentration compensation coefficient can be determined by analyzing the effect of hydrochloric acid concentration changes on coating quality. For example, by conducting electroplating experiments at different hydrochloric acid concentrations and measuring the coating thickness and uniformity, the value of the concentration compensation coefficient can be determined. The current density response factor can be determined by analyzing the effect of current density changes on coating quality. For example, by conducting electroplating experiments at different current densities and measuring the coating thickness and uniformity, the value of the current density response factor can be determined. In terms of data processing, the real-time detected hydrochloric acid concentration and current density values ​​can be filtered and calibrated to improve the accuracy and reliability of the data. Through these refined steps, the accuracy and dynamic adjustment effect of the anode plate area ratio correction equation can be further improved.

[0046] In some embodiments, the adaptive feedback mechanism is implemented through a coating defect detection subsystem, which includes: a laser confocal microscope for real-time scanning of the coating surface morphology; an image processing unit for calculating porosity, crack propagation index, and grain boundary segregation; and a parameter re-optimization module for regenerating a corrected set of process parameters when the porosity exceeds a preset porosity threshold or the grain boundary segregation exceeds a preset grain boundary segregation threshold.

[0047] The adaptive feedback mechanism monitors the morphology of the coating surface in real time through the coating defect detection subsystem, including parameters such as porosity, crack propagation index, and grain boundary segregation. When these parameters exceed preset thresholds, the adaptive feedback mechanism triggers a parameter re-optimization module to regenerate a corrected set of process parameters, thereby ensuring the formation of a uniform thermoelectric functional layer structure on the wafer surface. This mechanism effectively addresses fluctuations and uncertainties in the process, improving the stability and reliability of wafer surface treatment.

[0048] The coating defect detection subsystem includes a laser confocal microscope, an image processing unit, and a parameter re-optimization module. The laser confocal microscope is used to scan the coating surface morphology in real time, providing high-resolution images. The image processing unit analyzes these images, calculating parameters such as porosity, crack propagation index, and grain boundary segregation. Porosity refers to the volume proportion of pores in the coating, the crack propagation index measures the crack propagation trend, and grain boundary segregation reflects the compositional inhomogeneity at grain boundaries. The preset thresholds for these parameters are set according to process requirements and product quality standards. When the detected parameters exceed these thresholds, the parameter re-optimization module regenerates a corrected set of process parameters based on real-time data to adjust subsequent surface treatment processes.

[0049] The scanning parameters of a laser confocal microscope, such as scanning speed, resolution, and scanning range, need to be optimized based on the wafer's size and surface characteristics. The image processing unit can employ advanced image recognition algorithms, such as edge detection and feature extraction, to accurately calculate porosity, crack propagation index, and grain boundary segregation. The parameter re-optimization module can dynamically adjust process parameters based on historical and real-time monitoring data, using machine learning or optimization algorithms. For example, neural network algorithms can be used to train historical data, establishing a mapping relationship between parameters and coating quality, thereby quickly generating corrective parameters when defects are detected. In terms of data processing, real-time monitoring data can be filtered and calibrated to improve accuracy and reliability. Through these refined steps, the response speed and adjustment accuracy of the adaptive feedback mechanism can be further improved, ensuring high quality and high stability of wafer surface treatment.

[0050] In some embodiments, the parameter calculation model for the chemical degreasing process includes a coupling equation between sodium hydroxide concentration and temperature: In the formula, Where A is the optimal concentration of sodium hydroxide and A is the surface energy correction factor. denoted as the activation energy threshold, R as the gas constant, T as the actual solution temperature, B as the time response factor, and t as the degreasing treatment duration.

[0051] It should be noted that the parameter calculation model for the chemical degreasing process precisely controls the process by coupling the sodium hydroxide concentration with temperature. This coupling relationship considers factors such as the surface energy correction coefficient, activation energy threshold, gas constant, actual solution temperature, time response factor, and degreasing treatment duration, effectively removing oil contaminants from the wafer surface and providing a solid foundation for subsequent surface treatment processes. By optimizing these parameters, the stability and uniformity of the degreasing effect can be significantly improved.

[0052] The sodium hydroxide concentration-temperature coupling equation is a mathematical model describing the interaction between sodium hydroxide concentration and temperature during chemical degreasing. The surface energy correction coefficient A is a parameter used to adjust the solution surface energy, reflecting the strength of the interaction between the solution and the wafer surface. The activation energy threshold Ea is the minimum energy required for the chemical reaction to occur, determining the reaction rate. The gas constant R is a physical constant used to describe the thermodynamic properties of gases. The actual solution temperature T is the actual temperature of the solution during chemical degreasing, directly affecting the rate of the chemical reaction. The time response factor B is a parameter used to adjust the reaction time, reflecting the change in reaction rate over time. The degreasing treatment duration t is the duration of the chemical degreasing process, determining the thoroughness of the degreasing effect. These parameters are comprehensively calculated through the coupling equation to dynamically adjust the sodium hydroxide concentration and temperature to achieve the optimal degreasing effect.

[0053] When constructing the equations, the first step is to experimentally determine the reasonable ranges for the surface energy correction coefficient A and the activation energy threshold Ea. This can be achieved by conducting degreasing experiments at different concentrations and temperatures, measuring the amount of residual oil on the wafer surface. The gas constant R is a known physical constant, and its standard value can be used directly. The actual solution temperature T can be measured in real time using a temperature sensor to ensure precise temperature control. The time response factor B can be determined by analyzing the effect of time on the reaction rate during the degreasing process; for example, by measuring the degreasing effect at different time intervals to determine the value of the time response factor. In terms of data processing, the real-time measured temperature and concentration data can be filtered and calibrated to improve the accuracy and reliability of the data. Through these refined steps, the accuracy and dynamic adjustment effect of the sodium hydroxide concentration-temperature coupling equation can be further improved, ensuring the high efficiency and stability of the chemical degreasing process.

[0054] In some embodiments, the nickel plating process uses a two-layer neural network to predict the pH value fluctuation compensation amount. The input layer includes real-time current efficiency, nickel ion consumption rate and boric acid buffer capacity; the hidden layer sets neurons and uses the ReLU activation function; the output layer generates a pH adjuster addition instruction, and the control accuracy reaches a preset pH adjustment accuracy threshold.

[0055] The nickel plating process employs a dual-layer neural network to predict pH fluctuation compensation, precisely adjusting the pH of the solution during plating. Through its input layer, hidden layer, and output layer structure, the dual-layer neural network dynamically predicts and adjusts the pH based on parameters such as real-time current efficiency, nickel ion consumption rate, and borate buffer capacity, ensuring the stability of the nickel plating process and the quality of the plating layer. This technology effectively addresses fluctuations and uncertainties in the process, improving the automation and precision of the nickel plating procedure.

[0056] A two-layer neural network is a neural network structure consisting of an input layer, hidden layers, and an output layer. The input layer includes parameters such as real-time current efficiency, nickel ion consumption rate, and borate buffer capacity. These parameters are key factors affecting pH changes during nickel plating. Real-time current efficiency refers to the efficiency with which the actual current is used for coating deposition during nickel plating, reflecting the effective utilization rate of the current. Nickel ion consumption rate refers to the rate at which nickel ions are consumed from the solution during nickel plating, directly affecting the chemical equilibrium of the solution. Borate buffer capacity refers to the ability of borate in the solution to buffer pH changes, determining the solution's resistance to pH fluctuations. The hidden layer contains neurons and uses the ReLU activation function. The ReLU activation function is a commonly used nonlinear activation function that introduces nonlinearity, improving the network's expressive power. The output layer generates instructions for the amount of pH adjuster to be added, controlling the precision to a preset pH adjustment precision threshold. This means that the output layer can generate precise pH adjuster addition instructions based on the network's predictions, ensuring pH stability.

[0057] Preferably, when constructing the network, the parameters of the input layer need to be determined first. These parameters need to be collected in real time by sensors to ensure the accuracy and real-time nature of the data. The number of neurons in the hidden layer can be adjusted according to the complexity of the actual problem; for example, the optimal number of neurons can be determined through methods such as cross-validation. The parameters of the ReLU activation function can be optimized based on the training data to improve the network performance. The pH regulator addition command of the output layer needs to be accurately calculated based on the network's prediction results, which can be achieved by establishing a mapping relationship between pH value and regulator addition amount. In terms of data processing, the real-time input data can be normalized to improve the network's training efficiency and prediction accuracy. Through these refined steps, the prediction ability and control accuracy of the two-layer neural network can be further improved, ensuring the stability of pH value and the uniformity of coating quality during nickel plating.

[0058] In some embodiments, the cathode moving speed matching algorithm for the tin-bismuth plating process includes: establishing a mapping relationship between the cathode moving frequency f and the current density D: In the formula, As the reference moving frequency, The current density sensitivity coefficient is... For temperature compensation factor, This is the standard temperature value for the process. Given the current current density, As a reference current density, This represents the actual temperature of the solution.

[0059] The cathode moving speed matching algorithm in the tin-bismuth plating process is a crucial step in ensuring the quality of wafer surface treatment. By establishing a mapping relationship between cathode moving frequency and current density, the cathode moving speed is dynamically adjusted to adapt to the optimal process requirements under different current density conditions. This dynamic adjustment ensures the uniformity and quality stability of the plating layer, thereby improving the overall effect of wafer surface treatment.

[0060] Specifically, the cathode moving speed matching algorithm includes parameters such as the reference moving frequency, current density sensitivity coefficient, and temperature compensation factor. The reference moving frequency is the initial frequency of cathode movement, set based on process experience and experimental data, reflecting the optimal moving speed under standard process conditions. The current density sensitivity coefficient is a parameter used to adjust the impact of current density changes on the cathode moving speed; it reflects the degree of influence of current density changes on coating uniformity. The temperature compensation factor is a parameter used to adjust the impact of temperature changes on the cathode moving speed; it reflects the degree of influence of temperature changes on coating quality. The process standard temperature value is a temperature standard value set according to process requirements to ensure temperature stability and consistency. These parameters are comprehensively calculated by the cathode moving speed matching algorithm to dynamically adjust the cathode moving speed, ensuring the uniformity and quality stability of the coating.

[0061] When constructing the cathode moving speed matching algorithm, the first step is to experimentally determine a reasonable range for the reference moving frequency. This can be achieved by conducting tin-bismuth plating experiments at different moving frequencies and measuring the thickness uniformity and quality of the plating layer. The current density sensitivity coefficient can be determined by analyzing the impact of current density changes on plating quality. For example, by conducting experiments at different current densities and measuring the plating thickness and uniformity, the value of the current density sensitivity coefficient can be determined. The temperature compensation factor can be determined by analyzing the impact of temperature changes on plating quality. For example, by conducting experiments at different temperatures and measuring the plating thickness and uniformity, the value of the temperature compensation factor can be determined. In terms of data processing, the real-time measured current density and temperature data can be filtered and calibrated to improve the accuracy and reliability of the data. Through these refined steps, the accuracy and dynamic adjustment effect of the cathode moving speed matching algorithm can be further improved, ensuring the efficiency and stability of the tin-bismuth plating process.

[0062] In some embodiments, the parameter calculation system is deployed on an edge computing terminal, including: a multi-channel data acquisition card connected to various sensors; an FPGA chip executing optimization algorithms in real time; an industrial control module outputting PWM power adjustment signals to the electroplating power supply and temperature control equipment; and a human-machine interface displaying a three-dimensional coating quality cloud map and parameter correction suggestions.

[0063] It should be noted that the parameter calculation system mentioned in this invention is deployed on an edge computing terminal. This is a highly integrated system used for real-time processing and optimization of various parameters during wafer surface treatment. The system includes a multi-channel data acquisition card, an FPGA chip, an industrial control module, and a human-machine interface, enabling fully automated control from data acquisition to parameter optimization. Through this deployment method, the system can quickly respond to changes in the process and adjust process parameters in real time, ensuring high quality and high efficiency in wafer surface treatment.

[0064] Specifically, a multi-channel data acquisition card connects to various sensors, enabling simultaneous acquisition of multiple data points such as current density, temperature, solution concentration, and coating thickness. An FPGA chip, a programmable logic chip, executes optimization algorithms in real time, rapidly processing large amounts of data and generating optimized results. An industrial control module outputs PWM power control signals to the electroplating power supply and temperature control equipment, ensuring the stability and accuracy of process parameters through precise power control. A human-machine interface displays a 3D coating quality cloud map and parameter correction suggestions, allowing operators to intuitively understand the process status and adjust parameters according to system recommendations. These components work together to form a complete automated control system, effectively improving the efficiency and quality of wafer surface treatment.

[0065] In setting up a multi-channel data acquisition card, the appropriate number of channels and sampling rate need to be configured according to the type and quantity of sensors to ensure the accuracy and real-time performance of data acquisition. FPGA chip programming requires designing suitable logic circuits and algorithm implementations based on the specific needs of the optimization algorithm; for example, a parallel processing architecture can be used to improve algorithm execution speed. Parameter settings for the industrial control module require adjusting the frequency and duty cycle of the PWM signal according to the characteristics of the electroplating power supply and temperature control equipment to achieve precise power control. The design of the human-machine interface needs to consider the operator's usage habits, providing an intuitive graphical interface and a simple operating process. In terms of data processing, preprocessing of the acquired data, such as filtering and noise reduction, can improve data quality. Through these detailed steps, the performance and reliability of the parameter calculation system can be further improved, ensuring the efficiency and stability of the wafer surface treatment process.

[0066] The above embodiments of the present invention have the following beneficial effects: 1. Improved coating uniformity and thermoelectric performance stability: By collecting multi-dimensional data such as current density, temperature field, solution concentration and coating thickness in real time through sensor array, and combining iterative calculations of thermoelectric performance parameter optimization model, the anode plate area ratio, current density threshold and solution temperature fluctuation range are dynamically adjusted to ensure optimal matching of process parameters such as chemical degreasing, activation, nickel plating, and tin-bismuth plating, thereby forming a uniform and dense thermoelectric functional layer on the wafer surface, significantly improving thermoelectric conversion efficiency and material reliability.

[0067] 2. Reduce coating defect rate and optimize process tolerance: Based on the adaptive feedback mechanism, the coating porosity, crack propagation and grain boundary segregation are detected in real time by laser confocal microscope. When the parameters exceed the standard, the parameter re-optimization module is triggered to automatically correct the process deviation, reduce coating defects caused by solution concentration fluctuations, uneven current distribution or temperature changes, and improve product yield and batch consistency.

[0068] 3. Achieve intelligent and efficient production control: The parameter optimization system is deployed using edge computing terminals. Combined with FPGA real-time computing and industrial control module output PWM power adjustment signals, it can quickly respond to process changes. At the same time, through a dual-layer neural network to predict the nickel plating pH value compensation amount and cathode moving speed matching algorithm, the coating bonding strength and structural density are optimized, reducing manual intervention and improving production efficiency and process repeatability.

[0069] Furthermore, the storage medium in the embodiments of this application stores program instructions capable of implementing all the above methods. These program instructions can be stored in the storage medium in the form of a software product, including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, or terminal devices such as computers, servers, mobile phones, and tablets.

[0070] The above description is merely an explanation of some preferred embodiments of the present invention and the technical principles employed. Those skilled in the art should understand that the scope of the invention as described in the embodiments of the present invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. A method for surface treatment of bismuth telluride-based thermoelectric material wafers, characterized in that, include: Multi-dimensional data during the wafer surface treatment process are collected in real time using a sensor array. The data includes current density distribution, temperature field distribution, solution concentration gradient, and coating thickness increment. Based on the thermoelectric performance parameter optimization model, the collected data is iteratively calculated to generate the optimal parameter combination for chemical degreasing, activation, nickel plating, and tin-bismuth plating processes; the anode plate area ratio, current density threshold, and solution temperature fluctuation range of the electroplating equipment are dynamically adjusted according to the parameter combination; the parameter deviation is corrected based on the adaptive feedback mechanism to form a uniform thermoelectric functional layer structure on the wafer surface.

2. The surface treatment method for bismuth telluride-based thermoelectric material wafers according to claim 1, characterized in that, The sensor array includes a four-dimensional current probe matrix, an infrared thermal imaging unit, an electrochemical impedance spectroscopy analysis module, and an X-ray fluorescence thickness gauge; the time resolution of the real-time acquired data reaches a preset time accuracy threshold, and the spatial resolution reaches a preset spatial accuracy threshold; the parameter combination includes the fluctuation range of hydrochloric acid concentration in the activation solution, the dynamic correction coefficient of the nickel plating pH value, and the matching factor of the tin-bismuth cathode movement speed.

3. The surface treatment method for bismuth telluride-based thermoelectric material wafers according to claim 2, characterized in that, The objective function expression of the thermoelectric performance parameter optimization model is: In the formula, Let be the objective function. This is the weighting factor for temperature field uniformity. γ is the coating density optimization coefficient, and γ is the thermal stress suppression factor. This represents the measured value at the i-th temperature monitoring point. The standard temperature threshold for the process. The ion mobility in the solution, This is the critical value for lattice distortion. This is the upper limit of thermal stress. The starting time of the integration. This is the end time for integration.

4. The surface treatment method for bismuth telluride-based thermoelectric material wafers according to claim 3, characterized in that, The iterative calculation includes the following steps: generating multiple sets of process parameter combinations in the initial parameter space; calculating the predicted coating bonding strength corresponding to each set of parameters through finite element simulation; using a genetic algorithm to select a subset of parameters whose bonding strength is greater than a preset strength threshold; performing secondary optimization within the parameter subset based on a particle swarm optimization algorithm, and finally outputting the global optimal solution that satisfies the dual constraints of thermal conductivity and electrical conductivity.

5. The surface treatment method for bismuth telluride-based thermoelectric material wafers according to claim 4, characterized in that, The dynamic adjustment includes the anode plate area ratio correction equation: In the formula, This is the adjusted anode plate area ratio. This is the initial area ratio setting value. This is the concentration compensation coefficient for the activating solution. The current density response factor, This refers to the real-time measured concentration of hydrochloric acid. For standard process concentration, Given the current current density, The reference current density is used.

6. The surface treatment method for bismuth telluride-based thermoelectric material wafers according to claim 5, characterized in that, The adaptive feedback mechanism is implemented through a coating defect detection subsystem, which includes: A laser confocal microscope scans the surface morphology of the coating in real time; the image processing unit calculates porosity, crack propagation index and grain boundary segregation; when the porosity exceeds the preset porosity threshold or the grain boundary segregation exceeds the preset grain boundary segregation threshold, the parameter re-optimization module is triggered to regenerate the corrected set of process parameters.

7. The surface treatment method for bismuth telluride-based thermoelectric material wafers according to any one of claims 1-6, characterized in that, The parameter calculation model for the chemical degreasing process includes a coupled equation between sodium hydroxide concentration and temperature: In the formula, Where A is the optimal concentration of sodium hydroxide and A is the surface energy correction factor. denoted as the activation energy threshold, R as the gas constant, T as the actual solution temperature, B as the time response factor, and t as the degreasing treatment duration.

8. The surface treatment method for bismuth telluride-based thermoelectric material wafers according to claim 7, characterized in that, The nickel plating process uses a two-layer neural network to predict the pH value fluctuation compensation amount. The input layer includes real-time current efficiency, nickel ion consumption rate and boric acid buffer capacity; the hidden layer sets neurons and uses the ReLU activation function; the output layer generates pH adjuster addition instructions, and the control accuracy reaches the preset pH adjustment accuracy threshold.

9. The surface treatment method for bismuth telluride-based thermoelectric material wafers according to claim 8, characterized in that, The cathode moving speed matching algorithm for the tin-bismuth plating process includes: establishing the cathode moving frequency. Mapping relationship with current density D: In the formula, As the reference moving frequency, The current density sensitivity coefficient is... For temperature compensation factor, This is the standard temperature value for the process. Given the current current density, As a reference current density, This represents the actual temperature of the solution.

10. The surface treatment method for bismuth telluride-based thermoelectric material wafers according to claim 9, characterized in that, The parameter calculation system is deployed on an edge computing terminal and includes: a multi-channel data acquisition card connecting various sensors; an FPGA chip executing optimization algorithms in real time; an industrial control module outputting PWM power adjustment signals to the electroplating power supply and temperature control equipment; and a human-machine interface displaying a three-dimensional coating quality cloud map and parameter correction suggestions.

Images