Method and system for testing light color of semiconductor light-emitting device

Abstract

The invention discloses a light color test method for a semiconductor luminescent device, and the method comprises the steps: obtaining the electrical characteristic parameters of the semiconductor luminescent device under different current injection conditions, monitoring the internal voltage distribution, current density distribution and resistance change conditions of the device in real time, and building an electrical characteristic database; calculating the temperature field distribution generated by the heat effect in the device according to the change condition of the electrical characteristic parameters, and further obtaining the temperature gradient and heat flux density distribution data of each region of the device if the current density exceeds a preset threshold value; according to the temperature gradient and the heat flux density distribution data, light emitting wavelength, light intensity distribution and color coordinate changes of the semiconductor light emitting device under different temperature conditions are analyzed, and a mapping relation matrix between temperature and optical characteristics is obtained; according to a mapping relation matrix between the temperature and the optical characteristics, the rule that the optical characteristics change along with the temperature is analyzed, and a light color shift prediction model is established.

Description

Technical Field

[0001] This invention relates to the field of semiconductor performance testing technology, and in particular to a method and system for testing the light color of semiconductor light-emitting devices. Background Technology

[0002] Semiconductor light-emitting devices (LEDs) are a core technology in modern lighting, display, and communication fields, and their light color performance directly determines their performance in practical applications and their market competitiveness. Efficient and stable light color output is not only key to improving user experience but also crucial for energy conservation, environmental protection, and device lifespan. However, current research and applications still face significant technical bottlenecks.

[0003] Existing methods for designing and testing semiconductor light-emitting devices often fail to fully capture the complex behavior of devices in real-world operating environments. Many approaches focus on the analysis of single physical fields or static conditions, neglecting the dynamic interactions of various physical characteristics during device operation. This limitation leads to discrepancies between design parameters and actual performance, particularly in the issue of light color drift under high power or complex environments, which is difficult to accurately predict and optimize using traditional testing methods.

[0004] The core challenge lies in the difficulty of accurately simulating and controlling the interactions between the electrical, thermal, and optical properties within the device. During operation, current injection triggers thermal effects, and the resulting temperature rise alters the optical properties of the materials, leading to color shifts. For example, in high-brightness LEDs, increased current causes a rapid rise in chip temperature, altering the emission wavelength and causing the color to deviate from the design target. The complexity of this dynamic coupling effect makes optimizing device structure and material ratios exceptionally difficult. Furthermore, non-uniform material ratios or inappropriate structural design can further amplify color instability, especially during long-term operation or under extreme conditions, manifesting as decreased luminous efficiency and color temperature drift. Summary of the Invention

[0005] In order to solve the above-mentioned technical problems, the present invention provides a method and system for testing the light color of semiconductor light-emitting devices.

[0006] The technical solution of this invention is implemented as follows: A method and system for testing the light color of a semiconductor light-emitting device, comprising the following steps: Obtain the electrical characteristic parameters of semiconductor light-emitting devices under different current injection conditions, monitor the voltage distribution, current density distribution and resistance changes inside the devices in real time, and establish an electrical characteristic database; Based on the changes in electrical characteristic parameters, the temperature field distribution generated by the internal thermal effect of the device is calculated. If the current density exceeds the preset threshold, the temperature gradient and heat flux density distribution data of each region of the device are further obtained. Based on temperature gradient and heat flux density distribution data, the emission wavelength, light intensity distribution and color coordinate changes of semiconductor light-emitting devices under different temperature conditions are analyzed to obtain the mapping relationship matrix between temperature and optical properties; The mapping relationship matrix between temperature and optical properties is used to analyze the law of optical properties changing with temperature and establish a light color shift prediction model. If the temperature change exceeds the light color stability threshold, the light color shift amount and shift direction are calculated to determine the critical operating conditions of the device's light color stability. A dynamic coupling model of the device is constructed, and the interaction equations of the three physical fields of electricity, heat and optics are solved simultaneously to obtain the comprehensive performance of the device under real working environment.

[0007] Furthermore, the establishment of the electrical characteristic database specifically includes: The original set of electrical parameters is obtained by collecting data on voltage distribution, current density distribution and resistance change of semiconductor light-emitting devices under different current injection conditions through a sensor array; The original electrical parameter set is denoised and standardized using a preset signal processing algorithm to obtain a standardized electrical parameter set. This also includes accuracy optimization of the standardized set of electrical parameters: If the voltage distribution data in the standardized electrical parameter set exceeds the preset threshold range, then key electrical response features are extracted to obtain an electrical response feature set. Based on the electrical response feature set, the operating states of the devices are classified to obtain the operating state classification results. The working status classification results are grouped to obtain the grouping structure of the electrical characteristic database; Based on the grouping structure, the standardized set of electrical parameters and the set of electrical response characteristics are stored in the electrical property database to obtain a complete electrical property database; If the integrity of the data records in the complete electrical characteristic database meets the preset threshold requirements, the accuracy of the data in the database is verified by a data verification algorithm to obtain the final electrical characteristic database.

[0008] Furthermore, after acquiring the heat flux density distribution data of each region of the device, the process further includes: The heat flux density distribution data is divided into regions using a preset grid partitioning algorithm to obtain a set of heat flux partitions inside the device. If there are regions with abnormal heat flux density in the heat flux partition set, the heat flux density data of the abnormal regions are corrected through data standardization to obtain corrected heat flux density distribution data. Based on the corrected heat flux density distribution data, the internal heat flux zones of the device are classified to obtain the heat flux characteristic classification results. By using a preset data storage protocol, the heat flow characteristic classification results are associated with the temperature field distribution data and stored in the thermal characteristic database to obtain a complete thermal characteristic database.

[0009] Furthermore, the obtained mapping matrix between temperature and optical properties includes: The temperature change data of the device under different operating conditions is collected, and the collected data is preprocessed to obtain a normalized temperature data set. Based on the normalized temperature data set, the changes in the emission wavelength of the device under various temperature conditions are calculated to generate a wavelength distribution dataset. By using the wavelength distribution dataset, the light intensity distribution data of the device at different temperatures is extracted to obtain the light intensity distribution dataset; Based on the light intensity distribution dataset, the color coordinate changes of the device under various temperature conditions are analyzed to generate a color coordinate distribution dataset; If there are outliers in the color coordinate distribution dataset, the outliers are filtered out using a preset threshold judgment method to obtain the corrected color coordinate distribution dataset. Using the corrected chromaticity distribution dataset, the wavelength distribution dataset, light intensity distribution dataset, and chromaticity distribution dataset are jointly classified to generate a mapping matrix between temperature and optical properties. Based on the mapping matrix, the classification results are stored in the optical property database using a data storage protocol to obtain the associated temperature-optical property dataset.

[0010] Furthermore, the critical operating conditions for determining the optical color stability of the device include: Temperature change data of the device under different operating conditions are collected by sensors, and the data is preprocessed using a standardization method to obtain a normalized temperature dataset. Based on the normalized temperature dataset, the principal component analysis algorithm is used to extract key features of temperature changes and generate a temperature feature dataset. Using a temperature feature dataset, a light color shift prediction model is constructed using a support vector regression algorithm to obtain the light color shift prediction results. If the light color shift prediction result exceeds the preset stability threshold, the light color shift amount and shift direction are calculated by vector decomposition method to obtain the shift amount and direction dataset. Based on the offset and orientation datasets, cluster analysis is used to determine the stable interval of the light color offset, thus obtaining the range of light color stability. By using the range of optical color stability, the critical operating conditions of the device at different temperatures are identified using a boundary division method, and a critical condition dataset is obtained. Based on the critical condition dataset, the results are stored in the optical property database using a data storage protocol to obtain the associated temperature-photochromic stability dataset.

[0011] Furthermore, the comprehensive performance of the device under real-world operating conditions is specifically reflected by obtaining multiphysics coupling simulation results. The process of obtaining multiphysics coupling simulation results includes: The electrical, thermal, and optical signal data of the acquisition device under different operating conditions are collected, and the collected data are preprocessed to obtain a normalized multiphysics dataset. Based on the normalized multiphysics dataset, key features of electrical, thermal and optical signals are extracted to generate a multiphysics feature dataset. By using a multiphysics feature dataset, a finite element analysis framework is constructed to simultaneously solve the interaction equations of electrical, thermal, and optical physical fields, thereby obtaining multiphysics coupling simulation results.

[0012] Furthermore, after obtaining the multiphysics coupling simulation results, the method further includes evaluating the physical field output values ​​in the multiphysics coupling simulation results to obtain critical operating conditions, specifically including: If the output value of any physical field in the multiphysics coupling simulation results exceeds the preset performance threshold, the vector analysis method is used to decompose the physical field data exceeding the threshold and determine the offset direction and offset amount of the abnormal physical field. Based on the offset direction and offset amount of the anomalous physical field, cluster analysis is used to divide the performance stability range of the device under different operating conditions, and a performance stability range dataset is obtained. By using the performance stable interval dataset, the critical operating conditions of the device under multi-physics coupling are identified by the boundary division method, and a critical operating condition dataset is obtained. Based on the critical operating condition dataset, the results are stored in the multiphysics performance database using a data storage protocol to obtain the associated multiphysics performance dataset.

[0013] Furthermore, this method also includes: Based on the simulation results of the dynamic coupling model, the device structure parameters are optimized, including chip thickness, electrode layout, and heat dissipation structure size. If the light color shift still exceeds the design requirements, the material ratio optimization process is initiated.

[0014] Furthermore, the material proportioning optimization process is as follows: By using the results of structural parameter optimization to guide the adjustment of material ratios, the influence weight of different material components on the optical stability of the device is analyzed, and the optimal material ratio and doping concentration distribution scheme of the light-emitting layer are determined.

[0015] A semiconductor light-emitting device light color testing system, comprising: The data acquisition module is responsible for acquiring electrical, thermal, and optical signal data of semiconductor light-emitting devices under different operating conditions; The data preprocessing module preprocesses the collected raw data to ensure data quality and consistency. The feature extraction module extracts key features from the normalized multiphysics dataset for subsequent simulation and analysis. The multiphysics coupling simulation module constructs a finite element analysis framework, simultaneously solves the interaction equations of electrical, thermal, and optical physical fields, and obtains multiphysics coupling simulation results. The performance optimization module optimizes the structural parameters and material ratios of the device based on the results of multiphysics coupling simulation, so as to improve the optical and color stability and overall performance of the device.

[0016] Compared with the prior art, the present invention has the following advantages: 1. This invention constructs a dynamic coupling model of the device and simultaneously solves the interaction equations of the three physical fields of electricity, heat, and optics. It comprehensively analyzes the overall performance of the device in a real working environment. It not only considers the influence of a single physical field but also studies the dynamic interaction between different physical fields, which can more accurately simulate the actual operating state of the device. By monitoring the voltage distribution, current density distribution, and resistance changes inside the device in real time, as well as the temperature field distribution, temperature gradient, and heat flux density distribution data, this invention can capture various changes of the device during dynamic operation, thereby gaining a more comprehensive understanding of the device's behavior. 2. By constructing a dynamic coupling model and performing multiphysics coupling simulation, this invention can more accurately predict the optical and color performance of the device under different operating conditions. The simulation results can be used to optimize the device's design parameters and ensure the consistency between the design parameters and the actual performance. This invention establishes an optical and color shift prediction model by analyzing the mapping relationship matrix between temperature and optical properties. This model can predict the amount and direction of optical and color shift, thereby adjusting the device parameters in advance during the design stage and avoiding optical and color drift problems. 3. This invention solves the interaction equations of electrical, thermal, and optical physical fields simultaneously through a finite element analysis framework, enabling precise simulation of the dynamic behavior inside the device. This method can analyze in detail the thermal effects induced by current injection and the impact of temperature rise on optical properties, thereby achieving precise control over the complex physical processes inside the device. The simulation results guide the optimization of device structural parameters. Furthermore, by analyzing the influence weights of different material compositions on the device's optical color stability, the optimal luminescent layer material ratio and doping concentration distribution scheme are determined, thereby improving the device's optical color stability and overall performance. 4. This invention analyzes the influence weight of different material components on the light color stability of the device, determines the optimal material ratio and doping concentration distribution scheme of the light-emitting layer, and can effectively solve the problem of light color instability caused by uneven material ratio. By optimizing the structural parameters of the device, the heat dissipation performance and electrical performance of the device can be improved, thereby reducing light color shift. The optimized structural design can ensure the stability and reliability of the device under long-term operation or extreme conditions. 5. By constructing a dynamic coupling model, this invention can analyze the performance change patterns of the device at different operating stages, including thermal effects and optical property changes during long-term operation. It can identify factors that may lead to a decrease in luminous efficiency and color temperature drift in advance. By analyzing the light color shift prediction model, this invention can identify the critical operating conditions of the device at different temperatures. By optimizing the design and operating conditions of the device, it can prevent the device from entering a state of light color instability, thereby improving the performance of the device under long-term operation or extreme conditions. Attached Figure Description

[0017] Figure 1 This is a flowchart of a method for testing the light color of a semiconductor light-emitting device, as described in Example 1. Figure 2 This is a framework diagram of a semiconductor light-emitting device color testing system according to Example 2. Detailed Implementation

[0018] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. Example 1

[0019] like Figure 1 As shown, this embodiment provides a method for testing the light color of a semiconductor light-emitting device, including the following steps: Obtain the electrical characteristic parameters of semiconductor light-emitting devices under different current injection conditions, monitor the voltage distribution, current density distribution and resistance changes inside the devices in real time, and establish an electrical characteristic database; Based on the changes in electrical characteristic parameters, the temperature field distribution generated by the internal thermal effect of the device is calculated. If the current density exceeds the preset threshold, the temperature gradient and heat flux density distribution data of each region of the device are further obtained. Based on temperature gradient and heat flux density distribution data, the emission wavelength, light intensity distribution and color coordinate changes of semiconductor light-emitting devices under different temperature conditions are analyzed to obtain the mapping relationship matrix between temperature and optical properties; The mapping relationship matrix between temperature and optical properties is used to analyze the law of optical properties changing with temperature and establish a light color shift prediction model. If the temperature change exceeds the light color stability threshold, the light color shift amount and shift direction are calculated to determine the critical operating conditions of the device's light color stability. A dynamic coupling model of the device is constructed, and the interaction equations of the three physical fields of electricity, heat and optics are solved simultaneously to obtain the comprehensive performance of the device under real working environment.

[0020] Furthermore, the establishment of the electrical characteristic database specifically includes: The original set of electrical parameters is obtained by collecting data on voltage distribution, current density distribution and resistance change of semiconductor light-emitting devices under different current injection conditions through a sensor array; The original electrical parameter set is denoised and standardized using a preset signal processing algorithm to obtain a standardized electrical parameter set. This also includes accuracy optimization of the standardized set of electrical parameters: If the voltage distribution data in the standardized electrical parameter set exceeds the preset threshold range, then key electrical response features are extracted to obtain an electrical response feature set. Based on the electrical response feature set, the operating states of the devices are classified to obtain the operating state classification results. The working status classification results are grouped to obtain the grouping structure of the electrical characteristic database; Based on the grouping structure, the standardized set of electrical parameters and the set of electrical response characteristics are stored in the electrical property database to obtain a complete electrical property database; If the integrity of the data records in the complete electrical characteristic database meets the preset threshold requirements, the accuracy of the data in the database is verified by a data verification algorithm to obtain the final electrical characteristic database.

[0021] Specifically, examples can be given as follows: Multiple sensors are placed at key locations in semiconductor light-emitting devices (such as the chip surface, near electrodes, heat dissipation structures, etc.) to ensure comprehensive monitoring; Under different current injection conditions (e.g., from low current to high current, gradually increasing), data on voltage distribution, current density distribution, and resistance variation inside the device are collected. Use filtering algorithms (such as low-pass filters, wavelet transforms, etc.) to remove high-frequency noise from the data; normalize the collected electrical parameters so that the values ​​of all parameters are within the same dimension range (e.g., normalize to between 0 and 1). For example, the original voltage distribution data: [1.2V, 1.5V, 1.8V, 2.0V] → after standardization: [0.3, 0.5, 0.7, 1.0]; Original current density distribution data: [100 A / cm², 150 A / cm², 200 A / cm²] → Standardized: [0.33, 0.5, 0.67]; Original resistance variation data: [1.0Ω, 1.2Ω, 1.5Ω] → Standardized: [0.4, 0.6, 1.0]; Check whether the standardized voltage distribution data exceeds the preset threshold range (e.g., between 0 and 1). If it does, further processing is performed. Extract key electrical response features from the data that exceeds the threshold, such as voltage peaks and abrupt changes in current density. For example, the preset threshold range is 0 to 1; The standardized voltage distribution data: [0.3, 0.5, 1.2, 1.0] → 1.2 exceeds the threshold; Key electrical response features extracted: voltage peak value is 1.2, corresponding to the 3rd data point; Analyze the extracted key electrical response features to determine their correlation with the device's operating state, and use cluster analysis or machine learning algorithms (such as K-means clustering, support vector machine, etc.) to classify the operating state; For example, operating state 1: low current injection, stable voltage distribution, and low current density; Operating state 2: Medium current injection, slight fluctuations in voltage distribution, and moderate current density; Operating state 3: High current injection, significant fluctuations in voltage distribution, and high current density; Based on the classification results of the device's operating status, the data is divided into different groups, such as grouping by current injection range; For example, Group 1: Low current injection (0-100mA); Group 2: Medium current injection (100-500mA); Group 3: High current injection (500-1000mA); The standardized set of electrical parameters and the extracted set of electrical response features are stored in a database and organized according to a grouping structure; for example, group 1: low current injection. Standardized electrical parameter set: [0.3, 0.5, 0.7, 1.0]; Electrical response characteristic set: peak voltage is 0.7; Group 2: Medium current injection; Standardized electrical parameter set: [0.4, 0.6, 0.8, 1.0]; Electrical response characteristic set: the current density abrupt change point is 0.6; Group 3: High current injection; Standardized electrical parameter set: [0.5, 0.7, 0.9, 1.0]; Electrical response characteristic set: peak voltage is 0.9; Use methods such as data consistency checks and data integrity checks to verify whether the data in the database meets the preset accuracy requirements.

[0022] Furthermore, after acquiring the heat flux density distribution data of each region of the device, the process further includes: The heat flux density distribution data is divided into regions using a preset grid partitioning algorithm to obtain a set of heat flux partitions inside the device. If there are regions with abnormal heat flux density in the heat flux partition set, the heat flux density data of the abnormal regions are corrected through data standardization to obtain corrected heat flux density distribution data. Based on the corrected heat flux density distribution data, the internal heat flux zones of the device are classified to obtain the heat flux characteristic classification results. By using a preset data storage protocol, the heat flow characteristic classification results are associated with the temperature field distribution data and stored in the thermal characteristic database to obtain a complete thermal characteristic database.

[0023] Specifically, examples can be given as follows: When correcting heat flux density data in abnormal regions, data normalization formulas can be used. For example, suppose we use the min-maximum normalization method to correct the heat flux density values ​​to a preset normal range: The standardized heat flux density = original heat flux density − minimum / maximum value - minimum value × (target maximum value − target minimum value) + target minimum value; assuming the normal range of heat flux density is [100 W / m², 500 W / m²], and the heat flux density of a certain region is 600 W / m², it can be corrected using the following formula: Corrected heat flux density = 600 - 100 / (600 - 100 × (500 - 100) + 100 = 500 W / m 2 ; When using clustering algorithms (such as K-means) to classify heat flux zones, the Euclidean distance formula can be used. Assuming that clustering is performed based on the mean and variance of heat flux density, and the feature vectors of two regions are x=(250,50) and y=(300,60) respectively, then the Euclidean distance between them is d(x,y)≈50.99. When analyzing heat flux density distribution, it is necessary to calculate the gradient of heat flux density to understand the trend of heat flux density variation. The gradient can be calculated using the heat flux density gradient formula:

[0024] Assume that the heat flux density q varies with Δq in the x and y directions, respectively. x and Δq y The corresponding spacings are Δx and Δy, respectively, then the gradient is: ; For example, if the heat flux density in a certain region changes by 50 W / m² in the x-direction with a spacing of 1 mm (0.001 m), and the heat flux density changes by 30 W / m² in the y-direction with a spacing of 1 mm (0.001 m), then the gradient is ∇q = (50000, 30000) W / m 3 ; When storing heat flow characteristic classification results in association with temperature field distribution data, the correlation between them can be analyzed by calculating correlation coefficients, for example, using the Pearson correlation coefficient:

[0025] Suppose we have two sets of data: heat flux density q and temperature T. Calculate the correlation coefficient between them to understand the linear relationship between heat flux density and temperature. For example, heat flux density data q=[150,200,250,300,350] and corresponding temperature data T=[25,30,32,35,38]; Calculate the mean:

[0026]

[0027] Calculate the numerator:

[0028] Calculate the denominator:

[0029] Calculate the correlation coefficient:

[0030] This indicates a strong positive correlation between heat flux density and temperature.

[0031] Furthermore, the obtained mapping matrix between temperature and optical properties includes: The temperature change data of the device under different operating conditions is collected, and the collected data is preprocessed to obtain a normalized temperature data set. Based on the normalized temperature data set, the changes in the emission wavelength of the device under various temperature conditions are calculated to generate a wavelength distribution dataset. By using the wavelength distribution dataset, the light intensity distribution data of the device at different temperatures is extracted to obtain the light intensity distribution dataset; Based on the light intensity distribution dataset, the color coordinate changes of the device under various temperature conditions are analyzed to generate a color coordinate distribution dataset; If there are outliers in the color coordinate distribution dataset, the outliers are filtered out using a preset threshold judgment method to obtain the corrected color coordinate distribution dataset. Using the corrected chromaticity distribution dataset, the wavelength distribution dataset, light intensity distribution dataset, and chromaticity distribution dataset are jointly classified to generate a mapping matrix between temperature and optical properties. Based on the mapping matrix, the classification results are stored in the optical property database using a data storage protocol to obtain the associated temperature-optical property dataset.

[0032] Specifically, examples can be given as follows: Generate a mapping matrix between temperature and optical properties: Cluster analysis or machine learning algorithms are used to classify the data, and the classification results are organized into a matrix to represent the relationship between temperature and optical properties. For example, joint classification results: Temperature = 25℃, wavelength = 450nm, light intensity = 100mW, chromaticity coordinates = (0.14, 0.08) Temperature = 30℃, wavelength = 451nm, light intensity = 95mW, chromaticity coordinates = (0.14, 0.08) Temperature = 35℃, wavelength = 452nm, light intensity = 90mW, chromaticity coordinates = (0.14, 0.08) Temperature = 40℃, wavelength = 453nm, light intensity = 85mW, chromaticity coordinates = (0.14, 0.08) Temperature = 45℃, wavelength = 454nm, light intensity = 80mW, chromaticity coordinates = (0.14, 0.08) Mapping matrix: Temperature = 25℃, wavelength = 450nm, light intensity = 100mW, chromaticity coordinates = (0.14, 0.08) Temperature = 30℃, wavelength = 451nm, light intensity = 95mW, chromaticity coordinates = (0.14, 0.08) Temperature = 35℃, wavelength = 452nm, light intensity = 90mW, chromaticity coordinates = (0.14, 0.08) Temperature = 40℃, wavelength = 453nm, light intensity = 85mW, chromaticity coordinates = (0.14, 0.08) Temperature = 45℃, wavelength = 454nm, light intensity = 80mW, chromaticity coordinates = (0.14, 0.08) Based on the mapping matrix, the classification results are stored in the optical property database using a data storage protocol to obtain the associated temperature-optical property dataset.

[0033] Furthermore, the critical operating conditions for determining the optical color stability of the device include: Temperature change data of the device under different operating conditions are collected by sensors, and the data is preprocessed using a standardization method to obtain a normalized temperature dataset. Based on the normalized temperature dataset, the principal component analysis algorithm is used to extract key features of temperature changes and generate a temperature feature dataset. Using a temperature feature dataset, a light color shift prediction model is constructed using a support vector regression algorithm to obtain the light color shift prediction results. If the light color shift prediction result exceeds the preset stability threshold, the light color shift amount and shift direction are calculated by vector decomposition method to obtain the shift amount and direction dataset. Based on the offset and orientation datasets, cluster analysis is used to determine the stable interval of the light color offset, thus obtaining the range of light color stability. By using the range of optical color stability, the critical operating conditions of the device at different temperatures are identified using a boundary division method, and a critical condition dataset is obtained. Based on the critical condition dataset, the results are stored in the optical property database using a data storage protocol to obtain the associated temperature-photochromic stability dataset.

[0034] Specifically, examples can be given as follows: The normalized temperature dataset is converted into matrix form, and the principal components are extracted using the PCA algorithm to generate a temperature feature dataset; the covariance matrix is ​​then calculated.

[0035] Calculate eigenvalues ​​and eigenvectors, and select the first k principal components; For example, the normalized temperature data matrix:

[0036] Calculate the covariance matrix:

[0037] in, =0.5; A light color shift prediction model is constructed using the Support Vector Regression (SVR) algorithm. The temperature feature dataset is used as input, and the light color shift is used as output. The model is trained using the SVR algorithm to obtain the light color shift prediction model. Choose an appropriate kernel function (such as the RBF kernel) and parameters.

[0038] Training the model:

[0039] Where w is the weight vector, b is the bias term, ξi is the slack variable, and C is the regularization parameter; Calculate the light color shift amount and direction to determine whether the predicted light color shift exceeds the preset stability threshold. Light color offset:

[0040] Offset direction:

[0041] Where Δx and Δy are the changes in color coordinates; Using the light color offset and offset direction dataset as input, the K-means clustering algorithm is used to divide the data into different stable intervals.

[0042] Calculate the distance from each data point to the cluster center:

[0043] Assign data points to the nearest cluster center; Using the stable interval dataset as input, the critical operating conditions are identified using the boundary partitioning method, and the boundary of each stable interval is calculated: boundary = maximum value of interval 1 + minimum value of interval 2 / 2. The critical operating condition dataset is stored in a database to ensure data integrity and queryability. Temperature and light color stability data are stored together for easy subsequent analysis and querying.

[0044] Furthermore, the comprehensive performance of the device under real-world operating conditions is specifically reflected by obtaining multiphysics coupling simulation results. The process of obtaining multiphysics coupling simulation results includes: The electrical, thermal, and optical signal data of the acquisition device under different operating conditions are collected, and the collected data are preprocessed to obtain a normalized multiphysics dataset. Based on the normalized multiphysics dataset, key features of electrical, thermal and optical signals are extracted to generate a multiphysics feature dataset. By using a multiphysics feature dataset, a finite element analysis framework is constructed to simultaneously solve the interaction equations of electrical, thermal, and optical physical fields, thereby obtaining multiphysics coupling simulation results.

[0045] Specifically, examples can be given as follows: Under different current injection conditions, the electrical (voltage, current density, resistance), thermal (temperature, heat flux density), and optical (emission wavelength, light intensity, color coordinates) signal data of the device are collected using sensors. The collected data are then denoised and standardized to normalize the data to the range of [0, 1]. Use principal component analysis (PCA) or other feature extraction methods to extract key features for each physics field, and combine the extracted key features into a multiphysics feature dataset; Based on the multiphysics feature dataset, a finite element analysis framework is constructed, and the interaction equations of the electrical, thermal, and optical physical fields are defined. Electrical equations:

[0046] Thermal equation:

[0047] Optical equation:

[0048] The above equations were solved simultaneously using the finite element method to obtain multiphysics coupling simulation results. For example, the finite element model divides the device into multiple small units, each containing electrical, thermal, and optical characteristics.

[0049] Define the physical field equations:

[0050] Simultaneous solution: Simulations were performed using finite element software such as COMSOL Multiphysics. Simulation results: Voltage distribution, temperature distribution, and emission wavelength distribution of the device under different operating conditions; The simulation results are stored in a database to ensure data integrity and queryability.

[0051] Data association: Electrical, thermal, and optical performance data are associated and stored to facilitate subsequent analysis and retrieval.

[0052] Furthermore, after obtaining the multiphysics coupling simulation results, the method further includes evaluating the physical field output values ​​in the multiphysics coupling simulation results to obtain critical operating conditions, specifically including: If the output value of any physical field in the multiphysics coupling simulation results exceeds the preset performance threshold, the vector analysis method is used to decompose the physical field data exceeding the threshold and determine the offset direction and offset amount of the abnormal physical field. Based on the offset direction and offset amount of the anomalous physical field, cluster analysis is used to divide the performance stability range of the device under different operating conditions, and a performance stability range dataset is obtained. By using the performance stable interval dataset, the critical operating conditions of the device under multi-physics coupling are identified by the boundary division method, and a critical operating condition dataset is obtained. Based on the critical operating condition dataset, the results are stored in the multiphysics performance database using a data storage protocol to obtain the associated multiphysics performance dataset.

[0053] Specifically, examples can be given as follows: Extract the output values ​​of electrical, thermal, and optical physical fields from the multiphysics coupling simulation results, and check whether the output value of each physical field exceeds the preset performance threshold. For example, the electrical performance threshold is: voltage change does not exceed 10%, the thermal performance threshold is: temperature change does not exceed 20°C, and the optical performance threshold is: emission wavelength change does not exceed 5nm. Preset threshold: Electrical threshold: Voltage variation not exceeding 10% (reference voltage 2.5V, threshold range 2.25V to 2.75V); Thermal threshold: Temperature change not exceeding 20°C (reference temperature 25°C, threshold range 5°C to 45°C); Optical threshold: The emission wavelength variation does not exceed 5nm (reference wavelength 450nm, threshold range 445nm to 455nm). Simulation results: Operating conditions 1: Voltage = 2.5V, Temperature = 25°C, Wavelength = 450nm; Operating conditions 2: Voltage = 3.0V, Temperature = 30°C, Wavelength = 451nm; Operating conditions 3: Voltage = 3.5V, Temperature = 40°C, Wavelength = 452nm; Operating conditions 4: Voltage = 4.0V, Temperature = 50°C, Wavelength = 453nm; Evaluation results: Operating condition 1: All physical fields are within the threshold range; Operating condition 2: All physical fields are within the threshold range; Operating condition 3: Temperature exceeds the threshold (40°C > 35°C); Operating condition 4: Both temperature and voltage exceed the threshold (50°C > 45°C, 4.0V > 2.75V). Select physics data that exceeds the threshold, and calculate the offset direction and offset amount:

[0054] Offset direction:

[0055] Where ΔX and ΔY are the changes in the physical field data; Using offset and offset direction data as input, the K-means clustering algorithm is used to divide the data into different stable intervals. Calculate the distance from each data point to the cluster center, and assign the data point to the nearest cluster center; Using the performance stable interval dataset as input, the boundary partitioning method is used to identify critical operating conditions. The critical operating condition dataset is stored in a database to ensure data integrity and queryability. Electrical, thermal, and optical performance data are stored together for easy subsequent analysis and retrieval.

[0056] Furthermore, this method also includes: Based on the simulation results of the dynamic coupling model, the device structure parameters are optimized, including chip thickness, electrode layout, and heat dissipation structure size. If the light color shift still exceeds the design requirements, the material ratio optimization process is initiated.

[0057] Specifically, a genetic algorithm is used to iteratively optimize the chip thickness, electrode layout, and heat dissipation structure dimensions to generate an optimized set of structural parameters. By constructing a finite element simulation model using the optimized set of structural parameters, the light color shift of the device under operating conditions is calculated. If the light color shift exceeds the preset threshold, the shift data exceeding the threshold is extracted, and the key components of the light color shift are decomposed using the principal component analysis algorithm to determine the direction of material ratio adjustment. Based on the adjustment direction of the material ratio, a new material ratio scheme is generated using the particle swarm optimization algorithm, resulting in an updated material ratio dataset. By reconstructing the finite element simulation model using the updated material ratio dataset, the light color shift of the device under operating conditions is calculated to determine whether it meets the preset threshold. If the light color shift still exceeds the preset threshold, cluster analysis is performed on the material ratio dataset to divide the performance stable interval and obtain the performance stable interval dataset. Based on the performance stability interval dataset, the optimized structural parameters and material ratio schemes are stored in the multiphysics performance database using a data storage protocol, resulting in the associated performance optimization dataset.

[0058] Specifically, examples can be given as follows: A set of structural parameters (chip thickness, electrode layout, heat dissipation structure size) is randomly generated as the initial population; Define a fitness function, such as minimizing the light color shift; Select the best individuals to enter the next generation based on the fitness function; Perform a crossover operation on the selected individuals to generate new individuals; Mutation operations are performed on newly generated individuals to increase population diversity; Repeat the above steps until the termination condition (such as the maximum number of iterations or the fitness threshold) is met. Based on the optimized structural parameters, a finite element simulation model is constructed, the simulation is run, and the light color shift is calculated, for example: Optimized structural parameters: Chip thickness = 120μm; Electrode layout = 60μm; Heat dissipation structure size = 220μm; Simulation results: Light color shift = 0.5nm; Data with light color shifts exceeding a threshold are extracted, and principal component analysis (PCA) is used to decompose the key components of the light color shift, calculating the covariance matrix:

[0059] Calculate eigenvalues ​​and eigenvectors, select principal components, for example, preset threshold: 0.3 nm, light color shift data exceeding the threshold: light color shift amount = 0.5 nm; Principal component analysis results: Principal component 1: light color shift amount, Principal component 2: light color shift direction, material ratio adjustment direction: increase the doping concentration of the light-emitting layer material; Set the initial position and velocity of the particles, define a fitness function, such as minimizing the light color shift; update the position and velocity of the particles, find the optimal solution, and repeat the above steps until the termination condition is met; Based on the updated material ratio, the finite element simulation model was reconstructed, the simulation was run, and the light color shift was calculated. Using the material proportion dataset as input, the K-means clustering algorithm is used to divide the data into different stable intervals. The distance from each data point to the cluster center is calculated, and the data points are assigned to the nearest cluster center. For example: Material proportioning dataset: Material A = 55%; Material B = 25%; Doping concentration = 20%; Light color shift = 0.3nm; The K-means clustering algorithm was used to divide the data into two stable intervals: Interval 1: Color shift ≤ 0.2 nm; Interval 2: Light color shift > 0.2 nm; The optimized structural parameters and material proportions are stored in a database, and the electrical, thermal, and optical performance data are stored together for easy analysis and retrieval later. For example: Optimized structural parameters and material proportioning scheme: Chip thickness = 120μm; Electrode layout = 60μm; Heat dissipation structure size = 220μm; Material A = 55%; Material B = 25%; Doping concentration = 20%; Light color offset = 0.3nm.

[0060] Furthermore, the material proportioning optimization process is as follows: By using the results of structural parameter optimization to guide the adjustment of material ratios, the influence weight of different material components on the optical stability of the device is analyzed, and the optimal material ratio and doping concentration distribution scheme of the light-emitting layer are determined.

[0061] Furthermore, determining the optimal luminescent layer material ratio and doping concentration distribution scheme includes: Based on the results of structural parameter optimization, a multidimensional input dataset is constructed. A neural network algorithm is used to analyze the correlation between material composition and optical color stability, and the influence weights are ranked. Based on the order of influence weights, key material components are extracted, and principal component analysis is used to decompose their contribution to light color shift, thus determining the main regulating components. If the contribution of the light color shift of the main control component exceeds the preset threshold, the material ratio and doping concentration are adjusted by the particle swarm optimization algorithm to generate an updated ratio scheme. By using the updated ratio scheme, a finite element simulation model is constructed to calculate the light color shift of the device under operating conditions and determine whether it meets the preset threshold. If the light color shift still exceeds the preset threshold, cluster analysis is performed on the updated ratio scheme to divide the performance stable interval and obtain a stable ratio dataset. Based on the stable ratio dataset, the optimized ratio scheme and doping concentration distribution are stored in the performance database using a data storage protocol to obtain the associated optimized dataset; By using the associated optimized dataset, the convergence trend of light color shift is extracted, and a neural network algorithm is used to predict the direction of subsequent ratio adjustment to generate the final ratio scheme.

[0062] Specifically, examples can be given as follows: Based on the structural parameter optimization results, a multidimensional input dataset is constructed, including material composition, doping concentration, and light color shift, etc., and a neural network algorithm is used to analyze the correlation between material composition and light color stability. Input layer: material composition and doping concentration; Output layer: Light color offset; Train the network to obtain the influence weight of each input feature; for example: Multidimensional input dataset: Material A = 50%, Material B = 30%, Doping concentration = 20%, Color shift = 0.5nm; Material A = 55%, Material B = 25%, Doping concentration = 20%, Color shift = 0.3nm; Material A = 60%, Material B = 20%, Doping concentration = 20%, Color shift = 0.2nm; Neural network training results: Influence weight ranking: Material A > Material B > Doping concentration; Select the material components with the greatest influence weights, use Principal Component Analysis (PCA) to decompose the contribution of key components to the color shift, calculate the covariance matrix, calculate eigenvalues ​​and eigenvectors, and select the principal components; for example: Key material components extracted: Material A; Principal component analysis results: Principal Component 1: Contribution of Material A; Principal Component 2: Contribution of Material B; Main regulating component: Material A; Set the initial position and velocity of the particles, define a fitness function, such as minimizing the light color shift, update the position and velocity of the particles, find the optimal solution, and repeat the above steps until the termination condition is met. Based on the updated proportioning scheme, the finite element simulation model was reconstructed, the simulation was run, and the light color shift was calculated. Using the updated ratio scheme as input, the K-means clustering algorithm is used to divide the data into different stable intervals; Calculate the distance from each data point to the cluster center, and assign the data point to the nearest cluster center; The optimized formulation and doping concentration distribution are stored in the database, and the material ratio and light color shift are stored together for easy subsequent analysis and query. The convergence trend of light color shift is extracted from the optimized dataset, and a neural network algorithm is used to predict the direction of subsequent ratio adjustment. Example 2

[0063] like Figure 2 As shown, this embodiment provides a semiconductor light-emitting device light color testing system, including: The data acquisition module is responsible for acquiring electrical, thermal, and optical signal data of semiconductor light-emitting devices under different operating conditions; The data acquisition module monitors the voltage distribution, current density distribution, and resistance changes inside the device in real time. It acquires temperature distribution, temperature gradient, and heat flux density distribution data inside the device through sensors. It measures the changes in emission wavelength, light intensity distribution, and color coordinates of the device under different temperature conditions, generating a raw multiphysics dataset to provide basic data for subsequent data processing and analysis. The data preprocessing module preprocesses the collected raw data to ensure data quality and consistency. The data preprocessing module normalizes the collected electrical, thermal, and optical data to eliminate dimensional differences between different physical quantities, identifies and corrects outliers in the collected data, such as correcting data in areas of abnormal heat flux density through data standardization, removing invalid or noisy data, ensuring the accuracy of subsequent analysis, and generating a normalized multiphysics dataset to provide high-quality data for subsequent feature extraction and simulation analysis. The feature extraction module extracts key features from the normalized multiphysics dataset for subsequent simulation and analysis. The feature extraction module extracts key electrical features such as voltage distribution and current density distribution, key thermal features such as temperature field distribution, temperature gradient, and heat flux density distribution, and key optical features such as emission wavelength, light intensity distribution, and color coordinate changes. It integrates electrical, thermal, and optical features to generate a multi-physics feature dataset, which provides input data for finite element analysis and simulation. The multiphysics coupling simulation module constructs a finite element analysis framework, simultaneously solves the interaction equations of electrical, thermal, and optical physical fields, and obtains multiphysics coupling simulation results. The multiphysics coupling simulation module constructs a dynamic coupling model of the device based on a feature dataset, describes the interaction between electrical, thermal and optical physical fields, solves the interaction equations of the multiphysics fields using the finite element method, generates multiphysics coupling simulation results, evaluates the physical field output values ​​in the simulation results, identifies the offset direction and offset amount of abnormal physical fields, divides the performance stability range, and generates multiphysics coupling simulation results, including the comprehensive performance of the device under different operating conditions. The performance optimization module optimizes the structural parameters and material ratios of the device based on the results of multiphysics coupling simulation, so as to improve the optical and color stability and overall performance of the device. The performance optimization module optimizes chip thickness, electrode layout, and heat dissipation structure dimensions to improve the electrical and thermal performance of the device. It analyzes the influence weight of different material components on the optical stability of the device, determines the optimal luminescent layer material ratio and doping concentration distribution scheme, guides the design and manufacturing of the device based on the optimization results, ensures that the device performance in practical applications meets the design requirements, and generates optimized device structural parameters and material ratio schemes to provide guidance for device manufacturing and improvement.

[0064] This system, through data acquisition, preprocessing, feature extraction, multiphysics coupling simulation, and performance optimization, can comprehensively analyze and optimize the optical and color performance of devices in real working environments, ensuring their stability and reliability.

[0065] The specific embodiments of the invention have been described in detail above, but these are merely examples. The invention is not limited to the specific embodiments described above. Those skilled in the art should understand that the embodiments and descriptions in the specification are only illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for testing the light color of a semiconductor light-emitting device, characterized in that, Includes the following steps: To obtain the electrical characteristic parameters of semiconductor light-emitting devices under different current injection conditions, monitor the voltage distribution, current density distribution and resistance changes inside the devices in real time, and establish an electrical characteristic database; Based on the changes in electrical characteristic parameters, the temperature field distribution generated by the internal thermal effect of the device is calculated. If the current density exceeds the preset threshold, the temperature gradient and heat flux density distribution data of each region of the device are further obtained. Based on temperature gradient and heat flux density distribution data, the emission wavelength, light intensity distribution and color coordinate changes of semiconductor light-emitting devices under different temperature conditions are analyzed to obtain the mapping relationship matrix between temperature and optical properties; The mapping relationship matrix between temperature and optical properties is used to analyze the law of optical properties changing with temperature and establish a light color shift prediction model. If the temperature change exceeds the light color stability threshold, the light color shift amount and shift direction are calculated to determine the critical operating conditions of the device's light color stability. A dynamic coupling model of the device is constructed, and the interaction equations of the three physical fields of electricity, heat and optics are solved simultaneously to obtain the comprehensive performance of the device under real working environment.

2. The method for testing the light color of a semiconductor light-emitting device according to claim 1, characterized in that, The obtained temperature gradient and heat flux density distribution data for each region of the device include: The internal thermal effects of semiconductor light-emitting devices are modeled using the heat conduction equation, and the temperature field distribution data are calculated. If the temperature value in any region of the temperature field distribution data exceeds the preset threshold, the temperature gradient data inside the device is extracted using the finite element analysis algorithm. Based on the temperature gradient data, the heat flux density distribution data of each region of the device is obtained using the heat flux density calculation formula.

3. The method for testing the light color of a semiconductor light-emitting device according to claim 2, characterized in that, After obtaining the heat flux density distribution data of each region of the device, the process also includes: The heat flux density distribution data is divided into regions to obtain a set of heat flux partitions inside the device; If there are regions with abnormal heat flux density in the heat flux partition set, the heat flux density data of the abnormal regions are corrected to obtain the corrected heat flux density distribution data. Based on the corrected heat flux density distribution data, the internal heat flux zones of the device are classified to obtain the heat flux characteristic classification results. By using a preset data storage protocol, the heat flow characteristic classification results are associated with the temperature field distribution data and stored in the thermal characteristic database to obtain a complete thermal characteristic database.

4. The method for testing the light color of a semiconductor light-emitting device according to claim 1, characterized in that, The obtained mapping matrix between temperature and optical properties includes: The temperature change data of the device under different operating conditions is collected, and the collected data is preprocessed to obtain a normalized temperature data set. Based on the normalized temperature data set, the changes in the emission wavelength of the device under various temperature conditions are calculated to generate a wavelength distribution dataset. By using the wavelength distribution dataset, the light intensity distribution data of the device at different temperatures is extracted to obtain the light intensity distribution dataset; Based on the light intensity distribution dataset, the color coordinate changes of the device under various temperature conditions are analyzed to generate a color coordinate distribution dataset; If there are outliers in the color coordinate distribution dataset, the outliers are filtered out using a preset threshold judgment method to obtain the corrected color coordinate distribution dataset. Using the corrected chromaticity distribution dataset, the wavelength distribution dataset, light intensity distribution dataset, and chromaticity distribution dataset are jointly classified to generate a mapping matrix between temperature and optical properties. Based on the mapping matrix, the classification results are stored in the optical property database using a data storage protocol to obtain the associated temperature-optical property dataset.

5. The method for testing the light color of a semiconductor light-emitting device according to claim 1, characterized in that, The critical operating conditions for determining the optical color stability of the device include: By collecting temperature change data of the device under different operating conditions through sensors, the data is preprocessed to obtain a normalized temperature dataset. Based on the normalized temperature dataset, key features of temperature change are extracted to generate a temperature feature dataset. A light color shift prediction model was constructed using a temperature feature dataset to obtain the light color shift prediction results. If the light color shift prediction result exceeds the preset stability threshold, then calculate the light color shift amount and shift direction to obtain the shift amount and direction dataset; Based on the offset and direction dataset, the stable range of light color offset is determined, and the range of light color stability is obtained. By analyzing the range of optical color stability, the critical operating conditions of the device at different temperatures are identified, and a critical condition dataset is obtained. Based on the critical condition dataset, the results are stored in the optical property database to obtain the associated temperature-photochromic stability dataset.

6. The method for testing the light color of a semiconductor light-emitting device according to claim 1, characterized in that, The obtained comprehensive performance of the device in a real working environment is specifically represented by obtaining multiphysics coupling simulation results. The process of obtaining multiphysics coupling simulation results includes: The electrical, thermal, and optical signal data of the acquisition device under different operating conditions are collected, and the collected data are preprocessed to obtain a normalized multiphysics dataset. Based on the normalized multiphysics dataset, key features of electrical, thermal and optical signals are extracted to generate a multiphysics feature dataset. By using a multiphysics feature dataset, a finite element analysis framework is constructed to simultaneously solve the interaction equations of electrical, thermal, and optical physical fields, thereby obtaining multiphysics coupling simulation results.

7. The method for testing the light color of a semiconductor light-emitting device according to claim 6, characterized in that, After obtaining the multiphysics coupling simulation results, the method further includes evaluating the physical field output values ​​in the multiphysics coupling simulation results to obtain critical operating conditions, specifically including: If the output value of any physical field in the multiphysics coupling simulation results exceeds the preset performance threshold, the vector analysis method is used to decompose the physical field data exceeding the threshold and determine the offset direction and offset amount of the abnormal physical field. Based on the offset direction and offset amount of the abnormal physical field, the performance stability range of the device under different operating conditions is divided to obtain a performance stability range dataset. By using the performance stability interval dataset, the critical operating conditions of the device under multi-physics coupling are identified, and the critical operating condition dataset is obtained. Based on the critical operating condition dataset, the results are stored in the multiphysics performance database to obtain the associated multiphysics performance dataset.

8. The method for testing the light color of a semiconductor light-emitting device according to claim 1, characterized in that, Also includes: Based on the simulation results of the dynamic coupling model, the device structure parameters are optimized, including chip thickness, electrode layout, and heat dissipation structure size. If the light color shift still exceeds the design requirements, the material ratio optimization process is initiated.

9. The method for testing the light color of a semiconductor light-emitting device according to claim 8, characterized in that, The specific process for optimizing the material ratio is as follows: By using the results of structural parameter optimization to guide the adjustment of material ratios, the influence weight of different material components on the optical stability of the device is analyzed, and the optimal material ratio and doping concentration distribution scheme of the light-emitting layer are determined.

10. A semiconductor light-emitting device color testing system, used to implement the semiconductor light-emitting device color testing method according to any one of claims 1-9, characterized in that, include: The data acquisition module is responsible for acquiring electrical, thermal, and optical signal data of semiconductor light-emitting devices under different operating conditions; The data preprocessing module preprocesses the collected raw data to ensure data quality and consistency. The feature extraction module extracts key features from the normalized multiphysics dataset for subsequent simulation and analysis. The multiphysics coupling simulation module constructs a finite element analysis framework, simultaneously solves the interaction equations of electrical, thermal, and optical physical fields, and obtains multiphysics coupling simulation results. The performance optimization module optimizes the structural parameters and material ratios of the device based on the results of multiphysics coupling simulation, so as to improve the optical and color stability and overall performance of the device.

Images