A method for measuring resistivity uniformity of a VCSEL epitaxial wafer

Abstract

The application relates to the technical field of semiconductor performance testing, in particular to a VCSEL epitaxial wafer resistivity uniformity measurement method, which comprises the following steps: collecting the square resistance of each measurement point on a VCSEL epitaxial wafer; sequentially dividing all the measurement points on the epitaxial wafer from the center to the edge into multiple measurement point sets; constructing a resistance stability characteristic quantity based on the distribution discrete degree and the difference from the ideal square resistance of all the measurement points in each layer measurement point set and in combination with a preset resistivity fluctuation range; calculating the resistance data reliability of the epitaxial wafer based on the resistance stability characteristic quantity of all the measurement point sets and the change confusion situation, adjusting the response surface model coefficient in combination with the residual error of the initially fitted response surface model, re-fitting the response surface, and improving the response surface fitting accuracy; and then analyzing the epitaxial wafer resistivity uniformity according to the response surface, thereby improving the reliability of the uniformity evaluation result.

Description

Technical Field

[0001] This application relates to the field of semiconductor performance testing technology, specifically to a method for measuring the resistivity uniformity of VCSEL epitaxial wafers. Background Technology

[0002] A vertical-cavity surface-emitting laser (VCSEL) is a novel type of semiconductor laser in which light propagates perpendicularly to the surface of a semiconductor substrate. It possesses advantages such as a single longitudinal mode, small divergence angle, and a circularly symmetrical beam, and is commonly used in optical communication, lidar, and other fields. VCSEL epitaxial wafers are grown on GaAs substrates using MOCVD equipment. The structure and quality of the epitaxial layer directly determine the performance of the device, and the uniformity of resistance at various locations on the VCSEL epitaxial wafer is a key performance indicator. Therefore, it is necessary to measure the resistivity at various locations on the VCSEL epitaxial wafer to evaluate its resistivity uniformity and stability, thereby ensuring the quality of the VCSEL epitaxial wafer upon delivery.

[0003] Traditional methods for measuring the resistance of VCSEL epitaxial wafers typically involve using a four-point probe method to measure the resistance at multiple discrete points within the wafer. The uniformity of the wafer's resistance is then determined by analyzing the resistance at these discrete points. However, discrete points can only measure data at one or a few points. Traditionally, the overall state of the wafer is analyzed using response surface methodology. However, due to the growth process of VCSEL epitaxial wafers, the resistance fluctuation at the center is relatively small, while the fluctuation at the edges is larger. This causes a discrepancy between the response surface model fitted directly from the measured data and the actual resistance distribution of the epitaxial wafer, reducing the accuracy of the resistivity uniformity measurement results. Summary of the Invention

[0004] To address the aforementioned technical problems, this application provides a method for measuring the resistivity uniformity of VCSEL epitaxial wafers, thereby resolving the existing issues.

[0005] The method for measuring the resistivity uniformity of VCSEL epitaxial wafers in this application adopts the following technical solution: One embodiment of this application provides a method for measuring the resistivity uniformity of a VCSEL epitaxial wafer, the method comprising the following steps: A measurement point array is set on the VCSEL epitaxial wafer, and the resistance at each measurement point is collected, with the center of the epitaxial wafer as the center of the array; all measurement points are divided layer by layer from the center to the outermost layer to obtain the set of measurement points for each layer. Based on the preset resistivity fluctuation range and the degree of dispersion of resistance distribution at all measurement points within each measurement point set, and combined with the difference between the sheet resistance of each measurement point and the preset ideal sheet resistance, the resistance stability characteristic quantity of each measurement point set is constructed. Calculate the difference in resistance stability characteristics between two adjacent sets of measurement points, and construct the reliability of the resistance data of the epitaxial wafer based on the distribution dispersion of all the differences in resistance stability characteristics and the resistance stability characteristics of all sets of measurement points. Response surface fitting is performed on the sheet resistance of all measurement points to obtain the fitted sheet resistance value of each measurement point; based on the reliability of the resistance data and the difference between the fitted sheet resistance value and the measured sheet resistance value of each measurement point, the quadratic term coefficient correction index of the response surface model is calculated. Based on the quadratic term coefficient correction exponent, the quadratic term coefficients in the response surface model are adjusted. The response surface model with the quadratic term coefficients adjusted is used to refit the response surface of the sheet resistance of all measurement points. The change in epitaxial wafer resistance in the final fitted response surface is analyzed to detect the uniformity of epitaxial wafer resistivity.

[0006] In one embodiment, the process of obtaining the resistance stability characteristic is as follows: The ideal resistance fluctuation value of each set of measurement points is determined based on the distance of each set of measurement points from the center of the epitaxial wafer and the preset resistivity fluctuation range. The ideal resistance fluctuation value is positively correlated with the distance. Calculate the mean sheet resistance and coefficient of variation of all measurement points in each set of measurement points; calculate the difference between the mean resistance and the ideal sheet resistance, and denote it as the first difference. Based on the ideal resistance fluctuation value, the coefficient of variation, and the first difference, the resistance stability characteristic of each measurement point set is determined. The resistance stability characteristic is positively correlated with the ideal resistance fluctuation value and negatively correlated with the coefficient of variation and the first difference, respectively.

[0007] In one embodiment, the expression for the resistance stability characteristic is: In the formula, This represents the resistance stability characteristic of the set of k-th measurement points; This represents the preset ideal resistance fluctuation value for the k-th measurement point set; This represents the coefficient of variation of the sheet resistance of all measurement points in the k-th measurement point set; This represents the average sheet resistance of all measurement points in the k-th measurement point set; This represents the ideal sheet resistance of the VCSEL epitaxial wafer. , These are the preset first and second parameter tuning factors, respectively.

[0008] In one embodiment, the process of obtaining the reliability of the resistance data is as follows: Calculate the weighted sum of the resistance stability characteristics of all measurement point sets, where the weight of the resistance stability characteristics of each measurement point set is negatively correlated with the distance of each measurement point set from the center of the epitaxial wafer; The difference between the resistance stability characteristic of two adjacent sets of measurement points is recorded as the stability difference value; The reliability of the resistance data of the epitaxial wafer is determined based on the weighted sum and the coefficient of variation of the stable differences between all adjacent measurement point sets.

[0009] In one embodiment, the expression for the reliability of the resistance data is: In the formula, This indicates the reliability of the resistance data for VCSEL epitaxial wafers; Indicates the number of measurement points in the set; The weights of the resistance stability characteristics of the k-th measurement point set are represented. This represents the resistance stability characteristic of the set of k-th measurement points; This represents the maximum value among the resistance stability characteristic quantities of the set of all measurement points; The coefficient of variation represents all stable differences. This indicates a preset fixed parameter.

[0010] In one embodiment, the process of obtaining the term coefficient correction index is as follows: The absolute value of the difference between the fitted block resistance value and the measured block resistance value at each measurement point is divided by the measured block resistance value, and the result is recorded as the local resistance anomaly degree at each measurement point. The term coefficient correction index of the response surface model is determined based on the reliability of the resistance data and the local resistance anomaly at all measurement points.

[0011] In one embodiment, the term coefficient correction index is the product of the normalized value of the mean of the local resistance anomaly at all measurement points and the reliability of the resistance data.

[0012] In one embodiment, the process of adjusting the coefficients of the quadratic term in the response surface model is as follows: When the local resistance anomaly at any measurement point is greater than or equal to the preset anomaly threshold, the product of the quadratic term coefficient in the response surface model and the correction exponent of the quadratic term coefficient is used as the adjustment amount of the coefficient, and the sum of the coefficient and its adjustment amount is used as the adjusted value of the coefficient. When the local resistance anomaly at all measurement points is less than the preset anomaly threshold, the coefficients in the response surface model are not adjusted.

[0013] In one embodiment, the analysis of the epitaxial wafer resistivity variation in the final fitted response surface is used to detect the epitaxial wafer resistivity uniformity. The process is as follows: Calculate the maximum, minimum, and mean values ​​of the block resistance in the final fitted response surface. Calculate the difference between the maximum and minimum values ​​and the mean value of the block resistance in the response surface as the resistance distribution difference of the epitaxial wafer. The resistivity uniformity of the epitaxial wafer is detected based on the resistance distribution difference.

[0014] In one embodiment, the detection of epitaxial wafer resistivity uniformity based on the resistance distribution difference specifically involves: If the resistance distribution difference is greater than the preset uniformity threshold, the resistivity uniformity of the epitaxial wafer is deemed unqualified; otherwise, the resistivity uniformity of the epitaxial wafer is deemed qualified.

[0015] This application has at least the following beneficial effects: This application collects resistance data at various measurement points on a VCSEL epitaxial wafer, dividing all measurement points on the epitaxial wafer into multiple measurement point sets from the center to the edge. Based on the distribution dispersion of the sheet resistance measured at all measurement points in each set and its difference from the ideal sheet resistance, combined with a preset resistivity fluctuation range, a resistance stability characteristic is constructed to characterize the validity of the measured resistance data within the measurement point set, thereby reducing measurement deviations caused by edge airflow disturbances. Based on the resistance stability characteristic and its chaotic changes in all measurement point sets, the reliability of the epitaxial wafer resistance data is calculated, reflecting the authenticity of the measurement data. Combined with the residuals of the initially fitted response surface model, the response surface model coefficients are adjusted, and the response surface is refitted to improve the accuracy of the response surface fitting. Furthermore, the uniformity of the epitaxial wafer resistivity is analyzed based on the response surface, improving the reliability of the uniformity evaluation results. Attached Figure Description

[0016] Figure 1 A flowchart of a method for measuring the resistivity uniformity of VCSEL epitaxial wafers provided in this application; Figure 2 This is a schematic diagram illustrating the process of obtaining the characteristic quantities of resistance stability. Detailed Implementation

[0017] The following description, in conjunction with the accompanying drawings, details a specific scheme for measuring the resistivity uniformity of VCSEL epitaxial wafers provided in this application.

[0018] One embodiment of this application provides a method for measuring the resistivity uniformity of a VCSEL epitaxial wafer.

[0019] Specifically, the following method for measuring the resistivity uniformity of VCSEL epitaxial wafers is provided; please refer to [link / reference]. Figure 1The method includes the following steps: Step S1: Set up a measurement point array on the VCSEL epitaxial wafer, and collect the sheet resistance at each measurement point, with the center of the epitaxial wafer as the center of the array; divide all measurement points layer by layer from the center to the outermost layer to obtain the measurement point set of each layer.

[0020] For VCSEL epitaxial wafers, the sheet resistance of discrete points on the VCSEL epitaxial wafer is measured using the four-probe method. The four-probe test current is typically 10μA to 1mA; in this embodiment, it is 100μA to avoid exciting the epitaxial wafer to emit light and generate heat. The probe pressure is 50±5g. An N×N uniform dot matrix is ​​used to arrange the resistance measurement points on the VCSEL epitaxial wafer, where N is an odd number; in this embodiment, N is 7, resulting in 49 measurement points on the epitaxial wafer. The center point of the N×N dot matrix is ​​located at the center of the epitaxial wafer. An automatic probe station drives the four probes to complete the sheet resistance measurement at each point. In other embodiments of this application, the implementer can set the test current and the value of N according to the actual situation. It should be noted that the measurement should avoid the epitaxial wafer edge by 3mm.

[0021] A Cartesian coordinate system is constructed on the VCSEL epitaxial wafer with the center of the epitaxial wafer as the origin and any fixed direction as the positive direction of the horizontal coordinate. The sheet resistance values ​​of all measurement points are then arranged according to their location from the center to the edge and from the first quadrant to the fourth quadrant, resulting in a resistance data sequence. SG filtering is then used to denoise the data. SG filtering is a well-known technique, and its specific process will not be elaborated upon.

[0022] It should be noted that this application provides only one filtering method for filtering resistance data sequences. There are many existing filtering methods, and implementers may also use other filtering algorithms to filter resistance data sequences. This application does not impose any specific restrictions.

[0023] N The measurement points in the N-point matrix are divided according to their location distribution, and sets of measurement points are constructed. Specifically, N... The center point of the N-point matrix is ​​located at the center of the epitaxial wafer. The center measurement points are grouped into the first set of measurement points, and then the three measurement points centered on the center measurement point are... The set of measurement points forming the outermost ring within the 3rd neighborhood is used as the second set of measurement points, with the central measurement point as the center. The set of measurement points forming the outermost ring within the 5-neighborhood is taken as the third set of measurement points, and so on, until N is reached. The set of measurement points forming the outermost ring of the N-point lattice is taken as the last set of measurement points. Thus, a total of... A set of measurement points is used to analyze the resistance distribution in different regions of the epitaxial wafer.

[0024] Step S2: Based on the preset resistivity fluctuation range and the distribution dispersion of the sheet resistance of all measurement points in each measurement point set, and combined with the difference between the sheet resistance of each measurement point and the preset ideal sheet resistance, construct the resistance stability characteristic quantity of each measurement point set.

[0025] During the epitaxial growth of VCSEL epitaxial wafers, the central region is usually surrounded by the edge region and does not flow with the external airflow. Therefore, the airflow in the central region of the epitaxial wafer is more stable, while the airflow at the edge of the epitaxial wafer may interact with the airflow in the environment, generating lateral flow. The stable airflow allows the epitaxial metal to adhere uniformly to the substrate. As a result, the resistance uniformity in the region closer to the center of the epitaxial wafer should be higher, while the resistance uniformity in the edge region should be lower than that in the central region.

[0026] The resistance distribution of VCSEL epitaxial wafers typically exhibits radial symmetry and small, gentle fluctuations. Drastic fluctuations in local resistance can lead to abnormal photoelectric performance of the epitaxial wafer. During four-probe measurement, due to factors such as different contact states between the probes and the epitaxial wafer surface, and uneven local doping of the epitaxial wafer, the sheet resistance values ​​of each measurement point set may fluctuate to varying degrees. Therefore, the resistance stability characteristic of each measurement point set is calculated based on the sheet resistance value distribution within each measurement point set to characterize the stability of the measurement data in each measurement point set. Preferably, in this embodiment, the expression for the resistance stability characteristic is: In the formula, This represents the resistance stability characteristic of the set of k-th measurement points; This represents the preset ideal resistance fluctuation value for the k-th measurement point set. In this embodiment, the specific value is: Since the resistivity fluctuation range specified in GB / T14139-2019 is within... Therefore, the ideal resistance fluctuation values ​​for each set of measurement points from the center to the edge are 0%, 4%, 8%, and 12%, respectively. This represents the coefficient of variation of the sheet resistance of all measurement points in the k-th measurement point set; This represents the average sheet resistance of all measurement points in the k-th measurement point set; This represents the ideal sheet resistance of the VCSEL epitaxial wafer. The implementer can set this value according to the VCSEL epitaxial wafer design and process; this application does not impose specific limitations. In this embodiment... The value is 50Ω / □; , These are the first and second parameter tuning factors, respectively, both serving to prevent the denominator from being zero. In this embodiment... , Take values ​​of 0.1 and 1 respectively. Wherein, This is the first difference. This represents the ideal stability coefficient of the k-th measurement point set. The coefficient of variation is a well-known technique, and the specific process will not be elaborated further.

[0027] It should be noted that since there is only one measurement point in the first set of measurement points, the coefficient of variation of the sheet resistance cannot be calculated. Therefore, the ideal stability coefficient of the first set of measurement points is directly set to 1.

[0028] pass The inverse proportional mapping result characterizes the steady state of the sheet resistance data. The smaller the value, the more stable the sheet resistance data of all measurement points in the k-th measurement point set. Since the resistance distribution in the central region of the epitaxial wafer is more uniform and has smaller fluctuations during metal-organic chemical vapor deposition (MOCVD) epitaxial growth, while the resistance in the edge region fluctuates more, the ratio of the ideal resistance fluctuation value of each measurement point set to the coefficient of variation of the sheet resistance of all measurement points in that set is used to reflect the stability of the actual resistance measured at all measurement points in that set relative to the ideal condition. Represents an ideal state. Under ideal conditions, the smaller the fluctuation value in the actual state, the higher the stability of the data. Then, the resistance stability characteristic of the measurement point set is calculated by comparing the ideal stability coefficient with the first difference, which is used to characterize the validity of the measured resistance data within the measurement point set.

[0029] Step S3: Calculate the difference in resistance stability characteristics between two adjacent sets of measurement points, and construct the reliability of the resistance data of the epitaxial wafer based on the distribution dispersion of all the differences in resistance stability characteristics and the resistance stability characteristics of all sets of measurement points.

[0030] Ideally, the resistance distribution of a VCSEL epitaxial wafer should follow the radial doping pattern of MOCVD growth, exhibiting a single-trend characteristic of high resistance at the center and a slight, stable decrease towards the edges. Based on this, the resistance stability characteristic of each set of measurement points on the epitaxial wafer is calculated, and the difference between the resistance stability characteristic of two adjacent sets of measurement points is calculated and denoted as the stability difference value. This difference can be the absolute value of the difference, the square of the difference, etc. In this embodiment, the difference is the absolute value of the difference between the resistance stability characteristic values.

[0031] This value reflects the smoothness of resistance stability between adjacent regions; the greater the stability difference, the less smooth the resistance. Therefore, the reliability of the resistance data of the VCSEL epitaxial wafer is calculated to characterize the degree of confidence in the fit of the measurement data to the response surface model. Preferably, in this embodiment, the expression for the reliability of the resistance data is: In the formula, This indicates the reliability of the resistance data for VCSEL epitaxial wafers; This represents the number of measurement points in the set; in this embodiment, n=4. The weight of the resistance stability characteristic of the k-th measurement point set is given by the sum of the weights of all measurement point sets, where the weight of the measurement point set closer to the center of the epitaxial wafer is greater, in order to match the more critical characteristics of the central region of the VCSEL chip. This represents the resistance stability characteristic of the set of k-th measurement points; This represents the maximum value among the resistance stability characteristic quantities of the set of all measurement points; The coefficient of variation represents all stable differences. This represents a preset fixed parameter, which in this embodiment is set to 1, meaning that when the coefficient of variation of the stable difference value is 0, it ensures... .

[0032] It should be noted that when calculating the coefficient of variation for all stable differences, if the mean of all stable differences is less than 0.01, the coefficient of variation is set to 0. The implementer can set the weights for each set of measurement points according to the actual situation; this application does not impose specific restrictions. In this embodiment, the weights for each set of measurement points are set to 0.4, 0.3, 0.2, and 0.1 respectively, following the order from the center of the epitaxial wafer to the edge.

[0033] The value represents the uniformity of the epitaxial wafer data; the larger the value, the more stable and uniform the data. The reciprocal of the coefficient of variation of the resistance stability characteristic of adjacent measurement point sets represents the synchronicity of the changes in the resistance stability characteristic of the measurement point set. The product of the above two characteristics is used to obtain the reliability of the resistance data of the VCSEL epitaxial wafer, reflecting the global stability level of the epitaxial wafer resistance.

[0034] Step S4: Perform response surface fitting on the sheet resistance of all measurement points to obtain the fitted sheet resistance value of each measurement point; based on the reliability of the resistance data and the difference between the fitted sheet resistance value and the measured sheet resistance value of each measurement point, calculate the term coefficient correction index of the response surface model.

[0035] Traditional response surface models directly use second-order polynomial fitting, without considering local measurement artifacts caused by probe contact, resulting in large deviations in the resistance reconstruction of the model in abnormal regions.

[0036] First, the planar coordinates (a, b) and corresponding sheet resistance values ​​of all measurement points are collected and used as input to the second-order response surface model, where the coordinates are the independent variables and the sheet resistance value is the dependent variable. The response surface fitting result output by the algorithm is used as the initial fitting model, and the fitted sheet resistance value for each measurement point is calculated. Then, the absolute value of the difference between the fitted sheet resistance value and the measured sheet resistance value for each measurement point is calculated, and then divided by the measured sheet resistance value to obtain the local resistance anomaly of that measurement point, thereby quantifying the degree of deviation of the local resistance from the model. The fitting of the second-order response surface model is a well-known technique, and the specific process will not be elaborated further.

[0037] If the local resistance anomaly at any measurement point is greater than or equal to a preset anomaly threshold, the coefficients of the second-order response surface model are adjusted accordingly; otherwise, it indicates that the initial fitting model has a good fit and no adjustment is needed. In this embodiment, the anomaly threshold is set to 0.05. In other embodiments of this application, the implementer can set the anomaly threshold according to the actual situation.

[0038] The higher the reliability of the resistance data of the VCSEL epitaxial wafer, the more authentic and reliable the measurement data is, and the less obvious the interference. In this case, the deviation between the measured resistance and the initial model is not caused by measurement error, but by the real difference in the intrinsic resistance distribution of the epitaxial wafer. The correction range needs to be increased to make the model fit the real distribution. If Cred is too low, it means that the data contains interference noise. In this case, the correction range should be reduced to avoid misjudging the noise as the intrinsic characteristics of the epitaxial wafer and causing the model to be distorted.

[0039] Therefore, the expression for the correction exponent of the quadratic term coefficients in the response surface model is: In the formula, The quadratic term coefficient correction index represents the response surface model. This represents the mean of the local resistance anomaly at all measurement points; This indicates the reliability of the resistance data for VCSEL epitaxial wafers; This represents the normalization function. In this embodiment, the normalization function used is... Where x is the independent variable and y is the normalized result. It is an exponential function with the natural constant as the base.

[0040] The deviation of local resistance from the model was quantified by measuring local resistance anomalies, and the reliability of the resistance data reflected the global stability level of the epitaxial wafer resistance. The correction exponent of the quadratic term coefficients of the response surface model was obtained by combining the normalized values ​​of local resistance anomalies and the reliability of the resistance data, reflecting the adjustment range of the model from statistical fitting to adaptive approximation of the physical true distribution.

[0041] Step S5: Adjust the quadratic coefficients in the response surface model based on the quadratic coefficient correction exponent, and refit the response surface model of the sheet resistance of all measurement points using the adjusted quadratic coefficients. Analyze the epitaxial wafer resistance change in the final fitted response surface to detect the uniformity of epitaxial wafer resistivity.

[0042] The expression for updating the coefficients of the quadratic term is: ,in, This represents the updated value of the quadratic coefficient. For the term coefficient correction exponent of the response surface model, This represents the value of the quadratic term coefficients before the update. The updated values ​​of the quadratic term coefficients in the second-order response surface model are then used to refit the response surface for all measurement points. It should be noted that during the refitting process, the values ​​of these quadratic term coefficients remain unchanged from their updated values.

[0043] Furthermore, by defining the derivative of the function and performing integration, the maximum, minimum, and mean values ​​of the resistance response surface of the VCSEL epitaxial wafer are obtained. Finally, the resistance range of the VCSEL epitaxial wafer is obtained by subtracting the minimum value from the maximum value. The resistance range is then divided by the mean value to determine the resistance distribution variation of the VCSEL epitaxial wafer.

[0044] When the resistance distribution difference of the VCSEL epitaxial wafer resistors exceeds a preset uniformity threshold, it indicates a problem with the uniformity of the VCSEL epitaxial wafer resistors; when the resistance distribution difference of the VCSEL epitaxial wafer resistors is less than or equal to the uniformity threshold, it indicates good uniformity of the VCSEL epitaxial wafer resistors. In this embodiment, the uniformity threshold is set to 0.12 according to the resistivity specification in GB / T14139-2019. Implementers can set more precise values ​​according to specific requirements; this application does not impose specific limitations.

[0045] A schematic diagram of the process for obtaining the resistance stability characteristic is shown below. Figure 2 As shown.

[0046] In summary, this embodiment of the application collects the resistance at each measurement point on the VCSEL epitaxial wafer, dividing all measurement points on the epitaxial wafer into multiple measurement point sets from the center to the edge. Based on the distribution dispersion of the sheet resistance values ​​measured at all measurement points in each layer's measurement point set and the difference from the ideal sheet resistance, combined with a preset resistivity fluctuation range, a resistance stability characteristic quantity is constructed to characterize the validity of the sheet resistance data measured within the measurement point set, thereby reducing measurement deviations caused by edge airflow disturbances. Based on the resistance stability characteristic quantity of all measurement point sets and its chaotic changes, the reliability of the epitaxial wafer's resistance data is calculated, reflecting the authenticity of the measurement data. Combined with the residuals of the initially fitted response surface model, the response surface model coefficients are adjusted, and the response surface is refitted to improve the accuracy of the response surface fitting. Furthermore, the uniformity of the epitaxial wafer's resistivity is analyzed based on the response surface, improving the reliability of the uniformity evaluation results.

Claims

1. A method for measuring the resistivity uniformity of a VCSEL epitaxial wafer, characterized in that, The method includes the following steps: A measurement point matrix is ​​set on the VCSEL epitaxial wafer, and the sheet resistance at each measurement point is collected, with the center of the epitaxial wafer as the center of the matrix; all measurement points are divided layer by layer from the center to the outermost layer to obtain the set of measurement points for each layer. Based on the preset resistivity fluctuation range and the degree of dispersion of the sheet resistance of all measurement points in each measurement point set, and combined with the difference between the sheet resistance of each measurement point and the preset ideal sheet resistance, the resistance stability characteristic of each measurement point set is constructed. Calculate the difference in resistance stability characteristics between two adjacent sets of measurement points, and construct the reliability of the resistance data of the epitaxial wafer based on the distribution dispersion of all the differences in resistance stability characteristics and the resistance stability characteristics of all sets of measurement points. Response surface fitting is performed on the sheet resistance of all measurement points to obtain the fitted sheet resistance value of each measurement point; based on the reliability of the resistance data and the difference between the fitted sheet resistance value and the measured sheet resistance value of each measurement point, the quadratic term coefficient correction exponent of the second-order response surface model is calculated. Based on the quadratic term coefficient correction exponent, the quadratic term coefficients in the response surface model are adjusted, and the response surface model after the quadratic term coefficient adjustment is used to refit the sheet resistance of all measurement points. The epitaxial wafer resistance change in the finally fitted response surface is analyzed to detect the uniformity of epitaxial wafer resistivity. The process for obtaining the resistance stability characteristic is as follows: The ideal resistance fluctuation value of each set of measurement points is determined based on the distance of each set of measurement points from the center of the epitaxial wafer and the preset resistivity fluctuation range. The ideal resistance fluctuation value is positively correlated with the distance. Calculate the mean sheet resistance and coefficient of variation of all measurement points in each measurement point set; calculate the difference between the mean sheet resistance and the ideal sheet resistance, and denote it as the first difference. Based on the ideal resistance fluctuation value, the coefficient of variation, and the first difference, the resistance stability characteristic of each measurement point set is determined. The resistance stability characteristic is positively correlated with the ideal resistance fluctuation value and negatively correlated with the coefficient of variation and the first difference, respectively. The expression for the resistance stability characteristic is: In the formula, This represents the resistance stability characteristic of the set of k-th measurement points; This represents the preset ideal resistance fluctuation value for the k-th measurement point set; This represents the coefficient of variation of the sheet resistance of all measurement points in the k-th measurement point set; This represents the average sheet resistance of all measurement points in the k-th measurement point set; This represents the ideal sheet resistance of the VCSEL epitaxial wafer. , These are the preset first and second parameter tuning factors, respectively; The process for obtaining the reliability of the resistance data is as follows: Calculate the weighted sum of the resistance stability characteristics of all measurement point sets, where the weight of the resistance stability characteristics of each measurement point set is negatively correlated with the distance of each measurement point set from the center of the epitaxial wafer; The difference between the resistance stability characteristic of two adjacent sets of measurement points is recorded as the stability difference value; Based on the weighted summation value and the coefficient of variation of the stable differences between all adjacent measurement point sets, the reliability of the resistance data of the epitaxial wafer is determined. The expression for the reliability of the resistance data is: In the formula, This indicates the reliability of the resistance data for VCSEL epitaxial wafers; Indicates the number of measurement points in the set; The weights of the resistance stability characteristics of the k-th measurement point set are represented. This represents the resistance stability characteristic of the set of k-th measurement points; This represents the maximum value among the resistance stability characteristic quantities of the set of all measurement points; The coefficient of variation represents all stable differences. Indicates preset fixed parameters; The process of adjusting the coefficients of the quadratic term in the response surface model is as follows: The absolute value of the difference between the fitted block resistance value and the measured block resistance value at each measurement point is divided by the measured block resistance value, and the result is recorded as the local resistance anomaly degree at each measurement point. When the local resistance anomaly at any measurement point is greater than or equal to the preset anomaly threshold, the product of the quadratic term coefficient in the response surface model and the correction exponent of the quadratic term coefficient is used as the adjustment amount of the coefficient, and the sum of the coefficient and its adjustment amount is used as the adjusted value of the coefficient. When the local resistance anomaly at all measurement points is less than the preset anomaly threshold, the coefficients in the response surface model are not adjusted.

2. The method for measuring the resistivity uniformity of a VCSEL epitaxial wafer as described in claim 1, characterized in that, The process for obtaining the correction exponent for the quadratic term coefficient is as follows: The quadratic term coefficient correction index of the response surface model is determined based on the reliability of the resistance data and the local resistance anomaly at all measurement points.

3. The method for measuring the resistivity uniformity of a VCSEL epitaxial wafer as described in claim 1, characterized in that, The quadratic term coefficient correction index is the product of the normalized value of the mean of the local resistance anomaly at all measurement points and the reliability of the resistance data.

4. The method for measuring the resistivity uniformity of a VCSEL epitaxial wafer as described in claim 1, characterized in that, The analysis of the final fitted response surface reveals the change in epitaxial wafer resistance, which is used to detect the resistivity uniformity of the epitaxial wafer. The process is as follows: Calculate the maximum, minimum, and mean values ​​of the block resistance in the final fitted response surface. Calculate the difference between the maximum and minimum values ​​and the mean value of the block resistance in the response surface as the resistance distribution difference of the epitaxial wafer. The resistivity uniformity of the epitaxial wafer is detected based on the resistance distribution difference.

5. The method for measuring the resistivity uniformity of a VCSEL epitaxial wafer as described in claim 4, characterized in that, The process of detecting the resistivity uniformity of the epitaxial wafer based on the resistance distribution difference is specifically as follows: If the resistance distribution difference is greater than the preset uniformity threshold, the resistivity uniformity of the epitaxial wafer is deemed unqualified; otherwise, the resistivity uniformity of the epitaxial wafer is deemed qualified.

Images