Method for detecting and correcting center offset in wafer conveying process

Abstract

The invention belongs to the technical field of wafer conveying, and provides a method for detecting and correcting center offset in a wafer conveying process, which comprises the following steps of: acquiring three-dimensional coordinate data of a wafer, combining an ideal plane fitted by the wafer, preliminarily identifying a wafer warping form under a static working condition, identifying a slightly abnormal wafer, and correcting the center offset in the wafer conveying process. Determining a slightly abnormal area in the slightly abnormal wafer by using a sliding window local statistical anomaly detection method; extracting a slight warping area of the slightly abnormal wafer, and according to the static residual stress of the slight warping area, in combination with the additional stress generated on the slight warping area in the slightly abnormal wafer transmission process, performing coupling analysis on the vulnerability of the static defect of the slight warping area and the strength of external interference; the sudden warping risk of the slight warping area is evaluated; according to the invention, the deviation of the transfer center caused by periodic warping is effectively corrected, the accuracy of the wafer transfer process is ensured, and reliable support is provided for the stability of the subsequent processing link and the wafer quality.

Description

Technical Field

[0001] This invention belongs to the field of wafer transfer technology, specifically a method for detecting and correcting center offset during wafer transfer. Background Technology

[0002] In semiconductor manufacturing, the accuracy of wafer transfer directly affects the processing quality of subsequent key processes such as photolithography and etching. Wafer center offset is one of the important factors that leads to processing errors and reduced yield.

[0003] In existing technologies, the ability to identify periodic wavy warping is limited in warping morphology analysis, making it difficult to capture spatial periodic patterns and resulting in insufficient targeted processing of regular deformations. Furthermore, in the center offset correction stage, traditional methods typically perform overall fitting of wafer edge points, failing to distinguish between edge points in coupled and uncoupled warping regions, and ignoring the impact of local fluctuations caused by periodic warping on center positioning, leading to low correction accuracy. Simultaneously, existing technologies lack systematic analysis of the coupling effect between dynamic stress and static defects during transport, severely restricting the stability and efficiency of high-precision wafer manufacturing.

[0004] Therefore, the present invention provides a method for detecting and correcting center offset during wafer transfer. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0006] This invention provides a method for detecting and correcting center offset during wafer transport, comprising:

[0007] The three-dimensional coordinate data of the wafer is obtained and combined with the ideal plane of the wafer fitting to preliminarily identify the wafer warping morphology under static working conditions, identify slightly abnormal wafers, and use the sliding window local statistical anomaly detection method to determine the slightly abnormal regions in the slightly abnormal wafers.

[0008] The slight warping region of the slightly abnormal wafer is extracted. Based on the static residual stress of the slightly warping region and the additional stress generated by the slight abnormal wafer transfer process, the risk of sudden warping of the slightly warping region is assessed by coupling analysis of the fragility of static defects in the slightly warping region and the intensity of external interference.

[0009] If the risk of sudden warping is high, frequency domain analysis is used to extract the spatial periodic features of the coupled warping region in the slightly warped region, and periodic wave-shaped morphology is fitted to determine whether the coupled warping region exhibits periodic wave-shaped warping in spatial location.

[0010] If so, a fitting analysis is performed on the edge points of the slightly warped wafer with periodic wavy warping to evaluate the impact of the slightly warped wafer on the center offset detection during the transfer process, determine the final transfer center and final edge radius of the slightly warped wafer, and correct the center offset during the transfer process of the slightly warped wafer.

[0011] Furthermore, the process of identifying slightly abnormal wafers is as follows:

[0012] In static wafer warpage measurement, the ideal plane is usually defined as the XY reference plane of the measurement coordinate system;

[0013] Using the center of the starting position of the wafer datum as the origin, a real three-dimensional coordinate system is constructed. The XY reference plane of the real three-dimensional coordinate system is aligned with the ideal plane. The XY reference plane of the real three-dimensional coordinate system is divided into grids at fixed intervals, and measurement points are set at the intersections of the grids.

[0014] For all the measurement points set, the vertical height of the measurement points is calculated using the laser triangulation method;

[0015] The difference between the vertical height of the measurement point and the corresponding ideal height in the ideal plane is calculated, and the absolute value is taken to obtain the relative deviation of the height of the measurement point.

[0016] The relative height deviation of all measurement points is calculated using the coefficient of variation formula to obtain the relative height deviation fluctuation coefficient of the measurement points, which is denoted as the wafer height fluctuation value.

[0017] If the wafer height fluctuation value is less than the wafer height fluctuation threshold, the corresponding wafer is recorded as a slightly abnormal wafer.

[0018] Furthermore, the process of determining the slightly abnormal region in a slightly abnormal wafer using the sliding window local statistical anomaly detection method is as follows:

[0019] The sliding window size and sliding window step size are set according to the wafer grid spacing;

[0020] The local mean of the relative height deviation of all measurement points within the sliding window is obtained by summing and averaging the relative height deviations of the sliding window.

[0021] The standard deviation formula is used to calculate the standard deviation of the relative height deviation of all measurement points within the sliding window, thus obtaining the local standard deviation of the relative height deviation of the sliding window.

[0022] The anomaly determination coefficient is obtained by weighted summing of the local mean of the relative height deviation of the sliding window and the local standard deviation of the relative height deviation of the sliding window.

[0023] If the anomaly determination coefficient is less than or equal to the standard value of the anomaly determination coefficient, then all measurement points within the corresponding sliding window will be recorded as minor anomalies.

[0024] For all marked minor outliers, a 4-neighborhood connectivity analysis was performed to aggregate consecutively adjacent minor outliers into minor outlier regions.

[0025] Furthermore, the process for assessing the risk of sudden warping in the slightly warped area is as follows:

[0026] By coupling analysis of the vulnerability of static defects in slightly warped regions with the intensity of external disturbances, the vulnerability coefficient and the disturbance intensity coefficient of slightly warped regions are determined.

[0027] The vulnerability coefficient of the slightly warped region is multiplied with the interference intensity coefficient of the slightly warped region to obtain the sudden risk value of the slightly warped region.

[0028] If the sudden risk value of a slightly warped region is greater than or equal to the sudden risk value threshold, it indicates that the sudden warping risk of the slightly warped region is high, and the corresponding slightly warped region is recorded as a coupled warped region; otherwise, it indicates that the sudden warping risk of the slightly warped region is low.

[0029] Furthermore, the process for determining the fragility coefficient of the slightly warped region is as follows:

[0030] A monitoring period was set, and Raman spectroscopy was used to scan all slightly warped areas to obtain stress distribution cloud maps of all slightly warped areas;

[0031] Based on any slightly warped region, the stress values ​​corresponding to the stress distribution cloud map of the slightly warped region are summed and averaged to obtain the static residual stress of the slightly warped region. Introducing the yield strength of the wafer material, the static residual stress of the slightly warped region is compared with the yield strength of the wafer material to obtain the fragility coefficient of the slightly warped region.

[0032] Furthermore, the process for determining the interference intensity coefficient of the slightly warped region is as follows:

[0033] The additional stress generated by the slight abnormal wafer transfer process on the slight warping region is obtained, and the ratio of this stress to the static residual stress in the slight warping region is processed to obtain the interference intensity coefficient of the slight warping region.

[0034] Furthermore, the process of determining whether the coupled warped region exhibits periodic wavy warping in spatial location is as follows:

[0035] Frequency domain analysis was used to extract the spatial periodic features of the coupled warped region in the slightly warped region, and periodic wave-shaped morphology was fitted to determine the goodness of fit of the measurement points.

[0036] If the goodness of fit of the measurement point is greater than or equal to the goodness of fit threshold, it indicates that the coupled warped region exhibits periodic wavy warping in spatial location; if the goodness of fit of the measurement point is less than the goodness of fit threshold, it indicates that the coupled warped region does not exhibit periodic wavy warping in spatial location.

[0037] Furthermore, the process for determining the goodness of fit of the measurement points is as follows:

[0038] The coordinates of all measurement points in the coupled warped region are extracted to construct a spatial signal matrix. A two-dimensional fast Fourier transform is performed on the spatial signal matrix to obtain a frequency domain signal matrix. The power spectral density of the frequency domain signal is calculated to generate a power spectral density map. The peak points of the power values ​​are extracted from the power spectral density map. The radius of the neighborhood around the peak point is set. All power values ​​within the radius of the neighborhood around the peak point are extracted, summed, and averaged to obtain the average power of the neighborhood around the peak point. The standard deviation formula is used to calculate the standard deviation of all power values ​​within the radius of the neighborhood around the peak point. The signal-to-noise ratio of the peak point is then calculated.

[0039] If the signal-to-noise ratio of a peak point is greater than or equal to the standard signal-to-noise ratio value, then the corresponding peak point is recorded as a significant peak point; the reciprocal of the frequency in the X-axis direction corresponding to the significant peak point is used to obtain the candidate period in the X-axis direction of the significant peak point, and the reciprocal of the frequency in the Y-axis direction corresponding to the significant peak point is used to obtain the candidate period in the Y-axis direction of the significant peak point.

[0040] A two-dimensional sine fitting function is defined, and the least squares method is used to fit the coordinates of all measurement points in the measurement point set to the two-dimensional sine fitting function. The goodness of fit of the measurement points is calculated by the goodness of fit formula.

[0041] Furthermore, the process for evaluating the impact of slightly warped wafers on center offset detection during transport is as follows:

[0042] The wafer physical boundary is determined by the wafer diameter. Measurement points within a set grid spacing at a distance from the wafer physical boundary are extracted to form an edge point set. Measurement points in the edge point set that coincide with the coupled warped region are recorded as coupled edge points, and measurement points in the edge point set that coincide with the uncoupled warped region are recorded as uncoupled edge points.

[0043] The uncoupled edge points are fitted using the least squares circle fitting algorithm to minimize the sum of squared radial errors from all uncoupled edge points to the circle, thereby solving for the optimal reference center and reference radius and establishing a reference circle.

[0044] The center of the circle after directly fitting all the coupling edge points to a circle is taken as the center of disturbance, and the disturbance error between the center of disturbance and the optimal reference center is calculated by the Euclidean distance formula.

[0045] If the interference error is greater than or equal to the interference error threshold, it means that the slight warping of the wafer has a significant impact on the center offset detection during the transfer process, and the fluctuation of the coupling edge point needs to be corrected. If the interference error is less than the interference error threshold, the slight warping of the wafer has a small impact on the center offset detection during the transfer process, and the fluctuation of the coupling edge point can be ignored. The reference center can be directly used as the transfer center.

[0046] Furthermore, the process of determining the final transfer center and final edge radius of the slightly warped wafer is as follows:

[0047] Since slight warping of the wafer has a significant impact on center offset detection during the transfer process, a ripple radial conversion coefficient is introduced for each coupling edge point. The radial ripple offset corresponding to the relative height deviation of the coupling edge point is calculated. Combined with the phase of the X-axis and the Y-axis corresponding to the relative height deviation of the coupling edge point, it is determined whether the coupling edge point is at the peak, trough or middle of the periodic wavy warping, and the coupling edge point is corrected accordingly.

[0048] The corrected coupled edge points and uncoupled edge points are integrated to form a corrected full edge point set. Least square circle fitting is then performed on the full edge point set to solve for the final transmission center and the final edge radius.

[0049] The beneficial effects of this invention are as follows:

[0050] 1. This invention constructs a real three-dimensional coordinate system aligned with an ideal plane, and uses laser triangulation to accurately obtain the vertical height of measurement points and calculate the relative height deviation. This identifies slightly abnormal wafers under static conditions, effectively avoiding misjudgments caused by subjective parameter settings or measurement deviations, and achieving a clear distinction between slightly warped and non-slightly warped regions. Non-contact Raman spectroscopy is used to obtain the stress distribution in slightly warped regions to calculate static residual stress, avoiding wafer damage and accurately quantifying the region's fragility coefficient. Simultaneously, the interference intensity coefficient is calculated by combining the additional stress during wafer transport. The coupling of these two methods yields a sudden risk value, identifying sudden warping weak points in static slightly warped regions under transport forces, providing strong assurance for the stability and reliability of subsequent wafer processing and transport.

[0051] 2. This invention utilizes two-dimensional fast Fourier transform to extract spatial periodic features, combines signal-to-noise ratio analysis to screen significant peak points, and uses two-dimensional sine fitting and goodness-of-fit judgment to accurately identify whether the coupled warp region exhibits periodic wavy warping, avoiding ineffective processing of irregular deformations and improving the targeting and accuracy of warp morphology analysis. By distinguishing between coupled and uncoupled edge points, interference errors are calculated based on the reference circle and radial fluctuation offset. Combined with the peak and trough characteristics of periodic wavy warping, the coupled edge points are specifically corrected. Finally, the final transfer center is determined by fitting all edge points, effectively correcting the transfer center offset caused by periodic warping, ensuring the accuracy of the wafer transfer process, and providing reliable support for the stability and wafer quality of subsequent processing stages. Attached Figure Description

[0052] The invention will now be further described with reference to the accompanying drawings.

[0053] Figure 1 This is a flowchart illustrating the steps of a method for detecting and correcting center offset during wafer transfer according to an embodiment of the present invention.

[0054] Figure 2 This is a system block diagram of a system for detecting and correcting center offset during wafer transfer, according to an embodiment of the present invention. Detailed Implementation

[0055] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0056] Example 1

[0057] Please see Figure 1 As shown in the figure, a method for detecting and correcting center offset during wafer transfer according to an embodiment of the present invention includes the following steps:

[0058] Step 1: Obtain the three-dimensional coordinate data of the wafer, combine it with the ideal plane fitted to the wafer, preliminarily identify the wafer warping morphology under static operating conditions, identify wafers with slight abnormalities, and use the sliding window local statistical anomaly detection method to determine the slight abnormal regions in the wafers with slight abnormalities.

[0059] In static wafer warpage measurement, the ideal plane is usually defined as the XY reference plane of the measurement coordinate system (i.e., the z coordinate of all points on the ideal plane is 0 or a certain fixed reference value).

[0060] Using the center of the starting position of the wafer datum as the origin, a real three-dimensional coordinate system is constructed. The XY reference plane of the real three-dimensional coordinate system is aligned with the ideal plane. The XY reference plane of the real three-dimensional coordinate system is divided into grids at fixed intervals, and measurement points are set at the intersections of the grids.

[0061] It should be noted that the XY reference plane is a plane constructed by the X-axis and Y-axis of the actual three-dimensional coordinate system.

[0062] For all the measurement points set, a laser beam at a fixed angle is emitted towards the measurement point using laser triangulation. The laser beam is reflected by the point and received by the receiver. The vertical height of the measurement point is calculated using trigonometric relationships by using the angle between the emitted and reflected light and the fixed distance inside the device.

[0063] The ideal height corresponding to the ideal plane is set by those skilled in the art based on historical experience. The difference between the vertical height of the measurement point and the corresponding ideal height in the ideal plane is calculated, and the absolute value is taken to obtain the relative deviation of the height of the measurement point.

[0064] Based on the relative height deviation of the measurement points, the relative height deviation fluctuation coefficient of the measurement points is calculated using the coefficient of variation formula and recorded as the wafer height fluctuation value.

[0065] Compare the wafer height fluctuation value with the wafer height fluctuation threshold:

[0066] If the wafer height fluctuation value is greater than or equal to the wafer height fluctuation threshold, the corresponding wafer will be recorded as a severely abnormal wafer.

[0067] If the wafer height fluctuation value is less than the wafer height fluctuation threshold, the corresponding wafer is recorded as a slightly abnormal wafer.

[0068] The specific process for determining the minor anomalous region in a minor anomalous wafer using the sliding window local statistical anomaly detection method is as follows:

[0069] The sliding window size and sliding window step size are set according to the wafer grid spacing;

[0070] The local mean of the relative height deviation of all measurement points within the sliding window is obtained by summing and averaging the relative height deviations of the sliding window.

[0071] The standard deviation formula is used to calculate the standard deviation of the relative height deviation of all measurement points within the sliding window, thus obtaining the local standard deviation of the relative height deviation of the sliding window.

[0072] The anomaly determination coefficient is obtained by weighted summing of the local mean of the relative height deviation of the sliding window and the local standard deviation of the relative height deviation of the sliding window.

[0073] It should be noted that the weighting coefficient of the local mean of the relative height deviation of the sliding window is 1, and the weighting coefficient of the local standard deviation of the relative height deviation of the sliding window is 1.2.

[0074] The standard value of the anomaly judgment coefficient is set by those skilled in the art based on historical experience. If the anomaly judgment coefficient is greater than the standard value of the anomaly judgment coefficient, all measurement points in the corresponding sliding window will be recorded as serious anomalies.

[0075] If the anomaly determination coefficient is less than or equal to the standard value of the anomaly determination coefficient, then all measurement points within the corresponding sliding window will be recorded as minor anomalies.

[0076] For all marked minor outliers, perform 4-neighbor connectivity analysis (4-neighbor: adjacent in the top, bottom, left, and right directions) to aggregate consecutively adjacent minor outliers into minor outlier regions.

[0077] Step 2: Extract the slight warping region of the slightly abnormal wafer. Based on the static residual stress of the slightly warping region and the additional stress generated by the slight abnormal wafer transfer process, the risk of sudden warping in the slightly warping region is assessed by coupling analysis of the fragility of static defects in the slightly warping region and the intensity of external interference.

[0078] It should be noted that a static, slight warping area may become a weak point for sudden warping, and under the action of external force during transfer, it may induce local abnormal deformation. This local abnormal deformation will affect the detection of center offset during wafer transfer.

[0079] During a set monitoring period, Raman spectroscopy (non-contact, non-destructive to the wafer) was used to scan all slightly warped areas, obtaining stress distribution cloud maps of all slightly warped areas;

[0080] Based on any slightly warped region, the stress values ​​corresponding to the stress distribution cloud map of the slightly warped region are summed and averaged to obtain the static residual stress of the slightly warped region. The yield strength of the wafer material is introduced, and the ratio of the static residual stress of the slightly warped region to the yield strength of the wafer material is processed to obtain the fragility coefficient of the slightly warped region.

[0081] The additional stress generated by the slight abnormal wafer transfer process on the slight warping region is obtained, and the ratio is processed with the static residual stress of the slight warping region to obtain the interference intensity coefficient of the slight warping region.

[0082] It should be noted that the essence of additional stress is the dynamic stress generated in the slightly warped area by external forces such as airflow, negative pressure, clamping force, and temperature during the transmission process. During the transmission process (not at the moment of clamping, to avoid interference), the slightly warped area is scanned with a Raman spectrometer, and the static residual stress before transmission is compared with the total stress during transmission. The difference is the additional stress.

[0083] The vulnerability coefficient of the slightly warped region is multiplied with the interference intensity coefficient of the slightly warped region to obtain the sudden risk value of the slightly warped region.

[0084] In some embodiments, the sudden risk value of the slightly warped area is compared with the sudden risk value threshold. The specific comparison process is as follows:

[0085] If the sudden risk value of a slightly warped region is greater than or equal to the sudden risk value threshold, it indicates that the sudden warping risk of the slightly warped region is high, and the corresponding slightly warped region is recorded as a coupled warped region.

[0086] If the sudden risk value of a slightly warped region is less than the sudden risk value threshold, it indicates that the sudden warping risk of the slightly warped region is low, and the corresponding slightly warped region is recorded as a non-coupled warped region.

[0087] The technical solution of this embodiment is as follows: Obtain the three-dimensional coordinate data of the wafer, combine it with the ideal plane fitted to the wafer, preliminarily identify the wafer warpage morphology under static conditions, identify slightly abnormal wafers, and use the sliding window local statistical anomaly detection method to determine the slightly abnormal regions in the slightly abnormal wafers; extract the slightly warped regions of the slightly abnormal wafers, and based on the static residual stress of the slightly warped regions, combined with the additional stress generated by the slightly abnormal wafer transport process on the slightly warped regions, assess the sudden warping risk of the slightly warped regions by coupling analysis of the fragility of static defects in the slightly warped regions and the intensity of external interference; this invention constructs an actual three-dimensional coordinate system aligned with the ideal plane, combined with laser three-dimensional coordinates... The angle measurement method accurately obtains the vertical height of the measurement point and calculates the relative height deviation, identifying slightly abnormal wafers under static conditions. This effectively avoids misjudgments caused by subjective parameter settings or measurement deviations, achieving a clear distinction between slightly warped and non-slightly warped areas. Non-contact Raman spectroscopy is used to obtain the stress distribution in slightly warped areas to calculate static residual stress, avoiding wafer damage and accurately quantifying the region's vulnerability coefficient. Simultaneously, the interference intensity coefficient is calculated by combining the additional stress during wafer transport. By coupling these two methods, a sudden risk value is obtained, identifying sudden warping weak points in static slightly warped areas under transport external forces. This provides strong support for the stability and reliability of subsequent wafer processing and transport processes.

[0088] Example 2

[0089] Please see Figure 1 As shown in the figure, a method for detecting and correcting center offset during wafer transfer according to an embodiment of the present invention further includes the following steps:

[0090] Step 3: If the risk of sudden warping is high, frequency domain analysis is used to extract the spatial periodic features of the coupled warping region in the slight warping region, and periodic wave-shaped morphology fitting is performed to determine whether the coupled warping region exhibits periodic wave-shaped warping in spatial location.

[0091] Periodic wave warping is a deformation pattern with a spatial repetition and smooth undulation formed in the slightly warped area of ​​a wafer under the action of static residual stress and additional external force.

[0092] Extract the coordinates of the measurement points in all coupled warped regions and construct a spatial signal matrix. The rows of the spatial signal matrix correspond to the X-axis, the columns correspond to the Y-axis, and the elements of the spatial signal matrix are the relative height deviation values ​​corresponding to the measurement point positions.

[0093] Perform a two-dimensional fast Fourier transform on the spatial signal matrix to obtain the frequency domain signal matrix, calculate the power spectral density of the frequency domain signal, and generate a power spectral density map. The horizontal axis of the power spectral density map represents the frequency in the X-axis, the vertical axis represents the frequency in the Y-axis, and the color intensity represents the power level.

[0094] The peak power value is extracted from the power spectral density map. A radius is defined around the peak power value. All power values ​​within this radius are summed and averaged to obtain the average power of the power range around the peak power value. The standard deviation is then calculated using the standard deviation formula for all power values ​​within the radius around the peak power value. Calculate the signal-to-noise ratio (SNR) at the peak point, where, This represents the power value at the peak point itself. It is the average power of the neighborhood surrounding the peak point. It is the standard deviation of all power in the neighborhood surrounding the peak point;

[0095] If the signal-to-noise ratio of a peak point is greater than or equal to the standard value of the signal-to-noise ratio, then the corresponding peak point is recorded as a significant peak point.

[0096] If the signal-to-noise ratio of a peak point is less than the standard value of the signal-to-noise ratio, then the corresponding peak point is recorded as a non-significant peak point.

[0097] It should be noted that the standard value for signal-to-noise ratio was set by those skilled in the art based on historical experience;

[0098] The candidate period of the X-axis is obtained by taking the reciprocal of the frequency corresponding to the significant peak point. The candidate period of the Y-axis is obtained by taking the reciprocal of the frequency corresponding to the significant peak point.

[0099] Obtain the amplitude of the coupled warped region and set the two-dimensional sine fitting function. Specifically: ,in: The amplitude of the coupled warped region is represented by x and y, which represent the coordinates of the measurement point. Represents the candidate period in the X-axis direction of the coordinate system. Represents the candidate period in the Y-axis direction of the coordinate system. Indicates the phase in the X-axis direction of the coordinate system. This indicates the phase in the Y-axis direction, and C represents the global offset.

[0100] The least squares method is used to fit the coordinates of all measurement points in the measurement point set to a two-dimensional sine fitting function, and the solution is obtained. , , , , The optimal value of C is obtained, and the goodness of fit of the measurement points is calculated using the goodness-of-fit formula. This is used to measure the degree of fit between the actual data and the fitted function;

[0101] Set a goodness-of-fit threshold. If the goodness-of-fit of the measurement point is greater than or equal to the goodness-of-fit threshold, it indicates that the coupled warped region exhibits periodic wavy warping in spatial location. If the goodness-of-fit of the measurement point is less than the goodness-of-fit threshold, it indicates that the coupled warped region does not exhibit periodic wavy warping in spatial location, and no processing is performed.

[0102] Step 4: If so, perform fitting analysis on the edge points of the slightly warped wafer with periodic wavy warping, evaluate the impact of the slightly warped wafer on the center offset detection during the transfer process, determine the final transfer center and final edge radius of the slightly warped wafer, and correct the center offset during the transfer process of the slightly warped wafer.

[0103] The wafer physical boundary is determined by the wafer diameter. Measurement points within a set grid spacing at a distance from the wafer physical boundary are extracted to form an edge point set. Measurement points in the edge point set that coincide with the coupled warped region are recorded as coupled edge points, and measurement points in the edge point set that coincide with the uncoupled warped region are recorded as uncoupled edge points.

[0104] The uncoupled edge points are fitted using the least squares circle fitting algorithm to minimize the sum of squared radial errors from all uncoupled edge points to the circle, thereby solving for the optimal reference center and reference radius and establishing a reference circle.

[0105] The center of the circle after directly fitting all the coupling edge points to a circle is taken as the center of disturbance, and the disturbance error between the center of disturbance and the optimal reference center is calculated by the Euclidean distance formula.

[0106] If the interference error is greater than or equal to the interference error threshold, it means that the slight warping of the wafer has a significant impact on the center offset detection during the transfer process, and the fluctuation of the coupling edge point needs to be corrected. If the interference error is less than the interference error threshold, the slight warping of the wafer has a small impact on the center offset detection during the transfer process, and the fluctuation of the coupling edge point can be ignored. The reference center can be directly used as the transfer center.

[0107] Since slight wafer warping has a significant impact on center offset detection during transport, a ripple radial conversion coefficient is introduced for each coupling edge point, using the formula: Calculate the radial undulation offset corresponding to the relative height deviation of the coupling edge point. Where k is the wave radial conversion coefficient, which is determined by the stiffness and thickness of the wafer material and is calibrated experimentally;

[0108] Based on the phase of the X-axis and the Y-axis corresponding to the relative height deviation of the coupling edge point, determine whether the coupling edge point is at the crest, trough, or midpoint of the periodic wavy warp. Specifically:

[0109] The relative height deviation of the coupling edge points in the two-dimensional sinusoidal fitting function is denoted as the relative height deviation of the fitting. The relative height deviations of the fitting height of all coupling edge points are sorted from largest to smallest to obtain the sequence of fitting height deviations. The relative height deviation of the fitting height of the first coupling edge point in the sequence is extracted and denoted as the maximum relative height deviation. The relative height deviation of the fitting height of the first coupling edge point in the sequence is extracted and denoted as the minimum relative height deviation.

[0110] like , This represents the relative deviation of the maximum fit height, indicating that the coupling edge point is at the peak of the periodic wavy warp. The coupling edge point is then corrected to... ;

[0111] like , This represents the minimum fitting height relative deviation, indicating that the coupling edge point is at a trough in the periodic wavy warp. The coupling edge point is then corrected to... ;

[0112] like and , This is an empirical threshold used to quantify the near-equilibrium state. It is set by those skilled in the art based on historical experience. A threshold indicates that the coupling edge point is in the middle of the periodic wavy warp, with minimal offset, and is directly retained. ;

[0113] The corrected coupled edge points and uncoupled edge points are integrated to form a corrected full edge point set. Least square circle fitting is performed on the full edge point set again to solve for the final transport center and the final edge radius. The center offset of the slightly warped wafer during transport is corrected based on the final transport center and the final edge radius.

[0114] The technical solution of this embodiment is as follows: If the risk of sudden warping is high, frequency domain analysis is used to extract the spatial periodic features of the coupled warping region in the slight warping region, and periodic wavy morphology fitting is performed to determine whether the coupled warping region exhibits periodic wavy warping in spatial position; if so, fitting analysis is performed on the edge points of the slightly warped wafer with periodic wavy warping to evaluate the impact of the slightly warped wafer on center offset detection during the transfer process, determine the final transfer center and final edge radius of the slightly warped wafer, and correct the center offset during the transfer process of the slightly warped wafer; this invention utilizes two-dimensional fast Fourier transform to extract spatial periodic features, combined with signal-to-noise ratio analysis to screen significant peaks. By using two-dimensional sine fitting and goodness-of-fit judgment, the system can accurately identify whether the coupled warp region exhibits periodic wavy warping, avoiding ineffective processing of irregular deformations and improving the targeting and accuracy of warp morphology analysis. By distinguishing between coupled and uncoupled edge points, interference errors are calculated based on the reference circle and radial fluctuation offset. Combined with the peak and trough characteristics of periodic wavy warping, the coupled edge points are specifically corrected. Finally, the final transfer center is determined by fitting all edge points, effectively correcting the transfer center offset caused by periodic warping, ensuring the accuracy of the wafer transfer process, and providing reliable support for the stability and wafer quality of subsequent processing stages.

[0115] Example 3

[0116] Please see Figure 2 As shown in the figure, a system for detecting and correcting center offset during wafer transport according to an embodiment of the present invention includes the following modules:

[0117] Anomaly identification module: acquires the three-dimensional coordinate data of the wafer, combines it with the ideal plane fitted to the wafer, preliminarily identifies the wafer warping shape under static working conditions, identifies wafers with slight anomalies, and uses the sliding window local statistical anomaly detection method to determine the slight anomaly region in the wafer with slight anomalies.

[0118] Coupled Analysis Module: Extracts the slight warping region of the slightly abnormal wafer. Based on the static residual stress of the slightly warping region and the additional stress generated by the slight abnormal wafer transfer process on the slightly warping region, the module performs coupled analysis on the fragility of the static defects in the slightly warping region and the intensity of external interference to assess the risk of sudden warping in the slightly warping region.

[0119] Morphology fitting module: If the risk of sudden warping is high, frequency domain analysis is used to extract the spatial periodic features of the coupled warping region in the slightly warped region, and periodic wave-shaped morphology fitting is performed to determine whether the coupled warping region exhibits periodic wave-shaped warping in spatial location.

[0120] Center Offset Correction Module: If so, perform fitting analysis on the edge points of the slightly warped wafer with periodic wavy warping, evaluate the impact of the slightly warped wafer on center offset detection during the transfer process, determine the final transfer center and final edge radius of the slightly warped wafer, and correct the center offset during the transfer process of the slightly warped wafer.

[0121] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely 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 present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for detecting and correcting center offset during wafer transfer, characterized in that: include: The three-dimensional coordinate data of the wafer is obtained and combined with the ideal plane of the wafer fitting to preliminarily identify the wafer warping morphology under static working conditions, identify slightly abnormal wafers, and use the sliding window local statistical anomaly detection method to determine the slightly abnormal regions in the slightly abnormal wafers. The slight warping region of the slightly abnormal wafer is extracted. Based on the static residual stress of the slightly warping region and the additional stress generated by the slight abnormal wafer transfer process, the risk of sudden warping of the slightly warping region is assessed by coupling analysis of the fragility of static defects in the slightly warping region and the intensity of external interference. If the risk of sudden warping is high, frequency domain analysis is used to extract the spatial periodic features of the coupled warping region in the slightly warped region, and periodic wave-shaped morphology is fitted to determine whether the coupled warping region exhibits periodic wave-shaped warping in spatial location. If so, a fitting analysis is performed on the edge points of the slightly warped wafer with periodic wavy warping to evaluate the impact of the slightly warped wafer on the center offset detection during the transfer process, determine the final transfer center and final edge radius of the slightly warped wafer, and correct the center offset during the transfer process of the slightly warped wafer.

2. The method for detecting and correcting center offset during wafer transfer according to claim 1, characterized in that: The process for identifying wafers with minor anomalies is as follows: In static wafer warpage measurement, the ideal plane is usually defined as the XY reference plane of the measurement coordinate system; Using the center of the starting position of the wafer datum as the origin, a real three-dimensional coordinate system is constructed. The XY reference plane of the real three-dimensional coordinate system is aligned with the ideal plane. The XY reference plane of the real three-dimensional coordinate system is divided into grids at fixed intervals, and measurement points are set at the intersections of the grids. For all the measurement points set, the vertical height of the measurement points is calculated using the laser triangulation method; The difference between the vertical height of the measurement point and the corresponding ideal height in the ideal plane is calculated, and the absolute value is taken to obtain the relative deviation of the height of the measurement point. The relative height deviation of all measurement points is calculated using the coefficient of variation formula to obtain the relative height deviation fluctuation coefficient of the measurement points, which is denoted as the wafer height fluctuation value. If the wafer height fluctuation value is less than the wafer height fluctuation threshold, the corresponding wafer is recorded as a slightly abnormal wafer.

3. The method for detecting and correcting center offset during wafer transfer according to claim 2, characterized in that: The process of determining the minor anomalous region in a minor anomalous wafer using the sliding window local statistical anomaly detection method is as follows: The sliding window size and sliding window step size are set according to the wafer grid spacing; The local mean of the relative height deviation of all measurement points within the sliding window is obtained by summing and averaging the relative height deviations of the sliding window. The standard deviation formula is used to calculate the standard deviation of the relative height deviation of all measurement points within the sliding window, thus obtaining the local standard deviation of the relative height deviation of the sliding window. The anomaly determination coefficient is obtained by weighted summing of the local mean of the relative height deviation of the sliding window and the local standard deviation of the relative height deviation of the sliding window. If the anomaly determination coefficient is less than or equal to the standard value of the anomaly determination coefficient, then all measurement points within the corresponding sliding window will be recorded as minor anomalies. For all marked minor outliers, a 4-neighborhood connectivity analysis was performed to aggregate consecutively adjacent minor outliers into minor outlier regions.

4. The method for detecting and correcting center offset during wafer transfer according to claim 3, characterized in that: The process for assessing the risk of sudden warping in areas of slight warping is as follows: By coupling analysis of the vulnerability of static defects in slightly warped regions with the intensity of external disturbances, the vulnerability coefficient and the disturbance intensity coefficient of slightly warped regions are determined. The vulnerability coefficient of the slightly warped region is multiplied with the interference intensity coefficient of the slightly warped region to obtain the sudden risk value of the slightly warped region. If the sudden risk value of a slightly warped region is greater than or equal to the sudden risk value threshold, it indicates that the sudden warping risk of the slightly warped region is high, and the corresponding slightly warped region is recorded as a coupled warped region; otherwise, it indicates that the sudden warping risk of the slightly warped region is low.

5. The method for detecting and correcting center offset during wafer transfer according to claim 4, characterized in that: The process for determining the fragility coefficient of the slightly warped region is as follows: A monitoring period was set, and Raman spectroscopy was used to scan all slightly warped areas to obtain stress distribution cloud maps of all slightly warped areas; Based on any slightly warped region, the stress values ​​corresponding to the stress distribution cloud map of the slightly warped region are summed and averaged to obtain the static residual stress of the slightly warped region. Introducing the yield strength of the wafer material, the static residual stress of the slightly warped region is compared with the yield strength of the wafer material to obtain the fragility coefficient of the slightly warped region.

6. The method for detecting and correcting center offset during wafer transfer according to claim 4, characterized in that: The process for determining the interference intensity coefficient of the slightly warped region is as follows: The additional stress generated by the slight abnormal wafer transfer process on the slight warping region is obtained, and the ratio of this stress to the static residual stress in the slight warping region is processed to obtain the interference intensity coefficient of the slight warping region.

7. The method for detecting and correcting center offset during wafer transfer according to claim 6, characterized in that: The process of determining whether the coupled warped region exhibits periodic wavy warping in spatial location is as follows: Frequency domain analysis was used to extract the spatial periodic features of the coupled warped region in the slightly warped region, and periodic wave-shaped morphology was fitted to determine the goodness of fit of the measurement points. If the goodness of fit of the measurement point is greater than or equal to the goodness of fit threshold, it indicates that the coupled warped region exhibits periodic wavy warping in spatial location; if the goodness of fit of the measurement point is less than the goodness of fit threshold, it indicates that the coupled warped region does not exhibit periodic wavy warping in spatial location.

8. The method for detecting and correcting center offset during wafer transfer according to claim 7, characterized in that: The process for determining the goodness of fit of the measurement points is as follows: The coordinates of all measurement points in the coupled warped region are extracted to construct a spatial signal matrix. A two-dimensional fast Fourier transform is performed on the spatial signal matrix to obtain a frequency domain signal matrix. The power spectral density of the frequency domain signal is calculated to generate a power spectral density map. The peak points of the power values ​​are extracted from the power spectral density map. The radius of the neighborhood around the peak point is set. All power values ​​within the radius of the neighborhood around the peak point are extracted, summed, and averaged to obtain the average power of the neighborhood around the peak point. The standard deviation formula is used to calculate the standard deviation of all power values ​​within the radius of the neighborhood around the peak point. The signal-to-noise ratio of the peak point is then calculated. If the signal-to-noise ratio of a peak point is greater than or equal to the standard signal-to-noise ratio value, then the corresponding peak point is recorded as a significant peak point; the reciprocal of the frequency in the X-axis direction corresponding to the significant peak point is used to obtain the candidate period in the X-axis direction of the significant peak point, and the reciprocal of the frequency in the Y-axis direction corresponding to the significant peak point is used to obtain the candidate period in the Y-axis direction of the significant peak point. A two-dimensional sine fitting function is defined, and the least squares method is used to fit the coordinates of all measurement points in the measurement point set to the two-dimensional sine fitting function. The goodness of fit of the measurement points is calculated by the goodness of fit formula.

9. The method for detecting and correcting center offset during wafer transfer according to claim 8, characterized in that: The process for evaluating the impact of slightly warped wafers on center offset detection during transport is as follows: The wafer physical boundary is determined by the wafer diameter. Measurement points within a set grid spacing at a distance from the wafer physical boundary are extracted to form an edge point set. Measurement points in the edge point set that coincide with the coupled warped region are recorded as coupled edge points, and measurement points in the edge point set that coincide with the uncoupled warped region are recorded as uncoupled edge points. The uncoupled edge points are fitted using the least squares circle fitting algorithm to minimize the sum of squared radial errors from all uncoupled edge points to the circle, thereby solving for the optimal reference center and reference radius and establishing a reference circle. The center of the circle after directly fitting all the coupling edge points to a circle is taken as the center of disturbance, and the disturbance error between the center of disturbance and the optimal reference center is calculated by the Euclidean distance formula. If the interference error is greater than or equal to the interference error threshold, it means that the slight warping of the wafer has a significant impact on the center offset detection during the transfer process, and the fluctuation of the coupling edge point needs to be corrected. If the interference error is less than the interference error threshold, the slight warping of the wafer has a small impact on the center offset detection during the transfer process, and the fluctuation of the coupling edge point can be ignored. The reference center can be directly used as the transfer center.

10. The method for detecting and correcting center offset during wafer transfer according to claim 9, characterized in that: The process for determining the final transfer center and final edge radius of a slightly warped wafer is as follows: Since slight warping of the wafer has a significant impact on center offset detection during the transfer process, a ripple radial conversion coefficient is introduced for each coupling edge point. The radial ripple offset corresponding to the relative height deviation of the coupling edge point is calculated. Combined with the phase of the X-axis and the Y-axis corresponding to the relative height deviation of the coupling edge point, it is determined whether the coupling edge point is at the peak, trough or middle of the periodic wavy warping, and the coupling edge point is corrected accordingly. The corrected coupled edge points and uncoupled edge points are integrated to form a corrected full edge point set. Least square circle fitting is then performed on the full edge point set to solve for the final transmission center and the final edge radius.

Images