Detection method, device and equipment for wafer edge polishing

Abstract

The invention discloses a detection method, device and equipment for wafer edge polishing. The detection method comprises the following steps: positioning a wafer to be detected; obtaining plane part morphology data and curved surface morphology data of the positioned wafer to be detected; performing curve fitting on the plane part morphology data and the curved surface part morphology data to obtain a first curve and a second curve; performing feature extraction and feature point matching on the curve; calculating a homography transformation matrix between the two curves based on the matched feature point pairs; determining an overlapping area of the first curve and the second curve; determining an optimal stitching point; and butting the first curve and the second curve according to the optimal stitching point, and taking the butted curve as an edge detection curve of the wafer to be detected. Through the integrated design of the functions of centering positioning, height compensation, curve fitting, morphology observation and the like of the wafer, the data synchronism of multi-parameter measurement is ensured, and the accuracy and reliability of a measurement result are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of wafer detection technology, and in particular to a detection method, device and equipment for wafer edge polishing. Background Art

[0002] During wafer inspection, the mapping of the wafer is mainly concentrated on its internal area, while the mapping of the edge area is often ignored or not processed finely enough.

[0003] For edge area inspection, existing technologies often use thickness sensors and edge sensors to measure the horizontal and arc shapes of the wafer. These sensors can provide the edge profile of the wafer to help detect and analyze wafer quality issues.

[0004] However, since the thickness sensor and edge sensor can only detect the horizontal surface and arc part of the wafer edge respectively, the data detected by the thickness sensor and the edge sensor, and the detection of the wafer edge require the overall monitoring data of the wafer edge to evaluate the detection results. The existing method will choose to directly merge the data of the horizontal surface of the wafer edge and the data of the arc part together, but this simple merging will result in low accuracy at the data fusion point, and fail to fully consider the overall continuity and local morphological characteristics of the curve, so there will be obvious deviations or non-smooth transitions;

[0005] Therefore, existing methods generally lack accurate mapping methods for the edge of the wafer, resulting in the inability to effectively identify and control defects in the edge area and the amount of polishing during wafer processing and application. Summary of the invention

[0006] Therefore, the object of the present invention is to provide a detection method, device and equipment for wafer edge polishing, so as to achieve the optimal smooth docking of fitting curves between different morphological areas at the stitching points in wafer edge morphology mapping.

[0007] In order to achieve the above object, the present invention provides a method for detecting wafer edge polishing, wherein the wafer includes a flat portion and a curved portion at the edge, and comprises the following steps:

[0008] S1, obtaining the planar surface morphology data and the curved surface morphology data of the wafer to be inspected, wherein the planar surface morphology data includes the data of a straight line segment along the radial direction on the planar surface of the wafer, and the curved surface morphology data includes the data of a curve segment of the curved surface of the wafer at a radial cross section;

[0009] S2, performing curve fitting on the planar surface topography data and the curved surface topography data respectively to obtain a first curve and a second curve;

[0010] S3, extracting feature points from the first curve and the second curve respectively, and matching the feature points;

[0011] S4, based on the geometric correspondence between the matched feature point pairs, transforming the geometric position of one of the first curve and the second curve so that two feature points in each matched feature point pair overlap;

[0012] S5, screening the points in the overlapped area of ​​the first curve and the second curve after transformation based on the curvature characteristics to determine the best suture point;

[0013] S6. Butt the first curve and the second curve at the optimal stitching point, and use the butted curve as the edge detection curve of the wafer to be detected.

[0014] Further preferably, in S4, the geometric correspondence is obtained by calculating a homography transformation matrix between the matched feature point pairs.

[0015] Further preferably, the homography transformation matrix is ​​calculated using the following formula:

[0016]

[0017] Among them, M is the set of matched feature point pairs, p 1i is the characteristic point located on the first curve, p 2j is the characteristic point located on the second curve; p 1i and p 2j is a set of matching feature point pairs.

[0018] Further preferably, in S5, the points in the overlapped area are screened by constructing a cost function, wherein the cost function is configured to evaluate the quality of the points in the overlapped area as suture points based on curvature features.

[0019] Further, the cost function is constructed according to the following formula:

[0020]

[0021] Among them, R is the set of points in the overlapped area, p k is a point in R, ω k is assigned to p k The weight is set as a function of the curvature of the point or the change in curvature, d(p k， p) is point p k The distance to the stitching point p.

[0022] Furthermore, weights are set according to the curvature or curvature change rate of the points in the overlapping area, and the suture point when the cost function is minimized is the optimal suture point.

[0023] Further preferably, in S2, curve fitting is performed on the planar surface morphology data and the curved surface morphology data respectively to obtain the first curve and the second curve, which comprises:

[0024] Performing curve fitting on the planar surface topography data and the curved surface topography data obtained on the upper surface of the wafer to be inspected to obtain a first upper surface curve and a second upper surface curve, and stitching and docking the first upper surface curve and the second upper surface curve according to the steps described in S3-S6;

[0025] And / or curve fitting is performed on the planar surface morphology data and the curved surface morphology data obtained from the lower surface of the wafer to be inspected to obtain a first lower surface curve and a second lower surface curve; the first lower surface curve and the second lower surface curve are stitched and docked according to the steps described in S3-S6.

[0026] Further preferably, before S1, the step further includes S0: positioning the wafer to be inspected, and step S0 includes centering the wafer to be inspected and height positioning the wafer to be inspected.

[0027] Further preferably, centering the wafer to be inspected includes:

[0028] Placing a wafer to be inspected on a position to be inspected, and measuring the eccentricity of the wafer to be inspected;

[0029] The center position of the wafer to be inspected is adjusted according to the obtained eccentricity.

[0030] Further, the measuring of the eccentricity of the wafer to be inspected comprises the following steps:

[0031] The eccentric sensor is used to measure the amount of light leakage after the detection light emitted by the sensor passes through the wafer to be detected;

[0032] Determine the edge point of the wafer to be inspected according to the light leakage amount, and determine the position of the actual center of the wafer to be inspected according to the edge point;

[0033] The eccentricity is calculated according to the distance between the predetermined circle center and the actual circle center.

[0034] Further, the height positioning of the wafer to be inspected includes:

[0035] The upper thickness sensor is used to measure the distance A from the laser to the upper surface of the wafer to be inspected;

[0036] A lower thickness sensor is used to measure the distance B from the laser to the lower surface of the wafer to be inspected;

[0037] The height position of the wafer to be inspected is adjusted so that A and B are equal.

[0038] The present invention also provides a detection device for wafer edge polishing, comprising:

[0039] A data acquisition module is used to acquire the planar surface morphology data and the curved surface morphology data of the wafer to be inspected; the planar surface morphology data includes the data of a straight line segment along the radial direction on the planar surface of the wafer, and the curved surface morphology data includes the data of a curve segment at a radial cross section of the curved surface of the wafer;

[0040] A data processing module performs curve fitting on the planar surface morphology data and the curved surface morphology data respectively to obtain a first curve and a second curve; extracts feature points from the first curve and the second curve respectively; and matches the feature points; based on the geometric correspondence between the matched feature point pairs, transforms the geometric position of one of the first curve and the second curve so that the two feature points in each matched feature point pair overlap; for the points in the overlapped area of ​​the transformed first curve and the second curve, based on the curvature characteristics, screens to determine the best stitching point; docks the first curve and the second curve at the best stitching point, and uses the docked curve as the edge detection curve of the wafer to be detected.

[0041] Further preferably, in the data processing module, the geometric correspondence is obtained by calculating a homography transformation matrix between the matched feature point pairs; the homography transformation matrix is ​​calculated using the following formula:

[0042]

[0043] Among them, M is the set of matched feature point pairs, p 1i is the characteristic point located on the first curve, p 2j is the characteristic point located on the second curve; p 1i and p 2j is a set of matching feature point pairs.

[0044] Further preferably, the points in the overlapped area are screened by constructing a cost function, wherein the cost function is set to evaluate the quality of the points in the overlapped area as suture points based on curvature characteristics;

[0045] The cost function is calculated according to the following formula:

[0046]

[0047] Among them, R is the set of points in the overlap area, p k is a point in R, ω k is assigned to p k The weight is set as a function of the curvature of the point or the change in curvature, d(p k， p) is point p k The distance to the stitching point p.

[0048] Further preferably, it also includes a wafer positioning module, which is used to position the wafer to be inspected;

[0049] The wafer positioning module includes an eccentricity sensor, a four-axis motion platform, and a positioning judgment unit;

[0050] The four-axis motion platform is used to drive the wafer to be inspected located at the inspection position to perform linear motion in three directions of X, Y, and Z, and rotate during the motion;

[0051] The eccentricity sensor is used to obtain the amount of light leakage when the wafer to be detected rotates;

[0052] The positioning judgment unit measures the eccentricity of the wafer to be detected according to the amount of light leakage, adjusts the position of the wafer to be detected according to the obtained eccentricity, and obtains the thickness value of the wafer to be detected;

[0053] The position of the neutral plane is determined based on the obtained thickness value, and height compensation calibration is performed to position the wafer to be inspected.

[0054] Further preferably, the data acquisition module includes a thickness sensor; the thickness sensor is used to measure the planar surface morphology data of the wafer to be inspected; and includes: an upper thickness sensor and a lower thickness sensor;

[0055] The upper thickness sensor measures the distance A from the laser to the upper surface of the wafer to be inspected;

[0056] The lower thickness sensor measures the distance B from the laser to the lower surface of the wafer to be inspected;

[0057] The height position of the wafer to be inspected is adjusted so that A and B are equal, and height compensation calibration is performed.

[0058] Further, the data acquisition unit further includes an edge sensor, the edge sensor is used to measure the curved surface morphology data of the wafer to be detected, and the edge sensor includes an upper edge sensor and a lower edge sensor;

[0059] The upper edge sensor is used to collect the curved surface topography data of the upper surface during the rise compensation calibration process of the wafer to be inspected;

[0060] The lower edge sensor is used to collect the curved surface morphology data of the lower surface during the descent compensation calibration process of the wafer to be inspected.

[0061] Furthermore, curve fitting is performed on the planar surface morphology data and the curved surface morphology data respectively, and the first curve and the second curve obtained include:

[0062] Performing curve fitting on the planar surface morphology data and the curved surface morphology data obtained on the upper surface of the wafer to be inspected to obtain a first upper surface curve and a second upper surface curve, and stitching and docking the first upper surface curve and the second upper surface curve;

[0063] And / or curve fitting is performed on the planar surface morphology data and the curved surface morphology data obtained from the lower surface of the wafer to be inspected to obtain a first lower surface curve and a second lower surface curve; and the first lower surface curve and the second lower surface curve are stitched and butted.

[0064] Further preferably, it also includes a camera and a compensation light source; the camera is used to obtain pictures of the edge and each position of the positioning port of the wafer to be inspected.

[0065] The present invention also provides a detection device for wafer edge polishing, the device comprising

[0066] a memory storing computer program instructions;

[0067] A processor, when the computer program instructions are executed by the processor, implements the steps of the above-mentioned detection method for wafer edge polishing.

[0068] The wafer edge polishing detection method disclosed in the present application adjusts the horizontal position of the wafer by collecting the eccentricity, and adjusts the vertical position of the wafer by using a thickness sensor. After the adjustment, the wafer is in the optimal measurement position both in the horizontal position and the vertical position. At this time, the horizontal data and the curve data are scanned again, and the cross position is identified after the scanned data are fitted. The horizontal data and the curve data are stitched using the obtained optimal stitching point to ensure that the curve is smooth at the data fusion point; and the optimal smooth docking of the fitting curves between different morphological areas at the stitching point in the wafer edge morphology mapping is achieved.

[0069] The integrated design of wafer centering, height compensation, curve fitting and morphology observation ensures data synchronization of multi-parameter measurements and enhances the accuracy and reliability of measurement results.

[0070] By accurately mapping the edge of the wafer, the amount of wafer polishing can be quantified. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a schematic diagram of the flow of the detection method for wafer edge polishing of the present invention.

[0072] Figure 2 It is a schematic diagram of the region for obtaining the planar surface morphology data and the curved surface surface morphology data in the present invention.

[0073] Figure 3 It is the principle diagram of the eccentricity sensor in the present invention.

[0074] Figure 4 It is the principle diagram of eccentricity calculation in the present invention.

[0075] Figure 5 It is a schematic structural diagram of a wafer edge polishing detection device in the present invention.

[0076] Figure 6 It is a structural schematic diagram of the four-axis motion platform of the present invention.

[0077] Figure 7 This is a schematic diagram of the installation position of the eccentricity sensor of the present invention.

[0078] Figure 8 It is a schematic structural diagram of the eccentricity sensor of the present invention.

[0079] Fig. 9 It is a schematic diagram of the installation structure of the thickness measuring sensor in the present invention.

[0080] Fig.10 It is a schematic diagram of the structure of the edge sensor in the present invention.

[0081] Fig.11 It is a schematic diagram of the installation structure of the camera and the light source in the present invention.

[0082] In the figure:

[0083] 1. Four-axis motion platform; 1a. X-axis motion table; 1b. Y-axis motion table; 1c. Z-axis motion table; 1d. T-axis motion table; 2. Edge sensor, 201, upper sensor; 202, lower sensor; 3. Thickness sensor; 301, upper thickness sensor; 302, lower thickness sensor; 4. Eccentricity sensor; 401: light emitting device; 402: light receiving device; 5. Camera; 6. Light source; 7. Suction cup; W, wafer. DETAILED DESCRIPTION

[0084] The present invention is further described in detail below through the accompanying drawings and specific embodiments.

[0085] like Figure 1 As shown, the present invention also provides a method for detecting wafer edge polishing, comprising the following steps:

[0086] S1, obtaining the planar surface morphology data and the curved surface morphology data of the wafer to be inspected, wherein the planar surface morphology data includes the data of a straight line segment along the radial direction on the planar surface of the wafer, and the curved surface morphology data includes the data of a curve segment of the curved surface of the wafer at a radial cross section; in S1, Figure 2As shown, when the wafer is placed flat at the inspection position, the planar portion morphology data is the data of a straight line segment along the radial direction on the planar portion of the wafer, specifically the first plane and the second plane that are opposite to each other in the upper and lower directions of the wafer thickness, and the curved portion morphology data is the convex surface connecting the first plane and the second plane on the left and right sides of the wafer.

[0087] S2, performing curve fitting on the planar surface morphology data and the curved surface morphology data respectively to obtain a first curve and a second curve; comprising:

[0088] The planar part morphology data and the curved part morphology data obtained on the upper surface of the wafer to be inspected are curve fitted to obtain a first upper surface curve and a second upper surface curve; and / or the planar part morphology data and the curved part morphology data obtained on the lower surface of the wafer to be inspected are curve fitted to obtain a first lower surface curve and a second lower surface curve.

[0089] S3, extracting feature points from the first curve and the second curve respectively, and matching the feature points; wherein the feature points of the first curve may be points extracted from the end region of the first curve close to the plane portion; the feature points of the second curve may be points extracted from the end region of the second curve close to the edge of the wafer. Therefore, the extracted feature points must include points that appear simultaneously in the first curve and the second curve; therefore, when matching feature points, points that appear simultaneously in the first curve and the second curve are actually matched.

[0090] S4, based on the geometric correspondence between the matched feature point pairs, transforming the geometric position of one of the first curve and the second curve so that two feature points in each matched feature point pair overlap;

[0091] Feature point extraction and matching: Let P1 = {p 1i} and P2={p 2j} are the feature point sets on the curves L1 and L2, respectively, where i = 1, 2, ..., n and j = 1, 2, ..., m. By using the feature point extraction algorithm, we can get the feature point pairs (p 1i ,p 2j )

[0092] Furthermore, the geometric correspondence is obtained by calculating the homography transformation matrix between the matched feature point pairs; the homography transformation matrix is ​​calculated using the following formula:

[0093]

[0094] Among them, M is the set of matched feature point pairs, p 1i is the characteristic point located on the first curve, p 2j is the characteristic point located on the second curve; p 1iand p 2j is a set of matching feature point pairs. argmin represents the variable value when the function reaches the minimum value, that is, the formula is used to obtain the matrix H that makes the feature points on the second curve closest to the corresponding feature points on the first curve after transformation.

[0095] S5. Screening the points in the overlapping area of ​​the transformed first curve and the second curve based on the curvature characteristics to determine the best stitching points; screening the points in the overlapping area by constructing a cost function, wherein the cost function is set to evaluate the advantages and disadvantages of the points in the overlapping area as stitching points based on the curvature characteristics.

[0096] The cost function is constructed according to the following formula:

[0097]

[0098] Among them, R is the set of points in the overlapped area, p k is a point in R, ω k is assigned to p k The weight is set as a function of the curvature of the point or the change in curvature, d(p k， p) is point p k The distance to the stitching point p.

[0099] The weight is set according to the curvature or curvature change rate of the point in the overlap area. The weight ω k The calculation takes into account the curvature k of the point κ or rate of change of curvature

[0100] Where f is a function that maps curvature and its rate of change to weight. The suture point when the cost function is minimized is the optimal suture point. Determination of the optimal suture point: The optimal suture point p* is the point that minimizes the cost function C(p):

[0101] p*=argmin p∈R C(p); use an optimization algorithm to find p*;

[0102] According to the calculated homography transformation matrix H, the curve L2 is transformed to achieve seamless connection with L1 at p*. The transformed point p′ 2j The calculation is as follows:

[0103] p′ 2j =H·p 2j

[0104] S6. Butt the first curve and the second curve at the optimal stitching point, and use the butted curve as the edge detection curve of the wafer to be detected.

[0105] It should be noted that before S1, S0 is also included: positioning the wafer to be inspected, including centering the wafer to be inspected and height positioning the wafer to be inspected.

[0106] Among them, the wafer to be inspected is centered and positioned, and the positioning process includes the following steps:

[0107] S01, placing a wafer to be inspected on a position to be inspected, and measuring the eccentricity of the wafer to be inspected; further, measuring the eccentricity of the wafer to be inspected includes:

[0108] The eccentric sensor is used to measure the amount of light leakage after the detection light emitted by the sensor passes through the wafer to be detected;

[0109] Determine the edge point of the wafer to be inspected according to the light leakage amount, and determine the position of the actual center of the wafer to be inspected according to the edge point;

[0110] The eccentricity is calculated according to the distance d between the predetermined circle center O1 and the actual circle center O2.

[0111] S02. Adjust the center position of the wafer to be inspected according to the obtained eccentricity.

[0112] Height positioning of the wafer to be inspected, including:

[0113] The upper thickness sensor is used to measure the distance A from the laser to the upper surface of the wafer to be inspected;

[0114] A lower thickness sensor is used to measure the distance B from the laser to the lower surface of the wafer to be inspected;

[0115] The height position of the wafer to be inspected is adjusted so that A and B are equal. The time of the light beam traveling back and forth inside the wafer to be inspected when the upper or lower thickness sensor is measuring is obtained to calculate the thickness t of the wafer;

[0116] According to t=C-(A+B), the position of the neutral plane is determined and height compensation calibration is performed.

[0117] like Figure 5 As shown, the present invention also provides a detection device for wafer edge polishing, which is used to implement the steps of the above-mentioned wafer edge polishing detection method, including:

[0118] A data acquisition module is used to acquire the planar surface morphology data and the curved surface morphology data of the wafer to be inspected; the planar surface morphology data includes the data of a straight line segment along the radial direction on the planar surface of the wafer, and the curved surface morphology data includes the data of a curve segment at a radial cross section of the curved surface of the wafer;

[0119] A data processing module performs curve fitting on the planar surface morphology data and the curved surface morphology data respectively to obtain a first curve and a second curve; extracts feature points from the first curve and the second curve respectively; and matches the feature points; based on the geometric correspondence between the matched feature point pairs, transforms the geometric position of one of the first curve and the second curve so that the two feature points in each matched feature point pair overlap; for the points in the overlapped area of ​​the transformed first curve and the second curve, based on the curvature characteristics, screens to determine the best stitching point; docks the first curve and the second curve at the best stitching point, and uses the docked curve as the edge detection curve of the wafer to be detected.

[0120] The data acquisition module includes a side thickness sensor 3 for acquiring the topographic data of the planar portion of the wafer to be inspected and an edge sensor 2 for acquiring the topographic data of the curved portion;

[0121] The thickness sensor includes: an upper thickness sensor 301 and a lower thickness sensor 302; when the wafer to be inspected rises from the Z1 position to the Z1+ΔZ (compensation), the upper thickness sensor located at the top collects data, and when the wafer to be inspected descends from the Z1 position to the Z1-ΔZ (compensation), the lower thickness sensor located at the bottom collects data;

[0122] Similarly, the edge sensor is used to measure the morphological data of the curved surface. The edge sensor is used to measure the morphological data of the curved surface of the wafer to be inspected, and the edge sensor includes an upper edge sensor and a lower edge sensor;

[0123] The upper edge sensor is used to collect the curved surface topography data of the upper surface during the rise compensation calibration process of the wafer to be inspected;

[0124] The lower edge sensor is used to collect the curved surface morphology data of the lower surface during the descent compensation calibration process of the wafer to be inspected.

[0125] According to the collected data, calculation is performed to fit the curve. Curve fitting is performed on the planar surface morphology data and the curved surface morphology data respectively to obtain the first curve and the second curve including:

[0126] Performing curve fitting on the planar surface topography data obtained on the upper surface of the wafer to be inspected to obtain a first upper surface curve and a second upper surface curve;

[0127] Alternatively, curve fitting is performed on the planar portion topography data obtained on the lower surface of the wafer to be inspected to obtain a lower surface first curve and a lower surface second curve.

[0128] like Figure 5 Shown to Figure 6The four-axis motion platform 1 and the suction cup 7 are shown. When measuring in the measuring unit, the four-axis motion platform 1 and the suction cup 7 are used to rotate and fix the wafer W to be tested. Specifically, the four-axis motion platform 1 includes tables 1a, 1b, 1c and 1d that can act independently or in coordination, wherein 1a is an X-axis motion table, 1b is a Y-axis motion table, 1c is a Z-axis motion table and 1d is a T-axis motion table; these tables can be precisely moved in predetermined X, Y, Z-axis and T (rotation) directions. During the measurement process, the suction cup 7 firmly holds the wafer W to be tested through a precisely controlled adsorption force, ensuring that the wafer W to be tested remains stationary under the operation of the four-axis motion platform 1. On this basis, the four-axis motion platform 1 performs movement and rotation actions according to the control instructions to accurately position the edge portion to be tested of the wafer W to be tested to the measurement position. Edge sensor 2, thickness sensor 3 and eccentricity sensor 4. The three sensor mechanisms are positioned around the wafer W. It also includes a camera 5 and a light source 6 for observing the surface state of the edge portion of the wafer W to be inspected, so as to detect the polishing morphology of the wafer edge portion and the Notch (wafer notch).

[0129] It also includes a wafer positioning module for positioning the wafer to be inspected; the wafer positioning module includes an eccentricity sensor, a four-axis motion platform, and a positioning judgment unit;

[0130] The four-axis motion platform is used to drive the wafer to be inspected located at the inspection position to perform linear motion in three directions of X, Y, and Z, and rotate during the motion;

[0131] The eccentricity sensor is used to obtain the amount of light leakage when the wafer to be detected rotates;

[0132] The positioning judgment unit performs centering and height positioning on the wafer to be inspected.

[0133] When the wafer to be inspected is aligned and positioned, it includes: using an eccentric sensor to measure the amount of light leakage after the inspection light emitted by the sensor passes through the wafer to be inspected;

[0134] Determine the edge point of the wafer to be inspected according to the light leakage amount, and determine the position of the actual center of the wafer to be inspected according to the edge point;

[0135] The eccentricity is calculated according to the distance d between the predetermined circle center O1 and the actual circle center O2.

[0136] like Figure 3As shown, the eccentricity sensor 4 is a high-precision optical sensor that emits a beam of light to the edge of the wafer. When the light hits the edge of the wafer, ideally all the light will be absorbed by the wafer or reflected back to the sensor. If the wafer position deviates from the predetermined center of the circle, light will "leak" out of this edge area. The sensor receives the reflected light and measures the amount of light that leaks out. The amount of light leakage is related to the distance from the edge point to the predetermined center of the circle.

[0137] The eccentricity sensor 4 continuously measures the light leakage during the wafer rotation process, thereby obtaining a series of distance data from the edge points to the predetermined center of the circle. The actual center of the wafer can be calculated from these data. Figure 4 As shown, the actual center of the wafer is compared with the theoretical center, and the difference between the two is the eccentricity. The actual center position corresponding to each point on the actual circle is different, so the eccentricity corresponding to each point is different. When the wafer rotates, compensation is always performed according to the eccentricity. The compensation amount can be automatically given by the system and realized by moving the four-axis motion platform.

[0138] like Figure 7 and Figure 8 As shown, the eccentricity sensor 4 in the present invention includes a light emitting device 401 and a light receiving device 402, and the light emitting device 401 and the light receiving device 402 are installed in the tangent direction of the side of the wafer W, so that the edge portion of the wafer W is placed between the above two. The light emitting device 401 directs the laser beam to emit vertically downward, accurately aiming at the edge area of ​​the wafer W. The laser beam propagates in the radial direction of the wafer W and intersects with the wafer W in the edge area, causing part of the laser beam to be blocked by the edge of the wafer W. Therefore, only the laser beam that is not blocked can be captured by the light receiving device 402. The light receiving device 402 digitizes the received optical signal through the data processing unit to calculate the diameter parameter of the wafer W. During the measurement process, the four-axis motion platform 1 rotates and moves according to the predetermined control instructions to measure the diameter at multiple points on the edge portion of the wafer W, obtain continuous data of the wafer diameter, and thus calculate the eccentricity of the wafer.

[0139] When performing height positioning on the wafer to be inspected, it includes: determining the position of the neutral plane according to the obtained thickness value, and performing height compensation calibration to position the wafer to be inspected, specifically including:

[0140] The upper thickness sensor is used to measure the distance A from the laser to the upper surface of the wafer to be inspected;

[0141] A lower thickness sensor is used to measure the distance B from the laser to the lower surface of the wafer to be inspected;

[0142] Adjust the position of the wafer so that A=B; when adjusting the wafer, the thickness t of the wafer can be calculated by obtaining the round-trip propagation time of the light beam inside the wafer to be tested when the upper or lower thickness sensor is measuring;

[0143] According to t=C-(A+B), the position of the neutral plane is determined and height compensation calibration is performed.

[0144] like Fig. 9 As shown, after the centering step is performed, the point to be measured is rotated to align with the thickness sensor 3, and the thickness value at this time is recorded as a parameter for height compensation calibration. The upper thickness sensor 301 and the lower thickness sensor 302 are both independently operated units, each responsible for the measurement task of the upper surface and the lower surface of the wafer W. The upper thickness sensor 301 and the lower thickness sensor 302 both use laser beam focusing technology to accurately focus the laser beam on the surface of the wafer W. The thickness of the wafer is accurately calculated by measuring the time it takes for the laser beam to propagate back and forth inside the wafer W. For example, by measuring the distance between the upper and lower surfaces of the wafer W and the thickness sensor 3, the position of the neutral plane is determined, and the height compensation calibration is completed to ensure the positioning accuracy of the wafer W. During the calibration process, A=B is set to ensure that the distance from the upper and lower surfaces to the thickness sensor 3 is equal, thereby improving the calibration efficiency and ensuring the repeatability of the positioning accuracy of the wafer W during the measurement process.

[0145] like Fig.10 As shown, in order to improve the measurement accuracy and overcome the problem of measurement range limitation of a single sensor, the present application sets an upper sensor 201 and a lower sensor 202. The upper sensor 201 and the lower sensor 202 in the edge sensor 2 have the same function as the thickness sensor. Sensor 201 is arranged above the edge portion of the wafer W to be detected, and sensor 202 is arranged below the edge portion of the wafer W. In the process of the wafer W to be detected rising from the Z1 position to Z1+ΔZ (compensation), the upper edge data is collected, and according to the collected data, calculations are performed to fit the curve to obtain the edge features of the upper surface of the wafer W to be detected; in the process of the wafer W to be detected descending from the Z2 position to Z2+ΔZ, the lower edge data is collected, and similarly, according to the collected data, calculations are performed to fit the curve to obtain the edge features of the lower surface of the wafer W to be detected.

[0146] like Fig.11 As shown, it also includes a camera and a compensation light source; the camera is used to obtain images of the edge and each position of the positioning port of the wafer to be inspected.

[0147] In the data processing module, the step of calculating the homography transformation matrix between the two curves based on the matched feature point pairs includes: calculating the homography transformation matrix between the two curves according to the following formula:

[0148]

[0149] Among them, M is the set of matched feature point pairs, p 1i is the characteristic point located on the first curve, p 2j is the characteristic point located on the second curve; p 1i and p 2j is a set of matching feature point pairs.

[0150] The cost function is constructed according to the following formula:

[0151]

[0152] Among them, R is the set of points in the overlapped area, p k is a point in R, ω k is assigned to p k The weight of d(p k， p) is point p k The distance to the stitching point p.

[0153] Constructing a cost function in the overlapped area to determine the best suture point includes: matching weights according to the curvature or curvature change rate of the suture point according to the cost function, and the suture point when the cost function is minimized is the best suture point.

[0154] The data processing module also includes: docking the first curve and the second curve according to the optimal stitching point, including transforming the first curve to the second curve using a homography transformation matrix according to the obtained optimal stitching point, so as to achieve seamless docking of the first curve and the second curve at the optimal stitching point.

[0155] The present invention also provides a detection device for wafer edge polishing, which includes: a memory storing computer program instructions; a processor, which implements the steps of the above-mentioned detection method for wafer edge polishing when the computer program instructions are executed by the processor.

[0156] Obviously, the above embodiments are only examples for clear explanation, and are not intended to limit the implementation methods. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from them are still within the protection scope of the invention.

Claims

1. A method for detecting wafer edge polishing, wherein the wafer comprises a flat surface portion and a curved surface portion at the edge, characterized in that: The following steps are involved: S1, obtaining the planar surface morphology data and the curved surface morphology data of the wafer to be inspected, wherein the planar surface morphology data includes the data of a straight line segment along the radial direction on the planar surface of the wafer, and the curved surface morphology data includes the data of a curve segment of the curved surface of the wafer at a radial cross section; S2, performing curve fitting on the planar surface topography data and the curved surface topography data respectively to obtain a first curve and a second curve; S3, extracting feature points from the first curve and the second curve respectively, and matching the feature points; S4, based on the geometric correspondence between the matched feature point pairs, transforming the geometric position of one of the first curve and the second curve so that two feature points in each matched feature point pair overlap; S5, screening the points in the overlapped area of ​​the first curve and the second curve after transformation based on the curvature characteristics to determine the best suture point; S6. Butt the first curve and the second curve at the optimal stitching point, and use the butted curve as the edge detection curve of the wafer to be detected.

2. The detection method for wafer edge polishing according to claim 1, characterized in that: In S4, the geometric correspondence is obtained by calculating the homography transformation matrix between the matched feature point pairs.

3. The detection method for wafer edge polishing according to claim 2, characterized in that: The homography transformation matrix is ​​calculated using the following formula: Among them, M is the set of matched feature point pairs, p 1i is the characteristic point located on the first curve, p 2j is the characteristic point located on the second curve; p 1i and p 2j is a set of matching feature point pairs.

4. The detection method for wafer edge polishing according to any one of claims 1 to 3, characterized in that: In S5, the points in the overlapped area are screened by constructing a cost function, wherein the cost function is configured to evaluate the quality of the points in the overlapped area as suture points based on curvature features.

5. The detection method for wafer edge polishing according to claim 4, characterized in that: The cost function is constructed according to the following formula: Among them, R is the set of points in the overlapped area, p k is a point in R, ω k is assigned to p k The weight is set as a function of the curvature of the point or the change in curvature, d(p k， p) is point p k The distance to the stitching point p.

6. The detection method for wafer edge polishing according to claim 5, characterized in that: The weight is set according to the curvature or curvature change rate of the point in the overlapping area. The suture point when the cost function is minimized is the optimal suture point.

7. The detection method for wafer edge polishing according to claim 1, characterized in that: In S2, curve fitting is performed on the planar surface morphology data and the curved surface morphology data to obtain the first curve and the second curve respectively, including: Performing curve fitting on the planar surface topography data and the curved surface topography data obtained on the upper surface of the wafer to be inspected to obtain a first upper surface curve and a second upper surface curve, and stitching and docking the first upper surface curve and the second upper surface curve according to the steps described in S3-S6; And / or curve fitting is performed on the planar surface morphology data and the curved surface morphology data obtained from the lower surface of the wafer to be inspected to obtain a first lower surface curve and a second lower surface curve; the first lower surface curve and the second lower surface curve are stitched and docked according to the steps described in S3-S6.

8. The detection method for wafer edge polishing according to any one of claims 1 to 3, characterized in that: Before S1, the method further includes S0: positioning the wafer to be inspected. Step S0 includes centering the wafer to be inspected and height positioning the wafer to be inspected.

9. The detection method for wafer edge polishing according to claim 8, characterized in that: Centering the wafer to be inspected includes: Placing a wafer to be inspected on a position to be inspected, and measuring the eccentricity of the wafer to be inspected; The center position of the wafer to be inspected is adjusted according to the obtained eccentricity.

10. The detection method for wafer edge polishing according to claim 9, characterized in that: The method of measuring the eccentricity of the wafer to be inspected comprises the following steps: The eccentric sensor is used to measure the amount of light leakage after the detection light emitted by the sensor passes through the wafer to be detected; Determine the edge point of the wafer to be inspected according to the light leakage amount, and determine the position of the actual center of the wafer to be inspected according to the edge point; The eccentricity is calculated according to the distance between the predetermined circle center and the actual circle center.

11. The detection method for wafer edge polishing according to claim 8, characterized in that: The height positioning of the wafer to be inspected comprises: The upper thickness sensor is used to measure the distance A from the laser to the upper surface of the wafer to be inspected; A lower thickness sensor is used to measure the distance B from the laser to the lower surface of the wafer to be inspected; The height position of the wafer to be inspected is adjusted so that A and B are equal.

12. A detection device for wafer edge polishing, characterized in that: include: A data acquisition module is used to acquire the planar surface morphology data and the curved surface morphology data of the wafer to be inspected; the planar surface morphology data includes the data of a straight line segment along the radial direction on the planar surface of the wafer, and the curved surface morphology data includes the data of a curve segment at a radial cross section of the curved surface of the wafer; The data processing module performs curve fitting on the planar surface morphology data and the curved surface morphology data to obtain a first curve and a second curve; extracts feature points from the first curve and the second curve respectively; and matches the feature points; based on the geometric correspondence between the matched feature point pairs, transforms the geometric position of one of the first curve and the second curve so that two feature points in each matched feature point pair overlap; Points within the overlapping area of ​​the transformed first curve and the second curve are screened based on curvature characteristics to determine the best stitching point; the first curve and the second curve are docked at the best stitching point, and the docked curve is used as the edge detection curve of the wafer to be detected.

13. The detection device for wafer edge polishing according to claim 12, characterized in that: In the data processing module, the geometric correspondence is obtained by calculating the homography transformation matrix between the matched feature point pairs; the homography transformation matrix is ​​calculated using the following formula: Among them, M is the set of matched feature point pairs, p 1i is the characteristic point located on the first curve, p 2j is the characteristic point located on the second curve; p 1i and p 2j is a set of matching feature point pairs.

14. The detection device for wafer edge polishing according to claim 12, characterized in that: Screening the points in the overlapped area by constructing a cost function, wherein the cost function is set to evaluate the quality of the points in the overlapped area as suture points based on curvature characteristics; The cost function is calculated according to the following formula: Among them, R is the set of points in the overlapped area, p k is a point in R, ω k is assigned to p k The weight is set as a function of the curvature of the point or the change in curvature, d(p k， p) is point p k The distance to the stitching point p.

15. The detection device for wafer edge polishing according to any one of claims 12 to 14, characterized in that: It also includes a wafer positioning module for positioning the wafer to be inspected; The wafer positioning module includes an eccentricity sensor, a four-axis motion platform, and a positioning judgment unit; The four-axis motion platform is used to drive the wafer to be inspected located at the inspection position to perform linear motion in three directions of X, Y, and Z, and rotate during the motion; The eccentricity sensor is used to obtain the amount of light leakage when the wafer to be detected rotates; The positioning judgment unit measures the eccentricity of the wafer to be detected according to the amount of light leakage, adjusts the position of the wafer to be detected according to the obtained eccentricity, and obtains the thickness value of the wafer to be detected; The position of the neutral plane is determined based on the obtained thickness value, and height compensation calibration is performed to position the wafer to be inspected.

16. The detection device for wafer edge polishing according to any one of claims 12 to 14, characterized in that: The data acquisition module includes a thickness sensor; the thickness sensor is used to measure the planar surface morphology data of the wafer to be inspected; it includes: an upper thickness sensor and a lower thickness sensor; The upper thickness sensor measures the distance A from the laser to the upper surface of the wafer to be inspected; The lower thickness sensor measures the distance B from the laser to the lower surface of the wafer to be inspected; The height position of the wafer to be inspected is adjusted so that A and B are equal, and height compensation calibration is performed.

17. The detection device for wafer edge polishing according to claim 16, characterized in that: The data acquisition unit further includes an edge sensor, which is used to measure the curved surface morphology data of the wafer to be detected, and the edge sensor includes an upper edge sensor and a lower edge sensor; The upper edge sensor is used to collect the curved surface topography data of the upper surface during the rise compensation calibration process of the wafer to be inspected; The lower edge sensor is used to collect the curved surface morphology data of the lower surface during the descent compensation calibration process of the wafer to be inspected.

18. The detection device for wafer edge polishing according to claim 17, characterized in that: The planar surface morphology data and the curved surface morphology data are respectively subjected to curve fitting to obtain a first curve and a second curve including: Performing curve fitting on the planar surface morphology data and the curved surface morphology data obtained on the upper surface of the wafer to be inspected to obtain a first upper surface curve and a second upper surface curve, and stitching and docking the first upper surface curve and the second upper surface curve; And / or curve fitting is performed on the planar surface morphology data and the curved surface morphology data obtained from the lower surface of the wafer to be inspected to obtain a first lower surface curve and a second lower surface curve; and the first lower surface curve and the second lower surface curve are stitched and butted.

19. The detection device for wafer edge polishing according to any one of claims 12 to 14, characterized in that: It also includes a camera and a compensation light source; the camera is used to obtain pictures of the edge and each position of the positioning port of the wafer to be inspected.

20. A detection device for wafer edge polishing, characterized in that: The device includes a memory storing computer program instructions; A processor, when the computer program instructions are executed by the processor, implements the steps of the detection method for wafer edge polishing as claimed in any one of claims 1 to 11.

Images