Edge alignment method

Abstract

This invention provides an edge alignment method for accurately estimating the position of the edge of a workpiece. The edge alignment method includes the following steps: a coordinate calculation step, in which the coordinates of a point that may correspond to the edge of the workpiece are calculated at various locations on the outer periphery of the workpiece in the circumferential direction; an approximate circle generation step, in which an approximate circle is generated by using the least squares method on all calculated coordinates; a false detection position elimination step, in which the offset between the generated approximate circle and each coordinate in all coordinates is calculated, and if there are coordinates with an offset greater than a pre-set threshold, the coordinate with the largest offset is determined as a false detection position and the coordinate determined as a false detection position is excluded from consideration; and a processing area inference step, inferring the position of the edge of the workpiece based on the three or more remaining coordinates that have not been excluded, and inferring the processing area based on the position of the estimated edge.

Description

Technical Field

[0001] The present invention relates to an edge alignment method for inferring the processing area of ​​a processing area where processing is performed on the outer periphery of a disc-shaped workpiece. Background Technology

[0002] On disk-shaped workpieces such as semiconductor wafers, chamfered portions are sometimes formed by chamfering the corners of the outer periphery on the front side and the corners of the outer periphery on the back side, respectively. When the back side of the workpiece with the chamfered portion is ground to a thickness of, for example, less than half, a sharp area (so-called a sharp edge) is formed on the outer periphery of the back side.

[0003] Cracks and gaps are easily formed at sharp edges. In addition, there is a concern that cracks may spread from cracks formed at sharp edges, causing damage to the workpiece. Therefore, in order to prevent the formation of sharp edges, edge trimming techniques are known to remove the chamfered portion of the front side by removing a predetermined thickness of the outer periphery of the front side before grinding the back side (for example, see Patent Document 1).

[0004] When performing edge trimming, a cutting device is typically used. The cutting device has a chuck table that attracts and holds the workpiece. A cutting unit is mounted on the chuck table. The cutting unit includes a cylindrical spindle arranged substantially parallel to the upper surface of the chuck table and a cutting tool mounted on one end of the spindle. Furthermore, a camera is provided in the cutting unit for photographing the workpiece held by the chuck table.

[0005] When performing edge trimming, edge alignment is performed first. During edge alignment, the back side of the workpiece is held by suction using a chuck table, and multiple locations on the outer periphery of the front side of the workpiece are photographed using a camera. Then, the coordinates of a point on the edge are determined at each location, and the position of the edge of the workpiece is estimated based on the coordinates of multiple points. The annular area extending from this edge to a predetermined range on the center side of the front side is removed by cutting.

[0006] To achieve high-precision edge trimming, the position of the workpiece's edge needs to be accurately estimated during edge alignment. However, the accuracy of edge position estimation is reduced due to various factors. For example, in cases where there are foreign objects attached to the chamfer, water droplets or foreign objects attached to the holding surface of the chuck table, or foreign objects attached to the camera lens, the position of the foreign object may sometimes be mistakenly assumed to be a point on the edge.

[0007] Patent Document 1: Japanese Patent Application Publication No. 2000-173961

[0008] When the location of a foreign object is mistakenly identified as a point on the edge, the position of the edge of the workpiece cannot be properly estimated, resulting in a significant shift in the area to be cut. Summary of the Invention

[0009] The present invention was made in view of the above-mentioned problems, and its object is to be able to estimate the position of the edge of the workpiece with high accuracy even if the position of the foreign object is mistakenly identified as a point on the edge.

[0010] According to one aspect of the present invention, an edge alignment method for a disc-shaped workpiece is provided, wherein the edge alignment method comprises the following steps: a holding step, wherein the workpiece is held using a chuck table; a coordinate calculation step, wherein the coordinates of a point that may correspond to the edge of the workpiece are calculated at various locations among different portions of the outer periphery of the workpiece in the circumferential direction of the workpiece; an approximate circle generation step, wherein an approximate circle is generated by applying the least squares method to all coordinates calculated in the coordinate calculation step; a false detection position exclusion step, wherein the offset between the approximate circle generated by the approximate circle generation step and each of the coordinates is calculated, and if there is a coordinate with an offset greater than or equal to a predetermined threshold, the coordinate with the largest offset is determined as a false detection position, and the coordinate determined as a false detection position is excluded from consideration; and a processing area inference step, wherein after the false detection position exclusion step, the position of the edge of the workpiece is estimated based on the three or more remaining coordinates that have not been excluded, and the processing area of ​​the outer periphery of the workpiece is inferred based on the estimated edge position.

[0011] The preferred edge alignment method also has the following additional approximate circle generation step: after the coordinates with the largest offset are eliminated through the false detection position elimination step, the least squares method is used on all the remaining coordinates that were not eliminated from the consideration in the false detection position elimination step, thereby generating an approximate circle again.

[0012] Furthermore, the preferred edge alignment method also includes the following additional false detection location elimination step: calculating the offset between the approximate circle generated by the additional approximate circle generation step and each coordinate among all the remaining coordinates that were not excluded from the consideration object in the false detection location elimination step; if there are coordinates with an offset greater than or equal to a preset threshold, the coordinate with the largest offset is determined as a false detection location, and the coordinate determined as a false detection location is excluded from the consideration object; calculating the offset between the approximate circle generated by the additional approximate circle generation step and each coordinate among all the remaining coordinates that were not excluded from the consideration object in the false detection location elimination step; if there are no coordinates with an offset greater than or equal to a preset threshold, the processing area inference step is performed.

[0013] In one aspect of the edge alignment method of the present invention, an approximate circle is generated by applying the least squares method to all coordinates calculated in the coordinate calculation step (approximate circle generation step). Next, the offset between the approximate circle generated by the approximate circle generation step and each coordinate in the coordinate system is calculated. If there are coordinates with an offset exceeding a predetermined threshold, the coordinate with the largest offset is identified as a false detection position, and the coordinates identified as false detection positions are excluded from consideration (false detection position exclusion step).

[0014] Then, the position of the edge of the workpiece is estimated based on the three or more remaining coordinates that were not excluded, and the processing area at the outer periphery of the workpiece is calculated based on the estimated edge position (processing area estimation step). In this way, by excluding coordinates determined to be false detection positions from multiple coordinates, the position of the edge of the workpiece can be estimated with higher accuracy. Therefore, the processing area can be estimated with high accuracy. Attached Figure Description

[0015] Figure 1 It is a three-dimensional diagram of the cutting device.

[0016] Figure 2 (A) is a top view of the workpiece, etc. Figure 2 (B) is a partial cross-sectional side view of the workpiece, etc.

[0017] Figure 3 (A) is an example of a displayed screen. Figure 3 (B) is an enlarged view of an example of an image of a part of the outer periphery of the workpiece.

[0018] Figure 4 (A) is a diagram showing an approximate circle, etc. Figure 4 (B) is a diagram showing the second approximate circle formed by excluding the coordinate with the largest offset among multiple coordinates.

[0019] Figure 5 This is a diagram showing the cutting area.

[0020] Figure 6 This is a flowchart of the edge alignment method.

[0021] Label Explanation

[0022] 2: Cutting device; 4: Base; 6: X-axis moving mechanism; 8: X-axis guide rail; 10: X-axis moving table; 11: Workpiece; 11a: Front; 11b: Back; 11c: Edge; 11d: Specified position; 11e: Annular area; 12: X-axis ball screw; 14: X-axis pulse motor; 16: Theta stage; 18a: Table base; 18b: Table cover; 20: Chuck table; 20a: Rotary axis; 20b: Holding surface; 21, 21a: Approximate circle; 23: Inner threshold circle; 25: Outer threshold circle; 27, 27a: Offset; 29: Specified position; 30: Support structure; 31 32: Cutting area; 34: Y-axis guide rail; 36: Y-axis moving plate; 38: Y-axis ball screw; 40: Y-axis pulse motor; 42: Z-axis guide rail; 44: Z-axis moving plate; 46: Z-axis ball screw; 48: Z-axis pulse motor; 50a, 50b: Cutting unit; 52: Cutting tool; 54a, 54b: Camera unit; 56: Imaging element; 58: Display; 60: Tool position detection unit; 62: Control unit; 62a: Coordinate calculation unit; 62b: Approximate circle generation unit; 62c: Judgment unit; 62d: Cutting area inference unit; 64: Point; 66: Allowable range. Detailed Implementation

[0023] An embodiment of one aspect of the present invention will be described with reference to the accompanying drawings. Figure 1 This is a three-dimensional view of cutting device 2. Figure 1 The X-axis (machining feed direction, forward and backward direction), Y-axis (indexing feed direction), and Z-axis (vertical direction, height direction) shown are perpendicular to each other.

[0024] In addition, Figure 1 In this design, a functional block represents a part of the constituent elements. The cutting device 2 has a base 4 that mounts each constituent element. An X-axis moving mechanism 6 is provided on the upper surface of the base 4. The X-axis moving mechanism 6 has a pair of X-axis guide rails 8 that are approximately parallel to the X-axis direction.

[0025] An X-axis movable worktable 10 is slidably mounted on the X-axis guide rail 8. A nut (not shown) is mounted on the lower surface (back side) of the X-axis movable worktable 10, and an X-axis ball screw 12, which is substantially parallel to the X-axis guide rail 8, is rotatably screwed into the nut.

[0026] One end of the X-axis ball screw 12 is connected to the X-axis pulse motor 14. By using the X-axis pulse motor 14 to rotate the X-axis ball screw 12, the X-axis moving table 10 moves along the X-axis guide rail 8 in the X-axis direction.

[0027] A cylindrical θ-stage 16 is provided on the upper surface (front side) of the X-axis moving stage 10. The θ-stage 16 has a rotation drive source such as an electric motor (not shown). A disc-shaped worktable base 18a is provided on the θ-stage 16.

[0028] A worktable cover 18b is provided around the worktable base 18a. A retractable corrugated dustproof and dripproof cover (not shown) is provided on one side and the other side of the worktable cover 18b in the X-axis direction. The worktable cover 18b and the dustproof and dripproof cover cover the top of the X-axis moving mechanism 6.

[0029] A disc-shaped chuck table 20 is provided on the upper surface of the table base 18a. The lower part of the chuck table 20 is connected to the θ stage 16 via the table base 18a. The chuck table 20 can rotate around a rotation axis 20a (refer to) that is approximately parallel to the Z-axis direction. Figure 2 (A)) rotation.

[0030] The chuck worktable 20 has a disc-shaped frame made of a metal such as stainless steel. A recess is formed on the upper surface of the frame, and a disc-shaped perforated plate made of porous ceramic is fixed in the recess, having an outer diameter approximately the same as the inner diameter of the recess.

[0031] The perforated plate is connected to a suction source (not shown) such as a vacuum pump or an ejector via a flow path formed in the frame. When the suction source is activated, a negative pressure is generated on the upper surface (holding surface 20b) of the perforated plate. This negative pressure attracts and holds the workpiece 11 on the holding surface 20b.

[0032] The workpiece 11 is, for example, a disk-shaped wafer made of semiconductor material such as silicon, with a device region and a remaining peripheral region surrounding the device region on its front side. The device region is divided into multiple regions by predetermined dividing lines arranged in a grid pattern, and devices such as ICs (Integrated Circuits) and LSIs (Large Scale Integrations) are formed in each region.

[0033] A portal-shaped support structure 30 is provided on the upper surface of the base 4, and the support structure 30 is configured to span the X-axis moving mechanism 6. A cutting unit moving mechanism 32 is provided on the front surface of the support structure 30. The cutting unit moving mechanism 32 has a pair of Y-axis guide rails 34 disposed on the front surface of the support structure 30.

[0034] Each Y-axis guide rail 34 is arranged approximately parallel to the Y-axis direction. Two Y-axis movable plates 36 are mounted on the pair of Y-axis guide rails 34 in a manner that allows them to slide along the Y-axis direction. A nut portion (not shown) is provided on the back side of each Y-axis movable plate 36.

[0035] Different Y-axis ball screws 38 are rotatably connected to each nut. The Y-axis ball screws 38 are arranged approximately parallel to the Y-axis guide rails 34, and a Y-axis pulse motor 40 is connected to one end of each Y-axis ball screw 38.

[0036] If the Y-axis ball screw 38 is rotated by the Y-axis pulse motor 40, the Y-axis moving plate 36 moves along the Y-axis guide rail 34 in the Y-axis direction. A cutting unit is provided on the front surface of each Y-axis moving plate 36.

[0037] The cutting unit has a pair of Z-axis guide rails 42 disposed on the front surface of the Y-axis moving plate 36. Each Z-axis guide rail is arranged substantially parallel to the Z-axis direction. A Z-axis moving plate 44 is slidably mounted on the pair of Z-axis guide rails 42.

[0038] A nut portion (not shown) is provided on the back side of the Z-axis moving plate 44, and a Z-axis ball screw 46 parallel to the Z-axis guide rail 42 is rotatably connected to the nut portion. One end of the Z-axis ball screw 46 is connected to the Z-axis pulse motor 48.

[0039] If the Z-axis ball screw 46 is rotated by the Z-axis pulse motor 48, the Z-axis moving plate 44 moves along the Z-axis guide rail 42 in the Z-axis direction. A cutting unit 50a is fixed to the lower part of the Z-axis moving plate 44 located on the Y-axis side.

[0040] Additionally, a cutting unit 50b is fixed to the lower part of the Z-axis moving plate 44 located on the other side of the Y-axis direction. Each of the cutting units 50a and 50b has a prism-shaped spindle housing with its long side portion arranged approximately parallel to the Y-axis direction.

[0041] The spindle housing houses a cylindrical spindle (not shown), whose long side is roughly parallel to the Y-axis. The spindle is supported by the spindle housing in a rotatable manner.

[0042] A servo motor or other rotary drive source is connected to one end of the spindle. The other end of the spindle protrudes out of the spindle housing, and a cutting tool 52 with a ring-shaped cutting edge is mounted on this other end.

[0043] A camera unit 54a is disposed on the side of the front surface of the spindle housing of the cutting unit 50a. Similarly, a camera unit 54b is disposed on the side of the front surface of the spindle housing of the cutting unit 50b.

[0044] Camera units 54a and 54b each use visible light to photograph the workpiece 11 and other objects held by the holding surface 20b. Each camera unit 54a and 54b includes a light source such as an LED, an objective lens (not shown), and an imaging element 56 such as a CCD (Charge Coupled Device) image sensor or a CMOS (Complementary Metal Oxide Semiconductor) image sensor.

[0045] For example, when performing edge alignment, camera units 54a and 54b take pictures of multiple portions of the outer periphery of the front side 11a of the workpiece 11 in the circumferential direction. The images of the workpiece 11 taken by camera units 54a and 54b are displayed on display 58.

[0046] The display 58 is disposed on the front surface side of the cutting device 2. The display 58 is, for example, a touch panel that also serves as an input device for inputting instructions from the operator to the cutting device 2 and a display device for displaying images.

[0047] Below each of the cutting units 50a and 50b, there is a tool position detection unit 60 for detecting the position (height) of the lower end of the cutting tool 52. The tool position detection unit 60 has a rigid conductive component (not shown) formed of a metal plate or the like.

[0048] The conductive component, for example, has a cuboid shape. The position of the upper surface of the conductive component in the height direction is predetermined. When the lower end of the rotating cutting tool 52 contacts the upper surface of the conductive component, a closed circuit is formed by the cutting tool 52, the conductive component, etc., thus the position of the lower end of the cutting tool 52 can be determined.

[0049] The cutting device 2 is equipped with a control unit 62, which controls the operation of the X-axis moving mechanism 6, the θ stage 16, the cutting unit moving mechanism 32, the cutting units 50a and 50b, the camera units 54a and 54b, and the tool position detection unit 60.

[0050] The control unit 62 may be composed of, for example, a computer, which includes: a processing device such as a processor (CPU, represented by a CPU); a main storage device such as DRAM (Dynamic Random Access Memory), SRAM (Static Random Access Memory), and ROM (Read Only Memory); and an auxiliary storage device such as flash memory, hard disk drive, and solid-state drive.

[0051] The auxiliary storage device stores software containing a prescribed program. The control unit 62 functions by operating the processing device or similar equipment according to this software. The control unit 62 includes a coordinate calculation unit 62a, which calculates the coordinates of the edges 11c of the workpiece 11 (see reference 11c) from images of various parts of the outer periphery of the workpiece 11 obtained by the camera unit 54a or similar equipment. Figure 2 The coordinates of a point corresponding to (B).

[0052] The coordinate calculation unit 62a is configured by a program, for example, for an image whose display area is approximately square and consists of multiple gray levels (256 levels with pixel values ​​from 0 to 255) (see reference). Figure 3 (A) and Figure 3 (B) is binarized using a specified pixel value (e.g., 125) as a threshold.

[0053] The coordinate calculation unit 62a further considers the intersection of the boundary line in the image generated by binarization and the defined diagonal of the approximately square image as a point (a point) corresponding to the edge 11c 64 (refer to) Figure 4 (A), etc.), and calculate the coordinates of point 64. In addition, the coordinates of point 64 are calculated with the rotation axis 20a as the origin.

[0054] The control unit 62 also includes an approximate circle generation unit 62b, which generates an approximate circle 21 (see reference) by using the least squares method on the coordinates of multiple points 64. Figure 4 (A)). The approximate circle generation part 62b is constructed by a program.

[0055] A brief explanation of the algorithm for generating an approximate circle 21 using the least squares method is provided. Let the coordinates of the n points 64 be (Xi, Yi), the center of the approximate circle 21 calculated using the least squares method be (a, b), and the radius of the approximate circle 21 be r. Here, n is a natural number greater than 2, and i is a natural number greater than 1 and less than n.

[0056] Then, in ∑{(Xi-a) 2 +(Yi-b) 2 -r 2} 2 =0 is converted to ∑{Xi 2 +Yi 2 +AXi+BYi+C} 2 After =0, correspondingly, partial derivatives are applied to A, B, and C respectively. Here, ∑ represents the sum with respect to i.

[0057] Thus, we obtain three equations concerning A, B, and C. Solving these three equations yields the solutions for A, B, and C respectively. In the above transformations, since A = -2a, B = -2b, and C = a... 2 +b 2 -r 2 Therefore, based on the solutions of A, B and C, we can obtain the center coordinates (a, b) and radius r of the approximate circle 21.

[0058] The approximate circle generation unit 62b further generates an inner threshold circle 23 with a diameter smaller than the diameter of the approximate circle 21 by a predetermined threshold, and an outer threshold circle 25 with a diameter larger than the diameter of the approximate circle 21 by a predetermined threshold (see reference). Figure 4 (A, etc.).

[0059] The inner threshold circle 23 and the outer threshold circle 25 are circles arranged concentrically with the approximate circle 21. In this embodiment, the threshold is set to 50 μm, but the threshold is not limited to 50 μm, and can also be set to 40 μm, 30 μm, etc.

[0060] The control unit 62 also includes a decision unit 62c composed of a program. The decision unit 62c calculates the offset 27 between the generated approximate circle 21 and the coordinates of each of the coordinates of all the points 64 (see reference). Figure 4 (A)).

[0061] The offset 27 is defined by the radial distance from the approximate circle 21 to point 64. Next, if there are coordinates of point 64 where the offset 27 is greater than or equal to the aforementioned preset threshold, the determination unit 62c determines the coordinates of point 64 with the largest offset 27 as a false detection position.

[0062] Furthermore, the determination unit 62c excludes the coordinates of point 64, which is determined to be a false detection location, from the considerations of the approximate circle generation unit 62b when it generates the approximate circle 21 next time. For example, the determination unit 62c notifies the approximate circle generation unit 62b of the point 64 to be excluded, thereby excluding the false detection location from the considerations.

[0063] Additionally, the determination unit 62c can also notify the approximate circle generation unit 62b of all the remaining points 64 that were not excluded. Based on the notification from the determination unit 62c and the coordinates of the three or more remaining points 64 that were not excluded, the approximate circle generation unit 62b regenerates the approximate circle 21 (see reference). Figure 4 (B)

[0064] The control unit 62 also includes a cutting area estimation unit 62d for estimating the cutting area. The cutting area estimation unit 62d is configured by a program to make the latest approximate circle 21 as an estimation of the edge 11c of the outer periphery of the workpiece 11.

[0065] Then, the cutting area estimation unit 62d defines the annular region from the latest approximate circle 21 to a predetermined position on the center side of the latest approximate circle 21 as the cutting area (machining area). This estimates the cutting area to be cut. This cutting area roughly corresponds to the annular region 11e from the edge 11c to a predetermined position 11d on the front surface 11a (see reference). Figure 2 (B)

[0066] Furthermore, if there are no false detection positions in the coordinates of all points 64 from the beginning, the cutting area inference unit 62d makes a prediction that the approximate circle 21 generated based on the initial coordinates of all points 64 is the edge 11c of the workpiece 11, and infers the cutting area.

[0067] Next, refer to Figure 2 (A) Figure 2 (B) to Figure 6 The edge alignment method for the workpiece 11 will be explained. Additionally, Figure 6 This is a flowchart of the edge alignment method in this embodiment.

[0068] First, the back side 11b of the workpiece 11 is held by the holding surface 20b of the chuck table 20 (holding step S10). At this time, the center of the back side 11b is arranged in a manner that is approximately consistent with the center of the holding surface 20b.

[0069] After holding step S10, for example, a portion of the outer periphery of the front 11a side in a stationary state is photographed from above using camera unit 54a. Figure 2 (A) is a top view showing the workpiece 11, etc., in the peripheral imaging step S20.

[0070] Then, after the chuck table 20 is rotated around the rotation axis 20a by a predetermined angle, the chuck table 20 is brought to a stop, and the camera unit 54a is used again to take a picture of another part of the outer periphery of the front side 11a from above.

[0071] In this way, by repeatedly rotating at a specified angle and taking pictures of the outer periphery of the front side 11a, pictures are taken of multiple different parts of the workpiece 11 in the circumferential direction (outer periphery shooting step S20).

[0072] In this embodiment, a camera unit 54a is used to photograph 12 different locations on the outer periphery of the workpiece 11 by rotating the chuck table 20 by 30 degrees each time. Furthermore, the rotation angle is not limited to a constant angle; it can vary in a prescribed pattern, such as 40 degrees, 10 degrees, 25 degrees, 40 degrees, 10 degrees, 25 degrees, etc., or it can vary randomly.

[0073] Alternatively, camera unit 54b can be used instead of camera unit 54a, or both camera units 54a and 54b can be used. If both camera units 54a and 54b are used, the time required for the peripheral imaging step S20 can be shortened compared to using only one of them.

[0074] Figure 2 (B) is a partial cross-sectional side view of the workpiece 11 at its outer periphery. When the outer periphery is photographed using a camera unit 54a or the like, the brightness and darkness of the image are reversed with the edge 11c as the boundary due to differences in light reflectivity and other reasons.

[0075] For example, such as Figure 3 (A) Figure 3 As shown in (B), the area corresponding to the position further outward than the edge 11c in the radial direction of the front surface 11a has uniform brightness. Furthermore, in the chamfered portion, the area becomes darker closer to the center of the front surface 11a. Additionally, the area closer to the center of the front surface 11a than the inner periphery of the chamfered portion has a generally uniform darkness.

[0076] Figure 3 (A) is an example of a display screen shown on monitor 58. The display screen contains an image of the outer periphery of the workpiece 11. Figure 3 Image (B) is an enlarged view of an example of an image of a portion of the outer periphery of the workpiece 11. Furthermore, the brightness of the image is not limited to the example described above; it is also possible to make the area outside the edge 11c darker and the area inside the edge 11c brighter.

[0077] As described above, the coordinate calculation unit 62a performs binarization processing on the image using a predetermined pixel value as a threshold, and compares the boundary line generated by the binarization processing with the predetermined diagonal of the image. Figure 3 Let point 64 be the intersection of the dashed lines (B). Figure 3 In (B), point 64 is represented by the symbol “+”.

[0078] Then, the coordinate calculation unit 62a calculates the coordinates of point 64 of the workpiece 11 at each location photographed in the peripheral imaging step S20 (coordinate calculation step S30). Point 64 is a point that may correspond to the edge 11c of the workpiece 11.

[0079] Whenever an image is obtained in the peripheral imaging step S20, the coordinate calculation unit 62a of this embodiment calculates the coordinates of point 64 in that image. However, the coordinate calculation unit 62a may also summarize and calculate the coordinates of point 64 in each image after obtaining multiple or all images obtained in the peripheral imaging step S20.

[0080] After calculating the coordinates of all points 64, the approximate circle generation unit 62b uses the least squares method on the coordinates of all points 64 calculated through the coordinate calculation step S30 to generate an approximate circle 21 (approximate circle generation step S40).

[0081] In the approximate circle generation step S40 of this embodiment, if the center coordinates (a, b) and radius r of the approximate circle 21 can be calculated, then the approximate circle 21 is considered to have been generated. Furthermore, in the approximate circle generation step S40, the generated approximate circle 21 can also be actually displayed on the display 58.

[0082] For ease of explanation, the case where the approximate circle 21 is displayed in the approximate circle generation step S40 will be described below. Figure 4 (A) is a diagram showing the approximate circle 21 generated in the approximate circle generation step S40.

[0083] exist Figure 4 In (A), “×” represents each point 64 obtained from step S10 to step S30, and solid lines represent approximate circle 21. In addition, dashed lines represent inner threshold circle 23 and outer threshold circle 25, which are respectively configured as concentric circles with approximate circle 21.

[0084] Due to various factors, point 64 may deviate from the range (allowable range 66) between the inner threshold circle 23 and the outer threshold circle 25. For example, point 64 may sometimes deviate from the allowable range 66 due to foreign objects attached to the chamfer, water droplets or foreign objects attached to the retaining surface 20b, foreign objects attached to the lens of the camera unit 54a, inappropriate light during shooting, abnormal shape of the edge 11c, etc.

[0085] In this embodiment, point 64 located outside the allowable range 66 is considered a false detection location and is excluded from consideration in the subsequent approximate circle generation step S40. Therefore, the determination unit 62c first calculates the offset 27 between the approximate circle 21 and each coordinate of all points 64.

[0086] Then, the determination unit 62c determines whether there is a point 64 where the offset 27 is outside the allowable range 66 (i.e., the offset 27 is above a preset threshold) (offset comparison step S50).

[0087] When the condition is "yes" in step S50, the determination unit 62c excludes the point 64, which is determined to be a false detection position, from the considerations of the approximate circle generation unit 62b when it generates the approximate circle 21 next time (false detection position exclusion step S60). As a result, the position of the edge 11c of the workpiece 11 can be estimated with higher accuracy compared to the case where false detection positions are not excluded.

[0088] Furthermore, if all points 64 are within the allowable range 66 ("No" in step S50), the cutting area estimation unit 62d estimates the approximate circle 21 as the position of the edge 11c of the workpiece 11. Then, the cutting area estimation unit 62d estimates the cutting area 31 based on the estimated position of the edge 11c (cutting area estimation step (machining area estimation step) S70).

[0089] Specifically, the cutting area inference unit 62d defines the annular region from the approximate circle 21 (i.e., the estimated position of the edge 11c) to a predetermined position 29 on the center side of the approximate circle 21 as the cutting area 31. Figure 5 This is a diagram showing the cutting region 31. After step S70, the cutting region 31 is cut and removed to a predetermined depth using a cutting unit 50a, etc.

[0090] Additionally, as described above, if "yes" is indicated in step S50, after the false detection location exclusion step S60, the process returns to step S40 again. Then, the approximate circle generation unit 62b uses the least squares method on the coordinates of all the remaining points 64 that were not excluded from the consideration, thereby generating an approximate circle again (second (additional) approximate circle generation step S40).

[0091] Figure 4 (B) shows the coordinates of the point with the largest offset of 27 among multiple points 64 (in Figure 4 The diagram shows the second approximate circle 21a formed by removing point 64 at the 10 o'clock position in (B). Furthermore, the center and radius of the second approximate circle 21a are generally different from the center and radius of the previous approximate circle 21.

[0092] After the second approximate circle generation step S40, in step S50, the determination unit 62c calculates the offset 27a between the coordinates of the second approximate circle 21a and the coordinates of all the remaining points 64 that were not excluded. If there are coordinates of points 64 whose offset 27a is greater than or equal to a preset threshold (i.e., "yes" in step S50), the determination unit 62c determines the coordinates of the point 64 with the largest offset 27a as the false detection position.

[0093] Then, the determination unit 62c excludes the coordinates of the point 64, which is determined to be a false detection location, from the consideration objects (second (additional) false detection location exclusion step S60). The false detection location exclusion step S60 and the approximate circle generation step S40 are repeated until the false detection location disappears (that is, until "No" is achieved in step S50).

[0094] If there are no points 64 whose offset 27a is above the preset threshold (i.e., "No" in step S50), the cutting area inference unit 62d estimates the position of the edge 11c based on the coordinates of the three or more points 64 that are not excluded, and infers the cutting area 31 (cutting area inference step S70).

[0095] In this embodiment, by excluding the coordinates of points 64 with relatively large offsets such as 27 and 27a from the consideration of the approximate circle 21 as false detection locations, the edge 11c of the workpiece 11 can be estimated with higher accuracy. Therefore, the cutting area 31 can be inferred with high accuracy.

[0096] In addition, the structure and method of the above embodiments can be appropriately modified and implemented without departing from the scope of the present invention. For example, a laser displacement meter (not shown) can be used instead of camera units 54a and 54b.

[0097] The laser displacement meter has a light source such as a laser diode (not shown). The laser beam emitted from the light source is shaped by a projection lens (not shown) to extend in a predetermined direction perpendicular to the direction of travel of the laser beam, and illuminates the front side 11a in a linear manner across the edge 11c.

[0098] Light reflected from the subject is guided through a light-receiving lens (not shown) to a linear sensor (not shown), which has multiple photoelectric conversion elements arranged in a straight line at predetermined intervals of approximately 10 μm. Next, a method for edge alignment using this laser displacement meter will be described.

[0099] First, in holding step S10, the back side 11b is held. After step S10, a position detection step S25 (not shown) is performed using a laser displacement meter instead of the peripheral imaging step S20. In step S25, light is irradiated in a linear pattern across the edge 11c of the front side 11a of the stationary workpiece 11.

[0100] Then, by using a linear sensor to receive reflected light, the position of point 64 that may correspond to edge 11c is detected. Then, similarly to step S20 above, the chuck table 20 is rotated, and the positions of point 64 that may correspond to edge 11c are detected at different locations on the outer periphery of the workpiece 11 in the circumferential direction.

[0101] After the position detection step S25, the coordinate calculation unit 62a calculates the coordinates of point 64 at each location with the rotation axis 20a as the origin (coordinate calculation step S30). The processing after step S30 is the same as the edge alignment method described above.

Claims

1. A method for aligning the edges of a disc-shaped workpiece, characterized in that, This edge alignment method has the following steps: The holding step involves using a chuck table to hold the workpiece. The coordinate calculation step involves calculating the coordinates of a point that may correspond to the edge of the workpiece at various locations among multiple different parts of the outer periphery of the workpiece in the circumferential direction of the workpiece. The approximate circle generation step applies the least squares method to all coordinates calculated in the coordinate calculation step to generate an approximate circle; The false detection location exclusion step calculates the offset between the approximate circle generated in the approximate circle generation step and each coordinate in the total coordinate system. If any coordinate has an offset exceeding a pre-set threshold, the coordinate with the largest offset is identified as a false detection location, and this location is excluded from the consideration list. The processing area estimation step, following the false detection location elimination step, estimates the edge position of the workpiece based on the three or more remaining coordinates that were not eliminated, and infers the processing area of ​​the outer periphery of the workpiece based on the estimated edge position. The edge alignment method also includes an additional approximate circle generation step: after eliminating the coordinates with the largest offset through the false detection location elimination step, the least squares method is used on all remaining coordinates that were not eliminated from the consideration in the false detection location elimination step to generate an approximate circle again. The edge alignment method further includes an additional false detection location exclusion step: calculating the offset between the approximate circle generated by the additional approximate circle generation step and each of the remaining coordinates that were not excluded from the consideration object in the false detection location exclusion step; if there are still coordinates whose offset is above the threshold preset in the false detection location exclusion step, the coordinate with the largest offset is determined as a false detection location, and the coordinate with the largest offset that is determined as a false detection location is excluded from the consideration object. The offset between the approximate circle generated by the additional approximate circle generation step and each of the coordinates remaining in all the coordinates that were not excluded from the consideration in the false detection position exclusion step is calculated. If there are no coordinates whose offset is above the threshold set in the false detection position exclusion step, the processing area inference step is performed.

Images