Double-sided polishing apparatus and double-sided polishing method

Abstract

To provide a double-sided polishing device capable of suppressing a workpiece shape from deviating from a target shape at the end of final polishing, and obtaining a desired workpiece shape, a control section (30) of a polisher (10) based on a measurement result of a thickness meter (20) has: a first learner (31) that learns a correlation among a main polishing condition, a peripheral flatness, and an in-plane flatness; a second learner (32) that learns a correlation among an end-point polishing condition, a change degree of the in-plane flatness at a main polishing stop, a change degree of the peripheral flatness at the main polishing stop, and a change amount of the peripheral flatness during the end-point polishing; a main polishing condition correction section (33) that corrects a main polishing condition at a correction timing, using a main polishing condition obtained by inputting a correction target value of the in-plane flatness and a target value of the peripheral flatness obtained from a learning result of the first learner (31) into the first learner (31); and a main polishing stop condition setting section (34) that sets a main polishing stop condition using a change amount of the peripheral flatness obtained from a learning result of the second learner (32).

Description

Technical Field

[0001] This invention relates to a double-sided grinding device and method for grinding the front and back sides of a circular plate-shaped workpiece. Background Technology

[0002] Previously, a double-sided grinding apparatus was known that, when grinding the front and back sides of a circular plate-shaped workpiece such as a silicon wafer using an upper platform and a lower platform, controlled the distance between the platforms based on an optimal distance obtained, for example, based on the input desired workpiece flatness, using an artificial intelligence model that had learned the relationship between the distance between the platforms and the workpiece flatness (see, for example, Patent Document 1). Furthermore, a double-sided grinding apparatus is known that stops double-sided grinding of the workpiece when the overall shape index of the workpiece reaches a set value (see, for example, Patent Document 2).

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent Application Publication No. 2022-189524

[0006] Patent Document 2: WO2022 / 254856 Summary of the Invention

[0007] The problem the invention aims to solve

[0008] However, when grinding a workpiece in a double-sided grinding apparatus, the load applied to the workpiece is first gradually increased. If conditions such as the load reaching a specified load are met, the workpiece is then ground for a predetermined time while maintaining the load at the target load. Then, if the predetermined time has elapsed since the start of the main grinding, the main grinding is stopped. Afterward, the load applied to the workpiece is gradually reduced to complete the grinding process. Even when gradually reducing the load after the main grinding, the shape of the workpiece changes, and the amount of this shape change varies depending on the grinding conditions or the state of the auxiliary materials, and is therefore not constant. Therefore, the operator of the grinding apparatus will estimate the amount of shape change of the workpiece after the main grinding, and will manually adjust the timing of stopping the main grinding. However, because the timing of stopping the main grinding relies on the operator's intuition, deviations in the finished quality of the workpiece may occur.

[0009] On the other hand, in the existing double-sided grinding apparatus described in Patent Document 1 or Patent Document 2, only the grinding conditions during the main grinding are set. That is, in the existing double-sided grinding apparatus, the distance between the platforms is controlled during the main grinding, or the main grinding of the workpiece is stopped when the overall shape index of the workpiece reaches a set value. However, it does not take into account that the shape of the workpiece will change even when the load applied to the workpiece is gradually reduced after the main grinding. Therefore, sometimes the shape of the workpiece at the end of the final grinding is inconsistent with the target shape, deviating from the target shape and failing to obtain the desired workpiece shape.

[0010] The present invention is implemented in view of the above-mentioned problems. Its technical problem is to provide a double-sided grinding apparatus and a double-sided grinding method, which can suppress the workpiece shape from deviating from the target shape at the end of the final grinding and obtain the desired workpiece shape.

[0011] Means for solving technical problems

[0012] To achieve the above objectives, the double-sided grinding apparatus of the present invention is characterized by comprising: a grinding machine that holds a circular plate-shaped workpiece between a lower platform and an upper platform disposed opposite to the lower platform, and grinding the front and back sides of the workpiece by moving the lower platform and the upper platform relative to the workpiece while applying a load to the workpiece; a thickness measuring device that measures the thickness of the workpiece during the grinding process of the grinding machine; and a control unit that controls the grinding machine based on the measurement results of the thickness measuring device, wherein a predetermined peripheral region extending radially inward from the outer periphery and an in-plane region extending from the peripheral region to the center of the workpiece are defined on the workpiece, the control unit comprising: a first learner that learns the grinding conditions for main grinding while maintaining the load at a target load, and the correlation between the flatness of the peripheral region (i.e., peripheral flatness) and the flatness of the in-plane region (i.e., in-plane flatness); and a second learner that learns the grinding conditions for final grinding while gradually reducing the load, and the main grinding... The correlation between the degree of change in in-plane flatness at the stopping time of the main grinding, the degree of change in peripheral flatness at the stopping time of the main grinding, and the amount of change in peripheral flatness during the final grinding process; the main grinding condition correction unit calculates the correction target value of the in-plane flatness based on the predicted value of the in-plane flatness obtained by inputting the grinding conditions of the main grinding set at the correction time and the peripheral flatness at the correction time into the first learner, and corrects the grinding conditions of the main grinding set at the correction time according to the grinding conditions of the main grinding obtained by inputting the correction target value of the in-plane flatness and the target value of the peripheral flatness into the first learner; and the main grinding stopping condition setting unit inputs the grinding conditions of the final grinding, the degree of change in in-plane flatness at the calculation time, and the degree of change in peripheral flatness at the calculation time into the second learner to obtain the amount of change in peripheral flatness during the final grinding process, and sets the stopping condition of the main grinding based on the amount of change in peripheral flatness and the target value of peripheral flatness.

[0013] To achieve the above objectives, the double-sided grinding method of the present invention clamps a circular plate-shaped workpiece between a lower platform and an upper platform disposed opposite to the lower platform. While a load is applied to the workpiece, the lower platform and the upper platform are moved relative to the workpiece to grind the front and back sides of the workpiece. The double-sided grinding method is characterized by using: a first learner that learns the main grinding conditions for grinding the workpiece while maintaining the load at a target load, and sets the flatness of the outer peripheral region (i.e., the outer peripheral flatness) within a predetermined range from the outer peripheral end of the workpiece toward the radially inward side, and the flatness of the outer peripheral region from the workpiece... The correlation between the flatness of the in-plane region from the outer peripheral region to the center of the workpiece, i.e., the in-plane flatness; and a second learner, learning the correlation between the grinding conditions of the final grinding while gradually reducing the load on the workpiece, the degree of change of the in-plane flatness at the stop time of the main grinding, the degree of change of the outer peripheral flatness at the stop time of the main grinding, and the amount of change of the outer peripheral flatness during the final grinding process; the double-sided grinding method includes the steps of setting target values ​​for the outer peripheral flatness and the in-plane flatness; after setting the outer... After setting target values ​​for circumferential flatness and in-plane flatness, an initial grinding step is performed while gradually increasing the load on the workpiece. After the initial grinding is completed, the main grinding step begins. During the execution of the main grinding, a correction target value for in-plane flatness is calculated based on the predicted value of in-plane flatness obtained by inputting the grinding conditions of the main grinding at the correction time and the circumferential flatness at the correction time into the first learner. The grinding conditions of the main grinding set at the correction time are then corrected based on the grinding conditions of the main grinding obtained by inputting the correction target value of in-plane flatness and the target value of circumferential flatness into the first learner. During the execution of the main grinding, the grinding conditions of the final grinding, the degree of change of in-plane flatness at the calculation time, and the degree of change of circumferential flatness at the calculation time are input into the second learner to obtain the change in circumferential flatness during the final grinding process. Based on the change in circumferential flatness and the target value of circumferential flatness, a stop condition for the main grinding is set. After the stop condition for the main grinding is met, the final grinding step is performed.

[0014] Invention Effects

[0015] According to the double-sided grinding apparatus and double-sided grinding method of the present invention, it is possible to suppress the deviation of the workpiece shape from the target shape at the end of the final grinding, and the desired workpiece shape can be obtained. Attached Figure Description

[0016] Figure 1This is an explanatory diagram that schematically shows the overall structure of the double-sided grinding apparatus of Embodiment 1.

[0017] Figure 2 This is an explanatory diagram showing the positional relationship between the sun gear, internal gear, and planet carrier plate in Embodiment 1.

[0018] Figure 3A This is an explanatory diagram showing the trajectory of the measuring hole as it passes through the workpiece in the double-sided grinding apparatus of Embodiment 1.

[0019] Figure 3B This is an explanatory diagram showing the thickness data set for measuring the thickness of a workpiece ground by the double-sided grinding apparatus of Example 1, and the outer peripheral region and in-plane region defined on the workpiece.

[0020] Figure 4A This is an explanatory diagram showing an approximate straight line set for the thickness data set of a workpiece with a raised outer perimeter.

[0021] Figure 4B This is an explanatory diagram showing an approximate straight line set for the thickness data set of a workpiece with a planar outer perimeter.

[0022] Figure 5 This is an illustrative diagram showing the innermost and outermost partitions when the outer perimeter region is divided into multiple partitions.

[0023] Figure 6 This is an explanatory diagram showing the calculated values ​​and degree of change of in-plane flatness and peripheral flatness at each time point.

[0024] Figure 7 This is an illustrative diagram showing the dataset that the first learner of Embodiment 1 learns.

[0025] Figure 8 This is an illustrative diagram showing the dataset that the second learner of Embodiment 1 learns.

[0026] Figure 9 This is a flowchart illustrating the grinding process control flow executed by the control unit of Embodiment 1.

[0027] Figure 10 This is a table showing the main grinding conditions in the example workpiece and the peripheral flatness and in-plane flatness at this time.

[0028] Figure 11 This is a table showing the datasets that the learner used to learn the comparison examples.

[0029] Figure 12 This shows the grinding result of the workpiece shape at the end of the final grinding process in the double-sided grinding apparatus of Example 1, when the target value of the outer peripheral flatness is set to 0 (nm) for grinding the workpiece.

[0030] Figure 13 This shows the grinding result of the workpiece shape at the end of the final grinding process in the double-sided grinding apparatus of Example 1. Detailed Implementation

[0031] Hereinafter, based on Embodiment 1 shown in the accompanying drawings, the manner in which the double-sided grinding apparatus and double-sided grinding method are used to implement the present invention will be described.

[0032] The double-sided grinding apparatus 1 in Example 1 is a double-sided grinding apparatus for grinding both the front and back sides of thin, round workpieces W such as semiconductor wafers, crystal wafers, sapphire wafers, glass wafers, or ceramic wafers. Figure 1 As shown, the double-sided grinding apparatus 1 includes a grinding machine 10, a thickness measuring device 20, and a control unit 30.

[0033] The grinding machine 10 holds a workpiece W between a lower platform 11 and an upper platform 12 opposite to the lower platform 11. While a load is applied to the workpiece W, the lower platform 11 and the upper platform 12 move relative to the workpiece W, thereby simultaneously grinding both the front and back sides of the workpiece W. The grinding machine 10 includes: an annular circular plate-shaped lower platform 11 and upper platform 12 concentrically arranged around an axis L1; a sun gear 13 rotatably disposed at the center of the lower platform 11; an internal gear 14 disposed on the outer periphery of the lower platform 11; and a workpiece holding hole 15a formed between the lower platform 11 and the upper platform 12 (see reference). Figure 2 The planetary carrier plate 15. Additionally, a grinding pad 11a is attached to the upper surface of the lower platform 11, and a grinding pad 12a is attached to the lower surface of the upper platform 12. Furthermore, the upper platform 12 is provided with a plurality of supply holes (not shown) for supplying grinding slurry (hereinafter referred to as "slurry"). Moreover, it is sometimes possible to select whether to supply slurry for each supply hole or each group of supply holes.

[0034] Here, as Figure 2 As shown, the planetary carrier 15 meshes with the sun gear 13 and the internal gear 14. Furthermore, the planetary carrier 15 rotates on its own axis and revolves around the axis L1 while rotating through the sun gear 13 and the internal gear 14.

[0035] The workpiece W is positioned within the workpiece holding hole 15a of the planetary carrier plate 15. While the workpiece W is held between the grinding pad 11a attached to the rotating lower platform 11 and the grinding pad 12a attached to the rotating upper platform 12, the lower platform 11 and the upper platform 12 move relative to the workpiece W due to the rotation and revolution of the planetary carrier plate 15. The workpiece W is then ground by the grinding pads 11a and 12a. In other words, the surfaces of the grinding pads 11a and 12a become the grinding surfaces for grinding the workpiece W.

[0036] The upper platform 12 is fixed to the rod 16 via support studs 16a and mounting members 16b mounted on its upper surface. The rod 16 extends and retracts in the vertical direction via a fifth drive device M5, causing the upper platform 12 to move up and down. Furthermore, a predetermined load is applied to the workpiece W from the upper platform 12 according to the extension and retraction length of the rod 16. That is, the load applied to the workpiece W is adjusted by controlling the fifth drive device M5.

[0037] Furthermore, a first drive shaft 17a, erected along axis L1, is disposed at the center of the grinding machine 10. The first drive shaft 17a is a shaft that rotates via a first drive device M1. A driver 18 is fixed to the upper end of the first drive shaft 17a. Thus, the driver 18 rotates integrally with the first drive shaft 17a. On the other hand, a groove (not shown) is formed on the outer peripheral surface of the driver 18 to engage with a hook 12b provided on the upper platform 12. Furthermore, as the rod 16 extends and the upper platform 12 moves downward, the hook 12b engages with the groove of the driver 18, thereby causing the driver 18 and the upper platform 12 to rotate integrally. That is, since the upper platform 12 rotates integrally with the first drive shaft 17a that rotates via the first drive device M1, the rotational speed of the upper platform 12 can be adjusted by controlling the first drive device M1.

[0038] The second drive shaft 17b is fixed through the hole 13a in the center of the sun gear 13. The second drive shaft 17b is a hollow tube open at both ends, through which the first drive shaft 17a rotatably passes. Furthermore, the second drive shaft 17b rotates via the second drive device M2. That is, since the sun gear 13 and the second drive shaft 17b, which rotates via the second drive device M2, rotate as a unit, the rotational speed of the sun gear 13 can be adjusted by controlling the second drive device M2.

[0039] A third drive shaft 17c is formed at the lower part of the central portion of the lower platform 11. The third drive shaft 17c is a hollow tube open at both ends, through which the first drive shaft 17a and the second drive shaft 17b rotatably pass. Furthermore, the third drive shaft 17c rotates via a third drive device M3. That is, since the lower platform 11 and the third drive shaft 17c, which rotates via the third drive device M3, rotate as a unit, the rotational speed of the lower platform 11 can be adjusted by controlling the third drive device M3.

[0040] Furthermore, a fourth drive shaft 17d is formed in the internal gear 14. The fourth drive shaft 17d is a hollow tube open at both ends, through which the first drive shaft 17a, the second drive shaft 17b, and the third drive shaft 17c rotatably pass. The fourth drive shaft 17d rotates via a fourth drive device M4. That is, since the internal gear 14 and the fourth drive shaft 17d, which rotates via the fourth drive device M4, rotate as a unit, the rotational speed of the internal gear 14 can be adjusted by controlling the fourth drive device M4.

[0041] Furthermore, the double-sided grinding apparatus 1 of Embodiment 1, based on predetermined grinding conditions, controls the load applied to the workpiece W by extending and retracting the rod 16, while simultaneously controlling the rotational speeds of the lower platform 11, upper platform 12, sun gear 13, and internal gear 14 to grind the workpiece W. Since the extension and retraction of the rod 16 and the adjustment (increase or decrease) of the rotational speeds of the lower platform 11, etc., require time, the grinding process of the workpiece W is divided into initial grinding, main grinding, and final grinding.

[0042] Furthermore, "initial grinding" refers to the process of grinding workpiece W while gradually increasing the load applied to it after grinding begins. In initial grinding, the rotational speeds of the lower platform 11, upper platform 12, sun gear 13, and internal gear 14 are gradually increased. Moreover, the initial grinding conditions (initial grinding conditions), including grinding time and end conditions, are specified according to the prerequisite grinding conditions. Furthermore, the "preliminary grinding conditions" are grinding conditions predetermined based on the final processing target (grinding target) of workpiece W and the type of workpiece W.

[0043] Furthermore, "main grinding" refers to the process of grinding workpiece W under specified grinding conditions while maintaining the load applied to workpiece W at the target load after the initial grinding. In main grinding, the rotational speeds of the lower platform 11, upper platform 12, sun gear 13, and internal gear 14 are maintained at their target rotational speeds. Moreover, the grinding conditions (main grinding conditions), including the target load and target rotational speeds for each rotational speed, are specified at the start of main grinding based on the preliminary grinding conditions and are continuously and repeatedly modified during the execution of main grinding. Furthermore, the main grinding stop conditions (main grinding stop conditions), which specify the timing for stopping main grinding, are repeatedly specified during the execution of main grinding, and main grinding stops when the main grinding stop conditions are met. In addition, main grinding can consist of multiple divisions (multiple steps), and the main grinding conditions may also include appropriately adjusting the slurry temperature, flow rate, and supply destination based on monitoring the changing trends of the workpiece W's shape.

[0044] Furthermore, "final grinding" refers to the process of grinding workpiece W while gradually reducing the load applied to it after the main grinding has stopped. In final grinding, the rotational speeds of the lower platform 11, upper platform 12, sun gear 13, and internal gear 14 are gradually reduced. Moreover, the grinding conditions for final grinding, including the grinding time, are specified according to the preliminary grinding conditions. Additionally, the final grinding conditions may also include appropriately adjusting the slurry temperature, flow rate, and supply destination based on monitoring the changing trend of the workpiece W's shape.

[0045] Furthermore, a measuring hole 19 is formed on the upper platform 12 at a predetermined distance from the center along the radial direction. The measuring hole 19 penetrates the upper platform 12 and the grinding pad 12a, and a window member 19a is installed to allow the measuring light, i.e., the laser, to pass through.

[0046] The thickness measuring device 20 illuminates the workpiece W with measuring light and receives the measuring light reflected from the workpiece W to measure the thickness of the workpiece W (the distance between the front and back sides of the workpiece W) during the grinding process. Furthermore, the thickness measuring device 20 of Embodiment 1 quantifies the shape of the workpiece W based on the measured thickness data. The thickness measuring device 20 includes a measuring unit 21, a thickness measuring section 22, and a shape calculation section 23.

[0047] The measuring unit 21 is mounted on the upper platform 12 and rotates integrally with it. The measuring unit 21 includes: a laser source (not shown) that irradiates measuring light, i.e., a laser, toward the workpiece W via a window member 19a mounted through a measuring hole 19 on the upper platform 12; and a light-receiving unit (not shown) that receives reflected light from the front and back sides of the workpiece W as light signals. The light signals received by the light-receiving unit are transmitted to the thickness measuring unit 22 via the transmitting unit 21a.

[0048] The thickness measuring unit 22 measures the thickness of the workpiece W, for example, by means of light reflection interferometry. The thickness measuring unit 22 has a receiving unit 22a that receives a light signal transmitted from the measuring unit 21, and obtains the thickness data of the workpiece W based on the light signal received by the receiving unit 22a.

[0049] Here, as Figure 3A As shown, due to the rotation of the upper platform 12, during the period when the measuring hole 19 passes through the surface of the workpiece W, the laser from the measuring unit 21 continuously irradiates the surface of the workpiece W. Therefore, the thickness measuring unit 22 continuously measures the thickness of each in-plane position of the workpiece W along the passing trajectory Na to Nc of the measuring hole 19. Furthermore, during the period when the measuring hole 19 passes through each passing trajectory Na to Nc (during the passage of the measuring hole 19 from one end W1a to W3a of the workpiece W to the other end W1b to W3b), the thickness measuring unit 22 outputs a thickness data set consisting of multiple consecutive thickness data each time it passes through. Thus, whenever the measuring hole 19 passes through the surface of the workpiece W, the thickness measuring unit 22 outputs a thickness data set consisting of multiple consecutive data obtained by measuring the thickness at each in-plane position of the workpiece W (see reference). Figure 3B Furthermore, the thickness data set output from the thickness measurement unit 22 is input to the shape calculation unit 23.

[0050] Additionally, when inputting thickness data sets, the correlation between measurement data (e.g., GBIR, ESFQD, etc.) measured by an external measuring instrument (external measuring device) measuring the thickness (shape) of the workpiece W and the thickness data sets can be calculated separately, and correction values ​​obtained based on this correlation can be added as input. Furthermore, "GBIR (Global Backside IdealRange)" represents the difference between the maximum and minimum values ​​of the thickness distribution. Additionally, "ESFQD (EdgeSite Front Least Square)" outputs the value that is the larger of the absolute values ​​of the maximum and minimum distances from the reference plane on the outer periphery of the wafer.

[0051] Furthermore, the correlation between the thickness data set or the values ​​of in-plane flatness and peripheral flatness described below and the measurement data of the external measuring instrument is not necessarily high. Therefore, the thickness data set or the in-plane flatness and peripheral flatness can be corrected at any time or at a desired time based on values ​​obtained from a differential database, values ​​obtained by inputting the obtained values ​​into a conversion formula, or values ​​obtained by inputting measurement data measured by an external measuring instrument into a conversion formula. The differential database is a database obtained by database-izing the differences between the thickness data set and the measurement data of the external measuring instrument.

[0052] Furthermore, the differential database can be learned and updated at any time. This allows for understanding the correlation between the thickness data set measured by the thickness measurement unit 22 and the measurement data from an external measuring device. For example, the target value of the workpiece shape input to the control unit 30 can be input using the measurement standard of the external measuring device.

[0053] The shape calculation unit 23 transforms the shape of the workpiece W (hereinafter referred to as "workpiece shape") into numerical information based on the thickness data set of the workpiece W measured by the thickness measurement unit 22. In Embodiment 1, the workpiece shape is represented by in-plane flatness and peripheral flatness. Furthermore, the in-plane flatness and peripheral flatness information output from the shape calculation unit 23 are input to the control unit 30.

[0054] Here, "in-plane flatness" refers to the flatness of the in-plane region G defined on the workpiece W, denoted as an approximate straight line defined for the thickness data set in the in-plane region G, for example, using the least squares method. Figure 4A and Figure 4B The slope value of the approximate straight line (represented by A in the equation). Furthermore, in Figure 4A and Figure 4BIn this context, the approximate straight line represented by A' is also the approximate straight line set for the thickness data set of the in-plane region G. However, when the shape of the workpiece is transformed into numerical information in the shape calculation unit 23, when the workpiece W is viewed in cross-section cut by a straight line passing through the workpiece center Wo, the region to the right of the workpiece center Wo (in-plane region G', outer peripheral region E') is ignored.

[0055] In addition, the approximate straight line is represented by the following equation (1), and the slope value of the approximate straight line is "a" in the following equation (1).

[0056] Approximate straight line: Y = aX + b (1)

[0057] a: Slope value

[0058] b: Intercept

[0059] When the in-plane flatness is positive, the workpiece shape is a convex shape protruding from the center of the in-plane region G. Conversely, when the in-plane flatness is negative, the workpiece shape is a concave shape with the center of the in-plane region G recessed.

[0060] Furthermore, "peripheral flatness" refers to the flatness of the peripheral region E defined on the workpiece W. Peripheral flatness is obtained by performing a continuous processing of the slope value of the approximate straight line defined by the thickness data set of the peripheral region E, based on in-plane flatness. Here, "the continuous processing of the slope value of the approximate straight line defined by the thickness data set of the peripheral region E, based on in-plane flatness" refers to the process of adding or subtracting the slope value of the approximate straight line defined by the thickness data set of the peripheral region E from the in-plane flatness. In other words, "continuous processing" includes adding the in-plane flatness and the slope value of the approximate straight line defined by the thickness data set of the peripheral region E (calculating the sum), and calculating the difference between the in-plane flatness and the slope value of the approximate straight line defined by the thickness data set of the peripheral region E (calculating the difference).

[0061] The peripheral flatness of Example 1 is achieved by setting the thickness data set of the peripheral region E to an approximate straight line ( Figure 4A and Figure 4B The slope of the line (represented by B) is added to the in-plane flatness to obtain the sum of the in-plane flatness and the slope of the approximate line of the outer perimeter region E.

[0062] Furthermore, when the peripheral flatness is positive, the workpiece shape is a raised shape where the peripheral region E bounces up relative to the in-plane region G. Conversely, when the peripheral flatness is negative, the workpiece shape is a drooping shape where the peripheral region E droops down relative to the in-plane region G.

[0063] In addition, such as Figure 3A and Figure 3BAs shown, the "outer peripheral region E" is a defined area extending radially inward from the outer peripheral end We of the workpiece W. For example... Figure 3A and Figure 3B As shown, the "in-plane region G" is the area extending from the radially inner edge of the outer peripheral region E to the workpiece center Wo. That is, as... Figure 3A As shown, the outer peripheral region E is an annular region between the outer peripheral end We of workpiece W and the circle indicated by the dashed line within workpiece W, and is located at the periphery of workpiece W. Additionally, the in-plane region G is a circular region surrounded by the circle indicated by the dashed line within workpiece W, and is located at the center of workpiece W. Furthermore, as... Figure 3B As shown, when viewing a cross-section of workpiece W cut by a straight line passing through the workpiece center Wo, the in-plane region G and the outer peripheral region E are set to be symmetrical across the workpiece center Wo. Furthermore, the range of the outer peripheral region E can be arbitrarily determined.

[0064] and, Figure 4A The diagram shows an approximate straight line generated based on the thickness data set of workpiece W in the first example. Figure 4A In the case shown, the slope of the approximate straight line A in the in-plane region G is +50 (nm), and the slope of the approximate straight line B in the outer peripheral region E is 0 (nm). Therefore, the in-plane flatness of the workpiece W in the first example is +50 (nm), and the outer peripheral flatness is +50 (=+50+0) (nm).

[0065] in addition, Figure 4B The diagram shows an approximate straight line based on the thickness data set of workpiece W in the second example. Figure 4B In the case shown, the slope of the approximate straight line A in the in-plane region G is +30 (nm), and the slope of the approximate straight line B in the outer peripheral region E is -30 (nm). Therefore, the in-plane flatness of the workpiece W in the second example is +30 (nm), and the outer peripheral flatness is 0 (=+30+(-30)) (nm).

[0066] In addition, such as Figure 5 As shown, the outer peripheral region E can also be divided into multiple parts along the radial direction of the workpiece W (in... Figure 5 The example shown has 3 partitions. In this case, the perimeter flatness is calculated for each partition Ei, Eo1, Eo2 of the outer perimeter region E.

[0067] That is, in Figure 5 In the example shown, the designated partition in the outer peripheral region E that is adjacent to the in-plane region G is designated as the innermost partition Ei. Furthermore, the partition adjacent to the innermost partition Ei and radially outer from it is designated as the first outer partition Eo1. Further, the partition adjacent to the first outer partition Eo1 and radially outer from it is designated as the second outer partition Eo2.

[0068] Furthermore, the flatness of the innermost partition Ei is obtained by performing continuous processing on the slope value of the approximate straight line set for the thickness data group of the innermost partition Ei, based on in-plane flatness. Here, the flatness of the innermost partition Ei is obtained by adding the slope value of the approximate straight line set for the thickness data group of the innermost partition Ei to the in-plane flatness. That is, the flatness of the innermost partition Ei in Example 1 is the sum of the in-plane flatness and the slope value of the approximate straight line of the innermost partition Ei.

[0069] Furthermore, the flatness of the first outer partition Eo1 is obtained by performing a continuous process on the slope value of the approximate straight line set for the thickness data group of the first outer partition Eo1, based on the flatness of the partition (innermost partition Ei) adjacent to the radially inner side of the first outer partition Eo1. Here, the flatness of the first outer partition Eo1 is obtained by adding the slope value of the approximate straight line set for the thickness data group of the first outer partition Eo1 to the flatness of the innermost partition Ei. That is, the flatness of the first outer partition Eo1 in Embodiment 1 is the sum of the flatness of the innermost partition Ei and the slope value of the approximate straight line of the first outer partition Eo1.

[0070] Furthermore, the flatness of the second outer partition Eo2 is obtained by performing a continuous process on the slope value of the approximate straight line set for the thickness data group of the second outer partition Eo2, based on the flatness of the partition (first outer partition Eo1) that is radially inner to the second outer partition Eo2. Here, the flatness of the second outer partition Eo2 is obtained by adding the slope value of the approximate straight line set for the thickness data group of the second outer partition Eo2 to the flatness of the first outer partition Eo1. That is, the flatness of the second outer partition Eo2 in Embodiment 1 is the sum of the flatness of the first outer partition Eo1 and the slope value of the approximate straight line of the second outer partition Eo2.

[0071] Furthermore, during the grinding process of the workpiece W, the shape calculation unit 23 continuously calculates numerical information representing the shape of the workpiece (in-plane flatness and peripheral flatness) at arbitrary intervals (e.g., every 10 to 15 seconds). As a result, the control unit 30 can obtain the degree of change in in-plane flatness and the degree of change in peripheral flatness during the grinding process, respectively.

[0072] Here, the degree of change in in-plane flatness is used to... Figure 6 The data set of in-plane flatness, represented by ○, is expressed, for example, by the slope value of an approximate straight line defined by the least squares method; that is, it is expressed as the proportion of change in in-plane flatness over a specified time period. Furthermore, the degree of change in peripheral flatness is expressed as... Figure 6The data set of peripheral flatness, represented by ●, is expressed, for example, by the slope value of an approximate straight line set by the least squares method; that is, it is expressed as the percentage change in peripheral flatness over a specified time period. Furthermore, the control unit 30 determines the trend of workpiece shape change based on the degree of change in in-plane flatness and the degree of change in peripheral flatness.

[0073] That is, when the change in in-plane flatness is positive, the control unit 30 determines that the change trend of the in-plane region G is towards a convex direction. Conversely, when the change in in-plane flatness is negative, the control unit 30 determines that the change trend of the in-plane region G is towards a concave direction. For example, in Figure 6 In the example shown, the degree of change in in-plane flatness at time t0 (the slope of an approximate straight line defined for the in-plane flatness data set within a small time interval encompassed by time t0) is negative, as indicated by the dashed line. Therefore, the trend of change in the in-plane region G after time t0 is determined to be a trend of change towards concavity.

[0074] Furthermore, when the change in peripheral flatness is positive, the control unit 30 determines that the trend of change in peripheral region E is towards a lifting (bounce) direction. Conversely, when the change in peripheral flatness is negative, the control unit 30 determines that the trend of change in peripheral region E is towards a downward sag. For example, in... Figure 6 In the example shown, the degree of change in the perimeter flatness at time t0 (the slope of an approximate straight line defined for the perimeter flatness data set within a small time interval encompassed by time t0) is positive, as indicated by the double-dotted line. Therefore, the trend of change in the perimeter region E after time t0 is determined to be a trend of change towards the upturned direction.

[0075] The control unit 30 is composed of a CPU (Central Processing Unit) and includes a first learner 31, a second learner 32, a main grinding condition correction unit 33, a main grinding stop condition setting unit 34, a control calculation unit 35, and a memory 36. Furthermore, the control unit 30 is connected to an input device 41 operable by the operator of the double-sided grinding apparatus 1 and a display 42 visible to the operator.

[0076] The control unit 30 outputs control commands from the control calculation unit 35 to the first drive device M1 to the fifth drive device M5, based on the measurement results of the workpiece W by the thickness measuring device 20 (which may also include correction values ​​obtained by separately calculating the correlation with the measurement data of the shape of the workpiece W measured by an external measuring device), the processing target of the workpiece W, the preliminary grinding conditions, the auxiliary materials, and other conditions, information related to the device status of the grinding machine 10, the program stored in the memory 36, the main grinding conditions reset by the main grinding condition correction unit 33, and the main grinding stop conditions set by the main grinding stop condition setting unit 34, thereby controlling the operation of the grinding machine 10. Furthermore, the processing target of the workpiece W, the preliminary grinding conditions, the auxiliary materials, and other conditions, as well as information related to the device status of the grinding machine 10, can be input by the operator via the input device 41 or pre-stored in the memory 36. Additionally, during the operation of the grinding machine 10, the control unit 30 causes the display 42 to appropriately display the required information.

[0077] The first learner 31 is a learner that learns the correlation between the main grinding conditions at any point during the main grinding process, the peripheral flatness obtained at that point, and the in-plane flatness obtained at that point. That is, for example, as... Figure 7 As shown, the control unit 30 uses the main grinding conditions at any time during the main grinding process, the peripheral flatness included in the main grinding conditions, interference factors, and the in-plane flatness obtained at the same time as a dataset for the first learner 31 of Embodiment 1 to learn. Here, the main grinding conditions include, for example, the rotational speed of the lower platform 11, the rotational speed of the upper platform 12, the rotational speed of the sun gear 13, the rotational speed of the internal gear 14, the revolution speed of the planetary carrier plate 15, the rotational speed of the planetary carrier plate 15, the load applied to the workpiece W, the flow rate of the slurry, and the type of slurry. In addition, interference factors include, for example, the status of auxiliary materials, the service life of the planetary carrier plate 15 (planetary carrier life), the service life of the grinding pads 11a and 12a (pad life), and dressing conditions; and the device status, including the load rate of each drive device, the temperature of the area (drive chamber) where each drive device is arranged, the temperature of the grinding pads 11a and 12a, the temperature of the slurry, and the variation value of the load applied to the workpiece W. The main grinding conditions or interference factors can be input by the operator via the input device 41 or detected by sensors or the like.

[0078] The second learner 32 is a learner that learns the correlation between the endpoint grinding conditions, the degree of change in in-plane flatness when the main grinding stops, the degree of change in peripheral flatness when the main grinding stops, and the amount of change in peripheral flatness during the endpoint grinding process. That is, for example... Figure 8As shown, the control unit 30 uses the specified endpoint grinding conditions, the degree of change in in-plane flatness when the main grinding stops, the degree of change in peripheral flatness when the main grinding stops, and the amount of change in peripheral flatness during the endpoint grinding process under the endpoint grinding conditions based on the degree of change as a dataset for the second learner 32 of Embodiment 1 to learn. Here, the endpoint grinding conditions include, for example, the rotational speed of the lower platform 11, the rotational speed of the upper platform 12, the rotational speed of the sun gear 13, the rotational speed of the internal gear 14, the revolution speed of the planetary carrier 15, the rotational speed of the planetary carrier 15, the load applied to the workpiece W, the flow rate of the slurry, the type of slurry, and the endpoint grinding time (deceleration time), etc. The endpoint grinding conditions can be input by the operator via the input device 41, or detected by sensors, etc.

[0079] Furthermore, the "degree of change in in-plane flatness at the time of main grinding stop" is the slope value of an approximate straight line set for a set of in-plane flatness data within a small time period including the time when main grinding stops (the moment when main grinding stops). The "degree of change in peripheral flatness at the time of main grinding stop" is the slope value of an approximate straight line set for a set of peripheral flatness data within a small time period including the time when main grinding stops. Moreover, the "amount of change in peripheral flatness during the final grinding process" is the difference between the peripheral flatness at the time of main grinding stop and the peripheral flatness at the time of final grinding end.

[0080] During the main grinding process, the main grinding condition correction unit 33 calculates a correction target value for in-plane flatness based on the estimated value of in-plane flatness obtained by inputting the calculated value of the main grinding conditions set at the correction time and the peripheral flatness at the correction time into the first learner 31. This correction target value aims to make the calculated value of in-plane flatness approach the target value. Then, the main grinding condition correction unit 33 corrects the main grinding conditions set at the correction time based on the main grinding conditions obtained by inputting the correction target value of in-plane flatness and the target value of the final peripheral flatness at the end of grinding into the first learner 31. In other words, the main grinding condition correction unit 33 uses the first learner 31 to obtain main grinding conditions that satisfy the target value of peripheral flatness and the correction target value, and replaces the obtained main grinding conditions with the main grinding conditions set at the correction time, resetting them as new main grinding conditions. Furthermore, under the newly reset master grinding conditions, at least one of the following is changed: the rotational speed of the lower platform 11, the rotational speed of the upper platform 12, the rotational speed of the sun gear 13, the rotational speed of the internal gear 14, the revolution speed of the planetary carrier 15, the rotational speed of the planetary carrier 15, the load applied to the workpiece W, the flow rate of the slurry, the type of slurry, the destination of the slurry.

[0081] Here, the "correction target value for in-plane flatness used to make the calculated value of in-plane flatness close to the target value" is obtained through the following steps. Specifically, the main grinding condition correction unit 33 first inputs the main grinding conditions at the correction time and the calculated value of the peripheral flatness at the correction time into the first learner 31 to obtain a predicted value for in-plane flatness. Next, the main grinding condition correction unit 33 calculates the difference between the calculated value of in-plane flatness at the correction time and the predicted value of in-plane flatness. Furthermore, the main grinding condition correction unit 33 uses the value calculated by subtracting the difference value of in-plane flatness from the target value of final in-plane flatness at the end of grinding as the "correction target value".

[0082] During the main grinding process, the main grinding stop condition setting unit 34 inputs the final grinding condition, the degree of change in in-plane flatness at the calculation time, and the degree of change in in-plane flatness at the calculation time into the second learner 32 to obtain the amount of change in peripheral flatness during the final grinding process. Furthermore, based on the obtained amount of change in peripheral flatness and the target value of the final peripheral flatness at the grinding end time, the main grinding stop condition (main grinding stop condition) is set. Here, the main grinding stop condition is defined based on the peripheral flatness at the main grinding stop time.

[0083] That is, the main grinding stop condition setting unit 34 inputs the endpoint grinding condition, the degree of change in in-plane flatness at the calculation time (the slope value of an approximate straight line set for the data set of in-plane flatness within a small time period including the calculation time), and the degree of change in peripheral flatness at the calculation time (the slope value of an approximate straight line set for the data set of peripheral flatness within a small time period including the calculation time) into the second learner 32 to obtain the amount of change in peripheral flatness during the endpoint grinding process based on the degree of change. Next, the main grinding stop condition setting unit 34 performs inverse calculation based on the obtained amount of change in peripheral flatness, the target value of the final peripheral flatness at the end of grinding, and the endpoint grinding time to calculate the target value of the peripheral flatness at the endpoint grinding stop. Furthermore, the calculated target value of the peripheral flatness is defined as the main grinding stop condition.

[0084] Furthermore, during the initial grinding process, the control unit 30 outputs control commands from the control calculation unit 35 to the first drive device M1 to the fifth drive device M5, which correspond to the initial grinding conditions set in advance.

[0085] In addition, when the main grinding begins, the control unit 30 outputs control commands from the control calculation unit 35 to the first drive device M1 to the fifth drive device M5, which correspond to the main grinding conditions that were previously preset for the pre-grinding conditions.

[0086] Furthermore, during the execution of the main grinding, the control unit 30 corrects the main grinding conditions through the first learner 31 and the main grinding condition correction unit 33. When the main grinding conditions are corrected, the control unit 30 outputs control commands corresponding to the corrected main grinding conditions (the newly set main grinding conditions) to the first drive device M1 to the fifth drive device M5 from the control calculation unit 35.

[0087] Furthermore, during the execution of the main grinding, the control unit 30 sets the main grinding stop condition through the second learner 32 and the main grinding stop condition setting unit 34. When the main grinding stop condition is met, the control unit 30 outputs a control command from the control calculation unit 35 to the first drive device M1 to the fifth drive device M5 to stop the main grinding and execute the final grinding.

[0088] In addition, during the final grinding process, the control unit 30 outputs control commands from the control calculation unit 35 to the first drive device M1 to the fifth drive device M5, which correspond to the final grinding conditions set in advance for the pre-grinding conditions.

[0089] Figure 9 This is a flowchart illustrating the grinding process executed by the control unit 30 of the double-sided grinding apparatus 1 of Embodiment 1. Hereinafter, based on... Figure 9 The grinding process of Example 1 will be described. Furthermore, the grinding process is performed with the planetary carrier plate 15 and the workpiece W placed in the grinding machine 10. Additionally, during the grinding process (from the start of the initial grinding to the end of the final grinding), the thickness measuring device 20 continuously inputs the shape information (in-plane flatness and peripheral flatness information) of the workpiece W into the control unit 30.

[0090] In step S1, the control unit 30 sets the target values ​​for the final in-plane flatness and the final peripheral flatness at the end of the final grinding phase, which are the final machining targets for the workpiece W, and proceeds to step S2. Here, the target values ​​for in-plane flatness and peripheral flatness are input by the operator via the input device 41. Furthermore, the machining targets for the workpiece W can also be preset according to the type of workpiece W and stored in the memory 36. In this case, the control unit 30 reads the machining targets for the workpiece W from the memory 36 based on the type of workpiece W input by the operator and sets them.

[0091] In step S2, following the setting of the workpiece processing target in step S1, the control unit 30 sets the prerequisite grinding conditions and proceeds to step S3. Here, prerequisite grinding conditions refer to various conditions pre-set according to the processing target and type of workpiece W, which are prerequisites for grinding workpiece W. Prerequisite grinding conditions include, for example, initial grinding conditions including grinding time and grinding end conditions in initial grinding, main grinding conditions at the start time of main grinding, final grinding conditions including grinding time and grinding end conditions in final grinding, slurry-related information such as slurry flow rate and slurry type, slurry supply destination, and information related to the state of auxiliary material consumption such as planetary carrier life. Prerequisite grinding conditions are input by the operator via input device 41 or read from memory 36.

[0092] In step S3, following the setting of the prerequisite grinding conditions in step S2 or the determination in step S4 that the initial grinding has not ended, the control unit 30 outputs control commands from the control calculation unit 35 to the first drive device M1 to the fifth drive device M5 corresponding to the initial grinding conditions specified according to the prerequisite grinding conditions, executes the initial grinding, and proceeds to step S4. Furthermore, during the execution of the initial grinding, in the grinding machine 10, the load applied to the workpiece W is gradually increased, and the rotational speeds of the lower platform 11 and the upper platform 12, the rotational speed of the sun gear 13, and the rotational speed of the internal gear 14 are also gradually increased while grinding the workpiece W.

[0093] In step S4, following the initial grinding in step S3, the control unit 30 determines whether the initial grinding has ended. If the determination is yes (initial grinding has ended), the control unit 30 proceeds to step S5; if the determination is no (initial grinding has not ended), it returns to step S3. Here, the control unit 30 determines that the initial grinding has ended when the load applied to the workpiece W reaches a predetermined load, the rotational speed of the lower platform 11, etc., reaches a predetermined rotational speed, etc., thus confirming that the initial grinding has ended.

[0094] In step S5, following the determination that the initial grinding is complete in step S4, the control unit 30 outputs control commands from the control calculation unit 35 to the first drive device M1 to the fifth drive device M5 corresponding to the main grinding conditions specified according to the prerequisite grinding conditions, and begins to execute the main grinding, proceeding to steps S6 and S7. Furthermore, during the execution of the main grinding, in the grinding machine 10, the load applied to the workpiece W is adjusted to the target load specified by the main grinding conditions, and the rotational speeds of the lower platform 11 and upper platform 12, the rotational speed of the sun gear 13, and the rotational speed of the internal gear 14 are adjusted to the target rotational speeds while grinding the workpiece W.

[0095] In step S6, following the commencement of main grinding in step S5, the control unit 30 corrects the main grinding conditions via the main grinding condition correction unit 33. Furthermore, the correction of the main grinding conditions is performed in parallel with the setting of the main grinding stop condition in step S7 below, and is repeated at regular intervals (e.g., approximately 300 seconds) until it is determined in step S8 below that the main grinding has stopped.

[0096] Here, the main grinding condition correction unit 33 corrects the main grinding conditions according to the following steps.

[0097] (1) Input the calculated values ​​of the main grinding conditions at the correction time and the peripheral flatness at the correction time into the first learner 31 to obtain the predicted value of the in-plane flatness at the correction time under the main grinding conditions.

[0098] (2) Subtract the predicted value of the in-plane flatness obtained in (1) from the calculated value of the in-plane flatness at the correction time to calculate the difference value of the in-plane flatness (the difference between the predicted value of the in-plane flatness and the current value).

[0099] (3) Calculate the value obtained by subtracting the difference value of the in-plane flatness calculated in (2) from the target value of in-plane flatness set in step S1, and set it as the "corrected target value".

[0100] (4) Input the “correction target value” set in (3) and the target value of peripheral flatness set in step S1 into the first learner 31 to obtain the main grinding conditions.

[0101] (5) Replace the main grinding conditions obtained in (4) with the main grinding conditions set at the correction time and reset them to new main grinding conditions to correct the main grinding conditions.

[0102] In step S7, following the commencement of main grinding in step S5, the control unit 30 sets the main grinding stop condition via the main grinding stop condition setting unit 34. Furthermore, the setting of the main grinding stop condition is performed in parallel with the correction of the main grinding condition in step S6, as described above, and is repeated at regular intervals (e.g., approximately one second) until it is determined that the main grinding has stopped in step S8 below.

[0103] Here, the main grinding stop condition setting unit 34 sets the main grinding stop condition according to the following steps.

[0104] (1) Input the final grinding conditions determined by the premise grinding conditions set in step S2, the degree of change of in-plane flatness at the calculation time, and the degree of change of peripheral flatness at the calculation time into the second learner 32 to obtain the change in peripheral flatness during the final grinding process (in addition, the “calculation time” here refers to the setting time of the main grinding stop condition).

[0105] (2) Based on the change in peripheral flatness during the final grinding process obtained in step S1, the target value of peripheral flatness set in step S1 and the grinding time of the final grinding, perform inverse calculation to find the peripheral flatness required at the main grinding stop time (target value of peripheral flatness at the main grinding stop time) in order to satisfy the final target value of peripheral flatness at the end of grinding.

[0106] (3) The peripheral flatness obtained in (2) is set as the main grinding stop condition. In addition, the calculation of peripheral flatness in (2) is performed continuously at certain intervals. Therefore, the main grinding stop condition is updated each time peripheral flatness is calculated.

[0107] In step S8, following the setting of the main grinding stop condition in step S7, the control unit 30 determines whether to stop the main grinding. If the determination is yes (main grinding stops), the control unit 30 proceeds to step S9; if the determination is no (main grinding continues), it returns to step S7. Here, the main grinding stop condition is that the calculated value of the peripheral flatness reaches the peripheral flatness set as the main grinding stop condition in step S7. Therefore, in step S8, it is determined whether the current peripheral flatness is consistent with the peripheral flatness calculated in step S7.

[0108] In step S9, following the determination in step S8 that the main grinding has stopped, the control unit 30 stops the main grinding and outputs control commands from the control calculation unit 35 to the first drive device M1 to the fifth drive device M5 corresponding to the endpoint grinding conditions specified according to the prerequisite grinding conditions, and starts executing the endpoint grinding and proceeds to step S10. Furthermore, during the execution of the endpoint grinding, in the grinding machine 10, the load applied to the workpiece W is gradually reduced, and the workpiece W is ground while the rotational speeds of the lower platform 11 and upper platform 12, the rotational speed of the sun gear 13, and the rotational speed of the internal gear 14 are reduced respectively.

[0109] In step S10, after the final grinding begins in step S9, the control unit 30 determines whether the final grinding has ended. Then, if the determination is yes (final grinding has ended), the control unit 30 proceeds to step S11; if the determination is no (final grinding has not ended), it returns to step S9. Here, the control unit 30 determines that the final grinding has ended when the final grinding time has elapsed, the load applied to the workpiece W has become below a predetermined value, etc., because the conditions for ending the final grinding specified in the previous grinding conditions have been met.

[0110] In step S11, after determining that the grinding has ended in step S10, the control unit 30 stops the grinding machine 10 from grinding the workpiece W, records various grinding data into the memory 36 and enters the end state.

[0111] Hereinafter, the functions of the double-sided grinding apparatus 1 in Example 1 will be explained in terms of "correction and control of main grinding conditions", "control of main grinding stop conditions" and "other control functions".

[0112] [The effect of correcting and controlling the main grinding conditions]

[0113] In the double-sided grinding apparatus 1 of Example 1, the in-plane flatness and the peripheral flatness change separately during the main grinding process. Therefore, if the main grinding conditions are not set considering the correlation between the in-plane flatness and the peripheral flatness, it is difficult to control both the in-plane flatness and the peripheral flatness with high precision.

[0114] That is, in the double-sided grinding apparatus (hereinafter referred to as the "comparative double-sided grinding apparatus") that only learns the correlation between the in-plane flatness and the main grinding conditions when the peripheral flatness is 0 (nm), and inputs the target value of the in-plane flatness to the learner of the comparative example to obtain the main grinding conditions, the peripheral flatness is not considered. Therefore, a suitable workpiece shape cannot be formed.

[0115] To explain in more detail, for example, suppose there is a change in the shape of the workpiece under each master grinding condition as follows: Figure 10 The workpiece W shown is referred to as "Example Workpiece W'". Specifically, when Example Workpiece W' is subjected to main grinding under main grinding condition A, its in-plane flatness is +80 (nm) when the peripheral flatness is 0 (nm), +90 (nm) when the peripheral flatness is -20 (nm), and +100 (nm) when the peripheral flatness is -50 (nm). Furthermore, when main grinding is performed under main grinding condition B, its in-plane flatness is 0 (nm) when the peripheral flatness is 0 (nm), +10 (nm) when the peripheral flatness is -20 (nm), and +20 (nm) when the peripheral flatness is -50 (nm). In addition, when performing main grinding under main grinding condition C, the in-plane flatness is -20 (nm) when the peripheral flatness is 0 (nm), the in-plane flatness is 0 (nm) when the peripheral flatness is -20 (nm), and the in-plane flatness is +10 (nm) when the peripheral flatness is -50 (nm).

[0116] In contrast, as described above in the comparative example's double-sided polishing apparatus, the learner in the comparative example only learns the correlation between in-plane flatness and main polishing conditions when the peripheral flatness is 0 (nm). Therefore, as Figure 11As shown, the dataset that the learner of the comparison example learned is in-plane flatness +80 (nm) for main grinding condition A, in-plane flatness 0 (nm) for main grinding condition B, and in-plane flatness -20 (nm) for main grinding condition C.

[0117] Here, we consider the case where, during the main grinding process, the example workpiece W' is ground in a manner where the in-plane flatness is 0 (nm) when the peripheral flatness is -20 (nm). In this case, in the comparative example's double-sided grinding apparatus, since the peripheral flatness cannot be considered, it is ignored, and the result of inputting the target value of the in-plane flatness (0 (nm) in this case) to the learner of the comparative example is obtained as "Main Grinding Condition B". However, if the main grinding of the example workpiece W' is performed with the main grinding condition set to B, in the comparative example's double-sided grinding apparatus, during the main grinding process, it is predicted that the in-plane flatness is +10 (nm) when the peripheral flatness is -20 (nm). In addition, it is predicted that if the in-plane flatness is 0 (nm), the peripheral flatness will be 0 (nm). That is, in the comparative example's double-sided grinding apparatus, it is impossible to make both the in-plane flatness and peripheral flatness consistent with the target value with high precision.

[0118] In contrast, the double-sided grinding apparatus 1 of Embodiment 1 includes a first learner 31 and a main grinding condition correction unit 33. Here, the first learner 31 is a learner that learns the correlation between the main grinding conditions, peripheral flatness, and in-plane flatness. That is, as... Figure 10 As shown, the dataset learned by the first learner 31 in Embodiment 1 includes various master grinding conditions, peripheral flatness, and in-plane flatness under those conditions. Furthermore, the master grinding condition correction unit 33 corrects the master grinding conditions set at the correction time during the master grinding process based on the master grinding conditions obtained by inputting the correction target value and the target value of the final peripheral flatness into the first learner 31. The correction target value is a value calculated based on the predicted value of the in-plane flatness obtained by inputting the calculated values ​​of the master grinding conditions set at the correction time and the peripheral flatness at the correction time into the first learner 31.

[0119] That is, in the double-sided grinding apparatus 1 of Embodiment 1, during the execution of the main grinding, the main grinding conditions at the correction time and the calculated value of the peripheral flatness at the correction time are input into the first learner 31. Next, the double-sided grinding apparatus 1 of Embodiment 1 obtains the predicted value of the in-plane flatness at the correction time based on the input information input to the first learner 31 and the learning result of the first learner 31, and calculates the difference between the predicted and actual in-plane flatness at the correction time. Then, the double-sided grinding apparatus 1 of Embodiment 1 subtracts the difference value from the target value of the in-plane flatness to obtain the "correction target value". Then, the "correction target value" of the in-plane flatness and the target value of the peripheral flatness are input into the first learner 31 to obtain new main grinding conditions, and the main grinding conditions set at the correction time are corrected (step S6).

[0120] Hereinafter, we assume that the dataset for the first learner 31 to learn is... Figure 10 The content shown is specifically illustrated by taking the double-sided grinding apparatus 1 of Example 1 as an example, where the workpiece W is subjected to main grinding with the goal of achieving an in-plane flatness of 0 (nm) and an outer peripheral flatness of 0 (nm) at the end of the final grinding.

[0121] In this case, since the target values ​​at the final grinding end time are in-plane flatness of 0 (nm) and peripheral flatness of 0 (nm), the main grinding condition B is set to begin the main grinding. Then, during the main grinding process with main grinding condition B, when the calculated value of peripheral flatness is -20 (nm), the main grinding condition at the time of correction is adjusted, i.e., "main grinding condition B" and the calculated value of peripheral flatness at the time of correction, i.e., "-20 (nm)", are input into the first learner 31 (refer to...). Figure 10 The estimated value for in-plane flatness obtained from the result is "+10 (nm)".

[0122] In contrast, it is assumed that due to factors such as interference, the calculated value of in-plane flatness (actual in-plane flatness) is +30 (nm).

[0123] In this case, the estimated value of in-plane flatness is "+10 (nm)", while the calculated value of in-plane flatness is "+30 (nm)". Therefore, it can be inferred that the actual in-plane flatness during the main grinding process becomes a value that is +20 (nm) larger than the learning result of the first learner 31.

[0124] Therefore, in the double-sided polishing apparatus 1 of Embodiment 1, the difference (+20 (nm)) between the estimated value of in-plane flatness "+10 (nm)" and the calculated value of in-plane flatness "+30 (nm)" is calculated. Then, the "corrected target value (-20 (nm))" is obtained by subtracting the difference (+20 (nm)) from the target value of in-plane flatness (0 (nm)). Then, the double-sided polishing apparatus 1 of Embodiment 1 sends the result to the first learner 31 (refer to...). Figure 10 The input values ​​are the target values ​​for in-plane flatness correction (-20 nm) and peripheral flatness correction (0 nm). As a result, a new master polishing condition, "Master Polishing Condition C," is obtained. Then, instead of the master polishing condition (Master Polishing Condition B) set at the correction time, "Master Polishing Condition C" is set as the new master polishing condition.

[0125] Furthermore, if the main grinding conditions are modified to main grinding condition C and the main grinding is performed, the in-plane flatness is predicted to be -20 (nm) when the peripheral flatness is 0 (nm) in the learning result of the first learner 31. However, as described above, since the actual in-plane flatness in the main grinding process is predicted to be +20 (nm) greater than the learning result of the first learner 31, the double-sided grinding apparatus 1 can make the in-plane flatness 0 (nm) when the peripheral flatness is 0 (nm). Thus, in the double-sided grinding apparatus 1 of Embodiment 1, during the execution of the main grinding, the correlation between in-plane flatness and peripheral flatness can be considered to modify the main grinding conditions, thereby enabling high-precision control of both in-plane flatness and peripheral flatness.

[0126] "The controlling role of the main grinding stop condition"

[0127] It is known that in the double-sided grinding apparatus 1 of Embodiment 1, the workpiece shape, particularly the outer peripheral flatness, changes during the final grinding. Therefore, even if the workpiece shape reaches the target shape at the moment the main grinding stops, the outer peripheral flatness may change due to the final grinding, resulting in a final workpiece shape deviating from the target shape. Furthermore, if the degree of change in the workpiece at the moment the main grinding stops is different, the amount of change in outer peripheral flatness during the final grinding process will differ even if the same final grinding conditions are used. As a result, it is conceivable that the final workpiece shape will differ significantly from the workpiece shape at the moment the main grinding stops, deviating substantially from the target shape. Therefore, it is necessary to stop the main grinding based on the degree of change in the workpiece shape, taking into account the amount of change in outer peripheral flatness.

[0128] In contrast, the double-sided grinding apparatus 1 of Embodiment 1 includes a second learner 32 and a main grinding stop condition setting unit 34. Here, the second learner 32 learns the correlation between the endpoint grinding conditions, the degree of change in in-plane flatness at the start of main grinding, the degree of change in peripheral flatness at the start of main grinding, and the amount of change in peripheral flatness during the endpoint grinding process. Furthermore, the main grinding stop condition setting unit 34 sets the peripheral flatness of the workpiece W as the target peripheral flatness as the main grinding stop condition (main grinding stop condition). The target peripheral flatness is calculated based on the amount of change in peripheral flatness during the endpoint grinding process obtained by inputting the endpoint grinding conditions, the degree of change in in-plane flatness at the calculation time, and the degree of change in peripheral flatness at the calculation time into the second learner 32, and the target value of the peripheral flatness at the end of the endpoint grinding process.

[0129] That is, in the double-sided grinding apparatus 1 of Embodiment 1, when setting the main grinding stop condition, the end-point grinding condition, the degree of change in in-plane flatness at the calculation time, and the degree of change in peripheral flatness at the calculation time are input into the second learner 32. Then, the double-sided grinding apparatus 1 of Embodiment 1 obtains the change in peripheral flatness during the end-point grinding process based on the input information to the second learner 32 and the learning result of the second learner 32. Then, based on the change in peripheral flatness during the end-point grinding process, the target value of the final peripheral flatness, and the end-point grinding time, inverse calculation is performed to obtain the target value of peripheral flatness at the main grinding stop time (target peripheral flatness) required to meet the target value of peripheral flatness. Then, the main grinding stop condition is determined by the target peripheral flatness (step S7).

[0130] Thus, in the double-sided grinding apparatus 1 of Embodiment 1, the main grinding can be stopped such that the peripheral flatness at the time of stopping the main grinding is the desired shape. This desired shape takes into account the amount of change in peripheral flatness produced by the final grinding based on the degree of change in in-plane flatness and peripheral flatness. Therefore, even if the shape of the workpiece W changes due to the final grinding, the double-sided grinding apparatus 1 of Embodiment 1 can prevent the workpiece shape from deviating significantly from the target shape at the end of the final grinding (final grinding end time).

[0131] also, Figure 12 This is a summary of the shape of the workpiece W after final grinding, in the double-sided grinding apparatus 1 of Example 1, where the target value of the outer peripheral flatness at the end of grinding was set to 0 (nm). For example... Figure 12As shown, in the first sample workpiece, the in-plane flatness is +34 (nm) and the peripheral flatness is -2 (nm). In the second sample workpiece, the in-plane flatness is -4 (nm) and the peripheral flatness is -5 (nm). In the third sample workpiece, the in-plane flatness is +40 (nm) and the peripheral flatness is -3 (nm). In the fourth sample workpiece, the in-plane flatness is +62 (nm) and the peripheral flatness is -3 (nm).

[0132] Based on these results, the double-sided grinding apparatus 1 of Example 1 can reduce the difference between the actual peripheral flatness and the target value (0 (nm)) regardless of the magnitude of the in-plane flatness. Therefore, it can be seen that the double-sided grinding apparatus 1 of Example 1 can prevent the workpiece shape from deviating significantly from the target shape at the end of the final grinding (final grinding end time).

[0133] in addition, Figure 13 The image shows the grinding result of workpiece W at the final grinding end time when grinding workpiece W using the double-sided grinding apparatus 1 of Example 1. Figure 13 As shown, in the fifth sample workpiece, the target value for in-plane flatness was set to 0 (nm), and the target value for peripheral flatness was set to 0 (nm). The actual grinding result was -1 (nm) for in-plane flatness and +3 (nm) for peripheral flatness. In the sixth sample workpiece, the target value for in-plane flatness was set to 0 (nm), and the target value for peripheral flatness was set to -20 (nm). The actual grinding result was -4 (nm) for in-plane flatness and -20 (nm) for peripheral flatness. In the seventh sample workpiece, the target value for in-plane flatness was set to -30 (nm), and the target value for peripheral flatness was set to 0 (nm). The actual grinding result was -30 (nm) for in-plane flatness and +2 (nm) for peripheral flatness. In the eighth sample workpiece, the target value for in-plane flatness was set to +30 (nm), and the target value for peripheral flatness was set to 0 (nm). The actual grinding result was +30 (nm) for in-plane flatness and +3 (nm) for peripheral flatness. In addition, in the 9th sample workpiece, the target value of in-plane flatness was set to 0 (nm) and the target value of peripheral flatness was set to 0 (nm). The actual grinding results were -5 (nm) for in-plane flatness and -4 (nm) for peripheral flatness.

[0134] according to Figure 13 The results show that the double-sided grinding apparatus 1 of Example 1 does not cause both the in-plane flatness and the outer peripheral flatness to deviate significantly from the target value, and can suppress the workpiece shape from deviating from the target shape at the end of the final grinding.

[0135] [Other control functions]

[0136] In the double-sided grinding apparatus 1 of Embodiment 1, an in-plane region G and an outer peripheral region E are defined on the workpiece W, and the workpiece shape is numerically represented and displayed based on the in-plane flatness and the outer peripheral flatness. That is, in Embodiment 1, the workpiece shape is divided into multiple regions for control. On the other hand, the workpiece W ground by the double-sided grinding apparatus 1 of Embodiment 1 is sometimes used as a substrate for semiconductor elements on which fine electronic circuits are formed on the surface. Here, the electronic circuits formed on the surface of the workpiece W are sometimes formed across the boundary between the in-plane region G and the outer peripheral region E. Therefore, the boundary of the workpiece shape is preferably a smooth surface so that no inflection point is formed at the boundary between the in-plane region G and the outer peripheral region E. Furthermore, in order to control the workpiece shape so that no inflection point is formed at the boundary between the in-plane region G and the outer peripheral region E, it is necessary to numerically represent the flatness (smoothness of the workpiece shape) at the boundary between different regions such as the in-plane region G and the outer peripheral region E, that is, the slope change of the workpiece shape at the boundary between the in-plane region G and the outer peripheral region E.

[0137] In contrast, in the double-sided grinding apparatus 1 of Embodiment 1, the in-plane flatness is set as the slope value of an approximate straight line defined by the thickness data set of the in-plane region G. Furthermore, the peripheral flatness is obtained by performing a continuous processing of the slope value of the approximate straight line defined by the thickness data set of the peripheral region E, based on the in-plane flatness. That is, in the double-sided grinding apparatus 1 of Embodiment 1, the peripheral flatness is set as the sum of the slope value of the approximate straight line defined by the thickness data set of the peripheral region E and the in-plane flatness.

[0138] Thus, the double-sided grinding apparatus 1 of Embodiment 1 calculates both in-plane flatness and peripheral flatness. However, by performing continuous processing on the slope value of the approximate straight line set for the thickness data set of the peripheral region E, with in-plane flatness as the reference, the peripheral flatness is obtained. This allows the slope value of the approximate straight line set for the data set of the in-plane region G to be used as a reference line, and the slope value of the approximate straight line set for the data set of the peripheral region E relative to this reference line to be numerically converted into peripheral flatness. Therefore, peripheral flatness can be expressed as an index representing the degree of peripheral sag relative to in-plane flatness, and the flatness (smoothness of the workpiece shape) at the boundary of different regions such as the in-plane region G and the peripheral region E can be numerically represented. Furthermore, the double-sided grinding apparatus 1 of Embodiment 1 can numerically represent the shape of the workpiece W, making it easier for operators to grasp the workpiece shape.

[0139] Furthermore, the peripheral flatness can also be obtained by subtracting the slope of the approximate straight line defined by the thickness data set for the peripheral region E from the in-plane flatness. In other words, the peripheral flatness can be the difference between the in-plane flatness and the slope of the approximate straight line defined by the thickness data set for the peripheral region E.

[0140] Furthermore, the double-sided grinding apparatus 1 of Embodiment 1 can also divide the outer peripheral region E into multiple partitions along the radial direction of the workpiece W, and calculate the peripheral flatness for each partitioned outer peripheral region (innermost partition Ei, outermost partition Eo). By dividing the outer peripheral region E into multiple partitions, the double-sided grinding apparatus 1 of Embodiment 1 can correct the main grinding conditions or set the main grinding stop conditions according to the shape changes of each partition. Therefore, the shape of the workpiece at the end of the final grinding process can be made to match the target shape with higher precision.

[0141] In addition, in the double-sided grinding apparatus 1 of Embodiment 1, when the outer peripheral region E is divided into multiple partitions, the partition adjacent to the in-plane region G is designated as the innermost partition Ei, and the partition that is radially outward from the innermost partition Ei is designated as the outer partition Eo.

[0142] Furthermore, the flatness of the innermost partition Ei is obtained by performing a continuous processing of the slope value of an approximate straight line defined for the thickness data set of the innermost partition Ei, based on in-plane flatness. In Example 1, the flatness of the innermost partition Ei is the sum of the slope value of the approximate straight line defined for the thickness data set of the innermost partition Ei and the in-plane flatness. Additionally, the flatness of the outermost partition Eo is obtained by performing a continuous processing of the slope value of an approximate straight line defined for the thickness data set of the outermost partition Eo, based on the flatness of the partition radially inner to the outermost partition Eo (e.g., the innermost partition Ei). In Example 1, the flatness of the outermost partition Eo is the sum of the slope value of the approximate straight line defined for the thickness data set of the outermost partition Eo and the flatness of the partition radially inner to the outermost partition Eo (e.g., the innermost partition Ei).

[0143] Therefore, in the double-sided grinding apparatus 1 of Embodiment 1, even when the outer peripheral region E is divided into multiple partitions, the flatness (smoothness of the workpiece shape) of the outer peripheral region E can be appropriately expressed, and the workpiece shape can be controlled with high precision.

[0144] Furthermore, the flatness of the innermost partition Ei can be obtained by subtracting the slope of the approximate straight line set for the thickness data set of the innermost partition Ei from the in-plane flatness. In other words, the flatness of the innermost partition Ei can be the difference between the in-plane flatness and the slope of the approximate straight line set for the thickness data set of the innermost partition Ei.

[0145] Furthermore, the flatness of the outer partition Eo can be calculated by subtracting the slope of the approximate straight line defined for the thickness data set of the outer partition Eo from the flatness of the partition radially inner to the outer partition Eo (e.g., the innermost partition Ei). In other words, the flatness of the outer partition Eo can be the difference between the flatness of the partition radially inner to the outer partition Eo (e.g., the innermost partition Ei) and the slope of the approximate straight line defined for the thickness data set of the outer partition Eo.

[0146] In addition, in Example 1, the outer partition Eo is divided into two partitions: a first outer partition Eo1 and a second outer partition Eo2. However, the outer partition Eo may not be divided, or it may be divided into three or more partitions.

[0147] The double-sided grinding apparatus of the present invention has been described above based on Embodiment 1. However, the specific structure is not limited to this embodiment. As long as it does not depart from the spirit of the invention of each claim, design changes or additions are permitted.

[0148] In the double-sided grinding apparatus 1 of Example 1, Figure 9 In step S1 of the grinding process control shown, an example is illustrated where, when setting target values ​​for in-plane flatness and peripheral flatness, the target values ​​for final in-plane flatness and final peripheral flatness at the end of the final grinding phase, which are the final processing targets for workpiece W, are set. However, the target values ​​for in-plane flatness and peripheral flatness are not limited to these. For example, multiple target values ​​corresponding to the grinding condition of workpiece W can be set, and the target values ​​can be appropriately made different when modifying the main grinding conditions and setting the main grinding stop conditions.

[0149] Cross-references of related applications

[0150] This application claims priority based on Japanese Patent Application No. 2023-204053 filed with the Japan Patent Office on December 1, 2023, all of the disclosures of which are incorporated herein by reference in their entirety.

Claims

1. A double-sided grinding device, characterized in that, have: A grinding machine holds a circular plate-shaped workpiece between a lower platform and an upper platform disposed opposite to the lower platform. Under a load applied to the workpiece, the lower platform and the upper platform are moved relative to the workpiece to grind the front and back sides of the workpiece. A thickness measuring device is used to measure the thickness of the workpiece during the grinding process of the grinding machine. as well as The control unit controls the grinding machine based on the measurement results of the thickness measuring device. The workpiece has a defined outer peripheral region extending radially inward from its outer peripheral end and an in-plane region extending from the outer peripheral region to the center of the workpiece. The control unit has: The first learner learns the grinding conditions of the main grinding process while maintaining the load at the target load, and the correlation between the flatness of the outer peripheral region (i.e., outer peripheral flatness) and the flatness of the in-plane region (i.e., in-plane flatness). The second learner learns the correlation between the grinding conditions of the final grinding of the workpiece while gradually reducing the load, the degree of change of the in-plane flatness at the stop time of the main grinding, the degree of change of the peripheral flatness at the stop time of the main grinding, and the amount of change of the peripheral flatness during the final grinding process. The main grinding condition correction unit calculates a correction target value for the in-plane flatness based on the predicted value of the in-plane flatness obtained by inputting the grinding conditions of the main grinding set at the correction time and the peripheral flatness at the correction time into the first learner, and corrects the grinding conditions of the main grinding set at the correction time according to the grinding conditions of the main grinding obtained by inputting the correction target value of the in-plane flatness and the target value of the peripheral flatness into the first learner. as well as The main grinding stop condition setting unit inputs the grinding conditions of the endpoint grinding, the degree of change of the in-plane flatness at the calculation time, and the degree of change of the peripheral flatness at the calculation time into the second learner to obtain the amount of change of the peripheral flatness during the endpoint grinding process, and sets the stop condition of the main grinding based on the amount of change of the peripheral flatness and the target value of the peripheral flatness.

2. The double-sided grinding apparatus according to claim 1, characterized in that, The in-plane flatness is the slope value of an approximate straight line set by the thickness data set of the in-plane region, and the peripheral flatness is obtained by performing a continuous processing of the slope value of the approximate straight line set by the thickness data set of the peripheral region with the in-plane flatness as a reference.

3. The double-sided grinding apparatus according to claim 1 or 2, characterized in that, The outer peripheral region is divided into multiple partitions along the radial direction of the workpiece, and the outer peripheral flatness is calculated for each partition of the divided outer peripheral region.

4. The double-sided grinding apparatus according to claim 3, characterized in that, When the partition adjacent to the in-plane region is designated as the innermost partition, and the partition radially outer of the innermost partition is designated as the outer partition, The flatness of the innermost partition is determined by performing a continuous process on the slope of an approximate straight line set for the thickness data set of the innermost partition, based on the in-plane flatness. The flatness of the outer partition is determined by performing a continuous process on the slope of an approximate straight line set for the thickness data set of the outer partition, with the flatness of the partition adjacent to the radially inner side of the outer partition as a reference.

5. A double-sided grinding method, A circular plate-shaped workpiece is clamped between a lower platform and an upper platform opposite to the lower platform. While a load is applied to the workpiece, the lower platform and the upper platform are moved relative to the workpiece to grind the front and back sides of the workpiece. The double-sided grinding method is characterized by using: The first learner learns the grinding conditions of the main grinding process while maintaining the load at the target load, and the correlation between the flatness of the outer peripheral region (i.e., the outer peripheral flatness) set as a specified range from the outer peripheral end of the workpiece toward the radially inward side, and the flatness of the in-plane region (i.e., the in-plane flatness) from the outer peripheral region of the workpiece to the center of the workpiece. as well as The second learner learns the correlation between the grinding conditions of the final grinding of the workpiece while gradually reducing the load, the degree of change of the in-plane flatness at the stop time of the main grinding, the degree of change of the peripheral flatness at the stop time of the main grinding, and the amount of change of the peripheral flatness during the final grinding process. The double-sided grinding method has the following characteristics: The steps of setting the target value for the peripheral flatness and the target value for the in-plane flatness; After setting the target values ​​for the outer peripheral flatness and the in-plane flatness, an initial grinding step is performed while gradually increasing the load on the workpiece. After the initial grinding is completed, the main grinding step begins; During the execution of the main grinding, a step is taken to correct the grinding conditions set at the correction time based on the predicted value of the in-plane flatness obtained by inputting the grinding conditions of the main grinding at the correction time and the peripheral flatness at the correction time into the first learner. During the execution of the main grinding process, the grinding conditions of the final grinding, the degree of change of the in-plane flatness at the calculation time, and the degree of change of the peripheral flatness at the calculation time are input into the second learner to obtain the change in the peripheral flatness during the final grinding process. Based on the change in the peripheral flatness and the target value of the peripheral flatness, the step of setting the stop condition of the main grinding is performed. as well as After the main grinding stop condition is met, the final grinding step is performed.

Images