Pad surface determination method and pad surface determination system

Abstract

To provide a pad surface determination method capable of appropriately determining a surface property of a polishing pad including a state of a projection formed on a polishing surface of the polishing pad.SOLUTION: The present pad surface determination method irradiates the target region T in the polishing surface 2a of the polishing pad 2 with a plurality of lights from the plurality of light projection units 515253 at different incident angles, receives a plurality of reflected lights from the target region T by the imaging device 59, generates a plurality of images corresponding to the different incident angles by the imaging device 59, and determines surface properties of the polishing pad 2 based on at least one of the plurality of images. A projection 76 is formed at the 2a of the polishing surface in the target region T.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a pad surface evaluation method and a pad surface evaluation system for evaluating the surface properties of a polishing pad used to polish a substrate such as a wafer. [Background technology]

[0002] In the manufacturing process of semiconductor devices, planarization of semiconductor device surfaces is becoming increasingly important. The most important technology for this surface planarization is chemical mechanical polishing (CMP). Chemical mechanical polishing (hereinafter referred to as CMP) is a process in which a substrate such as a wafer is polished by supplying a polishing liquid containing abrasive grains such as silica (SiO2) onto the polishing surface of a polishing pad while sliding the substrate against the polishing surface.

[0003] A polishing apparatus for performing CMP includes a polishing table that supports a polishing pad having a polishing surface, and a polishing head that presses a substrate against the polishing pad. The polishing apparatus polishes a substrate as follows: While the polishing table and polishing pad are rotated together, a polishing liquid (typically a slurry) is supplied to the polishing surface of the polishing pad. While rotating the substrate, the polishing head presses the surface of the substrate against the polishing surface of the polishing pad. The substrate is brought into sliding contact with the polishing pad in the presence of the polishing liquid. The surface of the substrate is polished by the chemical action of the polishing liquid and the mechanical action of the abrasive grains contained in the polishing liquid and / or the polishing pad.

[0004] When polishing a substrate, abrasive grains and polishing debris adhere to the polishing surface of the polishing pad, reducing its polishing performance. Therefore, to restore the polishing surface of the polishing pad, the polishing pad is dressed (conditioned) using a dresser. The dresser has hard abrasive grains, such as diamond particles, fixed to its underside, and the polishing surface of the polishing pad is regenerated by scraping it away with the dresser.

[0005] A polishing pad gradually wears out as substrates are polished and dressed repeatedly. When the polishing pad wears out, the intended polishing performance cannot be achieved, so the polishing pad must be replaced periodically. Therefore, when the polishing pad has been in use for a predetermined time or when the number of substrates polished exceeds a predetermined number, the polishing pad is replaced with a new one. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2014-172153 [Patent Document 2] International Publication No. 2005 / 072910 Summary of the Invention [Problem to be solved by the invention]

[0007] However, the usage time of the polishing pad and the number of polished substrates only indirectly represent the wear of the polishing pad and may not properly reflect the wear of the polishing pad. As a result, polishing pads that have not yet reached the end of their life may be replaced, or polishing pads that have worn beyond their usage limit may continue to be used.

[0008] The life of a polishing pad is affected not only by wear of the polishing pad but also by the minute protrusions formed on the polishing surface of the polishing pad. These protrusions form the surface roughness of the polishing pad. If the protrusions on the polishing surface become smaller (i.e., if the surface roughness of the polishing pad becomes lower), the polishing rate (also called the removal rate) of the substrate may decrease.

[0009] Therefore, the present invention provides a pad surface determination method and a pad surface determination system that can appropriately determine the surface properties (surface condition) of a polishing pad, including the condition of the convex portions formed on the polishing surface of the polishing pad. [Means for solving the problem]

[0010] In one aspect, a pad surface evaluation method is provided for evaluating the surface properties of a polishing pad having a polishing surface for polishing a substrate, the method including irradiating a target area within the polishing surface with multiple light projecting units at different angles of incidence, forming a convex portion on the polishing surface within the target area, receiving multiple reflected lights from the target area with an imaging device, generating multiple images corresponding to the different angles of incidence with the imaging device, and evaluating the surface properties of the polishing pad based on at least one of the multiple images.

[0011] In one embodiment, determining the surface texture of the polishing pad includes generating a pad surface index value from the at least one image. In one embodiment, the pad surface index value is calculated based on brightness information of the image. In one embodiment, the step of generating the pad surface index value from the at least one image is a step of creating a difference image or a division image from the multiple images and calculating the pad surface index value based on brightness information of the difference image or the division image. In one embodiment, the multiple light beams from the multiple light-projecting units include a first light beam having a wavelength within a first wavelength range emitted from a first light-projecting unit and a second light beam having a wavelength within a second wavelength range emitted from a second light-projecting unit, and the first wavelength range and the second wavelength range are different from each other. In one aspect, the first light projecting unit and the second light projecting unit simultaneously project the first light and the second light onto the target area at the different angles of incidence. In one aspect, the plurality of images corresponding to the different angles of incidence are simultaneously generated by the imaging device. In one aspect, the method further includes repeatedly irradiating the target area with multiple lights and generating multiple images from the multiple reflected lights using the imaging device, thereby obtaining multiple groups of multiple time difference images corresponding to the different angles of incidence, and the step of generating the pad surface index value is a step of generating a difference image or a division image from multiple time difference images in a group with the same angle of incidence, and calculating the pad surface index value based on brightness information of the difference image or the division image. In one aspect, the method further includes repeatedly irradiating the target area with multiple lights, generating multiple images from multiple reflected lights using the imaging device, and generating the pad surface index values, thereby obtaining multiple pad surface index values ​​and calculating the variance of the multiple pad surface index values. In one embodiment, the process of determining the surface properties of the polishing pad based on the at least one image is a process of inputting the at least one image into a determination model constructed by machine learning and outputting a determination result of the surface properties of the polishing pad from the determination model. In one aspect, the multiple lights emitted from the multiple light-projecting units are irradiated onto the target area through a filter, and the filter is configured to allow the multiple lights to pass through but not allow light to pass from directions different from the directions of the multiple lights when viewed from a direction perpendicular to the polishing surface. In one aspect, the filter has a plurality of light-blocking walls that extend parallel to the direction of the plurality of light-projecting sections when viewed from a direction perpendicular to the polishing surface, and the plurality of light-blocking walls are arranged at intervals. In one embodiment, the process of generating the pad surface index value from the at least one image is a process of creating an evaluation image, which is either a difference image or a division image, from the multiple images, and calculating the pad surface index value using brightness values ​​within an acceptable range of the brightness values ​​of the pixels that make up the evaluation image. In one aspect, the pad surface evaluation method further includes creating a frequency distribution of brightness values ​​of pixels that constitute the evaluation image, calculating a standard deviation from the frequency distribution, and determining the tolerance range from the standard deviation. In one embodiment, the process of generating the pad surface index value from the at least one image is a process of correcting the at least one image by removing brightness values ​​outside an acceptable range from the brightness values ​​of pixels constituting at least one of the multiple images, generating an evaluation image, which is either a difference image or a division image, from the multiple images including the corrected image, and calculating the pad surface index value using the brightness values ​​of the pixels constituting the evaluation image. In one embodiment, the pad surface evaluation method further includes creating a frequency distribution of brightness values ​​of pixels constituting the at least one image, calculating a standard deviation from the frequency distribution, and determining the tolerance range from the standard deviation.

[0012] In one aspect, a pad surface evaluation system is provided for evaluating the surface properties of a polishing pad having a polishing surface for polishing a substrate, the pad surface evaluation system comprising: a plurality of light projecting units that irradiate a target area on the polishing surface where a convex portion is present with a plurality of light beams at different angles of incidence; an imaging device that receives a plurality of reflected lights from the target area and generates a plurality of images from the plurality of reflected lights; and a system processing unit that evaluates the surface properties of the polishing pad based on at least one of the plurality of images.

[0013] In one aspect, the system processor is configured to generate a pad surface index value from the at least one image. In one embodiment, the pad surface index value is calculated based on brightness information of the image. In one aspect, the system processing unit is configured to create a difference image or a division image from the plurality of images and calculate the pad surface index value based on brightness information of the difference image or the division image. In one embodiment, the plurality of light-projecting units include a first light-projecting unit that emits a first light having a wavelength within a first wavelength range and a second light-projecting unit that emits a second light having a wavelength within a second wavelength range, the first wavelength range and the second wavelength range being different from each other. In one embodiment, the system processing unit is configured to issue commands to the first light projecting unit and the second light projecting unit to simultaneously irradiate the first light and the second light onto the target area. In one aspect, the imaging device is configured to simultaneously generate the plurality of images from first reflected light and second reflected light corresponding to the first light and the second light, respectively. In one embodiment, the plurality of light-projecting elements include a first set of multiple light-projecting elements in a first direction from the target area and a second set of multiple light-projecting elements in a second direction from the target area when viewed from a direction perpendicular to the polishing surface of the polishing pad. In one embodiment, the system processing unit is configured to input the at least one image into a judgment model constructed by machine learning, and output a judgment result of the surface property of the polishing pad from the judgment model. In one aspect, the pad surface determination system further includes a filter arranged between the plurality of light-projecting units and the polishing surface, the filter being configured to allow the plurality of light beams to pass and not allow light beams to pass from directions different from the directions of the plurality of light beams when viewed from a direction perpendicular to the polishing surface. In one aspect, the filter has a plurality of light-blocking walls that extend parallel to the direction of the plurality of light-projecting sections when viewed from a direction perpendicular to the polishing surface, and the plurality of light-blocking walls are arranged at intervals. In one aspect, the system processing unit is configured to create an evaluation image, which is either a difference image or a division image, from the multiple images, and to calculate the pad surface index value using brightness values ​​of the pixels constituting the evaluation image that are within an acceptable range. In one aspect, the system processing unit is configured to create a frequency distribution of the brightness values ​​of the pixels that make up the evaluation image, calculate a standard deviation from the frequency distribution, and determine the acceptable range from the standard deviation. In one aspect, the system processing unit is configured to correct at least one of the plurality of images by removing brightness values ​​outside an acceptable range from the brightness values ​​of pixels constituting the at least one image, generate an evaluation image, which is either a difference image or a division image, from the plurality of images including the corrected image, and calculate the pad surface index value using the brightness values ​​of the pixels constituting the evaluation image. In one aspect, the system processing unit is configured to create a frequency distribution of luminance values ​​of pixels constituting the at least one image, calculate a standard deviation from the frequency distribution, and determine the acceptable range from the standard deviation. [Effects of the Invention]

[0014] The light emitted by the light-projecting unit is incident on a target area within the polishing surface. As the height of the protrusions within the target area decreases, the intensity of the light reflected from the target area changes. Therefore, the surface quality (surface condition) of the polishing pad can be accurately determined based on the image generated from the reflected light.

[0015] Because multiple light sources are irradiated onto the target area at different incident angles from multiple light projectors, at least one of the multiple images generated from the multiple reflected light sources exhibits a characteristic change in accordance with the change in height of the protrusions, regardless of the type of polishing pad and / or its position on the polishing surface. To determine the incident angle of light that can accurately reflect the change in the surface condition of the polishing pad (the change in the height of the protrusions), the polishing pad must be used from a new state to its end of life for each of the multiple different incident angles. According to the present invention, multiple images are generated from multiple reflected light sources corresponding to multiple incident angles, eliminating the need for a process for determining the optimal incident angle in advance. Furthermore, the surface condition of the polishing pad can be accurately determined based on at least one of the multiple images corresponding to the different incident angles. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 1 is a top view showing an embodiment of a polishing apparatus. [Figure 2] FIG. 2 is a side view of the polishing apparatus shown in FIG. [Figure 3] 2 is a schematic diagram showing an example of a convex portion on the polishing surface of a polishing pad. FIG. [Figure 4] 4(a) and 4(b) are diagrams illustrating how the intensity of reflected light changes in accordance with changes in the height of the convex portions formed on the polishing surface of the polishing pad. [Figure 5] 10A and 10B are schematic diagrams showing examples of images in which the luminance changes according to the change in height of the convex portions; [Figure 6] FIG. 10 is a diagram illustrating an example of a plurality of segment regions set in an image. [Figure 7] FIG. 2 is a diagram showing an example of a first region and a second region set in an image. [Figure 8] FIG. 10 is a diagram showing an example of a pad region and a reference region that appear in an image. [Figure 9] FIG. 10 is a side view showing another embodiment of the pad surface determination system. [Figure 10] 10 is a graph showing an example of a pad surface index value that changes with the use time of a polishing pad. [Figure 11] FIG. 2 is a schematic diagram showing an example of three images generated from a first reflected light, a second reflected light, and a third reflected light corresponding to three lights having different angles of incidence. [Figure 12] FIG. 12 is a schematic diagram showing an example of a difference image generated from two of the three images shown in FIG. 11. [Figure 13] FIG. 10 is a schematic diagram showing an example of black spots and white patterns on an image. [Figure 14] FIG. 10 is a diagram illustrating an example of a frequency distribution. [Figure 15] FIG. 10 is a side view showing still another embodiment of the pad surface determination system. [Figure 16] FIG. 16 is a top view of the pad surface determination system shown in FIG. [Figure 17] FIG. 10 is a side view showing still another embodiment of the pad surface determination system. [Figure 18]FIG. 1 is a schematic diagram illustrating an example of a determination model constructed using a deep learning method. [Figure 19] FIG. 10 is a side view showing still another embodiment of the pad surface determination system. [Figure 20] FIG. 10 is a top view showing an embodiment of a plurality of light projecting units and a filter as viewed from a direction perpendicular to the polishing surface. [Figure 21] FIG. 10 is a front view illustrating one embodiment of a plurality of light projectors and a filter. [Figure 22] 10A and 10B are diagrams illustrating an embodiment of the pitch and height of the light blocking walls of the filter. DETAILED DESCRIPTION OF THE INVENTION

[0017] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a top view showing one embodiment of a polishing apparatus. FIG. 2 is a side view of the polishing apparatus shown in FIG. 1. The polishing apparatus is an apparatus that chemically and mechanically polishes a substrate W such as a wafer. As shown in FIGS. 1 and 2, this polishing apparatus includes a polishing table 3 that supports a polishing pad 2 having a polishing surface 2a, a polishing head 1 that presses the substrate W against the polishing surface 2a, a polishing liquid supply nozzle 5 that supplies a polishing liquid (e.g., a slurry containing abrasive grains) to the polishing surface 2a, and a pad surface evaluation system 40 that evaluates the properties of the polishing surface 2a of the polishing pad 2.

[0018] The polishing apparatus further includes a support shaft 14, a polishing head swing arm 16 connected to the upper end of the support shaft 14, and a polishing head shaft 10 rotatably supported at the free end of the polishing head swing arm 16. The polishing head 1 is fixed to the lower end of the polishing head shaft 10. The polishing head 1 is configured to be able to hold a substrate W on its lower surface. The substrate W is held with the surface to be polished facing downward.

[0019] A polishing head rotating mechanism (not shown) equipped with an electric motor and the like is disposed inside the polishing head swing arm 16. This polishing head rotating mechanism is connected to the polishing head shaft 10 and is configured to rotate the polishing head shaft 10 and the polishing head 1 around the axis of the polishing head shaft 10.

[0020] The polishing head shaft 10 is connected to a polishing head lifting mechanism (including, for example, a ball screw mechanism) not shown. This polishing head lifting mechanism is configured to move the polishing head shaft 10 up and down relative to the polishing head swing arm 16. The up and down movement of the polishing head shaft 10 allows the polishing head 1 to move up and down relative to the polishing head swing arm 16 and the polishing table 3.

[0021] The polishing apparatus further includes a table motor 6 that rotates the polishing table 3 together with the polishing pad 2. The table motor 6 is disposed below the polishing table 3, and the polishing table 3 is connected to the table motor 6 via a table shaft 3a. The polishing table 3 and the polishing pad 2 are rotated around the axis of the table shaft 3a by the table motor 6. The polishing pad 2 is affixed to the upper surface of the polishing table 3. The exposed surface of the polishing pad 2 forms a polishing surface 2a that polishes a substrate W such as a wafer.

[0022] The polishing apparatus further includes a polishing control unit 50 that controls its operation. The polishing head 1, polishing head rotation mechanism, polishing head lifting mechanism, polishing liquid supply nozzle 5, table motor 6, and pad surface determination system 40 are electrically connected to the polishing control unit 50, and the operations of the polishing head 1, polishing head rotation mechanism, polishing head lifting mechanism, polishing liquid supply nozzle 5, table motor 6, and pad surface determination system 40 are controlled by the polishing control unit 50.

[0023] The dressing control unit 50 is composed of at least one computer. The dressing control unit 50 includes a storage device 50a that stores a program for controlling the operation of the dressing machine, and an arithmetic unit 50b that executes calculations according to instructions included in the program. The storage device 50a includes a main storage device such as a random access memory (RAM) and an auxiliary storage device such as a hard disk drive (HDD) or a solid state drive (SSD). Examples of the arithmetic unit 50b include a CPU (central processing unit) and a GPU (graphics processing unit). However, the specific configuration of the dressing control unit 50 is not limited to these examples.

[0024] The substrate W is polished as follows. While the polishing table 3 and polishing head 1 are rotated in the directions indicated by the arrows in FIGS. 1 and 2, a polishing liquid is supplied from a polishing liquid supply nozzle 5 to the polishing surface 2a of the polishing pad 2 on the polishing table 3. While the substrate W is rotated by the polishing head 1, the polishing head 1 presses the substrate W against the polishing surface 2a of the polishing pad 2 with the polishing liquid present on the polishing pad 2. The surface of the substrate W is polished by the chemical action of the polishing liquid and the mechanical action of the abrasive grains contained in the polishing liquid and / or the polishing pad 2. The substrate W may then be water-polished while pure water is supplied onto the polishing pad 2 from a pure water nozzle (not shown).

[0025] After polishing of the substrate W is completed, the substrate W is moved outside the polishing pad 2 and transported to a device for the next processing. Then, a dresser (not shown) is used to dress the polishing surface 2a of the polishing pad 2. The dresser has hard abrasive grains such as diamond particles fixed to its underside, and the polishing surface 2a of the polishing pad 2 is slightly scraped away by this dresser, thereby regenerating the polishing surface 2a of the polishing pad 2.

[0026] A plurality of minute protrusions are formed on the polishing surface 2a of the polishing pad 2. The plurality of minute protrusions constitute the surface roughness of the polishing pad 2. If the protrusions on the polishing surface 2a become smaller (i.e., if the surface roughness of the polishing pad 2 becomes lower), the polishing rate of the substrate W may decrease. Therefore, the polishing apparatus of this embodiment is equipped with a pad surface determination system 40 that determines the surface texture of the polishing pad 2. The surface texture of the polishing pad 2 includes the state of the protrusions on the polishing surface 2a.

[0027] FIG. 3 is a schematic diagram showing an example of convex portions 76 on the polishing surface 2a. The multiple convex portions 76 may be arranged regularly or irregularly. The convex portions 76 on the polishing surface 2a may include not only the convex portions of the polishing surface 2a itself, but also foreign matter (abrasive grains, polishing debris, etc.) on the polishing surface 2a. Examples of convex portions on the polishing surface 2a itself include both protrusions that are the initial surface shape of the polishing pad 2 formed during the manufacture of the polishing pad 2, and minute protrusions formed by diamond particles of a dresser during dressing (conditioning) of the polishing pad 2.

[0028] The pad surface evaluation system 40 includes a plurality of light projecting units 51, 52, 53 that irradiate a target area T in the polishing surface 2a of the polishing pad 2 with a plurality of light beams at different angles of incidence, an imaging device 59 that receives a plurality of light beams reflected from the target area T and generates a plurality of images from the plurality of reflected light beams, and a system processing unit 70 that evaluates the surface texture of the polishing pad 2 based on at least one of the plurality of images. As will be described later, the system processing unit 70 is configured to generate a pad surface index value from at least one of the plurality of images corresponding to the different angles of incidence, and evaluate the surface texture of the polishing pad 2 based on the pad surface index value.

[0029] The system processing unit 70 is composed of at least one computer. The system processing unit 70 includes a storage device 70a that stores a program for determining the properties of the polishing surface 2a of the polishing pad 2, and an arithmetic unit 70b that executes calculations according to instructions included in the program. The storage device 70a includes a main storage device such as a random access memory (RAM) and an auxiliary storage device such as a hard disk drive (HDD) or a solid state drive (SSD). Examples of the arithmetic unit 70b include a CPU (central processing unit) and a GPU (graphics processing unit). However, the specific configuration of the system processing unit 70 is not limited to these examples.

[0030] The light projecting units 51, 52, 53 and the image capturing device 59 are disposed above the polishing surface 2a of the polishing pad 2. The light projecting units 51, 52, 53 and the image capturing device 59 are fixed to a support member (not shown), and the relative positions of the light projecting units 51, 52, 53 and the image capturing device 59 with respect to the polishing table 3 and the polishing pad 2 are fixed. The system processing unit 70 is electrically connected to the light projecting units 51, 52, 53 and the image capturing device 59, and the operations of the light projecting units 51, 52, 53 are controlled by the system processing unit 70.

[0031] The light projecting units 51, 52, and 53 include light sources such as lasers, light-emitting diodes (LEDs), and strobe flash light projecting units (e.g., xenon flash lamps). In addition to the light sources, each of the light projecting units 51 and 52 may further include optical elements such as lenses and fiber optic cables that guide the light generated from the light sources to the target area T within the polishing surface 2a.

[0032] In this embodiment, the multiple light-projecting units 51, 52, and 53 include a first light-projecting unit 51, a second light-projecting unit 52, and a third light-projecting unit 53 that are arranged vertically. These light-projecting units 51, 52, and 53 are installed at different elevation angles with respect to the polishing surface 2a of the polishing pad 2 and are configured to irradiate the polishing surface 2a with light at different incident angles. In this embodiment, the elevation angle of the first light-projecting unit 51 is larger than that of the second light-projecting unit 52, which is also larger than that of the third light-projecting unit 53. In one embodiment, only two light-projecting units may be provided, or four or more light-projecting units may be provided, as long as they can irradiate the polishing surface 2a with light at different incident angles.

[0033] The multiple light-projecting units 51, 52, and 53 irradiate the target area T on the polishing surface 2a with light at different times. More specifically, the system processing unit 70 issues commands to the multiple light-projecting units 51, 52, and 53 to cause them to emit light continuously (i.e., at different times). The target area T is an area on the polishing surface 2a where the light emitted from the multiple light-projecting units 51, 52, and 53 is incident. The arrangement pitch of the convex portions is smaller than the width of the target area T. Therefore, at least one convex portion described with reference to FIG. 3 exists within the target area T. Therefore, the light emitted from each of the light-projecting units 51, 52, and 53 is incident on the convex portion within the target area T.

[0034] Grooves and / or holes (not shown) having a predetermined pattern for holding a polishing liquid are formed on the polishing surface 2a of the polishing pad 2. In this embodiment, no grooves or holes (collectively referred to as recesses) exist in the target area T.

[0035] The imaging device 59 includes an image sensor 58. The imaging device 59 is disposed in a position where it can receive reflected light from the target area T, which includes the convex portion. The imaging device 59 may further include an optical element such as a lens. Examples of the image sensor 58 include a CCD sensor and a CMOS sensor. The multiple light projecting units 51, 52, and 53 and the imaging device 59 face the target area T. The imaging device 59 is disposed above the target area T. The imaging device 59 generates multiple images from the multiple reflected lights from the target area T and transmits these images to the system processing unit 70.

[0036] 4(a) and 4(b) are diagrams illustrating how the intensity of reflected light changes with the change in height of the convex portions 76 formed on the polishing surface 2a of the polishing pad 2. When light is incident on the polishing surface 2a of the polishing pad 2 at an incident angle θ, part of the light is blocked by the convex portions 76. As a result, the shadows of the convex portions 76 are cast on the polishing surface 2a. The area of ​​these shadows decreases as the height of the convex portions 76 decreases. As a result, the intensity of the reflected light increases.

[0037] 5 is a schematic diagram showing an example of an image in which the brightness changes according to the change in height of the convex portion 76. The low-brightness areas in each image in FIG. 5 correspond to the shadows of the convex portion 76 on the polishing surface 2a of the polishing pad 2. As shown in FIG. 5, the overall brightness of the image changes according to the change in height of the convex portion 76.

[0038] Thus, the intensity of the reflected light from the target area T including the protrusions 76 changes with changes in the height of the protrusions 76, i.e., with changes in the properties of the polishing surface 2a of the polishing pad 2. Therefore, the imaging device 59 generates multiple images from the multiple reflected lights corresponding to the multiple light-projecting units 51, 52, and 53, and the system processing unit 70 is configured to generate a pad surface index value from one of the multiple images and generate an alarm signal when the pad surface index value changes beyond a threshold. The image used to generate the pad surface index value may be an image generated from the reflected light corresponding to a pre-selected one of the multiple light-projecting units 51, 52, and 53. The image used to generate the pad surface index value may be all or a part of the original image.

[0039] The pad surface index value is a numerical value calculated based on the luminance information of the image. The luminance information of the image may be luminance information of an image that has been shading-corrected to correct the luminance. The luminance information of the image may be luminance information of an image that has been cut (trimmed) from the original image. Furthermore, image filtering (spatial filtering), such as smoothing to reduce noise, may be performed as preprocessing. In one example, shading correction may be performed on the image, and the shading-corrected image may be trimmed, or the image filtering may be performed on the shading-corrected image.

[0040] Examples of brightness information of an image include the brightness of the entire image, the brightness of a portion of the image, and the brightness distribution within the image. An example of a pad surface index value calculated based on the brightness information of an image may be the sum or statistical value (e.g., average, variance, standard deviation) of the brightness values ​​of the pixels that make up the image. In another example, the system processing unit 70 may perform image processing to identify a convex portion 76 of the polishing pad 2 in the image and calculate the pad surface index value based on the brightness of the area including the convex portion 76.

[0041] In another example, as shown in FIG. 6, the pad surface index value may be a variance calculated from a plurality of brightness values ​​of a plurality of segment areas S in an image. The plurality of segment areas S may be, for example, grid-shaped areas predefined in the image. In another example, the plurality of segment areas S may be a plurality of areas of other shapes. The variance of a plurality of brightness values ​​of a plurality of segment areas S in an image is found as follows: First, a representative value such as the sum or a statistical value (for example, an average) of a plurality of brightness values ​​of a plurality of pixels in each segment area S is calculated, and then the variance of the plurality of representative values ​​calculated for each of the plurality of segment areas S is calculated. The variance calculated in this way is the pad surface index value of the image.

[0042] The variance calculated from the multiple brightness values ​​of the multiple segment regions S in the image represents the clarity of the shadow of the convex portion 76 that appears in the image. More specifically, when the shadow of the convex portion 76 appears clearly in the image (i.e., when the convex portion 76 is high), the variance is large. In contrast, when the shadow of the convex portion 76 does not appear clearly in the image (i.e., when the convex portion 76 is low), the variance is small. Therefore, the system processing unit 70 may generate an alarm signal when the variance as the pad surface index value falls below a threshold. In one embodiment, the pad surface index value may be calculated based on the brightness values ​​of the image obtained by performing smoothing processing, rather than the above variance.

[0043] In yet another example, as shown in FIG. 7, the pad surface index value may be the ratio of the luminance value of the first region S1 to the luminance value of the second region S2 in the image. The first region S1 is a region that includes the convex portion 76. In the example shown in FIG. 7, the first region S1 is located within the second region S2. In another example, the second region S2 may be located within the first region S1. In yet another example, the first region S1 and the second region S2 may partially overlap. In yet another example, the first region S1 and the second region S2 may be separated. The second region S2 may not include the convex portion 76. The ratio of the luminance value of the first region S1 to the luminance value of the second region S2 can be calculated by dividing the luminance value of the first region S1 by the luminance value of the second region S2. By performing such a division, it is possible to cancel out changes in the light intensity of the light-emitting unit over time and differences in the light intensity among multiple light-emitting units.

[0044] In yet another example, as shown in FIG. 8, the pad surface index value may be the ratio of the brightness value of a pad region S3 in the image to the brightness value of a reference region S4. The pad region S3 is the region of the polishing surface 2a of the polishing pad 2 that appears in the image. More specifically, the pad region S3 is the region including the convex portion 76 formed on the polishing surface 2a of the polishing pad 2. The reference region S4 is the region of a reflector 80, which appears in the image and will be described below. The reflector 80 will be described with reference to FIG. 9. FIG. 9 is a side view showing another embodiment of the pad surface determination system 40. The configuration and operation of this embodiment, which are not specifically described, are the same as those of the embodiment described above with reference to FIGS. 1 and 2, and therefore, redundant description will be omitted.

[0045] As shown in FIG. 9, the pad surface determination system 40 includes a reflector 80 disposed above the polishing surface 2a of the polishing pad 2. The reflector 80 is located above a target area T in the polishing surface 2a of the polishing pad 2, which is irradiated with light emitted from the multiple light-projecting units 51, 52, and 53. The imaging device 59 is disposed above the target area T and the reflector 80, and faces the target area T and the reflector 80. As shown in FIG. 8, the imaging device 59 is disposed so as to generate an image including a pad area S3 where the polishing surface 2a of the polishing pad 2 appears, and a reference area S4 where the reflector 80 appears.

[0046] The system processing unit 70 is configured to calculate a pad surface index value, which is the ratio of the luminance value of the pad area S3 to the luminance value of the reference area S4 in the image. The luminance value of the pad area S3 is a representative value such as the sum or statistical value (e.g., average) of multiple luminance values ​​of multiple pixels in the pad area S3. Similarly, the luminance value of the reference area S4 is a representative value such as the sum or statistical value (e.g., average) of multiple luminance values ​​of multiple pixels in the reference area S4. The ratio of the luminance value of the pad area S3 to the luminance value of the reference area S4 is calculated by dividing the luminance value of the pad area S3 by the luminance value of the reference area S4. By performing such division, it is possible to cancel changes in the light intensity of the light-projecting units over time and differences in the light intensity between the multiple light-projecting units.

[0047] FIG. 10 is a graph showing an example of a pad surface index value that changes with the usage time of the polishing pad 2. In FIG. 10, the vertical axis represents the pad surface index value, and the horizontal axis represents the usage time of the polishing pad 2. In the example shown in FIG. 10, the pad surface index value increases as the height of the convex portion 76 decreases and exceeds the threshold value. When the pad surface index value changes beyond the threshold value, the system processing unit 70 can generate an alarm signal to notify that the polishing pad 2 has reached the end of its life. Depending on the algorithm for calculating the pad surface index value, the pad surface index value may decrease as the height of the convex portion 76 decreases. In that case, the system processing unit 70 is configured to generate an alarm signal when the pad surface index value falls below the threshold value.

[0048] 5, the shadow of the convex portion 76 appears clearly on each image, but if the exposure time of the image sensor 58 or the light emission time of the light projector is too long compared to the rotation speed of the polishing table 3, the shadow of the convex portion 76 may not appear clearly in each image. Even in such cases, the image is generated from the reflected light from the target area T, which includes the shadow of the convex portion 76, so the overall brightness information of the image changes according to the height of the convex portion 76.

[0049] In this embodiment, the system processing unit 70 is configured to calculate a pad surface index value based on the brightness information of an image generated from reflected light corresponding to one pre-selected light-emitting unit from the plurality of light-emitting units 51, 52, and 53, and to generate an alarm signal when the pad surface index value changes beyond a threshold value. In one embodiment, the system processor 70 may be configured to calculate a plurality of pad surface index values ​​based on brightness information of a plurality of images sent from the imaging device 59, and generate an alarm signal when at least one of the plurality of pad surface index values ​​changes beyond a threshold value.

[0050] The timing of irradiating the target area T with light and generating an image of the target area T is not particularly limited. However, in one embodiment, to accurately measure the intensity of the reflected light, the irradiation of the target area T with light and generation of an image of the target area T are performed when a polishing liquid such as a slurry is removed from the polishing pad 2. For example, the irradiation of the target area T with light and generation of an image of the target area T are performed during the dressing of the polishing pad 2, which is performed between the end of polishing of a substrate and the start of polishing of the next substrate. More specifically, the system processing unit 70 issues commands to the multiple light-projecting units 51, 52, and 53 to generate light during the dressing of the polishing pad 2, and the imaging device 59 receives the multiple reflected lights from the target area T and generates multiple images from the multiple reflected lights corresponding to the multiple light-projecting units 51, 52, and 53. Since the light-projecting units 51, 52, and 53 emit light while the polishing pad 2 is rotating, the light is irradiated onto different areas on the polishing pad 2.

[0051] The height of the convex portions 76 formed on the polishing surface 2a of the polishing pad 2 varies depending on the type of polishing pad 2. According to this embodiment, multiple light sources are irradiated onto the target area T from multiple light projectors 51, 52, and 53 at different incident angles. Therefore, regardless of the type of polishing pad 2 and / or its position on the polishing surface 2a, the pad surface index value exhibits a characteristic change in accordance with the change in the height of the convex portions 76. In order to determine the incident angle of light that can accurately reflect the change in the surface condition of the polishing pad 2 (the change in the height of the convex portions 76), it is necessary to use the polishing pad from a new state to its end of life for each of the multiple different incident angles. According to this embodiment, multiple images are generated from multiple reflected lights corresponding to multiple incident angles, so there is no need for a process of determining the optimal incident angle in advance.

[0052] In one embodiment, the angle of incidence of light is selected based on the height and / or arrangement pitch of the protrusions 76 of the polishing pad 2. For example, if the arrangement pitch of the multiple protrusions 76 shown in FIG. 3 is small (i.e., if the distance between the multiple protrusions 76 is small), if the angle of incidence of light is too small, the shadow of one protrusion 76 will extend to an adjacent protrusion 76, and the pad surface index value may not accurately reflect the change in height of the protrusions 76. Therefore, in this case, a large angle of incidence is selected. Conversely, if the arrangement pitch of the multiple protrusions 76 is large, a small angle of incidence is selected. In another example, as a result of dressing the polishing pad 2, depending on the dressing conditions, the polishing surface 2a may become wavy and fine streaks may form. In this case, relatively tall streaks are formed, so a large angle of incidence is desirable. In this way, the angle of incidence of light is selected so that the pad surface index value can accurately reflect the change in height of the protrusions 76.

[0053] In one embodiment, the plurality of light projecting units 51, 52, 53 may irradiate the target area T with a plurality of lights, the imaging device 59 may generate a plurality of images from the plurality of reflected lights, and the system processing unit 70 may generate a pad surface index value by repeating these steps within a predetermined period of time to obtain a plurality of pad surface index values, and the system processing unit 70 may calculate the variance of the plurality of pad surface index values ​​and generate an alarm signal when the variance is greater than a variance threshold value. Examples of the predetermined period include the period for polishing a predetermined number of substrates and the period corresponding to one rotation of the polishing table 3.

[0054] As wear of the polishing pad 2 progresses, local differences in the height of the protrusions 76 within the polishing surface 2a become greater, and as a result, the variance in the pad surface index values ​​acquired within the predetermined period is expected to increase. For example, there is a difference in the amount of wear of the polishing pad 2 between the area of ​​the polishing pad 2 that contacts the center of the substrate and the area of ​​the polishing pad 2 that contacts the edge of the substrate. Therefore, the variance in the multiple pad surface index values ​​acquired within the predetermined period is expected to increase. The system processing unit 70 generates an alarm signal when the variance in the multiple pad surface index values ​​is greater than a variance threshold, thereby notifying the user that the polishing pad 2 is significantly worn.

[0055] In one embodiment, multiple pad surface determination systems 40 may be provided at different positions above the polishing surface 2a of the polishing pad 2. Any one of the multiple pad surface determination systems 40 may acquire multiple pad surface index values ​​generated by these multiple pad surface determination systems 40, calculate the variance of the multiple pad surface index values, and generate an alarm signal when the variance is greater than a variance threshold. Furthermore, in one embodiment, the pad surface determination system 40 may move above the polishing pad 2 and emit light at different positions to generate multiple pad surface index values.

[0056] In the above-described embodiment, one imaging device 59 is provided, but a plurality of imaging devices 59 provided at different elevation angles may be provided in correspondence with a plurality of light projecting units 51, 52, 53 having different elevation angles.

[0057] In one embodiment, the system processing unit 70 may acquire multiple groups of time difference images corresponding to different angles of incidence by repeatedly irradiating the target area T with multiple light beams from the multiple light projectors 51, 52, and 53 and generating multiple images using the imaging device 59 from multiple reflected light beams from the target area T. The system processing unit 70 may be configured to generate a difference image or a division image from the multiple time difference images in the same group of incident angles and calculate a pad surface index value based on brightness information of the difference image or the division image. The difference image is generated by calculating the difference in brightness values ​​between the multiple time difference images in the same group of incident angles. The division image is generated by dividing the brightness value of one time difference image by the brightness value of another time difference image in the same group of incident angles.

[0058] Next, another embodiment of the pad surface determination system 40 will be described. The configuration and operation of this embodiment, unless otherwise specifically described, are the same as those of the embodiment described above with reference to FIGS. 1 to 10 , and therefore, redundant description will be omitted. In this embodiment, the multiple light-projecting units 51, 52, and 53 shown in FIG. 2 include a first light-projecting unit 51 that emits first light having a wavelength within a first wavelength range, a second light-projecting unit 52 that emits second light having a wavelength within a second wavelength range, and a third light-projecting unit 53 that emits third light having a wavelength within a third wavelength range. The first wavelength range, the second wavelength range, and the third wavelength range are different from one another.

[0059] In one embodiment, the first light, the second light, and the third light are lights having different colors (i.e., different wavelengths). For example, the first light is any one of red light, green light, and blue light, the second light is any other one of red light, green light, and blue light, and the third light is any other one of red light, green light, and blue light. In one example, the wavelength of the first light is any one of 470 nm, 525 nm, and 625 nm, the second light is any other one of 470 nm, 525 nm, and 625 nm, and the third light is any other one of 470 nm, 525 nm, and 625 nm.

[0060] The system processing unit 70 is configured to issue commands to the first light-projecting unit 51, the second light-projecting unit 52, and the third light-projecting unit 53 to simultaneously project the first light, the second light, and the third light onto the target area T within the polishing surface 2a. Therefore, the first light, the second light, and the third light overlap within the target area T.

[0061] In one embodiment, the system processing unit 70 may be configured to issue commands to the first light-projecting unit 51, the second light-projecting unit 52, and the third light-projecting unit 53 to constantly generate the first light, the second light, and the third light, while issuing commands to the imaging device 59 to generate a first image, a second image, and a third image at a predetermined timing (e.g., simultaneously or at different timings) from the first reflected light, the second reflected light, and the third reflected light corresponding to the first light, the second light, and the third light, respectively.

[0062] In this embodiment, the imaging device 59 including the image sensor 58 is configured to simultaneously generate a first image, a second image, and a third image from the first reflected light, the second reflected light, and the third reflected light, which correspond to the first light, the second light, and the third light, respectively. The configuration of the imaging device 59 that generates three images from the superimposed first reflected light, the second reflected light, and the third reflected light is not particularly limited.

[0063] In one embodiment, the imaging device 59 may include a prism that separates the superimposed first reflected light, second reflected light, and third reflected light into the first reflected light, second reflected light, and third reflected light, and three image sensors that receive the separated first reflected light, second reflected light, and third reflected light. In another embodiment, the imaging device 59 may include color filters that separately extract the first reflected light, second reflected light, and third reflected light from the superimposed first reflected light, second reflected light, and third reflected light, and an image sensor that receives the first reflected light, second reflected light, and third reflected light that have passed through the color filters. In yet another embodiment, the imaging device 59 may include an image sensor having three groups of pixels that receive the superimposed first reflected light, second reflected light, and third reflected light, respectively.

[0064] The wavelengths of light emitted at different angles of incidence are different, and the imaging device 59 can selectively receive light of each wavelength to generate an image. Therefore, the imaging device 59 can generate multiple images simultaneously, thereby reducing the time required to generate a pad surface index value.

[0065] 11 is a schematic diagram showing an example of three images generated from first reflected light, second reflected light, and third reflected light corresponding to three lights with different incident angles. Three light projectors 51, 52, and 53 simultaneously irradiate target area T with the first light, second light, and third light, and imaging device 59 simultaneously receives the first reflected light, second reflected light, and third reflected light from target area T and simultaneously generates a first image, a second image, and a third image from these reflected lights. Therefore, as shown in FIG. 11, a convex portion 76 appears at the same position in the first image, the second image, and the third image.

[0066] The system processing unit 70 is configured to calculate a pad surface index value based on brightness information of at least one of the three images generated from the first reflected light, the second reflected light, and the third reflected light. In one embodiment, the system processing unit 70 may be configured to calculate the pad surface index value based on brightness information of an image generated from reflected light corresponding to a pre-selected one of the plurality of light projectors 51, 52, and 53, and to generate an alarm signal when the pad surface index value changes beyond a threshold. In another embodiment, the system processing unit 70 may be configured to calculate multiple pad surface index values ​​based on brightness information of multiple images sent from the imaging device 59, and to generate an alarm signal when at least one of the multiple pad surface index values ​​changes beyond a threshold.

[0067] In one embodiment, the system processor 70 may be configured to create a difference image or a division image from two of the first, second, and third images corresponding to three different angles of incidence, and calculate the pad surface index value based on brightness information of the difference image or the division image. The difference image is generated by calculating the difference in brightness values ​​between two of the first, second, and third images. The division image is generated by dividing the brightness value of one of the first, second, and third images by the brightness value of the other image.

[0068] FIG. 12 is a schematic diagram illustrating an example of a difference image generated from two of the three images shown in FIG. 11 . As can be seen from FIG. 12 , the difference image is generated from two images generated simultaneously, eliminating noise from the two images and canceling the brightness of areas other than the shadow of the convex portion 76. Therefore, the difference image can represent minute changes in the brightness of the shadow of the convex portion 76. The system processing unit 70 compares the pad surface index value calculated based on the brightness information of the difference image with a threshold value to accurately detect changes in the properties of the polishing surface 2 a of the polishing pad 2. Although not shown, a division image is also generated in a similar manner. In one embodiment, the system processing unit 70 compares the pad surface index value calculated based on the brightness information of the division image with a threshold value to accurately detect changes in the properties of the polishing surface 2 a of the polishing pad 2.

[0069] To cancel the brightness of areas other than the shadows of the convex portions 76, the system processing unit 70 may change the density of the image in advance. For example, the system processing unit 70 may add or subtract an offset value to the brightness value of a pixel, or normalize the brightness value of a pixel. By adjusting the image density in this manner, it is possible to correct for differences in the overall intensity of reflected light due to the angle of incidence and differences in the wavelength sensitivity of the image sensor 58.

[0070] In the above-described embodiment, the pad surface determination system 40 includes three light projecting units 51, 52, and 53 that project first, second, and third lights onto the target area T, and the imaging device 59 receives the first, second, and third reflected lights from the target area T and generates a first image, a second image, and a third image from these reflected lights. In one embodiment, the pad surface determination system 40 may include two light projecting units that emit lights at different incident angles, and the imaging device 59 receives the first and second reflected lights from the target area T and generates a first image and a second image from these reflected lights.

[0071] If there is foreign matter such as dust on the image sensor 58 of the imaging device 59, black spots M1 may appear in the image generated by the imaging device 59, as shown in FIG. 13. Furthermore, if there are air bubbles in the liquid on the polishing surface 2a of the polishing pad 2, white patterns M2 may appear in the image generated by the imaging device 59, as shown in FIG. 13. Such black spots M1 and white patterns M2 appear in the image as brightness deviation areas that have a brightness significantly different from other areas in the image. These brightness deviation areas prevent the pad surface evaluation system 40 from accurately evaluating the properties of the polishing surface 2a of the polishing pad 2.

[0072] Therefore, in one embodiment, the system processing unit 70 removes brightness deviation areas from the image as follows: The system processing unit 70 creates an evaluation image, which is either a difference image or a division image, from two of the multiple images generated from the reflected light of the three light-projecting units 51, 52, and 53, and calculates the pad surface index value using brightness values ​​within an acceptable range of the brightness values ​​of the pixels constituting the evaluation image. The lower limit of the acceptable range is a threshold value for removing low brightness areas caused by foreign matter on the image sensor 58, and the upper limit of the acceptable range is a threshold value for removing high brightness areas caused by air bubbles in the liquid on the polishing surface 2a.

[0073] In one embodiment, the allowable range is determined as follows. As shown in FIG. 14, the system processing unit 70 creates a frequency distribution of the luminance values ​​of the pixels constituting the evaluation image, which is either a difference image or a division image, calculates a standard deviation σ from the frequency distribution, and determines the allowable range from the standard deviation σ. In one embodiment, the system processing unit 70 may determine the allowable range with a lower limit of −σ and an upper limit of +σ. In another embodiment, the system processing unit 70 may determine the allowable range with a lower limit of −3σ and an upper limit of +3σ. The center of the allowable range is, for example, the luminance value of peak P1 in the frequency distribution or the median of the overall luminance values ​​of the evaluation image.

[0074] The system processing unit 70 calculates a pad surface index value using the brightness values ​​within the tolerance range determined as described above. Examples of the pad surface index value include the sum or statistical value (e.g., average, variance, standard deviation) of the brightness values ​​within the tolerance range. Since the brightness values ​​within the tolerance range do not include the brightness values ​​of pixels within the brightness deviation region, the pad surface evaluation system 40 can accurately evaluate the properties of the polishing surface 2a of the polishing pad 2.

[0075] In another embodiment, the system processing unit 70 may remove brightness deviation areas from an image before generating an evaluation image, which is either a difference image or a division image, by following the same process as described above. For example, the system processing unit 70 corrects a first image generated from reflected light from one of the three light-projecting units 51, 52, and 53 by removing brightness values ​​outside an allowable range from the brightness values ​​of pixels constituting the first image, and generates an evaluation image, which is either a difference image or a division image, from the corrected first image and a second image generated from reflected light from the other one of the three light-projecting units 51, 52, and 53.

[0076] The acceptable range can be determined in the same manner as in the embodiment described with reference to Fig. 14. That is, the system processing unit 70 creates a frequency distribution of the luminance values ​​of the pixels constituting the first image, calculates a standard deviation from the frequency distribution, and determines the acceptable range from the standard deviation. The center of the acceptable range is, for example, the peak luminance value in the frequency distribution of the first image or the median of the overall luminance values ​​of the first image.

[0077] The system processing unit 70 calculates a pad surface index value using the brightness values ​​of the pixels that make up the evaluation image generated from the corrected first image and second image. Examples of the pad surface index value include the sum or statistical value (e.g., mean, variance, standard deviation) of the brightness values ​​of the pixels that make up the evaluation image. Since the brightness values ​​of the pixels that make up the evaluation image do not include the brightness values ​​of pixels in the brightness deviation region, the pad surface evaluation system 40 can accurately evaluate the properties of the polishing surface 2a of the polishing pad 2.

[0078] In one embodiment, the system processing unit 70 may correct the second image by removing brightness deviation areas from the second image before generating the evaluation image, similar to the first image. Specifically, the system processing unit 70 corrects the second image by removing brightness values ​​outside the acceptable range from the brightness values ​​of the pixels that make up the second image. The system processing unit 70 generates an evaluation image, which is either a difference image or a division image, from the corrected first image and the corrected second image, and calculates a pad surface index value using the brightness values ​​of the pixels that make up the evaluation image.

[0079] The tolerance range used for correcting the second image can be determined in the same manner as in the embodiment described with reference to Fig. 14. That is, the system processing unit 70 creates a frequency distribution of the luminance values ​​of the pixels that make up the second image, calculates a standard deviation from the frequency distribution, and determines the tolerance range from the standard deviation. The center of the tolerance range is, for example, the peak luminance value in the frequency distribution of the second image or the median of the overall luminance values ​​of the second image.

[0080] Next, another embodiment of the pad surface determination system 40 will be described with reference to FIGS. 15 and 16. FIG. 15 is a side view showing another embodiment of the pad surface determination system 40, and FIG. 16 is a top view of the pad surface determination system 40 shown in FIG. 15. The configuration and operation of this embodiment, which are not particularly described, are the same as those of the embodiment described above with reference to FIGS. 1 to 12, and therefore redundant description will be omitted. In this embodiment, as shown in FIG. 16, the multiple light-projecting units include a first set of multiple light-projecting units 51A, 52A, and 53A located in a first direction from the target area T, and a second set of multiple light-projecting units 51B, 52B, and 53B located in a second direction from the target area T, when viewed from a direction perpendicular to the polishing surface 2a of the polishing pad 2.

[0081] The first direction and the second direction are directions that intersect with each other when viewed from a direction perpendicular to the polishing surface 2a of the polishing pad 2. In the embodiment shown in Figures 15 and 16, the first direction and the second direction are perpendicular to each other when viewed from a direction perpendicular to the polishing surface 2a of the polishing pad 2, the first direction is perpendicular to the radial direction of the polishing pad 2, and the second direction is the radial direction of the polishing pad 2. However, the first direction and the second direction are not limited to this embodiment as long as they are directions that intersect with each other when viewed from a direction perpendicular to the polishing surface 2a of the polishing pad 2.

[0082] The first set of multiple light-projecting units 51A, 52A, and 53A are arranged at different elevation angles relative to the polishing surface 2a and are configured to emit first light at different angles of incidence onto a target area T within the polishing surface 2a. Each of the light-projecting units 51A, 52A, and 53A is configured to emit first light having a wavelength within a first wavelength range. The second set of multiple light-projecting units 51B, 52B, and 53B are also arranged at different elevation angles relative to the polishing surface 2a and are configured to emit second light at different angles of incidence onto a target area T within the polishing surface 2a. Each of the light-projecting units 51B, 52B, and 53B is configured to emit second light having a wavelength within a second wavelength range. The first wavelength range and the second wavelength range are different from each other.

[0083] In one embodiment, the first light and the second light are lights having different colors (i.e., different wavelengths). For example, the first light is any one of red light, green light, and blue light, and the second light is any other one of red light, green light, and blue light.

[0084] The system processing unit 70 is configured to synchronize the first set of multiple light-projecting units 51A, 52A, and 53A and the second set of multiple light-projecting units 51B, 52B, and 53B to emit light continuously. More specifically, the system processing unit 70 issues a command to the first light-projecting unit 51A of the first set and the first light-projecting unit 51B of the second set to simultaneously emit first and second lights at the same incident angle, then issues a command to the second light-projecting unit 52A of the first set and the second light-projecting unit 52B of the second set to simultaneously emit first and second lights at the same incident angle, and then issues a command to the third light-projecting unit 53A of the first set and the third light-projecting unit 53B of the second set to simultaneously emit first and second lights at the same incident angle. The first and second lights at three different incident angles are emitted continuously.

[0085] In one embodiment, the system processing unit 70 may be configured to issue commands to a first set of multiple light-projecting units 51A, 52A, 53A and a second set of multiple light-projecting units 51B, 52B, 53B to emit light at all times, while issuing commands to the imaging device 59 to generate multiple images at predetermined timings (for example, simultaneously or at different timings) from multiple reflected lights corresponding to the multiple lights emitted from the light-projecting units 51A, 52A, 53A, 51B, 52B, 53B.

[0086] The imaging device 59 is disposed in a position where it can receive first reflected light and second reflected light from the target area T, which correspond to the first light and second light emitted from the first set of multiple light-projecting units 51A, 52A, and 53A and the second set of multiple light-projecting units 51B, 52B, and 53B. The imaging device 59 is configured to simultaneously generate two images from the first reflected light and second reflected light corresponding to the first light and second light at the same angle of incidence. The specific configuration of the imaging device 59 is the same as that of the embodiment already described, and therefore a redundant description thereof will be omitted.

[0087] In one embodiment, the system processing unit 70 is configured to acquire two images generated from two reflected lights corresponding to the first light and the second light at the same incident angle from the imaging device 59, generate a difference image or a division image of the two images, calculate a pad surface index value based on brightness information of the difference image or the division image, and generate an alarm signal when the pad surface index value changes beyond a threshold value.

[0088] The first set of multiple light-projecting units 51A, 52A, 53A and the second set of multiple light-projecting units 51B, 52B, 53B can irradiate the target area T with light from different directions. Such an arrangement of the light-projecting units can eliminate the influence of uneven distribution of the convex portions 76 on the polishing surface 2a of the polishing pad 2 and also reduce the influence of the direction of the arrangement pattern of the convex portions 76 of the polishing pad 2. The system processing unit 70 can accurately detect changes in the surface texture of the polishing pad 2 based on the pad surface index value.

[0089] Next, another embodiment of the pad surface determination system 40 will be described with reference to FIGS. 17 and 18. In this embodiment, the system processing unit 70 is configured to input at least one of the multiple images generated by the imaging device 59 into a determination model 77 constructed by machine learning, and output a determination result of the surface texture of the polishing pad 2 from the determination model 77. The irradiation of the target area T with light and the generation of an image of the target area T are performed between the end of polishing of the substrate and the start of polishing of the next substrate. For example, the irradiation of the target area T with light and the generation of an image of the target area T are performed while the polishing pad 2 is being dressed.

[0090] The system processing unit 70 has a determination model 77 stored in its storage device 70a. This determination model 77 is a trained model constructed by machine learning. Examples of machine learning include the SVR (support vector regression) method, the PLS (partial least squares) method, the deep learning method, the random forest method, and the decision tree method. In one example, the determination model 77 is composed of a neural network constructed by the deep learning method.

[0091] The training data used for machine learning of the determination model 77 includes images of the polishing surface of the polishing pad and also includes the surface texture corresponding to the image of the polishing surface of the polishing pad, which is the correct label. The surface texture is a numerical value that indicates the degree of the surface texture of the polishing pad, and can be expressed in a predetermined manner, such as 0 or 1, a percentage from 0 to 100%, a numerical value from 1 to 10, or a scale from 1 to 5.

[0092] For example, an image generated when the worker determines that the polishing pad needs to be replaced is associated with a corresponding correct label of 0, and an image generated when the worker determines that the polishing pad does not need to be replaced is associated with a corresponding correct label of 1. Whether the polishing pad needs to be replaced is determined based on factors such as a decrease in the polishing rate and the amount of wear on the polishing pad.

[0093] In another example, when the surface texture as a correct label is expressed as a percentage of 0 to 100%, a surface texture of 0% indicates that the polishing pad is in a new state, and a surface texture of 100% indicates that the polishing pad needs to be replaced. An image generated when the operator determines that the polishing pad needs to be replaced is associated with a corresponding correct label of 100%. If the number of substrates polished when it is determined that the pad needs to be replaced is 1,000 (i.e., if the number of substrates polished is 1,000, which corresponds to a correct label of 100%), an image of the polished surface generated when 900 substrates have been polished is associated with a corresponding correct label of 90%. An image of the polished surface generated when 800 substrates have been polished is associated with a corresponding correct label of 80%. Similarly, correct labels corresponding to images of intermediate states of the polishing surface of the polishing pad are determined. In this way, training data including correct labels of 0 to 100% and multiple images corresponding to these correct labels is obtained.

[0094] In the embodiment shown in FIG. 17, a plurality of light projecting units 51, 52, and 53 corresponding to a plurality of different angles of incidence are provided. A plurality of training data may be created corresponding to a plurality of different angles of incidence. A plurality of judgment models corresponding to the plurality of training data (i.e., corresponding to a plurality of different angles of incidence) may be created by machine learning. In one embodiment, a single judgment model may be created by machine learning using a plurality of training data corresponding to a plurality of different angles of incidence. In another embodiment, a single training data set including a plurality of images generated from a plurality of reflected lights corresponding to a plurality of different angles of incidence may be created, and a single judgment model may be created by machine learning using the training data.

[0095] In one embodiment, a difference image or a division image may be created from multiple images generated from multiple reflected lights corresponding to multiple different angles of incidence, training data including the difference image or the division image may be created, and a single judgment model may be created by machine learning using the training data. For example, a difference image or a division image may be created from a first image corresponding to a first angle of incidence and a second image corresponding to a second angle of incidence, training data including the difference image or the division image may be created, and a judgment model may be created by machine learning using the training data. In this case, the difference image or the division image may be input to the judgment model (trained model) constructed by machine learning, similar to the training data.

[0096] In one embodiment, a first judgment model may be created by machine learning using training data including a difference image or division image created from a first image corresponding to a first angle of incidence and a second image corresponding to a second angle of incidence, and a second judgment model may be created by machine learning using training data including a difference image or division image created from a second image corresponding to the second angle of incidence and a third image corresponding to a third angle of incidence.

[0097] 18 is a schematic diagram showing an example of a determination model 77 constructed using a deep learning method. The determination model 77 has an input layer 101, multiple hidden layers (also called intermediate layers) 102, and an output layer 103. An image is input to the input layer 101, and a determination result of the surface texture of the polishing pad is output from the output layer 103. The determination result of the surface texture of the polishing pad output from the output layer 103 is, for example, a numerical value or a combination of multiple numerical values ​​indicating the surface texture of the polishing pad.

[0098] The determination model 77 is constructed using a deep learning method as follows. An image of the polishing surface of a polishing pad included in the training data is input to the input layer 101 shown in FIG. 18. The determination model 77 is configured so that when an image is input to the input layer 101, a numerical value indicating the surface texture of the polishing pad corresponding to the image is output from the output layer 103. In the machine learning for constructing the determination model 77, the system processing unit 70 compares the numerical value indicating the surface texture output from the output layer 103 with the correct label corresponding to the input image and adjusts the parameters (weights, thresholds, etc.) of each node (neuron) to minimize the error. In this way, the determination model 77 is trained to output an appropriate surface texture determination result from the output layer 103 based on the image input to the input layer 101. The determination model 77 may be configured to output the determination result of the surface texture of the polishing pad along with the confidence level of the determination result.

[0099] In one embodiment, the determination of the surface texture of the polishing pad 2 using the determination model 77 is performed as follows: The imaging device 59 generates a first image, a second image, and a third image from first reflected light, second reflected light, and third reflected light, which correspond to first light, second light, and third light, respectively, having different angles of incidence, and the system processing unit 70 inputs one of the first image, second image, and third image to the determination model 77 and outputs the determination result of the surface texture of the polishing pad 2 from the determination model 77.

[0100] In another embodiment, the system processing unit 70 has a plurality of judgment models 77 corresponding to different angles of incidence, inputs a plurality of images corresponding to different angles of incidence into the corresponding judgment models 77, respectively, and outputs a plurality of judgment results of the surface texture of the polishing pad 2 from the judgment models 77. When at least one of the obtained judgment results indicates that the polishing pad 2 has reached the end of its life, the polishing pad 2 is replaced with a new polishing pad.

[0101] In yet another embodiment, the system processing unit 70 creates a difference image or a division image from two of multiple images corresponding to different angles of incidence, inputs the difference image or the division image into a judgment model 77, and outputs a judgment result of the surface texture of the polishing pad 2 from the judgment model 77.

[0102] According to these embodiments, it is possible to monitor the surface texture of the polishing pad 2 based on the output surface texture determination result. For example, when the output surface texture is 50%, it can be determined that the current surface texture of the polishing pad 2 is half of the surface texture at the time of replacement of the polishing pad 2.

[0103] Figure 19 is a side view showing yet another embodiment of the pad surface determination system 40. The configuration and operation of this embodiment, which will not be particularly described, are the same as those of the embodiment described above with reference to Figures 1 and 2, and therefore redundant description will be omitted. As shown in Figure 19, the pad surface determination system 40 of this embodiment includes a filter 90 arranged between a plurality of light projecting units 51, 52, 53 and the polishing surface 2a.

[0104] The filter 90 is located directly above the target area T within the polishing surface 2a. The filter 90 is held by a filter holder (not shown), and the position of the filter 90 is fixed. The filter 90 is configured to allow the passage of multiple light beams emitted from the multiple light-projecting units 51, 52, and 53, while not allowing the passage of light beams coming from directions different from the directions of the light beams from the multiple light-projecting units 51, 52, and 53 when viewed from a direction perpendicular to the polishing surface 2a. The light beams emitted from the multiple light-projecting units 51, 52, and 53 pass through the filter 90 and reach the target area T within the polishing surface 2a.

[0105] The imaging device 59 is disposed above the target area T and the filter 90 and faces the target area T and the filter 90. The imaging device 59 generates an image from the reflected light that has passed through the filter 90.

[0106] FIG. 20 is a top view showing one embodiment of the light-transmitting units 51, 52, and 53 and the filter 90, as viewed perpendicular to the polishing surface 2a. As shown in FIG. 20, the filter 90 has a plurality of light-blocking walls 91 extending parallel to the light-transmitting units 51, 52, and 53 when viewed perpendicular to the polishing surface 2a. That is, when viewed perpendicular to the polishing surface 2a, the longitudinal direction of the light-blocking walls 91 is parallel to the direction of the light-transmitting units 51, 52, and 53. The direction of the light-transmitting units 51, 52, and 53 corresponds to the optical axis direction of the light emitted from the light-transmitting units 51, 52, and 53. Each light-blocking wall 91 is made of a material that does not allow light to pass through. In one embodiment, the light-blocking walls 91 are a plurality of louvers or a plurality of plates extending parallel to the direction of the light-transmitting units 51, 52, and 53 when viewed perpendicular to the polishing surface 2a.

[0107] FIG. 21 is a front view showing one embodiment of the plurality of light-projecting units 51, 52, and 53 and the filter 90 shown in FIG. 20. As shown in FIG. 21, both side surfaces of each light-blocking wall 91 are perpendicular to the polishing surface 2a. The plurality of light-blocking walls 91 are arranged at intervals. The plurality of light beams emitted from the plurality of light-projecting units 51, 52, and 53 reach the polishing surface 2a of the polishing pad 2 through light paths 92 between the plurality of light-blocking walls 91. In this embodiment, the light paths 92 are the spaces between the light-blocking walls 91. In one embodiment, a light-transmitting material such as glass or transparent resin may be disposed between the plurality of light-blocking walls 91.

[0108] 22 is a diagram illustrating one embodiment of the pitch and height of the light blocking walls 91 of the filter 90. If the pitch Pf of the light blocking walls 91 is too large, the filter 90 cannot block light that is obliquely incident on the light blocking walls 91 when viewed from a direction perpendicular to the polishing surface 2 a. Similarly, if the height Hf of the light blocking walls 91 is too low, the filter 90 cannot block light that is obliquely incident on the light blocking walls 91 when viewed from a direction perpendicular to the polishing surface 2 a.

[0109] 22 , in one embodiment, the ratio (Hf / Pf) of the height Hf to the pitch Pf of the light blocking walls 91 of the filter 90 is equal to or greater than the aspect ratio Lb / La of the convex portions 76 formed on the polishing surface 2a of the polishing pad 2. The pitch Pf is the distance between two adjacent light blocking walls 91. La represents the width of the convex portions 76, and Lb represents the height of the convex portions 76. A filter 90 having light blocking walls 91 with such dimensions can block light that is obliquely incident on the light blocking walls 91.

[0110] Light from light-projecting units other than the multiple light-projecting units 51, 52, and 53 can become undesirable noise when determining the surface texture of the polishing pad 2. The filter 90 allows the multiple light beams emitted from the light-projecting units 51, 52, and 53 to pass, while not allowing the passage of light beams emitted in a direction different from the direction of the light from the light-projecting units 51, 52, and 53 (e.g., oblique or vertical directions) when viewed from a direction perpendicular to the polishing surface 2a. Furthermore, the filter 90 can convert the multiple light beams emitted from the multiple light-projecting units 51, 52, and 53 into collimated light. As a result, the system processing unit 70 can accurately determine the surface texture of the polishing pad 2 based on at least one of the multiple images corresponding to the multiple light beams emitted from the multiple light-projecting units 51, 52, and 53.

[0111] The above-described embodiments have been described for the purpose of enabling a person of ordinary skill in the art to practice the present invention. Various modifications of the above-described embodiments would be obvious to a person skilled in the art, and the technical concept of the present invention may be applied to other embodiments. Therefore, the present invention is not limited to the described embodiments, but is to be interpreted in the broadest scope in accordance with the technical concept defined by the claims. [Explanation of symbols]

[0112] 1 polishing head 2 polishing pads 2a Polished surface 3 Polishing table 3a Table axis 5 Polishing liquid supply nozzle 6 Table Motor 10 Grinding head shaft 14 Spindle 16 Polishing head swing arm 40 Pad Surface Judgment System 50 Polishing control section 51, 52, 53, 51A, 52A, 53A, 51B, 52B, 53B Light projecting section 58 Image Sensor 59 Imaging device 70 System Processing Unit 76 Convex part 77 Decision Model 80 Reflector 90 filters 91 Light-blocking wall 92 Light passage T target region

Claims

1. A pad surface evaluation method for evaluating the surface quality of a polishing pad having a polishing surface for polishing a substrate, comprising: A plurality of light beams are irradiated onto a target area in the polishing surface at different angles of incidence from a plurality of light projecting units, and a convex portion is formed on the polishing surface in the target area, receiving a plurality of reflected beams from the target area with an imaging device; generating a plurality of images corresponding to the different angles of incidence with the imaging device; A pad surface evaluation method for evaluating the surface condition of the polishing pad based on at least one of the plurality of images.

2. 2. The method of claim 1, wherein determining the surface texture of the polishing pad comprises generating a pad surface index value from the at least one image.

3. 3. The pad surface determination method according to claim 2, wherein the pad surface index value is calculated based on brightness information of the image.

4. generating the pad surface index value from the at least one image, creating a difference image or a division image from the plurality of images; 4. The pad surface determination method according to claim 3, further comprising the step of calculating the pad surface index value based on brightness information of the difference image or the division image.

5. the plurality of light beams from the plurality of light-projecting units include a first light beam having a wavelength within a first wavelength range emitted from a first light-projecting unit, and a second light beam having a wavelength within a second wavelength range emitted from a second light-projecting unit; 2. The pad surface determination method according to claim 1, wherein the first wavelength range and the second wavelength range are different from each other.

6. 6. The pad surface determination method according to claim 5, wherein the first light projecting unit and the second light projecting unit simultaneously project the first light and the second light onto the target area at the different angles of incidence.

7. The pad surface determination method according to claim 5 , wherein the plurality of images corresponding to the different angles of incidence are simultaneously generated by the imaging device.

8. and acquiring a plurality of groups of time-lapse images corresponding to the different angles of incidence by repeatedly irradiating the target area with a plurality of lights and generating a plurality of images from the plurality of reflected lights with the imaging device; 3. The pad surface determination method according to claim 2, wherein the step of generating the pad surface index value is a step of generating a difference image or a division image from a plurality of time difference images in a group having the same incident angle, and calculating the pad surface index value based on brightness information of the difference image or the division image.

9. acquiring a plurality of pad surface index values ​​by repeating the steps of irradiating the target area with a plurality of lights, generating a plurality of images from the plurality of reflected lights with the imaging device, and generating the pad surface index value; 3. The method of claim 2, further comprising calculating a variance of the plurality of pad surface index values.

10. The step of determining the surface texture of the polishing pad based on the at least one image includes: inputting the at least one image into a decision model constructed by machine learning; 2. The pad surface evaluation method according to claim 1, further comprising the step of outputting a result of evaluation of the surface quality of the polishing pad from the evaluation model.

11. The plurality of light beams emitted from the plurality of light projecting units are irradiated onto the target area through a filter, 2. The pad surface evaluation method according to claim 1, wherein the filter is configured to allow the plurality of light beams to pass and not allow light beams to pass from directions different from the directions of the plurality of light beams when viewed from a direction perpendicular to the polishing surface.

12. 12. The pad surface determination method according to claim 11, wherein the filter has a plurality of light-blocking walls extending parallel to the direction of the plurality of light-projecting units when viewed from a direction perpendicular to the polishing surface, and the plurality of light-blocking walls are arranged at intervals.

13. generating the pad surface index value from the at least one image, creating an evaluation image, which is either a difference image or a division image, from the plurality of images; 3. The pad surface evaluation method according to claim 2, further comprising the step of calculating the pad surface index value using brightness values ​​within an allowable range of brightness values ​​of pixels constituting the evaluation image.

14. creating a frequency distribution of luminance values ​​of pixels constituting the evaluation image; Calculating a standard deviation from the frequency distribution; The method for determining a pad surface according to claim 13, further comprising determining the tolerance range from the standard deviation.

15. generating the pad surface index value from the at least one image, correcting at least one image among the plurality of images by removing luminance values ​​outside an allowable range from luminance values ​​of pixels constituting the at least one image; generating an evaluation image, which is either a difference image or a division image, from the plurality of images including the corrected image; 3. The pad surface evaluation method according to claim 2, further comprising the step of calculating the pad surface index value using brightness values ​​of pixels constituting the evaluation image.

16. creating a frequency distribution of luminance values ​​of pixels constituting said at least one image; Calculating a standard deviation from the frequency distribution; 16. The method for determining a pad surface according to claim 15, further comprising determining the tolerance range from the standard deviation.

17. A pad surface evaluation system for evaluating the surface quality of a polishing pad having a polishing surface for polishing a substrate, comprising: a plurality of light projecting units that project a plurality of lights at different angles of incidence onto a target area on the polishing surface where a convex portion is present; an imaging device that receives a plurality of reflected beams from the target area and generates a plurality of images from the plurality of reflected beams; A pad surface determination system comprising a system processor that determines a surface texture of the polishing pad based on at least one of the plurality of images.

18. 18. The pad surface determining system according to claim 17, wherein the system processor is configured to generate a pad surface index value from the at least one image.

19. The pad surface determination system according to claim 18, wherein the pad surface index value is calculated based on brightness information of an image.

20. 20. The pad surface determination system according to claim 19, wherein the system processing unit is configured to create a difference image or a division image from the plurality of images, and calculate the pad surface index value based on brightness information of the difference image or the division image.

21. the plurality of light-projecting units include a first light-projecting unit that emits a first light having a wavelength within a first wavelength range and a second light-projecting unit that emits a second light having a wavelength within a second wavelength range; 18. The pad surface determining system according to claim 17, wherein the first wavelength range and the second wavelength range are different from each other.

22. 22. The pad surface determination system according to claim 21, wherein the system processing unit is configured to issue a command to the first light projecting unit and the second light projecting unit to simultaneously irradiate the first light and the second light onto the target area.

23. 22. The pad surface determining system according to claim 21, wherein the imaging device is configured to simultaneously generate the plurality of images from first reflected light and second reflected light corresponding to the first light and the second light, respectively.

24. 18. The pad surface determination system of claim 17, wherein the plurality of light projecting units include a first set of a plurality of light projecting units in a first direction from the target area and a second set of a plurality of light projecting units in a second direction from the target area when viewed from a direction perpendicular to the polishing surface of the polishing pad.

25. The system processing unit inputting the at least one image into a decision model constructed by machine learning; 18. The pad surface determination system according to claim 17, configured to output a determination result of the surface texture of the polishing pad from the determination model.

26. The pad surface determination system further includes a filter disposed between the plurality of light projecting units and the polishing surface, 18. The pad surface determination system according to claim 17, wherein the filter is configured to allow the plurality of light beams to pass and not allow light beams to pass from directions different from the directions of the plurality of light beams when viewed from a direction perpendicular to the polishing surface.

27. 27. The pad surface determination system according to claim 26, wherein the filter has a plurality of light-blocking walls extending parallel to the direction of the plurality of light-projecting units when viewed from a direction perpendicular to the polishing surface, and the plurality of light-blocking walls are arranged at intervals.

28. The system processing unit creating an evaluation image, which is either a difference image or a division image, from the plurality of images; 19. The pad surface determination system according to claim 18, wherein the pad surface index value is calculated using brightness values ​​within an allowable range among brightness values ​​of pixels constituting the evaluation image.

29. The system processing unit creating a frequency distribution of luminance values ​​of pixels constituting the evaluation image; Calculating a standard deviation from the frequency distribution; 29. The pad surface determination system according to claim 28, configured to determine the tolerance range from the standard deviation.

30. The system processing unit correcting at least one image among the plurality of images by removing luminance values ​​outside an allowable range from luminance values ​​of pixels constituting the at least one image; generating an evaluation image, which is either a difference image or a division image, from the plurality of images including the corrected image; 19. The pad surface determination system according to claim 18, wherein the pad surface index value is calculated using brightness values ​​of pixels that form the evaluation image.

31. The system processing unit creating a frequency distribution of luminance values ​​of pixels constituting said at least one image; Calculating a standard deviation from the frequency distribution; 31. The pad surface determination system according to claim 30, configured to determine the tolerance range from the standard deviation.

Images