Polishing method, polishing apparatus, and computer-readable recording medium

Abstract

Provided is a polishing method, a polishing apparatus, and a computer-readable recording medium that can measure the film thickness of a substrate such as a semiconductor wafer having various configuration elements on a surface with high precision. The polishing method generates a plurality of light spectra of reflected light from a plurality of measurement points on a substrate (W), classifies the plurality of light spectra into a plurality of primary light spectra belonging to a first group and a secondary light spectrum belonging to a second group based on the shape of each light spectrum, and determines a plurality of film thicknesses of the substrate (W) from the plurality of primary light spectra, and determines the film thickness at a detection point corresponding to the secondary light spectrum using the primary light spectrum or the plurality of film thicknesses.

Description

Technical Field

[0001] This invention relates to a method and apparatus for polishing substrates such as wafers, and more particularly to a technique for detecting film thickness based on optical information contained in reflected light from the substrate. Background Technology

[0002] A polishing apparatus for substrates such as semiconductor wafers is configured such that, while rotating the substrate and polishing pad via a polishing head and a polishing table respectively, the substrate is pressed against the polishing pad on the polishing table, thereby polishing the surface of the substrate. A representative example of a polishing apparatus is a CMP (chemical mechanical polishing) apparatus, which supplies a slurry to the polishing pad while pressing the substrate against the pad in the presence of the slurry. The surface of the substrate is polished by the chemical action of the slurry and the mechanical action of the abrasive particles contained in the slurry.

[0003] In such a polishing apparatus, an in-situ spectrophotometer is used for measuring the thickness of the insulating layer (transparent layer) on a substrate. This spectrophotometer includes a light source and a beam splitter mounted on the polishing table, as well as optical fiber cables for transmitting and receiving light, respectively connected to the light source and the beam splitter. The tips of these optical fiber cables function as optical sensor heads.

[0004] Each time the polishing table rotates, the optical sensor head scans the wafer surface. That is, as the optical sensor head traverses the substrate, it illuminates multiple measurement points on the substrate and receives the reflected light from these measurement points. A spectrometer decomposes the reflected light from each measurement point according to the wavelength and generates light intensity data. The spectrum generation unit of the spectrophotometer generates the spectrum of the reflected light using the light intensity data. Since this spectrum varies according to the substrate's film thickness, the spectrophotometer can determine the current film thickness of the substrate based on the spectrum.

[0005] Existing technical documents

[0006] Patent documents

[0007] Patent Document 1: International Publication No. 2003 / 083522

[0008] Patent Document 2: Japanese Patent Application Publication No. 2015-8303

[0009] Patent Document 3: Japanese Patent Application Publication No. 2008-244335

[0010] The technical problem that the invention aims to solve

[0011] However, because such in-situ spectrophotometers measure the film thickness of a rotating substrate while moving the optical sensor head, the position of the measurement point on the substrate is not fixed. The surface of the substrate is composed of various structural elements such as components and scribing lines. The intensity of reflected light from the substrate varies not only due to the film thickness but also due to these structural elements constituting the substrate surface. For example, the intensity of reflected light from components differs from the intensity of reflected light from scribing lines. In other cases, the intensity of reflected light varies due to differences in the underlying structure of the components. As a result, the measured film thickness will vary depending on the measurement point. Summary of the Invention

[0012] Therefore, the present invention provides a polishing method and polishing apparatus that can measure the film thickness of a substrate such as a semiconductor wafer having various structural elements on its surface with high accuracy.

[0013] Technical means for solving technical problems

[0014] In one embodiment, a polishing method is provided in which, while polishing a substrate, multiple spectra of reflected light from multiple measurement points on the substrate are generated. Based on the shape of each spectrum, the multiple spectra are classified into multiple primary spectra belonging to a first group and secondary spectra belonging to a second group. Multiple film thicknesses of the substrate are determined by the multiple primary spectra, and the film thickness at the measurement point corresponding to the secondary spectrum is determined by using the primary spectra or the multiple film thicknesses.

[0015] In one embodiment, the process of determining the film thickness at the measurement point corresponding to the secondary spectrum is as follows: generating an estimated spectrum that corresponds to the secondary spectrum and belongs to the first group, and determining the film thickness of the substrate based on the estimated spectrum.

[0016] In one embodiment, the process of generating the estimated spectrum is the process of generating the estimated spectrum from the plurality of primary spectra by interpolation or extrapolation.

[0017] In one embodiment, the process of generating the estimated spectrum involves inputting the plurality of primary spectra into a spectrum generation model and outputting the estimated spectrum from the spectrum generation model.

[0018] In one embodiment, the process of determining the film thickness at the measurement point corresponding to the secondary spectrum is the process of determining the film thickness at the measurement point corresponding to the secondary spectrum from the plurality of film thicknesses by interpolation or extrapolation.

[0019] In one embodiment, the process of classifying the plurality of spectra into primary spectra belonging to the first group and secondary spectra belonging to the second group is as follows: inputting the plurality of spectra generated during the grinding process of the substrate into a classification model, and classifying the plurality of spectra into primary spectra belonging to the first group and secondary spectra belonging to the second group based on the classification results output from the respective classification models.

[0020] In one embodiment, the grinding method further includes the following steps: while grinding the sample substrate, generating multiple training spectra of reflected light from the sample substrate, classifying the multiple training spectra into a first group and a second group, and using classification training data containing the multiple training spectra and the classification results of the multiple training spectra, determining the parameters of the classification model through machine learning.

[0021] In one embodiment, a polishing method is provided in which, while polishing a reference substrate, multiple spectra of reflected light from multiple measurement points on the reference substrate are generated; based on the shape of each spectrum, the multiple spectra are classified into multiple primary spectra belonging to a first group and secondary spectra belonging to a second group; an estimated spectrum corresponding to the secondary spectrum and belonging to the first group is generated; the multiple primary spectra and the estimated spectrum are associated with film thickness respectively; the multiple primary spectra and the estimated spectrum are added as reference spectra to a database having multiple reference spectra including the multiple primary spectra and the estimated spectrum; while polishing the substrate, a spectrum of reflected light from the substrate is generated; a reference spectrum with the spectral shape closest to the reflected light from the substrate is determined; and a film thickness associated with the determined reference spectrum is determined.

[0022] In one embodiment, a polishing apparatus is provided, comprising: a polishing table supporting a polishing pad; a polishing head pressing a substrate against the polishing pad and polishing the substrate; an optical sensor head directing light to a plurality of measurement points on the substrate and receiving reflected light from the plurality of measurement points; and a processing system generating a plurality of spectra of the reflected light, the processing system being configured to: classify the plurality of spectra into a plurality of primary spectra belonging to a first group and a plurality of secondary spectra belonging to a second group based on the shape of each spectrum; determine a plurality of film thicknesses of the substrate using the plurality of primary spectra; and determine the film thickness at a measurement point corresponding to the secondary spectrum using the primary spectrum or the plurality of film thicknesses.

[0023] In one embodiment, the processing system is configured to generate an estimated spectrum corresponding to the secondary spectrum and belonging to the first group, and to determine the film thickness of the substrate by means of the estimated spectrum.

[0024] In one embodiment, the processing system is configured to generate the estimated spectrum from the plurality of primary spectra by interpolation or extrapolation.

[0025] In one embodiment, the processing system has a spectral generation model, the processing system being configured to input the plurality of primary spectra into the spectral generation model, and output the estimated spectrum from the spectral generation model.

[0026] In one embodiment, the processing system is configured to determine the film thickness at a measurement point corresponding to the secondary spectrum from the plurality of film thicknesses by interpolation or extrapolation.

[0027] In one embodiment, the processing system includes a classification model, the processing system being configured to: input the plurality of spectra generated during the grinding process of the substrate into the classification model respectively, and classify the plurality of spectra into primary spectra belonging to a first group and secondary spectra belonging to a second group based on the classification results output from the classification model.

[0028] In one embodiment, the processing system includes a storage device that stores multiple training spectra of reflected light from a sample substrate. The processing system is configured to classify the multiple training spectra into a first group and a second group, and use the classification results, including the multiple training spectra, to determine the parameters of the classification model through machine learning.

[0029] In one embodiment, a polishing apparatus is provided, comprising: a polishing table supporting a polishing pad; a polishing head pressing a substrate against the polishing pad and polishing the substrate; an optical sensor head directing light to a plurality of measurement points on the substrate and receiving reflected light from the plurality of measurement points; and a processing system having a storage device storing a database including a plurality of reference spectra and a plurality of spectra of reflected light from the plurality of measurement points on the reference substrate, the processing system being configured to: classify the plurality of spectra of the reflected light into a plurality of primary spectra belonging to a first group and a secondary spectra belonging to a second group based on the shape of each spectrum; generate an estimated spectrum corresponding to the secondary spectrum and belonging to the first group; associate the plurality of primary spectra and the estimated spectrum with film thickness respectively; and add the plurality of primary spectra and the estimated spectrum as reference spectra to the database.

[0030] In one embodiment, a computer-readable recording medium is provided, recording a program for causing a computer to perform the following steps: generating multiple spectra of reflected light from multiple measurement points on a substrate during a substrate polishing process; classifying the multiple spectra into multiple primary spectra belonging to a first group and secondary spectra belonging to a second group based on the shape of each spectrum; determining multiple film thicknesses of the substrate using the multiple primary spectra; and determining the film thickness at a measurement point corresponding to the secondary spectrum using the primary spectra or the multiple film thicknesses.

[0031] In one embodiment, the step of determining the film thickness at the measurement point corresponding to the secondary spectrum is as follows: generating an estimated spectrum corresponding to the secondary spectrum and belonging to a first group, and determining the film thickness of the substrate by means of the estimated spectrum.

[0032] In one embodiment, the step of generating the estimated spectrum is the step of generating the estimated spectrum from the plurality of primary spectra by interpolation or extrapolation.

[0033] In one embodiment, the step of generating the estimated spectrum is to input the plurality of primary spectra into a spectral generation model and output the estimated spectrum from the spectral generation model.

[0034] In one embodiment, the step of determining the film thickness at the measurement point corresponding to the secondary spectrum is the step of determining the film thickness at the detection point corresponding to the secondary spectrum from the plurality of film thicknesses by interpolation or extrapolation.

[0035] In one embodiment, the step of classifying the plurality of spectra into primary spectra belonging to the first group and secondary spectra belonging to the second group comprises the following steps: inputting the plurality of spectra generated during the grinding process of the substrate into a classification model, and classifying the plurality of spectra into primary spectra belonging to the first group and secondary spectra belonging to the second group based on the classification results output from the classification model.

[0036] In one embodiment, the program is configured to further cause the computer to perform the following steps: generating multiple training spectra of reflected light from the sample substrate during the polishing process of the sample substrate; classifying the multiple training spectra into a first group and a second group; and using classification training data including the multiple training spectra and the classification results of the multiple training spectra, determining the parameters of the classification model by machine learning.

[0037] In one embodiment, a computer-readable recording medium is provided, recording a program for causing a computer to perform the following steps: generating multiple spectra of reflected light from multiple measurement points on a reference substrate during the polishing process of the reference substrate; classifying the multiple spectra into multiple primary spectra belonging to a first group and secondary spectra belonging to a second group based on the shape of each spectrum; generating an estimated spectrum corresponding to the secondary spectrum and belonging to the first group; associating the multiple primary spectra and the estimated spectrum with film thickness respectively; adding the multiple primary spectra and the estimated spectrum as reference spectra to a database having multiple reference spectra including the multiple primary spectra and the estimated spectrum; generating a spectrum of reflected light from the substrate while polishing the substrate; determining a reference spectrum whose spectral shape is closest to that of the reflected light from the substrate; and determining a film thickness associated with the determined reference spectrum.

[0038] The effects of the invention

[0039] The estimated spectrum is a primary spectrum that is expected to accurately reflect the film thickness of the substrate. Therefore, the processing system can determine the correct film thickness of the substrate using the estimated spectrum. In particular, the processing system can determine the correct film thickness at all measurement points across multiple measurement points on the substrate. Attached Figure Description

[0040] Figure 1 This is a schematic diagram illustrating one embodiment of the grinding apparatus.

[0041] Figure 2 This is a diagram showing an example of a spectrum generated by the processing system.

[0042] Figure 3 (a)~ Figure 3 (c) is a schematic diagram representing an example of a processing system.

[0043] Figure 4 It means Figure 1 A cross-sectional view of one embodiment of the detailed structure of the grinding apparatus shown.

[0044] Figure 5 This is a schematic diagram illustrating the principle of an optical film thickness measuring device.

[0045] Figure 6 This is a top view showing the positional relationship between the substrate and the polishing table.

[0046] Figure 7 This figure illustrates an example of a method for determining film thickness using the spectrum of reflected light.

[0047] Figure 8This is a schematic diagram illustrating an example of multiple measurement points on the surface (polished surface) of a substrate.

[0048] Figure 9 This diagram illustrates one implementation method for generating an estimated spectrum using multiple primary spectra from adjacent measurement points.

[0049] Figure 10 This is a diagram illustrating one embodiment of generating an estimated spectrum by using a primary spectrum along the temporal sequence of the grinding time.

[0050] Figure 11 This is a flowchart illustrating the operation of an optical film thickness measuring device in determining the film thickness of a substrate.

[0051] Figure 12 This is a schematic diagram illustrating an example of a spectral production model.

[0052] Figure 13 This is a flowchart illustrating one implementation of updating a database of reference spectra.

[0053] Figure 14 This is a flowchart illustrating the automatic classification of spectra and the creation of classification models that constitute the classification methods for spectra.

[0054] Figure 15 This is a schematic diagram illustrating an example of a classification model.

[0055] Figure 16 This is a flowchart illustrating the operation of an optical film thickness measuring device with a classification model to determine the film thickness of a substrate.

[0056] Symbol Explanation

[0057] 1 Grinding head

[0058] 2 Grinding pad

[0059] 2a Grinding surface

[0060] 3 Grinding table

[0061] 5. Grinding fluid supply nozzle

[0062] 6 Grinding table motor

[0063] 7. Optical sensor head

[0064] 9 Grinding Control Department

[0065] 10-head shaft

[0066] 17 Connecting structural components

[0067] 18 Grinding head motor

[0068] 31. Fiber optic cable for light transmission

[0069] 32 Optical fiber cable for receiving light

[0070] 40 Optical film thickness measuring device

[0071] 44 Light Source

[0072] 47 beam splitter

[0073] 48. Photodetector

[0074] 49 Processing System

[0075] 49a Storage device

[0076] 49b Processing device

[0077] 50A First Hole

[0078] 50B Second Hole

[0079] 51 Through Hole

[0080] 60 databases Detailed Implementation

[0081] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

[0082] Figure 1 This is a schematic diagram illustrating one embodiment of the grinding apparatus. (For example...) Figure 1 As shown, the polishing apparatus includes a polishing table 3 that supports the polishing pad 2, a polishing head 1 that presses a substrate W, such as a wafer with a film, onto the polishing pad 2, a polishing table motor 6 that rotates the polishing table 3, and a polishing liquid supply nozzle 5 for supplying polishing liquid such as slurry to the polishing pad 2. The upper surface of the polishing pad 2 forms a polishing surface 2a for polishing the substrate W.

[0083] The grinding head 1 is connected to the head shaft 10, which is connected to a grinding head motor (not shown). The grinding head motor causes the grinding head 1 and the head shaft 10 to rotate together in the direction indicated by the arrow. The grinding table 3 is connected to a grinding table motor 6, which is configured to rotate the grinding table 3 and the grinding pad 2 in the direction indicated by the arrow.

[0084] The substrate W is polished as follows. While the polishing table 3 and polishing head 1 are moving towards... Figure 1 The polishing head rotates in the direction indicated by the arrow, while polishing slurry is supplied from the polishing slurry supply nozzle 5 to the polishing surface 2a of the polishing pad 2 on the polishing table 3. As the substrate W rotates through the polishing head 1, with polishing slurry present on the polishing pad 2, the substrate W is pressed against the polishing surface 2a of the polishing pad 2 by the polishing head 1. The surface of the substrate W is polished by the chemical action of the polishing slurry and the mechanical action of the abrasive particles contained in the polishing slurry.

[0085] The polishing apparatus includes an optical film thickness measuring device 40 for determining the film thickness of a substrate W. The optical film thickness measuring device 40 includes a light source 44, a beam splitter 47, an optical sensor head 7 connected to the light source 44 and the beam splitter 47, and a processing system 49 connected to the beam splitter 47. The optical sensor head 7, the light source 44, and the beam splitter 47 are mounted on a polishing table 3 and rotate integrally with the polishing table 3 and the polishing pad 2. The optical sensor head 7 is positioned to traverse the surface of the substrate W on the polishing pad 2 during each revolution of the polishing table 3 and the polishing pad 2.

[0086] The processing system 49 includes: a storage device 49a storing a program for performing the spectrum generation and film thickness detection of the substrate W (described later); and a processing device 49b performing calculations according to instructions contained in the program. The processing system 49 comprises at least one computer. The storage device 49a includes a main storage device such as RAM and an auxiliary storage device such as a hard disk drive (HDD) or a solid-state drive (SSD). Examples of the processing device 49b include a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit). However, the specific structure of the processing system 49 is not limited to these examples.

[0087] Light emitted from light source 44 is transmitted to optical sensor head 7 and guided from optical sensor head 7 to the surface of substrate W. The light is reflected from the surface of substrate W, and the reflected light from the surface of substrate W is received by optical sensor head 7 and sent to beam splitter 47. Beam splitter 47 decomposes the reflected light according to wavelength and measures the intensity of the reflected light at each wavelength. The intensity measurement data of the reflected light is sent to processing system 49.

[0088] The processing system 49 is configured to generate a spectrum of reflected light based on intensity measurement data. The spectrum of reflected light is represented as a graph (i.e., a spectroscopic waveform) showing the relationship between the wavelength and intensity of the reflected light. The intensity of the reflected light can also be represented as a relative value such as reflectance or relative reflectance.

[0089] Figure 2 This is a diagram showing an example of a spectrum generated by the processing system 49. The spectrum is represented as a graph (i.e., a spectroscopic waveform) showing the relationship between the wavelength and intensity of light. Figure 2 In the diagram, the horizontal axis represents the wavelength of light reflected from the substrate, and the vertical axis represents the relative reflectivity derived from the intensity of the reflected light. Relative reflectivity is an index representing the intensity of reflected light, and is the ratio of the light intensity to a specified reference intensity. By dividing the intensity of light at each wavelength (measured intensity) by the specified reference intensity, unwanted noise such as light learning of the device and inherent intensity fluctuations of the light source can be removed from the measured intensity.

[0090] The reference intensity is the intensity of light measured in advance at each wavelength, and the relative reflectance is calculated for each wavelength. Specifically, the relative reflectance is obtained by dividing the intensity (measured intensity) of light at each wavelength by the corresponding reference intensity. For example, the reference intensity can be obtained by directly measuring the intensity of light emitted from the optical sensor head 7, or by illuminating light from the optical sensor head 7 onto a mirror and measuring the intensity of the reflected light from the mirror. Alternatively, the reference intensity can also be the intensity of reflected light from the silicon substrate measured by the beam splitter 47 when the uncoated silicon substrate (bare substrate) is water-polished on the polishing pad 2, or when the aforementioned silicon substrate (bare substrate) is placed on the polishing pad 2.

[0091] In actual grinding, the corrected measured intensity is obtained by subtracting the dark level (background intensity obtained under shading conditions) from the measured intensity, and then the corrected reference intensity is obtained by subtracting the dark level from the reference intensity. Finally, the relative reflectance is obtained by dividing the corrected measured intensity by the corrected reference intensity. Specifically, the relative reflectance R(λ) can be obtained by the following formula (1).

[0092] (Equation 1)

[0093]

[0094] Here, λ is the wavelength of light reflected from the substrate, E(λ) is the intensity at wavelength λ, B(λ) is the reference intensity at wavelength λ, and D(λ) is the background intensity (dark level) at wavelength λ measured under light-shielding conditions.

[0095] The optical sensor head 7 directs light to the surface of the substrate W (the surface to be polished) with each rotation of the polishing table 3, and receives reflected light from the substrate W. The reflected light is sent to a beam splitter 47. The beam splitter 47 decomposes the reflected light according to wavelength and measures the intensity of the reflected light at each wavelength. The intensity measurement data of the reflected light is sent to a processing system 49, which generates a data processing algorithm based on the intensity measurement data. Figure 2 The spectrum shown. Furthermore, the processing system 49 determines the film thickness of the substrate W by analyzing the spectrum of the reflected light. Figure 2 In the example shown, the spectrum of the reflected light is a spectroscopic waveform that represents the relationship between the relative reflectivity and the wavelength of the reflected light. However, the spectrum of the reflected light can also be a spectroscopic waveform that represents the relationship between the intensity of the reflected light itself and the wavelength of the reflected light.

[0096] like Figure 1As shown, the storage device 49a of the processing system 49 has a database 60 that contains multiple reference spectra. These multiple reference spectra are spectra of reflected light from multiple substrates that have been previously polished; in other words, they are spectra of reflected light generated when a substrate different from substrate W is polished. In the following description, the substrate used in generating the reference spectra is referred to as the reference substrate.

[0097] The processing system 49 consists of at least one computer. The aforementioned at least one computer may also be one or more servers. The processing system 49 may be an edge server connected to the optical splitter 47 via a communication line, or a cloud server connected to the optical splitter 47 via a communication network such as the Internet or a local area network, or a fog computing element (gateway, fog server, router, etc.) set up in the network connected to the optical splitter 47.

[0098] The processing system 49 may be multiple servers connected via a communication network such as the Internet or a local area network. For example, the processing system 49 may be a combination of an edge server and a cloud server. In one embodiment, the database 60 is located in a data server (not shown) at a location separate from the processing device 49b.

[0099] Figure 3 (a) to Figure 3 (c) is a schematic diagram representing an example of processing system 49. Figure 3 (a) represents an example where the entire processing system 49 is configured as a controller within a factory equipped with a grinding table 3 and a grinding head 1. In this example, the processing system 49, together with the grinding table 3 and the grinding head 1, constitutes a single device.

[0100] Figure 3 (b) represents an example where the processing system 49 is configured within a fog server 500 located within a factory. The fog server 500 is connected to the beam splitter 47 via a gateway 400. Examples of the gateway 400 include communication connection devices such as routers. The gateway 400 can be connected to the beam splitter 47 and / or the fog server 500 via a wired connection or wireless connection. In one embodiment, the processing system 49 may also be located within the gateway 400. The implementation where the processing system 49 is located within the gateway 400 is suitable for high-speed processing of intensity measurement data of reflected light transmitted from the beam splitter 47. On the other hand, the implementation where the processing system 49 is located within the fog server 500 is suitable for situations where high-speed processing is not required. In one embodiment, multiple computers constituting the processing system 49 may also be located on both the gateway 400 and the fog server 500.

[0101] Figure 3(c) indicates an example where the processing system 49 is located within a cloud server 600 configured outside the factory. The cloud server 600 is connected to the splitter 47 via a fog server 500 and a gateway 400. The fog server 500 can also be omitted. Figure 3 The embodiment shown in (c) is applied to a situation where multiple grinding devices are connected to a cloud server 600 via a communication network, and the processing system 49 processes a large amount of data.

[0102] Return to Figure 1 The processing system 49 is connected to a polishing control unit 9 for controlling the polishing operation of the substrate W. The polishing control unit 9 controls the polishing operation of the substrate W based on the film thickness of the substrate W determined by the processing system 49. For example, the polishing control unit 9 is configured to determine the polishing endpoint as the time point when the film thickness of the substrate W reaches the target film thickness, or to change the polishing conditions of the substrate W when the film thickness of the substrate W reaches a predetermined value.

[0103] Figure 4 It means Figure 1 The diagram shows a cross-sectional view of one embodiment of the detailed structure of the grinding apparatus. The head shaft 10 is connected to the grinding head motor 18 via a connecting structure 17 and rotates. The rotation of the head shaft 10 causes the grinding head 1 to rotate in the direction indicated by the arrow.

[0104] The beam splitter 47 includes a photodetector 48. In one embodiment, the photodetector 48 is composed of a photodiode, CCD, or CMOS, etc. The optical sensor head 7 is optically connected to the light source 44 and the photodetector 48. The photodetector 48 is electrically connected to the processing system 49.

[0105] The optical film thickness measuring device 40 includes a light-emitting fiber optic cable 31 that guides light emitted from a light source 44 to the surface of a substrate W, and a light-receiving fiber optic cable 32 that receives reflected light from the substrate W and sends the reflected light to a beam splitter 47. The tips of the light-emitting fiber optic cable 31 and the light-receiving fiber optic cable 32 are located inside the polishing table 3.

[0106] The top ends of the optical fiber cable 31 for transmitting light and the top ends of the optical fiber cable 32 for receiving light constitute an optical sensor head 7 that guides light to the surface of the substrate W and receives reflected light from the substrate W. The other end of the optical fiber cable 31 for transmitting light is connected to the light source 44, and the other end of the optical fiber cable 32 for receiving light is connected to the beam splitter 47. The beam splitter 47 is configured to decompose the reflected light from the substrate W according to wavelength and measure the intensity of the reflected light across a predetermined wavelength range.

[0107] Light source 44 transmits light through optical fiber cable 31 to optical sensor head 7, which emits light towards substrate W. Optical sensor head 7 receives reflected light from substrate W and transmits it through optical fiber cable 32 to beam splitter 47. Beam splitter 47 decomposes the reflected light according to its wavelength and measures the intensity of each wavelength. Beam splitter 47 sends the intensity measurement data of the reflected light to processing system 49. Processing system 49 generates a spectrum of the reflected light based on the intensity measurement data.

[0108] The polishing table 3 has a first hole 50A and a second hole 50B opening on its upper surface. Furthermore, a through hole 51 is formed on the polishing pad 2 at a position corresponding to these holes 50A and 50B. Holes 50A and 50B communicate with the through hole 51, which opens onto the polishing surface 2a. The first hole 50A is connected to a liquid supply line 53, and the second hole 50B is connected to a drainage line 54. An optical sensor head 7, composed of the tip of a light-emitting optical fiber cable 31 and the tip of a light-receiving optical fiber cable 32, is disposed in the first hole 50A and located below the through hole 51.

[0109] During the polishing process of substrate W, pure water is supplied as a washing solution to the first hole 50A via liquid supply line 53, and then to the through hole 51 through the first hole 50A. The pure water fills the space between the surface of substrate W (the polished surface) and the optical sensor head 7. The pure water flows into the second hole 50B and is discharged through drainage line 54. The pure water flowing in the first hole 50A and the through hole 51 prevents polishing solution from entering the first hole 50A, thereby ensuring the optical path.

[0110] The optical fiber cable 31 for projection is a light transmission unit that guides light emitted by the light source 44 to the surface of the substrate W. The tips of the optical fiber cable 31 for projection and the optical fiber cable 32 for receiving are located within the first hole 50A and near the polished surface of the substrate W. The optical sensor head 7, formed by the tips of the optical fiber cable 31 for projection and the optical fiber cable 32 for receiving, is positioned facing the substrate W held by the polishing head 1. Whenever the polishing table 3 rotates, light is irradiated onto the surface (polished surface) of the substrate W. In this embodiment, only one optical sensor head 7 is provided in the polishing table 3, but multiple optical sensor heads 7 may also be provided in the polishing table 3.

[0111] Figure 5 This is a schematic diagram illustrating the principle of the optical film thickness measuring device 40. Figure 6 This is a top view showing the positional relationship between the substrate W and the polishing table 3. Figure 5In the example shown, the substrate W has a lower film and an upper film formed on top of the lower film. The upper film is, for example, a silicon layer or an insulating film. An optical sensor head 7, consisting of the tips of a light-emitting optical fiber cable 31 and a light-receiving optical fiber cable 32, is disposed opposite to the surface of the substrate W. The optical sensor head 7 illuminates light onto the surface of the substrate W each time the polishing table 3 rotates once.

[0112] The light irradiated onto the substrate W is carried by the medium ( Figure 5 In the example, the light reflected from the interfaces between water and the upper film, and between the upper and lower films, interferes with each other. The manner of this interference varies depending on the thickness of the upper film (i.e., the optical path length). Therefore, the spectrum generated by the reflected light from the substrate W varies depending on the thickness of the upper film.

[0113] During the polishing process of substrate W, the optical sensor head 7 moves transversely across substrate W for each revolution of the polishing table 3. When the optical sensor head 7 is below substrate W, the light source 44 emits light. Light from the optical sensor head 7 is guided to the surface of substrate W (the polished surface). The reflected light from substrate W is received by the optical sensor head 7 and sent to the beam splitter 47. The beam splitter 47 measures the intensity of reflected light at each wavelength across a specified wavelength range and sends the intensity measurement data to the processing system 49. The processing system 49 generates a spectrum of reflected light representing the intensity of light at each wavelength based on the intensity measurement data.

[0114] like Figure 7 As shown, the processing system 49 compares the spectrum of the reflected light with multiple reference spectra in the database 60 to determine the reference spectrum whose shape is closest to that of the reflected light spectrum. Specifically, the processing system 49 calculates the difference between the spectrum of the reflected light and each reference spectrum to determine the reference spectrum with the smallest calculated difference. Then, the processing system 49 determines the film thickness associated with the determined reference spectrum.

[0115] Each reference spectrum is pre-associated with the film thickness at the time the reference spectrum was acquired. That is, each reference spectrum is acquired at a different film thickness, and multiple reference spectra correspond to multiple different film thicknesses. Therefore, the current film thickness of the substrate W during polishing can be determined by using a reference spectrum whose shape is closest to that of the reflected light spectrum.

[0116] In this embodiment, when the optical sensor head 7 traverses the substrate W once, the optical sensor head 7 continuously emits light to multiple measurement points on the substrate W and receives reflected light from these multiple measurement points. Figure 8 This is a schematic diagram illustrating an example of multiple measurement points on the surface (polished surface) of substrate W. For example... Figure 8As shown, each time the optical sensor head 7 traverses the substrate W, it directs light to multiple measurement points MP and receives reflected light from these multiple measurement points MP. Therefore, each time the optical sensor head 7 traverses the substrate W (i.e., each time the polishing table 3 rotates once), the processing system 49 generates multiple spectra corresponding to the reflected light from the multiple measurement points MP. The generated multiple spectra are stored in the storage device 49a.

[0117] The spectrum of reflected light depends not only on the film thickness of the substrate W, but also on the structural elements (e.g., components, scribing lines, etc.) constituting the surface of the substrate W. Therefore, in this embodiment, in order to improve the accuracy of film thickness measurement of the substrate W, the optical film thickness measuring device 40 determines the film thickness of the substrate W as follows.

[0118] Figure 9 This is a schematic diagram illustrating an example of the spectrum of reflected light from multiple measurement points on substrate W. Figure 9 In the example shown, the first measurement point MP1 and the third measurement point MP3 are on the first structural element R1 (e.g., an element) of the substrate W, and the second measurement point MP2 is on the second structural element R2 (e.g., a scribing line) of the substrate W. The first structural element R1 and the second structural element R2 have different surface structures. Due to this difference in surface structure, the shapes of the spectra S1 and S3 of the reflected light from the first measurement point MP1 and the third measurement point MP3 are significantly different from the shape of the spectrum S2 of the reflected light from the second measurement point MP2.

[0119] Processing system 49 classifies these spectra S1, S2, and S3 based on their shape into primary spectra belonging to the first group and secondary spectra belonging to the second group. Figure 9 In the example shown, the processing system 49 classifies the spectra S1 and S3 of the reflected light from the first measurement point MP1 and the third measurement point MP3 as primary spectra belonging to the first group, and classifies the spectrum S2 of the reflected light from the second measurement point MP2 as secondary spectra belonging to the second group.

[0120] Processing system 49 preferentially uses primary spectra S1 and S3 to determine the film thickness of substrate W compared to secondary spectra S2. Processing system 49 determines the film thickness at the first measurement point MP1 and the third measurement point MP3 using primary spectra S1 and S3, which are the spectra of reflected light from the first measurement point MP1 and the spectra of reflected light from the third measurement point MP3. The film thickness is determined based on a reference... Figure 7 Perform the procedures as described.

[0121] Since the secondary spectrum S2, belonging to the second group, differs significantly in shape from the primary spectra S1 and S3, belonging to the first group, the film thickness determined by the secondary spectrum S2 may be relatively inaccurate. Therefore, the processing system 49 determines the film thickness at the second measurement point MP2, where the secondary spectrum S2 has been acquired, as follows.

[0122] like Figure 9 As shown, the processing system 49 generates a presumed spectrum S2' that corresponds to the secondary spectrum S2 and belongs to the first group using primary spectra S1 and S3. More specifically, the processing system 49 uses the primary spectra S1 and S3 of the first measurement point MP1 and the third measurement point MP3 to generate the presumed spectrum S2' of the second measurement point MP2 by interpolation.

[0123] The estimated spectrum S2' has the shape of a primary spectrum that would be classified as belonging to the first group. That is, the estimated spectrum S2' corresponds to the primary spectrum of reflected light from the second measurement point MP2, assuming the second measurement point MP2 is located on the first structural element R1 (e.g., an element) of the substrate W. The processing system 49 can also generate the estimated spectrum S2' from the primary spectra S1 and S3 by interpolation through the arrangement of the first measurement points MP1, MP2, and MP3. The processing system 49 determines the film thickness using the generated estimated spectrum S2'. This film thickness is determined based on a reference... Figure 7 The process is performed according to the instructions. The primary spectra S1, S3 and the estimated spectrum S2' are stored in the storage device 49a of the processing system 49.

[0124] The estimated spectrum S2' is a primary spectrum that is expected to accurately reflect the film thickness of the substrate W. Therefore, the optical film thickness measuring device 40 can determine the correct film thickness of the second structural element R2, which has a structure different from the first structural element R1 of the substrate W, by using the estimated spectrum S2'. In particular, the optical film thickness measuring device 40 can determine Figure 8 The correct film thickness at all measurement points MP shown.

[0125] In this embodiment, for the sake of simplicity, although the estimated spectrum of a measurement point is generated from the primary spectra of two measurement points, the present invention is not limited to this embodiment. It is also possible to generate the estimated spectrum of a measurement point from the primary spectra of three or more measurement points. Depending on the surface structure of the substrate, the film thickness may also be determined preferentially using the spectrum of reflected light from the scribing lines. In such cases, the spectrum of reflected light from the scribing lines is classified as a primary spectrum belonging to the first group.

[0126] exist Figure 9In the illustrated embodiment, the processing system 49 uses the primary spectra of reflected light from the first measurement point MP1 and the third measurement point MP3, which are located near the second measurement point MP2, to generate the estimated spectrum S2' of the second measurement point MP2. However, in one embodiment, the processing system 49 may also generate the estimated spectrum of the second measurement point MP2 using multiple primary spectra of the second measurement point MP2 along the temporal sequence of the grinding time. Hereinafter, refer to... Figure 10 This implementation method will be described.

[0127] During the polishing process of the substrate W, the processing system 49 generates a spectrum of reflected light from the second measurement point MP2 each time the polishing table 3 rotates, thereby acquiring multiple spectra of the second measurement point MP2. The processing system 49 arranges these multiple spectra along the polishing time, i.e., according to the number of rotations of the polishing table 3, and classifies the multiple spectra based on their shape into primary spectra belonging to a first group and secondary spectra belonging to a second group. Figure 10 In the example shown, the spectrum S2n-1 generated during the (n-1)th rotation of the grinding table 3 is classified as a primary spectrum, the spectrum S2n generated during the nth rotation of the grinding table is classified as a primary spectrum, and the spectrum S2n+1 generated during the (n+1)th rotation of the grinding table 3 is classified as a secondary spectrum.

[0128] Processing system 49 generates a presumed spectrum S2n+1', which is expected to be generated during the (n+1)th rotation of the grinding table 3, using primary spectra S2n-1 and S2n. The presumed spectrum S2n+1' is a primary spectrum corresponding to the secondary spectrum S2n+1 and belonging to the first group. Specifically, processing system 49 uses primary spectra S2n-1 and S2n and extrapolates to generate the presumed spectrum S2n+1'. Processing system 49 determines the film thickness using the generated presumed spectrum S2n+1'. This film thickness is determined based on a reference... Figure 7 The described process is performed. The primary spectra S2n-1, S2n and the estimated spectrum S2n+1' are stored in the storage device 49a of the processing system 49.

[0129] The processing system 49 can also generate an estimated spectrum from multiple primary spectra through interpolation. For example, if spectra S2n-1 and S2n+1 are classified as primary spectra and spectrum S2n is classified as a secondary spectrum, the processing system 49 can also generate an estimated spectrum S2n' that corresponds to the secondary spectrum S2n and belongs to the first group from the primary spectra S2n-1 and S2n+1 through interpolation. Alternatively, the estimated spectrum can be generated from three or more temporally sequential primary spectra through interpolation or extrapolation.

[0130] Figure 11 This is a flowchart illustrating the operation of the optical film thickness measuring device 40 in determining the film thickness of the substrate W.

[0131] In step 1-1, during the polishing process of substrate W, whenever the optical sensor head 7 traverses substrate W (i.e., whenever the polishing table 3 rotates once), the optical sensor head 7 illuminates multiple measurement points on substrate W and receives reflected light from these measurement points.

[0132] In steps 1-2, during the polishing process of the substrate W, the processing system 49 generates multiple spectra of reflected light from multiple measurement points. The generated multiple spectra are stored in the storage device 49a of the processing system 49.

[0133] In steps 1-3, the processing system 49 classifies the multiple spectra of the reflected light into primary spectra belonging to the first group and secondary spectra belonging to the second group based on the shape of each spectrum.

[0134] In steps 1-4, the processing system 49 determines multiple film thicknesses of the substrate W by using multiple primary spectra belonging to the first group.

[0135] In steps 1-5, the processing system 49 uses the aforementioned multiple primary spectra to generate a presumed spectrum. This presumed spectrum is a primary spectrum that corresponds to the secondary spectra classified in steps 1-3 and belongs to the first group. The presumed spectrum is based on a reference... Figure 9 or Figure 10 The process described is generated.

[0136] In steps 1-6, the processing system determines the film thickness of substrate W by estimating the spectrum.

[0137] Steps 1-4 described above can also be performed after steps 1-5 described above. Specifically, after generating the estimated spectrum through steps 1-5, the processing system 49 can determine multiple film thicknesses of the substrate W by using multiple primary spectra and estimated spectra.

[0138] Processing device 49 sends the determined film thickness of substrate W to... Figure 1 and Figure 4 The polishing control unit 9 is shown. The polishing control unit 9 controls the polishing operation of the substrate W based on the film thickness of the substrate W. For example, the polishing control unit 9 determines the polishing endpoint as the time when the film thickness of the substrate W reaches the target film thickness, or changes the polishing conditions of the substrate W when the film thickness of the substrate W reaches a predetermined value.

[0139] The processing system 49 can further calculate the moving average of the film thickness determined by the primary spectrum and the estimated spectrum. The grinding control unit 9 can also determine the grinding endpoint based on the moving average of the film thickness, or it can change the grinding conditions. The moving average of the film thickness can be a moving average over time of multiple film thicknesses along a time sequence, or it can be a moving average over space of multiple film thicknesses at multiple adjacent measurement points. According to the above embodiments, since the multiple film thicknesses along the time sequence and the multiple film thicknesses in adjacent spaces fluctuate less, the values ​​of these moving averages of film thicknesses represent the correct representative values ​​of the multiple film thicknesses.

[0140] In one embodiment, the processing system 49 may also generate an inferred spectrum from a single spectrum instead of using a spectral generation model for interpolation or extrapolation. Figure 9 In the example shown, the processing system 49 inputs the primary spectra S1 and S3 of the reflected light from the first measurement point MP1 and the third measurement point MP3 into the spectral generation model, and outputs the estimated spectrum S2' of the second measurement point MP2 from the spectral generation model. Figure 10 In the example shown, the processing system 49 inputs the primary spectrum S2n-1 generated in the (n-1)th rotation of the grinding table 3 and the primary spectrum S2n generated in the nth rotation of the grinding table 3 into the spectrum generation model, and outputs the estimated spectrum S2n+1' from the spectrum generation model.

[0141] The spectrum generation model is a learning model composed of a neural network that learns to generate a spectrum based on artificial intelligence algorithms. Examples of artificial intelligence algorithms include support vector regression, deep learning, random forests, and decision trees, but this embodiment uses deep learning as an example of machine learning. Deep learning is a learning method based on neural networks with multiple layers of intermediate layers (also called hidden layers). In this specification, machine learning using a neural network consisting of an input layer, two or more intermediate layers, and an output layer will be referred to as deep learning.

[0142] The spectral generation model is stored in the storage device 49a of the processing system 49. The processing system 49 executes machine learning using training data according to instructions contained in a program electrically stored in the storage device 49a, thereby constructing the spectral generation model. The training data used for machine learning includes multiple primary spectra generated when grinding multiple substrates having the same layered structure as the substrate W to be ground. More specifically, the training data includes one of the multiple primary spectra generated when grinding the multiple substrates as the objective variable (forward data) and the other primary spectra as explanatory variables.

[0143] exist Figure 9In the example shown, the variable is the primary spectrum of the first measurement point MP1 and the third measurement point MP3 generated when grinding a substrate, and the objective variable is the primary spectrum of the second measurement point MP2 generated when grinding the substrate. Alternatively, the variable is the primary spectrum of the first measurement point MP1 and the third measurement point MP3 generated when grinding a first substrate, and the objective variable is the primary spectrum of the second measurement point MP2 generated when grinding a second substrate.

[0144] exist Figure 10 In the example shown, the explanatory variables are multiple primary spectra at specified measurement points generated during the grinding of the first and second substrates, and the objective variable is the primary spectrum at the same specified measurement point generated during the grinding of the third substrate. In this example, the primary spectrum used as both the explanatory and objective variables is a time-series primary spectrum.

[0145] The processing system 49 inputs the primary spectrum generated during the grinding of the substrate W to the spectral generation model, and outputs the estimated spectrum from the spectral generation model. Figure 12 This is a schematic diagram illustrating an example of a spectral generation model. For example... Figure 12 As shown, the spectrum generation model consists of a neural network with an input layer 200, multiple intermediate layers 201, and an output layer 202.

[0146] Reference Figures 9 to 12 The generation of the estimated spectrum for explanation can be used Figure 7 The database 60 showing the reference spectrum has been updated. The following refers to... Figure 13 The flowchart shown illustrates one embodiment of updating the database 60 of the reference spectrum.

[0147] In step 2-1, while the reference substrate having the same laminated structure as the substrate W to be polished is polished by the polishing apparatus described above, the optical sensor head 7 illuminates light to multiple measurement points on the reference substrate and receives reflected light from these measurement points. More specifically, whenever the optical sensor head 7 traverses the reference substrate (i.e., whenever the polishing table 3 rotates one revolution), the optical sensor head 7 illuminates light to multiple measurement points on the reference substrate and receives reflected light from these measurement points.

[0148] In step 2-2, the processing system 49 generates multiple spectra of reflected light from the aforementioned multiple measurement points on the reference substrate. The generated multiple spectra are stored in the storage device 49a of the processing system 49.

[0149] In steps 2-3, the processing system 49 classifies the multiple spectra into multiple primary spectra belonging to the first group and secondary spectra belonging to the second group based on the shape of each spectrum.

[0150] In steps 2-4, the processing system 49 generates a presumed spectrum that corresponds to the secondary spectra classified in steps 2-3 and belongs to the first group, based on the multiple primary spectra described above. The presumed spectrum is determined according to a reference... Figure 9 , Figure 10 or Figure 12 It is generated by the described process.

[0151] In steps 2-5, the processing system 49 correlates multiple primary spectra and estimated spectra with multiple film thicknesses at multiple measurement points. The measurement points corresponding to the estimated spectra are the measurement points where light is reflected as represented by the secondary spectra classified in steps 2-3 above.

[0152] In steps 2-6, the processing system adds multiple primary spectra and estimated spectra as reference spectra to the database 60, thereby updating the database 60. The multiple primary spectra and estimated spectra are added to the database 60 in a state associated with their respective film thicknesses.

[0153] Since the substrate W, which is the object of grinding, and the reference substrate have the same stacked structure, the primary spectrum and the estimated spectrum generated during the grinding of substrate W can also be added to database 60 as reference spectra. The primary spectrum and the estimated spectrum generated during the grinding of substrate W can be used as reference spectra for determining the film thickness of other substrates with the same stacked structure.

[0154] In the above embodiments, although an estimated spectrum is generated using multiple primary spectra, in one embodiment, the processing system 49 can determine the film thickness at the measurement point corresponding to the secondary spectrum from multiple film thicknesses determined by multiple primary spectra through interpolation or extrapolation without generating an estimated spectrum. Figure 9 In the example shown, the processing system 49 calculates the film thickness at the second measurement point MP2 by interpolation or extrapolation, based on the film thickness determined by the primary spectrum S1 at the first measurement point MP1 and the primary spectrum S3 at the third measurement point MP3. Figure 10 In the example shown, the processing system 49 calculates the film thickness at a specific time point of the specified measurement point from multiple film thicknesses at different time points of the aforementioned measurement point by interpolation or extrapolation. This embodiment, which does not generate an estimated spectrum, reduces the load on the processing system 49.

[0155] Next, a classification method for classifying spectra into primary and secondary spectra based on their shape will be described. This classification method includes the steps of automatically classifying training spectra, creating a separation model, and inputting the spectrum of reflected light generated during substrate polishing into the classification model.

[0156] Automatic spectral classification involves classifying multiple pre-prepared training spectra into multiple groups (clusters) based on a classification algorithm (clustering algorithm), and further classifying these groups (clusters) into a first cluster and a second cluster. Examples of classification algorithms (clustering algorithms) include k-means and Gaussian mixture model (GMM). The multiple pre-prepared training spectra are the spectra of reflected light obtained during the polishing of multiple sample substrates. These training spectra are stored in storage device 49a.

[0157] The creation of a classification model involves using multiple training spectra classified into a first group and a second group, along with the classification results of these training spectra, to construct a classification model composed of neural networks through machine learning. The construction of the classification model includes determining the parameters of the classification model (weighting coefficients, bias, etc.).

[0158] Figure 14 This is a flowchart illustrating the automatic classification of training spectra and the creation of classification models that constitute the classification methods for spectra.

[0159] In step 3-1, while the sample substrate is being ground using the aforementioned grinding apparatus, the processing system 49 receives intensity measurement data of reflected light from the sample substrate and generates multiple training spectra using this intensity measurement data. The sample substrate may have the same layered structure as the substrate W being ground, or it may not have the same layered structure. Multiple sample substrates are prepared, and the grinding and generation of training spectra for each sample substrate are repeated. The training spectra are stored in the storage device 49a.

[0160] In step 3-2, the processing system 49 classifies the multiple training spectra into multiple groups (clusters) according to the classification algorithm. As mentioned above, well-known clustering algorithms such as k-means and Gaussian mixture model (GMM) are used in the classification algorithm.

[0161] In step 3-3, the processing system 49 further classifies the multiple groups into a first group and a second group. There are also cases where multiple training spectra are classified into three or more groups according to the classification algorithm. In this case, at least one of these groups is classified (selected) as the first group, and at least one of the other groups is classified (selected) as the second group. For example, in the case where multiple training spectra are classified into three groups, one group is classified (selected) as the first group, and the other two groups are classified (selected) as the second group.

[0162] The selection of which of the multiple groups classified according to the classification algorithm is designated as the first group can be preset by the processing system 49 or by the user. For example, the processing system 49 may classify multiple spectra into multiple groups according to the classification algorithm, select the group to which the most spectra belong as the first group, and designate the spectra belonging to the selected first group as the first spectrum. In other examples, the processing system 49 may select the group to which the spectrum with the film thickness profile most consistent with the film thickness profile obtained by an external film thickness measuring instrument belongs as the first group, and designate the spectra belonging to the selected first group as the primary spectrum. In yet another example, the processing system 49 may create a hypothetical model of the laminated structure of the substrate W to be polished, perform a light reflection simulation, generate a hypothetical spectrum (or theoretical spectrum) of the reflected light from the hypothetical model, determine the group to which the spectrum with a shape close to the hypothetical spectrum belongs, select the determined group as the first group, and designate the spectra belonging to the selected first group as the primary spectrum. In another example, the processing system 49 may select the group with the smallest fluctuation in spectral shape as the first group and designate the spectra belonging to the selected first group as primary spectra.

[0163] In steps 3-4, the processing system 49 generates classification training data containing multiple training spectra classified into a first group and a second group, and the classification results of these spectra. The classification training data includes multiple training spectra as explanatory variables and the respective classification results of these training spectra as target variables. For example, the training spectra classified into the first group (explanatory variable) and the numerical values ​​representing the first group (target variable) as classification results are combined. Similarly, the training spectra classified into the second group (explanatory variable) and the numerical values ​​representing the second group (target variable) as classification results are combined. The classification training data is stored in the storage device 49a of the processing system 49.

[0164] In steps 3-5, the processing system 49 uses the above-mentioned classification training data and machine learning to determine the parameters of the classification model (weighting coefficients, bias, etc.).

[0165] Figure 15 This is a schematic diagram illustrating an example of a classification model. For example... Figure 15 As shown, the classification model consists of a neural network having an input layer 250, multiple intermediate layers 251, and an output layer 252. In one embodiment, deep learning is used as the machine learning algorithm for constructing the classification model. The processing system 49 inputs the training spectrum into the input layer 250 of the classification model. Specifically, the processing system 49 inputs the intensity (e.g., relative reflectance) of reflected light at each wavelength constituting the training spectrum into the input layer 250 of the classification model.

[0166] The processing system 49 adjusts the parameters (weights, biases, etc.) of the classification model to output from the output layer 252 a classification result (representing the value of the first or second group) corresponding to the training spectrum input to the input layer 250. This machine learning process results in a classification model that has been fully learned. The classification model is stored in the storage device 49a of the processing system 49.

[0167] During the polishing process of the substrate W, the processing system 49, equipped with a classification model, inputs multiple spectra of reflected light from the substrate W one by one into the classification model and outputs classification results from the model. Based on the classification results output from the model, the processing system 49 classifies the multiple spectra of reflected light into primary spectra belonging to a first group and secondary spectra belonging to a second group. The processing system 49 uses the primary spectra belonging to the first group to determine the film thickness of the substrate W and generates the aforementioned estimated spectrum. The processing system 49 determines the film thickness of the substrate W using the estimated spectrum.

[0168] Figure 16 This is a flowchart illustrating the operation of the optical film thickness measuring device 40 with a classification model in determining the film thickness of the substrate W.

[0169] In step 4-1, during the polishing process of substrate W, whenever the optical sensor head 7 traverses substrate W (i.e., whenever the polishing table 3 rotates once), the optical sensor head 7 illuminates multiple measurement points on substrate W and receives reflected light from these measurement points.

[0170] In step 4-2, during the polishing process of the substrate W, the processing system 49 generates multiple spectra of reflected light from multiple measurement points. The generated multiple spectra are stored in the storage device 49a of the processing system 49.

[0171] In step 4-3, the processing system 49 inputs multiple spectra one by one into the classification model, performs calculations according to the calculation algorithm defined by the classification model, and outputs the classification result from the classification model.

[0172] In step 4-4, the processing system 49 classifies the multiple spectra of the reflected light into primary spectra belonging to the first group and secondary spectra belonging to the second group based on the classification results output from the classification model.

[0173] In steps 4-5, the processing system 49 determines multiple film thicknesses of the substrate W by using multiple primary spectra belonging to the first group.

[0174] In steps 4-6, the processing system generates a presumed spectrum corresponding to the secondary spectrum and belonging to the first group using the multiple primary spectra described above. The presumed spectrum is based on a reference... Figure 9 , Figure 10 or Figure 12 It was generated according to the described implementation method.

[0175] In steps 4-7, the processing system 49 determines the film thickness of the substrate W by estimating the spectrum.

[0176] Steps 4-5 described above can also be performed after steps 4-6 described above. Specifically, after generating the estimated spectrum through steps 4-6, the processing system 49 can determine multiple film thicknesses of the substrate W using multiple primary spectra and estimated spectra.

[0177] The processing device 49 can also replace steps 4-6 and 4-7 above to determine the film thickness at the measurement point corresponding to the secondary spectrum by interpolation or extrapolation from multiple film thicknesses determined by multiple primary spectra.

[0178] The processing system 49 can further calculate the moving average of the film thickness as determined above. Additionally, the polishing control unit 9 can determine the polishing endpoint based on the moving average of the film thickness, or it can change the polishing conditions. The moving average of the film thickness can be a moving average over time of multiple film thicknesses along a time sequence, or it can be a moving average over space of the film thickness at multiple adjacent measurement points. According to the embodiments described above, since the multiple film thicknesses along the time sequence and the multiple film thicknesses in adjacent spaces fluctuate less, the values ​​of these moving averages of the film thickness represent accurate representative values ​​of the multiple film thicknesses.

[0179] A processing system 49, comprising at least one computer, operates according to instructions contained in a program electrically stored in the storage device 49a. That is, the processing system 49 executes the various operational steps of each of the above embodiments according to the instructions contained in the program. The program used to cause the processing system 49 to execute these steps is recorded in a computer-readable recording medium, which is a non-transitory tangible object, and is provided to the processing system 49 via the recording medium. Alternatively, the program may also be input into the processing system 49 via a communication network such as the Internet or a local area network.

[0180] The above embodiments are described with the aim of enabling those skilled in the art to carry out the present invention. Various modifications of the above embodiments can obviously be implemented by those skilled in the art, and the technical concept of the present invention can also be applied to other embodiments. Therefore, the present invention is not limited to the described embodiments, and the technical concept is interpreted in the broadest sense according to the scope of the claims.

Claims

1. A grinding method, characterized in that, While grinding the substrate, multiple spectra of reflected light from two or more measurement points located on a first structural element of the substrate and a second measurement point located on a second structural element of the substrate are generated. The surface structures of the first structural element and the second structural element are different. Based on the shape of each spectrum, the multiple spectra are classified into multiple primary spectra belonging to a first group and secondary spectra belonging to a second group. The film thickness at the measurement point on the substrate is determined by the multiple primary spectra. The primary spectrum is used to determine the film thickness at the second measurement point corresponding to the secondary spectrum.

2. The grinding method as described in claim 1, characterized in that, The process of determining the film thickness at the second measurement point corresponding to the secondary spectrum is as follows: Generate a presumed spectrum that corresponds to the secondary spectrum and belongs to the first group, the presumed spectrum having a shape that would be classified as a primary spectrum belonging to the first group. The film thickness at the second measurement point of the substrate is determined based on the estimated spectrum.

3. The grinding method as described in claim 2, characterized in that, The process of generating the estimated spectrum is the process of generating the estimated spectrum from the plurality of primary spectra by interpolation or extrapolation.

4. The grinding method as described in claim 2, characterized in that, The process of generating the estimated spectrum involves inputting the plurality of primary spectra into a spectrum generation model and outputting the estimated spectrum from the spectrum generation model.

5. The grinding method as described in claim 1, characterized in that, The process of determining the film thickness at the second measurement point corresponding to the secondary spectrum is a process of determining the film thickness at the second measurement point corresponding to the secondary spectrum by interpolation or extrapolation from the film thickness determined by the primary spectrum.

6. The grinding method according to any one of claims 1 to 5, characterized in that, The process of classifying the multiple spectra into primary spectra belonging to the first group and secondary spectra belonging to the second group is as follows: The multiple spectra generated during the grinding process of the substrate are respectively input into the classification model. The plurality of spectra are classified into primary spectra belonging to the first group and secondary spectra belonging to the second group based on the classification results output from the classification model.

7. The grinding method as described in claim 6, characterized in that, It also includes the following processes: While the sample substrate is being polished, multiple training spectra of reflected light from the sample substrate are generated. The multiple training spectra are classified into the first group and the second group. The parameters of the classification model are determined by machine learning using classification training data that includes the multiple training spectra and the classification results of the multiple training spectra.

8. A grinding method, characterized in that, While grinding a reference substrate, multiple spectra of reflected light from two or more measurement points located on a first structural element of the reference substrate and a second measurement point located on a second structural element of the reference substrate are generated. The first structural element and the second structural element have different surface structures. Based on the shape of each spectrum, the multiple spectra are classified into multiple primary spectra belonging to a first group and secondary spectra belonging to a second group. Generate a presumed spectrum that corresponds to the secondary spectrum and belongs to the first group, the presumed spectrum having a shape that would be classified as a primary spectrum belonging to the first group. The plurality of primary spectra and the estimated spectra are respectively correlated with the film thickness. The plurality of primary spectra and the estimated spectra are added as reference spectra to a database, which contains a plurality of reference spectra including the plurality of primary spectra and the estimated spectra. While the substrate is being polished, a spectrum of reflected light from the substrate is generated. Determine a reference spectrum that best matches the spectral shape of the reflected light from the substrate. The film thickness associated with the determined reference spectrum is determined.

9. A grinding apparatus, characterized in that, have: A grinding table that supports a grinding pad; A grinding head that presses a substrate onto the grinding pad and grinds the substrate; An optical sensor head that directs light to two or more measurement points located on a first structural element of the substrate and a second measurement point located on a second structural element of the substrate, and receives reflected light from the measurement points and the second measurement point; as well as A processing system that generates multiple spectra of the reflected light. The processing system is configured as follows: Based on the shape of each spectrum, the multiple spectra are classified into multiple primary spectra belonging to a first group and secondary spectra belonging to a second group. The film thickness at the measurement point on the substrate is determined by using the multiple primary spectra. The primary spectrum is used to determine the film thickness at the second measurement point corresponding to the secondary spectrum.

10. The grinding apparatus as described in claim 9, characterized in that, The processing system is configured as follows: Generate a presumed spectrum that corresponds to the secondary spectrum and belongs to the first group, the presumed spectrum having a shape that would be classified as a primary spectrum belonging to the first group. The film thickness at the second measurement point on the substrate is determined by the estimated spectrum.

11. The grinding apparatus as described in claim 10, characterized in that, The processing system is configured to generate the estimated spectrum from the plurality of primary spectra by interpolation or extrapolation.

12. The grinding apparatus as described in claim 10, characterized in that, The processing system has a spectral generation model. The processing system is configured to input the plurality of primary spectra into the spectral generation model and output the estimated spectrum from the spectral generation model.

13. The grinding apparatus as described in claim 9, characterized in that, The processing system is configured to determine the film thickness at the second measurement point corresponding to the secondary spectrum by interpolation or extrapolation from the film thickness determined by the primary spectrum.

14. The grinding apparatus according to any one of claims 9 to 13, characterized in that, The processing system has a classification model. The processing system is configured as follows: The multiple spectra generated during the grinding process of the substrate are respectively input into the classification model. The plurality of spectra are classified into primary spectra belonging to the first group and secondary spectra belonging to the second group based on the classification results output from the classification model.

15. The grinding apparatus as described in claim 14, characterized in that, The processing system includes a storage device that stores multiple training spectra of reflected light from the sample substrate. The processing system is configured as follows: The multiple training spectra are classified into the first group and the second group. The parameters of the classification model are determined by machine learning using the classification results, which include the plurality of training spectra.

16. A grinding apparatus, characterized in that, have: A grinding table that supports a grinding pad; A grinding head that presses a substrate onto the grinding pad and grinds the substrate; An optical sensor head that directs light to multiple measurement points on the substrate and receives reflected light from the multiple measurement points; as well as The processing system has a storage device. The storage device internally stores a database including multiple reference spectra and multiple spectra of reflected light from two or more measurement points located on a first structural element of the reference substrate and a second measurement point located on a second structural element of the reference substrate, wherein the first structural element and the second structural element have different surface structures. The processing system is configured as follows: Based on the shape of each spectrum, the multiple spectra of the reflected light are classified into multiple primary spectra belonging to the first group and secondary spectra belonging to the second group. Generate a presumed spectrum that corresponds to the secondary spectrum and belongs to the first group, the presumed spectrum having a shape that would be classified as a primary spectrum belonging to the first group. The plurality of primary spectra and the estimated spectra are respectively correlated with the film thickness. The plurality of primary spectra and the estimated spectra are added to the database as reference spectra.

17. A computer-readable recording medium, characterized in that, The program is documented to cause the computer to perform the following steps: The step of generating multiple spectra of reflected light from two or more measurement points located on a first structural element of the substrate and a second measurement point located on a second structural element of the substrate during the substrate polishing process, wherein the surface structures of the first structural element and the second structural element are different; Based on the shape of each spectrum, the steps are as follows: classifying the multiple spectra into multiple primary spectra belonging to the first group and secondary spectra belonging to the second group. The step of determining the film thickness at the measurement point of the substrate using the plurality of primary spectra; as well as The step of using the primary spectrum to determine the film thickness at the second measurement point corresponding to the secondary spectrum.

18. The computer-readable recording medium as described in claim 17, characterized in that, The steps for determining the film thickness at the second measurement point corresponding to the secondary spectrum are as follows: The step of generating a presumed spectrum corresponding to the secondary spectrum and belonging to the first group, wherein the presumed spectrum has a shape that would be classified as a primary spectrum belonging to the first group. The step of determining the film thickness at the second measurement point of the substrate using the estimated spectrum.

19. The computer-readable recording medium as described in claim 18, characterized in that, The step of generating the estimated spectrum is the step of generating the estimated spectrum from the plurality of primary spectra by interpolation or extrapolation.

20. The computer-readable recording medium as described in claim 18, characterized in that, The step of generating the estimated spectrum is to input the plurality of primary spectra into the spectrum generation model and output the estimated spectrum from the spectrum generation model.

21. The computer-readable recording medium as described in claim 17, characterized in that, The step of determining the film thickness at the second measurement point corresponding to the secondary spectrum is to determine the film thickness at the second measurement point corresponding to the secondary spectrum by interpolation or extrapolation from the film thickness determined by the primary spectrum.

22. The computer-readable recording medium as claimed in any one of claims 17 to 21, characterized in that, The step of classifying the plurality of spectra into primary spectra belonging to the first group and secondary spectra belonging to the second group is as follows: The step of inputting the multiple spectra generated during the grinding process of the substrate into the classification model. The step of classifying the plurality of spectra into primary spectra belonging to the first group and secondary spectra belonging to the second group based on the classification results output from the classification model.

23. The computer-readable recording medium as described in claim 22, characterized in that, The program is configured to further cause the computer to perform the following steps: The step of generating multiple training spectra of reflected light from the sample substrate during the grinding process of the sample substrate; The step of classifying the plurality of training spectra into the first group and the second group; as well as The step of determining the parameters of the classification model using classification training data that includes the plurality of training spectra and the classification results of the plurality of training spectra.

24. A computer-readable recording medium, characterized in that, The program is documented to cause the computer to perform the following steps: In the grinding process of the reference substrate, a step is taken to generate multiple spectra of reflected light from two or more measurement points located on a first structural element of the reference substrate and a second measurement point located on a second structural element of the reference substrate, wherein the surface structures of the first structural element and the second structural element are different. Based on the shape of each spectrum, the steps are as follows: classifying the multiple spectra into multiple primary spectra belonging to the first group and secondary spectra belonging to the second group. The step of generating a presumed spectrum that corresponds to the secondary spectrum and belongs to the first group, wherein the presumed spectrum has a shape that would be classified as a primary spectrum belonging to the first group; The step of relating the plurality of primary spectra and the estimated spectra to the film thickness respectively; The step of adding the plurality of primary spectra and the estimated spectra as reference spectra to a database, the database having a plurality of reference spectra including the plurality of primary spectra and the estimated spectra; A step of generating a spectrum of reflected light from a substrate while polishing it; The step of determining a reference spectrum that most closely resembles the spectral shape of the reflected light from the substrate; as well as The step of determining the film thickness associated with the determined reference spectrum.

Images