Method for detecting incorrect workpieces that are not to be polished, and optical film thickness measuring device.

Abstract

To provide a method capable of detecting a wrong workpiece (for example, a wafer) not subjected to polishing.SOLUTION: A method comprises the steps of: preparing spectral data for inspection of reflected light from a workpiece W before polishing or after starting polishing of the workpiece W; inputting the spectral data for inspection to an autoencoder; calculating difference between output data output from the autoencoder and the spectral data for inspection; and, when the difference is larger than a threshold, determining that the workpiece W used to prepare the spectrum data for inspection is a wrong workpiece not subjected to polishing.SELECTED DRAWING: Figure 5

Description

Technical Field

[0001] The present invention relates to a technique for measuring film thickness based on the spectrum of reflected light from a workpiece used in the manufacture of semiconductor devices such as wafers, substrates, and panels, and particularly relates to a technique for detecting an erroneous workpiece that is not a target for polishing.

Background Art

[0002] The manufacturing process of semiconductor devices includes various steps such as a step of polishing insulating films such as SiO₂ and a step of polishing metal films such as copper and tungsten. The manufacturing processes of back-illuminated CMOS sensors and through-silicon vias (TSVs) include a step of polishing a silicon layer (silicon wafer) in addition to the steps of polishing insulating films and metal films. Polishing of the wafer is terminated when the thickness of the film (insulating film, metal film, silicon layer, etc.) constituting the surface thereof reaches a predetermined target value.

[0003] Wafer polishing is performed using a polishing apparatus. In order to measure the film thickness of non-metallic films such as insulating films and silicon layers, polishing apparatuses are generally equipped with an optical film thickness measuring device. This optical film thickness measuring device is configured to guide light emitted from a light source to the surface of the wafer and analyze the spectrum of reflected light from the wafer to detect the film thickness of the wafer. When the film thickness of the wafer reaches the target value, the polishing of the wafer is terminated.

[0004] Wafer polishing is executed according to a polishing recipe set in the polishing apparatus. Since this polishing recipe differs for each type of wafer, it is necessary to change the polishing recipe based on the type of wafer that is the polishing target. Furthermore, different types of wafers produce different spectra of reflected light from the wafers, so the conditions for polishing end point detection also differ.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

[0006] If a wafer that is not intended for polishing (i.e., a wafer of a different type that is not compatible with the polishing recipe set in the polishing machine) is brought into the polishing machine and the machine polishes the wafer, the polishing recipe will not be suitable for that wafer, and the polishing endpoint detection will not be performed properly, making it impossible to perform normal polishing on that wafer.

[0007] Therefore, the present invention provides a method and an optical film thickness measuring apparatus that can detect an incorrect workpiece (e.g., a wafer) that is not intended for polishing. [Means for solving the problem]

[0008] In one embodiment, a method is provided for detecting an incorrect workpiece that is not to be polished, comprising: creating inspection spectral data of reflected light from the workpiece before or after polishing the workpiece; inputting the inspection spectral data into an autoencoder, the autoencoder being a trained model constructed by machine learning using training data including multiple indicator spectral data of multiple reflected lights from a correct workpiece that is to be polished; calculating the difference between the output data output from the autoencoder and the inspection spectral data; and determining that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished when the difference is greater than a threshold.

[0009] In one embodiment, the step of creating the inspection spectral data is a step of creating inspection spectral data of reflected light from the workpiece during the water polishing of the workpiece, which is performed before the chemical mechanical polishing of the workpiece, or in the initial stage of the chemical mechanical polishing of the workpiece. In one embodiment, the method further includes classifying a plurality of spectral data acquired in the past before or after polishing of a workpiece into a plurality of groups according to a clustering algorithm, creating training data including a plurality of index spectral data belonging to one of the plurality of groups, and performing the machine learning using the training data to construct the trained model, which is the autoencoder, wherein the plurality of index spectral data belonging to one of the plurality of groups are a plurality of spectral data of a plurality of reflected light from the correct workpiece to be polished.

[0010] In one embodiment, a method is provided for detecting an incorrect workpiece that is not to be polished, comprising: determining reference spectral data from multiple indicator spectral data of multiple reflected light from a correct workpiece that is to be polished; creating inspection spectral data of reflected light from the workpiece before or after polishing; calculating the difference between the reference spectral data and the inspection spectral data; and determining that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished if the difference is greater than a threshold.

[0011] In one embodiment, the step of creating the inspection spectral data is a step of creating inspection spectral data of reflected light from the workpiece during the water polishing of the workpiece, which is performed before the chemical mechanical polishing of the workpiece, or in the initial stage of the chemical mechanical polishing of the workpiece. In one embodiment, the difference is the Euclidean distance. In one embodiment, the step of determining the reference spectral data includes classifying a plurality of spectral data acquired before or after polishing a past workpiece into a plurality of groups according to a clustering algorithm, and determining the reference spectral data from a plurality of index spectral data belonging to one of the plurality of groups, wherein the plurality of index spectral data belonging to one of the plurality of groups are a plurality of spectral data of multiple reflected light from the correct workpiece to be polished. In one embodiment, the method further includes the steps of normalizing the plurality of indicator spectral data to create a plurality of normalized indicator spectral data, and normalizing the test spectral data to create normalized test spectral data.

[0012] In one embodiment, an optical film thickness measuring device is provided for optically measuring the film thickness of a workpiece, comprising: a light source that emits light; an optical sensor head that irradiates the workpiece with light emitted from the light source and receives reflected light from the workpiece; and a processing system that determines the film thickness of the workpiece based on spectral measurement data of the reflected light from the workpiece, wherein the processing system creates inspection spectral data of the reflected light from the workpiece before or after polishing the workpiece, inputs the inspection spectral data to an autoencoder, the autoencoder is a trained model constructed by machine learning using training data including multiple indicator spectral data of multiple reflected lights from a correct workpiece to be polished, calculates the difference between output data output from the autoencoder and the inspection spectral data, and determines that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished when the difference is greater than a threshold, thus providing an optical film thickness measuring device.

[0013] In one embodiment, the processing system is configured to create inspection spectral data of reflected light from the workpiece during the wet polishing of the workpiece, which is performed before the chemical mechanical polishing of the workpiece, or in the initial stages of the chemical mechanical polishing of the workpiece. In one embodiment, the processing system is configured to classify multiple spectral data acquired before or after polishing of past workpieces into multiple groups according to a clustering algorithm, create training data including multiple index spectral data belonging to one of the multiple groups, and perform machine learning using the training data to construct the trained model, which is the autoencoder, wherein the multiple index spectral data belonging to one of the multiple groups are multiple spectral data of multiple reflected light from the correct workpiece to be polished.

[0014] In one embodiment, an optical film thickness measuring device is provided for optically measuring the film thickness of a workpiece, comprising: a light source that emits light; an optical sensor head that irradiates the workpiece with light emitted from the light source and receives reflected light from the workpiece; and a processing system that determines the film thickness of the workpiece based on spectral measurement data of the reflected light from the workpiece, wherein the processing system determines reference spectral data from a plurality of index spectral data of a plurality of reflected light from a correct workpiece to be polished, creates inspection spectral data of the reflected light from the workpiece before or after polishing the workpiece, calculates the difference between the reference spectral data and the inspection spectral data, and determines that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished when the difference is greater than a threshold.

[0015] In one embodiment, the processing system is configured to create inspection spectral data of reflected light from the workpiece during the wet polishing of the workpiece, which is performed before the chemical mechanical polishing of the workpiece, or in the initial stages of the chemical mechanical polishing of the workpiece. In one embodiment, the difference is the Euclidean distance. In one embodiment, the processing system is configured to classify multiple spectral data acquired before or after polishing of a past workpiece into multiple groups according to a clustering algorithm, and to determine the reference spectral data from multiple index spectral data belonging to one of the multiple groups, wherein the multiple index spectral data belonging to one of the multiple groups are multiple spectral data of multiple reflected light from the correct workpiece to be polished. In one embodiment, the processing system is configured to normalize the plurality of indicator spectral data to create a plurality of normalized indicator spectral data, and to normalize the test spectral data to create normalized test spectral data. [Effects of the Invention]

[0016] An autoencoder is a pre-trained model built using machine learning with training data containing multiple indicator spectral data of reflected light from the correct workpiece to be polished. More specifically, when indicator spectral data is input to the autoencoder, machine learning is performed so that the autoencoder outputs data that is substantially the same as the input indicator spectral data. If spectral data of reflected light from an incorrect workpiece that is not to be polished is input to the autoencoder, the autoencoder will output data that is significantly different from the input spectral data. Therefore, if the difference between the input inspection spectral data and the output data (also called the reconstruction error) is greater than a threshold, it can be determined that the workpiece used to generate that inspection spectral data is an incorrect workpiece that is not to be polished.

[0017] Similarly, when the difference between predetermined reference spectrum data and inspection spectrum data is larger than a threshold value, it is possible to determine that the workpiece used to generate the inspection spectrum data is an incorrect workpiece that is not a polishing target. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] [Figure 1] It is a schematic diagram showing an embodiment of a polishing apparatus. [Figure 2] It is a cross-sectional view showing an embodiment of the detailed configuration of an optical film thickness measuring apparatus. [Figure 3] It is a graph showing an example of spectrum measurement data. [Figure 4] It is a schematic diagram showing an example of an autoencoder. [Figure 5] It is a diagram showing an example when spectrum data of reflected light from a correct workpiece that is a polishing target is input to the autoencoder. [Figure 6] It is a diagram showing an example when spectrum data of reflected light from an incorrect workpiece that is not a polishing target is input to the autoencoder. [Figure 7] It is a flowchart illustrating an embodiment for determining whether a workpiece is an incorrect workpiece that is not a polishing target. [Figure 8] It is a diagram showing a state in which a plurality of spectrum data are divided into three groups by clustering. [Figure 9] It is a flowchart illustrating another embodiment for determining whether a workpiece is an incorrect workpiece that is not a polishing target. [Figure 10] FIG. 10(a) is a graph showing an example of a plurality of index spectrum data before normalization, and FIG. 10(b) is a graph showing an example of a plurality of index spectrum data after normalization. [Figure 11]Figure 11(a) is a graph showing the results of periodically calculating the difference between the reference spectral data determined from multiple indicator spectral data before normalization and the test spectral data before normalization. Figure 11(b) is a graph showing the results of periodically calculating the difference between the reference spectral data determined from multiple indicator spectral data after normalization and the test spectral data after normalization. [Modes for carrying out the invention]

[0019] Embodiments of the present invention will be described below with reference to the drawings. Figure 1 is a schematic diagram showing one embodiment of a polishing apparatus. As shown in Figure 1, the polishing apparatus includes a polishing table 3 that supports a polishing pad 2, a polishing head 1 that presses a workpiece W having a film onto the polishing pad 2, a table motor 6 that rotates the polishing table 3, a liquid supply nozzle 5 for supplying a polishing liquid such as slurry onto the polishing pad 2, and an operation control unit 9 for controlling the operation of the polishing apparatus. The upper surface of the polishing pad 2 constitutes a polishing surface 2a for polishing the workpiece W. Examples of workpieces W include wafers used in the manufacture of semiconductor devices, substrates (circular substrates, polygonal substrates), panels, etc.

[0020] The polishing head 1 is connected to the head shaft 10, which is connected to a polishing head motor (not shown). The polishing head motor rotates the polishing head 1 together with the head shaft 10 in the direction indicated by the arrow. The polishing table 3 is connected to the table motor 6, which is configured to rotate the polishing table 3 and the polishing pad 2 in the direction indicated by the arrow. The polishing head 1, the polishing head motor, the liquid supply nozzle 5, and the table motor 6 are connected to the operation control unit 9.

[0021] The workpiece W is polished as follows: While the polishing table 3 and polishing head 1 are rotated in the direction indicated by the arrows in Figure 1, polishing fluid is supplied from the liquid supply nozzle 5 to the polishing surface 2a of the polishing pad 2 on the polishing table 3. As the workpiece W is rotated by the polishing head 1, the workpiece W is pressed against the polishing surface 2a of the polishing pad 2 by the polishing head 1 while the polishing fluid is present on the polishing pad 2. The surface of the workpiece W is polished by the chemical action of the polishing fluid and the mechanical action of the abrasive particles contained in the polishing fluid and / or the polishing pad 2. Polishing of a workpiece W using such a polishing fluid (e.g., slurry) is called chemical mechanical polishing of the workpiece W.

[0022] The operation control unit 9 comprises a storage device 9a in which a program is stored, and an arithmetic unit 9b that performs calculations according to the instructions contained in the program. The operation control unit 9 is composed of at least one computer. The storage device 9a comprises a main memory such as random access memory (RAM) and an auxiliary storage device such as a hard disk drive (HDD) or solid state drive (SSD). Examples of arithmetic units 9b include a CPU (central processing unit) and a GPU (graphics processing unit). However, the specific configuration of the operation control unit 9 is not limited to these examples.

[0023] The polishing apparatus includes an optical film thickness measuring device 20 for measuring the thickness of the film on the workpiece W. The optical film thickness measuring device 20 includes a light source 22 that emits light, an optical sensor head 25 that irradiates the workpiece W with light from the light source 22 and receives reflected light from the workpiece W, a spectrometer 27 connected to the optical sensor head 25, and a processing system 30 that determines the thickness of the film on the workpiece W based on the reflected light from the workpiece W. The optical sensor head 25 is located in the polishing table 3 and rotates together with the polishing table 3. Although not shown, a plurality of optical sensor heads 25 connected to the spectrometer 27 and the light source 22 may be provided.

[0024] The processing system 30 comprises a storage device 30a in which a program is stored, and an arithmetic unit 30b that performs calculations according to the instructions contained in the program. The processing system 30 consists of at least one computer. The storage device 30a comprises a main memory such as random access memory (RAM) and an auxiliary storage device such as a hard disk drive (HDD) or solid-state drive (SSD). Examples of arithmetic units 30b include a CPU (central processing unit) and a GPU (graphics processing unit). However, the specific configuration of the processing system 30 is not limited to these examples.

[0025] The motion control unit 9 and the processing system 30 may each be composed of multiple computers. For example, the motion control unit 9 and the processing system 30 may each be composed of a combination of edge servers and cloud servers. In one embodiment, the motion control unit 9 and the processing system 30 may be composed of a single computer.

[0026] Figure 2 is a cross-sectional view showing one embodiment of the detailed configuration of the optical film thickness measuring device 20. The optical film thickness measuring device 20 comprises a light source 22 that emits light, a light-emitting optical fiber cable 31 connected to the light source 22, and a light-receiving optical fiber cable 32 connected to a spectrometer 27. The tip 31a of the light-emitting optical fiber cable 31 and the tip 32a of the light-receiving optical fiber cable 32 constitute an optical sensor head 25. That is, the light-emitting optical fiber cable 31 guides the light from the light source 22 to the workpiece W on the polishing pad 2, and the light-receiving optical fiber cable 32 receives the reflected light from the workpiece W and transmits it to the spectrometer 27.

[0027] The spectrometer 27 is connected to the processing system 30. The light-emitting optical fiber cable 31, the light-receiving optical fiber cable 32, the light source 22, and the spectrometer 27 are mounted on the polishing table 3 and rotate together with the polishing table 3 and the polishing pad 2. The optical sensor head 25, consisting of the tip 31a of the light-emitting optical fiber cable 31 and the tip 32a of the light-receiving optical fiber cable 32, is positioned facing the surface of the workpiece W on the polishing pad 2. The position of the optical sensor head 25 is such that it crosses the surface of the workpiece W on the polishing pad 2 each time the polishing table 3 and the polishing pad 2 rotate once. The polishing pad 2 has a through hole 2b located above the optical sensor head 25. The optical sensor head 25 irradiates the workpiece W with light through the through hole 2b and receives reflected light from the workpiece W through the through hole 2b each time the polishing table 3 rotates once.

[0028] The light source 22 is a flash light source that emits light repeatedly at short time intervals. An example of the light source 22 is a xenon flash lamp. The light source 22 is electrically connected to the motion control unit 9 and emits light in response to a trigger signal sent from the motion control unit 9. More specifically, while the optical sensor head 25 moves across the surface of the workpiece W on the polishing pad 2, the light source 22 receives multiple trigger signals and emits light multiple times. Therefore, each time the polishing table 3 rotates once, light is irradiated onto multiple measurement points on the workpiece W.

[0029] Light emitted by the light source 22 is transmitted to the optical sensor head 25 through the light-emitting optical fiber cable 31 and emitted from the optical sensor head 25. The light enters the workpiece W on the polishing pad 2 through the through hole 2b of the polishing pad 2. The light reflected from the workpiece W passes through the through hole 2b of the polishing pad 2 again and is received by the optical sensor head 25. The reflected light from the workpiece W is transmitted to the spectrometer 27 through the light-receiving optical fiber cable 32.

[0030] The spectrometer 27 is configured to decompose light according to its wavelength and measure the intensity of the reflected light at each wavelength over a predetermined wavelength range. That is, the spectrometer 27 decomposes the reflected light from the workpiece W according to its wavelength, measures the intensity of the reflected light at each wavelength over a predetermined wavelength range, and generates spectral measurement data. The spectral measurement data is sent to the processing system 30.

[0031] Figure 3 is a graph showing an example of spectral measurement data. In Figure 3, the vertical axis represents the intensity of reflected light from the workpiece W, and the horizontal axis represents the wavelength of the reflected light. As shown in Figure 3, spectral measurement data can be represented as a spectrum (spectral waveform) that shows the relationship between the wavelength and intensity of light.

[0032] The processing system 30 is configured to determine the film thickness of a workpiece W based on spectral measurement data of reflected light from the workpiece W. Known techniques are used to determine the film thickness of the workpiece W based on spectral measurement data. For example, the processing system 30 determines the reference spectral data that most closely matches the shape of the spectral measurement data from a reference spectral library and determines the film thickness associated with this determined reference spectral data. In another example, the processing system 30 performs a Fourier transform on the spectral measurement data and determines the film thickness from the resulting frequency spectrum.

[0033] The polishing of the workpiece W is performed according to a polishing recipe set in the polishing apparatus. This polishing recipe includes the type of polishing fluid used, the rotation speed of the polishing table, the rotation speed of the polishing head, the pressure of the polishing head against the workpiece W, the film thickness measurement conditions, and the polishing endpoint detection conditions. The polishing recipe differs for each type of workpiece (i.e., the laminated structure that makes up the surface of the workpiece). In particular, different types of workpieces result in different spectra of reflected light from the workpiece, and therefore the conditions for polishing endpoint detection differ.

[0034] If a workpiece that is not intended for polishing (i.e., a workpiece of a different type that is not compatible with the polishing recipe set in the polishing machine) is brought into the polishing machine and the machine attempts to polish the workpiece, the polishing recipe will not be suitable for that workpiece, and the polishing endpoint detection will not be performed properly, making it impossible to perform normal polishing on that workpiece.

[0035] Therefore, in this embodiment, the processing system 30 is configured to determine, before or after polishing the workpiece W, whether the workpiece W is the correct workpiece to be polished or an incorrect workpiece that is not to be polished. A correct workpiece to be polished means a workpiece that corresponds to (is suitable for) the polishing recipe set in the polishing device, and an incorrect workpiece that is not to be polished means a workpiece that does not correspond to (is not suitable for) the polishing recipe set in the polishing device.

[0036] The processing system 30 creates inspection spectral data of reflected light from the workpiece W before or after polishing the workpiece W, inputs this inspection spectral data into an autoencoder, and calculates the difference between the output data output from the autoencoder and the inspection spectral data. The inspection spectral data is generated in the same way as the spectral measurement data described with reference to Figures 2 and 3, so a detailed explanation is omitted.

[0037] In one embodiment, the timing for creating the inspection spectral data is immediately before or immediately after polishing the workpiece W. In one embodiment, the processing system 30 creates inspection spectral data of the reflected light from the workpiece W during the water polishing of the workpiece W, which is performed before chemical mechanical polishing of the workpiece W using an abrasive liquid such as slurry. Water polishing of the workpiece W is performed with water supplied from the liquid supply nozzle 5 onto the polishing surface 2a of the polishing pad 2, instead of an abrasive liquid (e.g., slurry), while the polishing head 1 holds the workpiece W on the polishing surface 2a of the polishing pad 2. Unlike abrasive liquids such as slurry, water has no etching function and does not contain abrasive particles. In addition, the polishing head 1 does not press the workpiece W strongly against the polishing pad 2 during water polishing. Therefore, polishing of the workpiece W does not progress during water polishing, and substantial polishing of the workpiece W does not begin.

[0038] In one embodiment, the processing system 30 may create inspection spectral data of reflected light from the workpiece W during the initial stage of chemical mechanical polishing of the workpiece W using an abrasive liquid such as a slurry. Examples of the initial stage of chemical mechanical polishing include the period from when chemical mechanical polishing of the workpiece W is started until a predetermined time has elapsed, or the period from when chemical mechanical polishing of the workpiece W is started until the polishing table 3 has rotated a predetermined number of times. For example, after chemical mechanical polishing of the workpiece W has started, and while the polishing table 3 is making its first rotation, the processing system 30 creates inspection spectral data of reflected light from the workpiece W.

[0039] The processing system 30 inputs the inspection spectral data generated immediately before or immediately after the start of chemical mechanical polishing of the workpiece W, as described above, into the autoencoder. The autoencoder is a trained model constructed by machine learning using training data that includes multiple indicator spectral data of multiple reflected light from at least one correct workpiece to be polished. The multiple indicator spectral data included in the training data are multiple past spectral data acquired before or after the start of chemical mechanical polishing of the correct workpiece to be polished. For example, the multiple indicator spectral data included in the training data are multiple spectral data obtained during past water polishing or past chemical mechanical polishing of the correct workpiece to be polished.

[0040] The correct workpiece used to generate multiple metric spectral data included in the training data may be one or more. In one embodiment, multiple metric spectral data are generated using multiple correct workpieces to improve the accuracy of machine learning, and training data containing these metric spectral data is created.

[0041] The arithmetic unit 30b of the processing system 30 executes machine learning using training data according to instructions contained in the program stored in the storage device 30a, and constructs an autoencoder. The autoencoder, which is a trained model constructed by machine learning, is stored in the storage device 30a.

[0042] When index spectral data is input to an autoencoder, machine learning is used to ensure that the autoencoder outputs data that is essentially the same as the input index spectral data. However, if spectral data of reflected light from an incorrect workpiece that is not intended for polishing is input to the autoencoder, the autoencoder will output data that is significantly different from the input spectral data.

[0043] The processing system 30 is configured to determine that if the difference (also called the reconstruction error) between the input inspection spectral data and the output data of the autoencoder is greater than a threshold, the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished.

[0044] Figure 4 is a schematic diagram showing an example of an autoencoder. The autoencoder has an encoder 41 that extracts feature quantities from the input spectral data and a decoder 42 that reconstructs the input spectral data from the extracted feature quantities.

[0045] Figure 5 shows an example of what happens when spectral data of reflected light from the correct workpiece to be polished is input to the autoencoder. Since the autoencoder is a trained model built using machine learning with training data that includes multiple indicator spectral data using the correct workpiece to be polished, as shown in Figure 5, when spectral data of the correct workpiece to be polished is input to the autoencoder, the autoencoder outputs output data that is substantially the same as the input spectral data. Therefore, the difference (reconstruction error) between the input spectral data and the output data is small.

[0046] Figure 6 shows an example of what happens when spectral data of reflected light from an incorrect workpiece (not intended for polishing) is input to the autoencoder. As shown in Figure 6, when spectral data of reflected light from an incorrect workpiece (not intended for polishing) is input to the autoencoder, the autoencoder outputs output data that is significantly different from the input spectral data. Therefore, the difference (reconstruction error) between the input spectral data and the output data is large.

[0047] In this way, the processing system 30 determines whether workpiece W is a correct workpiece to be polished or an incorrect workpiece to be polished, based on the difference between the inspection spectral data of workpiece W and the output data obtained from the autoencoder. If the processing system 30 determines that workpiece W is an incorrect workpiece to be polished, the processing system 30 sends the determination result to the operation control unit 9. Upon receiving the determination result, the operation control unit 9 either prevents the polishing device from performing chemical mechanical polishing of workpiece W, or stops the chemical mechanical polishing of workpiece W by the polishing device. Alternatively, upon receiving the determination result, the operation control unit 9 may change the polishing recipe set in the polishing device to a polishing recipe prepared in advance for workpiece W.

[0048] Figure 7 is a flowchart illustrating one embodiment for determining whether or not workpiece W is an incorrect workpiece that is not intended for polishing. In step 101, the processing system 30 creates spectral data for inspection of reflected light from the workpiece W immediately before or immediately after the start of polishing. For example, the processing system 30 creates spectral data for inspection of reflected light from the workpiece W during the water polishing of the workpiece W or in the initial stages of chemical mechanical polishing of the workpiece W.

[0049] In step 102, the processing system 30 inputs the spectral data generated in step 101 to the autoencoder and outputs output data from the autoencoder. In step 103, the processing system 30 calculates the difference between the spectral data generated in step 101 and the output data output from the autoencoder. In step 104, the processing system 30 determines whether the above difference is greater than a threshold. In step 105, the processing system 30 determines that workpiece W is an incorrect workpiece that is not to be polished when the above difference is greater than a threshold.

[0050] In step 106, the processing system 30 sends the determination result of the workpiece W to the operation control unit 9. Upon receiving the determination result, the operation control unit 9 either prevents the polishing of the workpiece W by the polishing device, stops the polishing of the workpiece W by the polishing device, or changes the polishing recipe set in the polishing device to a polishing recipe prepared in advance for the workpiece W. In step 107, if the above difference is less than the threshold, the processing system 30 polishes the workpiece W according to the polishing recipe set in the polishing device.

[0051] The training data used for machine learning in autoencoders should include only the indicator spectral data of the correct workpiece being polished, from among the numerous spectral data obtained during past hydropolishing or chemical mechanical polishing. If the numerous spectral data obtained during past hydropolishing or chemical mechanical polishing includes spectral data of incorrect workpieces that were not being polished, such spectral data should not be used in the training data. In other words, it is necessary to extract the indicator spectral data of the correct workpiece being polished from the numerous spectral data. However, such a task is extremely laborious.

[0052] Therefore, in one embodiment, the processing system 30 classifies multiple spectral data acquired before or after chemical mechanical polishing of a past workpiece into multiple groups according to a clustering algorithm. For example, the processing system 30 classifies a large number of spectral data obtained during past water polishing or past chemical mechanical polishing into multiple groups according to a clustering algorithm. The processing system 30 is then configured to create training data containing multiple index spectral data belonging to one of the multiple groups, and to perform machine learning using the training data to build an autoencoder, which is a trained model. The multiple index spectral data belonging to one of the multiple groups are multiple spectral data of multiple reflected light from the correct workpiece to be polished.

[0053] Clustering, also known as cluster analysis, is a type of artificial intelligence algorithm that classifies multiple spectral data based on certain characteristics. The memory device 30a of the processing system 30 stores a program for dividing multiple spectral data into multiple groups according to the clustering algorithm. The arithmetic unit 30b of the processing system 30 divides the multiple spectral data into multiple groups by performing calculations according to the instructions contained in the program.

[0054] Figure 8 shows how multiple spectral data sets are divided into three groups through clustering. In the example shown in Figure 8, the processing system 30 performs machine learning using spectral data included in the second group, which is the group of index spectral data, as training data. Clustering can quickly divide a vast amount of spectral data acquired in the past into index spectral data and other spectral data. In the example shown in Figure 8, multiple spectral data sets are divided into three groups, but they may also be divided into two groups, or four or more groups.

[0055] Next, another embodiment for determining whether workpiece W is an incorrect workpiece that is not to be polished will be described. The configuration and operation of the optical film thickness measuring device 20 used in this embodiment are the same as those described in the above embodiment with reference to Figures 1 to 8, so a redundant explanation will be omitted.

[0056] In this embodiment, the processing system 30 determines reference spectral data from multiple indicator spectral data of multiple reflected light from the correct workpiece to be polished, creates inspection spectral data of reflected light from the workpiece W before or after polishing the workpiece W, calculates the difference between the reference spectral data and the inspection spectral data, and determines that the workpiece W used to create the inspection spectral data is an incorrect workpiece that is not to be polished if the difference is greater than a threshold.

[0057] In one embodiment, the timing for creating the inspection spectral data is immediately before or immediately after polishing the workpiece W. In one embodiment, the processing system 30 creates the inspection spectral data of the reflected light from the workpiece W during the water polishing of the workpiece W, which is performed before chemical mechanical polishing of the workpiece W. In another embodiment, the processing system 30 creates the inspection spectral data of the reflected light from the workpiece W in the initial stage of chemical mechanical polishing of the workpiece W.

[0058] The reference spectral data may be one or the average of multiple index spectral data of multiple reflected light from the correct workpiece being polished. The multiple index spectral data are multiple past spectral data obtained before or after the start of polishing of the correct workpiece being polished. For example, the multiple index spectral data are multiple spectral data obtained during past water polishing or chemical mechanical polishing of the correct workpiece being polished.

[0059] The correct workpiece used to generate multiple index spectral data may be one or more. In one embodiment, to improve the accuracy of the reference spectral data, multiple index spectral data are generated using multiple correct workpieces, and the reference spectral data is determined from these index spectral data. The reference spectral data is stored in the storage device 30a.

[0060] The processing system 30 calculates the difference between the reference spectral data and the test spectral data. In this embodiment, the processing system 30 calculates the Euclidean distance as the difference between the reference spectral data and the test spectral data.

[0061] When the difference between the reference spectral data and the inspection spectral data is less than a threshold, the processing system 30 determines that the workpiece W used to generate the inspection spectral data is a correct workpiece to be polished. Conversely, when the difference between the reference spectral data and the inspection spectral data is greater than a threshold, the processing system 30 determines that the workpiece W used to generate the inspection spectral data is an incorrect workpiece that is not to be polished.

[0062] Figure 9 is a flowchart illustrating one embodiment for determining whether or not workpiece W is an incorrect workpiece that is not intended for polishing. In step 201, the processing system 30 determines reference spectral data from multiple previously acquired index spectral data. The multiple previously acquired index spectral data are multiple spectral data obtained before or after polishing of the correct workpiece to be polished. For example, the multiple previously acquired index spectral data are multiple spectral data obtained during past water polishing or chemical mechanical polishing of the correct workpiece to be polished. In one embodiment, the reference spectral data is one of the multiple previously acquired index spectral data, or the average of the multiple previously acquired index spectral data.

[0063] In step 202, the processing system 30 creates spectral data for inspection of reflected light from the workpiece W immediately before or after the start of polishing. For example, the processing system 30 creates spectral data for inspection of reflected light from the workpiece W during the water polishing of the workpiece W or in the initial stages of chemical mechanical polishing of the workpiece W. In step 203, the processing system 30 calculates the difference between the reference spectral data and the test spectral data. In step 204, the processing system 30 determines whether the above difference is greater than a threshold. In step 205, the processing system 30 determines that workpiece W is an incorrect workpiece that is not to be polished when the above difference is greater than a threshold.

[0064] In step 206, the processing system 30 sends the determination result of the workpiece W to the operation control unit 9. Upon receiving the determination result, the operation control unit 9 either prevents the polishing of the workpiece W by the polishing device, stops the polishing of the workpiece W by the polishing device, or changes the polishing recipe set in the polishing device to a polishing recipe prepared in advance for the workpiece W. In step 207, if the above difference is less than the threshold, the processing system 30 polishes the workpiece W according to the polishing recipe set in the polishing device.

[0065] The clustering described with reference to Figure 8 can also be applied to this embodiment. That is, the processing system 30 may be configured to classify multiple spectral data acquired in the past into multiple groups according to a clustering algorithm, and to determine a reference spectral data from multiple index spectral data belonging to one of the multiple groups.

[0066] In one embodiment, the processing system 30 may be configured to normalize a plurality of previously acquired index spectral data to create a plurality of normalized index spectral data, and to determine reference spectral data from the plurality of normalized index spectral data. Furthermore, the processing system 30 may be configured to normalize test spectral data to create normalized test spectral data.

[0067] Figure 10(a) is a graph showing an example of multiple index spectral data before normalization, and Figure 10(b) is a graph showing an example of multiple index spectral data after normalization. As can be seen from the comparison between Figure 10(a) and Figure 10(b), normalization can reduce the variability in the levels of index spectral data.

[0068] Figure 11(a) is a graph showing the results of periodically calculating the difference between reference spectral data determined from multiple indicator spectral data before normalization and test spectral data before normalization. Figure 11(b) is a graph showing the results of periodically calculating the difference between reference spectral data determined from multiple indicator spectral data after normalization and test spectral data after normalization. In Figures 11(a) and 11(b), the vertical axis represents the difference (Euclidean distance) between the reference spectral data and the test spectral data, and the horizontal axis represents time.

[0069] As can be seen from the comparison between Figure 11(a) and Figure 11(b), the normalization process can reduce the variability in the difference between the reference spectral data and the inspection spectral data. As a result, the processing system 30 can correctly determine whether or not the workpiece W is the correct workpiece to be polished.

[0070] In the embodiments described so far, one optical sensor head 25 is provided, but the present invention is not limited to the above embodiments, and multiple optical sensor heads 25 may be provided in the polishing table 3. For example, multiple optical sensor heads 25 may be arranged so that they traverse different parts of the workpiece W (e.g., the central region and the edge region) as the polishing table 3 rotates.

[0071] The embodiments described above are intended to enable persons with ordinary skill in the art to implement the present invention. Various modifications of the above embodiments can be made naturally by those skilled in the art, and the technical idea of ​​the present invention can be applied to other embodiments as well. Therefore, the present invention is not limited to the embodiments described, but is to be interpreted in the broadest sense according to the technical idea defined by the claims. [Explanation of Symbols]

[0072] 1 Polishing head 2 polishing pads 2a Polished surface 3 Polishing Table 5. Liquid supply nozzle 6 Table motors 9. Operation Control Unit 10 Head Shaft 20 Optical film thickness measuring device 22 Light source 25 Optical sensor heads 27 Spectrometer 30 Processing Systems 31. Floodlight Fiber Optic Cable 32 Optical fiber receiving cable 41 encoders 42 Decoders W Workpiece

Claims

1. A method for detecting incorrect workpieces that are not intended for polishing, Before polishing the workpiece or after the start of polishing, spectral data for inspecting the reflected light from the workpiece is created. The aforementioned inspection spectral data is input to an autoencoder, which is a trained model constructed by machine learning using training data that includes multiple indicator spectral data of multiple reflected lights from the correct workpiece to be polished. The difference between the output data output from the autoencoder and the inspection spectrum data is calculated. A method for determining that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished, when the difference is greater than a threshold.

2. The method according to claim 1, wherein the step of creating the inspection spectral data is a step of creating inspection spectral data of reflected light from the workpiece during the water polishing of the workpiece which is performed before the chemical mechanical polishing of the workpiece, or in the initial stage of the chemical mechanical polishing of the workpiece.

3. Multiple spectral data obtained from past workpieces before or after polishing are classified into multiple groups according to a clustering algorithm. The training data is created, which includes multiple indicator spectral data belonging to one of the aforementioned multiple groups. The process further includes performing the machine learning using the training data to construct the trained model, which is the autoencoder. The method according to claim 1, wherein the plurality of index spectral data belonging to one of the plurality of groups are plurality of spectral data of plurality of reflected light from the correct workpiece to be polished.

4. A method for detecting incorrect workpieces that are not intended for polishing, The reference spectral data is determined from multiple indicator spectral data of multiple reflected light from the correct workpiece to be polished. During the water polishing of the workpiece, which is performed before the chemical mechanical polishing of the workpiece, or in the initial stages of the chemical mechanical polishing of the workpiece, spectral data for inspection of reflected light from the workpiece is created. The difference between the reference spectral data and the inspection spectral data is calculated. A method for determining that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished, when the difference is greater than a threshold.

5. A method for detecting an incorrect workpiece that is not to be polished, The reference spectral data is determined from multiple indicator spectral data of multiple reflected light from the correct workpiece to be polished. Before polishing the workpiece or after the start of polishing, spectral data for inspecting the reflected light from the workpiece is created. The difference between the reference spectral data and the test spectral data is calculated, and the difference is the Euclidean distance. A method for determining that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished, when the difference is greater than a threshold.

6. A method for detecting an incorrect workpiece that is not to be polished, The reference spectral data is determined from multiple indicator spectral data of multiple reflected light from the correct workpiece to be polished. Before polishing the workpiece or after the start of polishing, spectral data for inspecting the reflected light from the workpiece is created. The difference between the reference spectral data and the inspection spectral data is calculated. The process includes determining that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished, when the difference is greater than a threshold. The step of determining the aforementioned reference spectral data is: Multiple spectral data obtained from past workpieces before or after polishing are classified into multiple groups according to a clustering algorithm. The process includes determining the reference spectral data from a plurality of index spectral data belonging to one of the plurality of groups, The method wherein the multiple index spectral data belonging to one of the multiple groups are multiple spectral data of multiple reflected light from the correct workpiece to be polished.

7. A method for detecting an incorrect workpiece that is not to be polished, Multiple index spectral data of reflected light from the correct workpiece to be polished are normalized to create multiple normalized index spectral data. A reference spectral data is determined from the aforementioned multiple normalized index spectral data. Before polishing the workpiece or after the start of polishing, spectral data for inspecting the reflected light from the workpiece is created. The aforementioned test spectral data is normalized to create normalized test spectral data. The difference between the reference spectral data and the normalized test spectral data is calculated. A method for determining that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished, when the difference is greater than a threshold.

8. An optical film thickness measuring device for optically measuring the film thickness of a workpiece, A light source that emits light, An optical sensor head that irradiates the workpiece with light emitted from the light source and receives reflected light from the workpiece, The system includes a processing system that determines the film thickness of the workpiece based on spectral measurement data of reflected light from the workpiece. The processing system is Before polishing the workpiece or after the start of polishing, spectral data for inspecting the reflected light from the workpiece is created. The aforementioned inspection spectral data is input to an autoencoder, which is a trained model constructed by machine learning using training data that includes multiple indicator spectral data of multiple reflected lights from the correct workpiece to be polished. The difference between the output data output from the autoencoder and the inspection spectrum data is calculated. An optical film thickness measuring device configured to determine that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished when the difference is greater than a threshold.

9. The optical film thickness measuring apparatus according to claim 8, wherein the processing system is configured to produce spectral data for inspection of reflected light from the workpiece during the wet polishing of the workpiece, which is performed before the chemical mechanical polishing of the workpiece, or in the initial stages of the chemical mechanical polishing of the workpiece.

10. The processing system is Multiple spectral data obtained from past workpieces before or after polishing are classified into multiple groups according to a clustering algorithm. The training data is created, which includes multiple indicator spectral data belonging to one of the aforementioned multiple groups. The system is configured to perform machine learning using the training data to construct the trained model, which is the autoencoder. The optical film thickness measuring apparatus according to claim 8, wherein the plurality of index spectral data belonging to one of the plurality of groups are plurality of spectral data of plurality of reflected light from the correct workpiece to be polished.

11. An optical film thickness measuring device for optically measuring the film thickness of a workpiece, A light source that emits light, An optical sensor head that irradiates the workpiece with light emitted from the light source and receives reflected light from the workpiece, The system includes a processing system that determines the film thickness of the workpiece based on spectral measurement data of reflected light from the workpiece. The processing system is The reference spectral data is determined from multiple indicator spectral data of multiple reflected light from the correct workpiece to be polished. During the water polishing of the workpiece, which is performed before the chemical mechanical polishing of the workpiece, or in the initial stages of the chemical mechanical polishing of the workpiece, spectral data for inspection of reflected light from the workpiece is created. The difference between the reference spectral data and the inspection spectral data is calculated. An optical film thickness measuring device configured to determine that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished when the difference is greater than a threshold.

12. An optical film thickness measuring device for optically measuring the film thickness of a workpiece, A light source that emits light, An optical sensor head that irradiates the workpiece with light emitted from the light source and receives reflected light from the workpiece, The system includes a processing system that determines the film thickness of the workpiece based on spectral measurement data of reflected light from the workpiece. The processing system is The reference spectral data is determined from multiple indicator spectral data of multiple reflected light from the correct workpiece to be polished. Before polishing the workpiece or after the start of polishing, spectral data for inspecting the reflected light from the workpiece is created. The difference between the reference spectral data and the test spectral data is calculated, and the difference is the Euclidean distance. An optical film thickness measuring device configured to determine that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished when the difference is greater than a threshold.

13. An optical film thickness measuring device for optically measuring the film thickness of a workpiece, A light source that emits light, An optical sensor head that irradiates the workpiece with light emitted from the light source and receives reflected light from the workpiece, The system includes a processing system that determines the film thickness of the workpiece based on spectral measurement data of reflected light from the workpiece. The processing system is The reference spectral data is determined from multiple indicator spectral data of multiple reflected light from the correct workpiece to be polished. Before polishing the workpiece or after the start of polishing, spectral data for inspecting the reflected light from the workpiece is created. The difference between the reference spectral data and the inspection spectral data is calculated. When the difference is greater than a threshold, the workpiece used to create the inspection spectral data is determined to be an incorrect workpiece that is not to be polished. The processing system is Multiple spectral data obtained from past workpieces before or after polishing are classified into multiple groups according to a clustering algorithm. The system is configured to determine the reference spectral data from multiple index spectral data belonging to one of the aforementioned multiple groups, An optical film thickness measuring device in which the multiple index spectral data belonging to one of the multiple groups are multiple spectral data of multiple reflected light from the correct workpiece to be polished.

14. An optical film thickness measuring device for optically measuring the film thickness of a workpiece, A light source that emits light, An optical sensor head that irradiates the workpiece with light emitted from the light source and receives reflected light from the workpiece, The system includes a processing system that determines the film thickness of the workpiece based on spectral measurement data of reflected light from the workpiece. The processing system is Multiple index spectral data of reflected light from the correct workpiece to be polished are normalized to create multiple normalized index spectral data. A reference spectral data is determined from the aforementioned multiple normalized index spectral data. Before polishing the workpiece or after the start of polishing, spectral data for inspecting the reflected light from the workpiece is created. The aforementioned test spectral data is normalized to create normalized test spectral data. The difference between the reference spectral data and the normalized test spectral data is calculated. An optical film thickness measuring device configured to determine that the workpiece used to create the inspection spectral data is an incorrect workpiece that is not to be polished when the difference is greater than a threshold.

Images