JP2000033561A - End point detecting device and end point detecting method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To surely detect the end point in working a thin film by detecting the end by using the signal strengths corresponding to plural wavelengths. SOLUTION: Plural images A1, A2 are picked up with the image pickup operation shifted before and after in time. A pixel PE1 corresponding to a designated position is selected from each of the picked-up images A1, A2. The corresponding pixels such as PE1 or the like have a signal reflective of the relative reflectance of the position on a sample to each wavelength. Therefore, according to the signal strength, the working end point in a designated position is determined by an end point determining part 42. In the case where residual error is within a designated allowable error range, an end point signal is output. The similar operation is repeated for all pixelss in the image A1 (A2) to create an end point distribution map. An end point distribution output part 44 displays the end point distribution map on an input/output part 50.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a semiconductor wafer, a glass substrate for a liquid crystal display, a glass substrate for a photomask,
The present invention relates to a technique for detecting the end point of a process for changing the thickness of a thin film on a substrate (hereinafter, referred to as a “substrate”) such as an optical disk substrate.

[0002]

2. Description of the Related Art In the manufacture of semiconductor devices, a complete planarization process is required to achieve high integration and miniaturization. At present, those developed and operated at the practical level include CMP (Chemical Mechanical Politics).
shing). this is,
A metal film is formed after a groove is formed in the oxide film, and the metal film is polished using an abrasive that selectively acts on the metal film. With this technique, wiring formation and planarization can be achieved simultaneously.

[0003] However, this CMP technique also has a number of problems, one of which is the "metal residual film" in which the metal that should be originally removed is left on the oxide film surface. There is a problem with defects due to

[0004] In order to prevent such defects due to the residual metal film, an operator observes the surface of the substrate with a microscope to determine whether or not the result is acceptable. However, due to a request for automation, the residual film defect is detected using a CCD image sensor. Inspection devices are also appearing. In such an inspection apparatus, it is required to detect that no residual film defect exists, in other words, that the thin film processing has surely progressed to its end point.

[0005]

However, CCDs
When a conventional inspection device that detects a residual film defect using an image sensor is used, the contrast of the image on the substrate surface is poor, and the defect detection accuracy is not always good. Also,
In some cases, there is a problem such as an erroneous determination that the end point has been reached even though the defect due to the metal residual film exists, that is, the state where the thin film processing has not reached the end point.

In view of the above problems, it is an object of the present invention to provide an end point detecting device and an end point detecting method capable of reliably detecting an end point during thin film processing.

[0007]

According to a first aspect of the present invention, there is provided an end point detecting apparatus, comprising: an object having a thin film formed on a surface of a predetermined base material; An end point detection device that detects an end point of processing involving, illuminating means for illuminating the object with light having a plurality of specific wavelengths, and a captured image of the object by receiving reflected light from the object And the signal intensity corresponding to the plurality of specific wavelengths in a predetermined pixel in the captured image is the signal intensity corresponding to the plurality of specific wavelengths at the time of reaching an end point obtained in advance. And end point determination means for outputting an end point signal indicating that processing at a position on the object corresponding to the predetermined pixel has reached an end point when all of the above conditions are satisfied.

According to a second aspect of the present invention, in the end point detecting apparatus according to the first aspect, the end point determining means determines each of the signal intensities of the predetermined pixels corresponding to the plurality of specific wavelengths. The magnitude of a difference vector between the first signal strength vector as each element and the second signal strength vector having each of the signal intensities corresponding to the plurality of specific wavelengths at the time of reaching the end point determined in advance is predetermined. The end point signal is output when the value falls within the allowable error range.

According to a third aspect of the present invention, in the end point detecting apparatus according to the first aspect, the illuminating means includes a predetermined spectral reflectance before the start of processing and a spectral reflectance at the end of the processing, which are obtained in advance for the object. Based on the reflectance, the light is illuminated with light of the plurality of specific wavelengths, the difference between the two reflectances satisfying a predetermined criterion.

An end point detecting device according to a fourth aspect of the present invention is the end point detecting device according to the first aspect, wherein an end point signal distribution displaying means for displaying a distribution of an occurrence state of an end point signal in a predetermined area of the captured image, Is further provided.

According to a fifth aspect of the present invention, in the end point detecting apparatus according to the first aspect, the processing is performed when the end point signals are output for all the pixels included in the predetermined area of the captured image. It is characterized by terminating.

According to a sixth aspect of the present invention, there is provided an end point detecting method for an object on which a thin film is formed on a surface of a predetermined base material, the end point of a process involving a change in the thickness of the thin film. An end point detection method for detecting, wherein an imaging step of illuminating the object with light of a plurality of specific wavelengths, receiving reflected light from the object to obtain a captured image of the object, and the captured image When each of the signal intensities corresponding to the plurality of specific wavelengths in a predetermined pixel within all coincides with each of the signal intensities corresponding to the plurality of specific wavelengths at the time of reaching a previously determined end point, the predetermined And an end point determining step of outputting an end point signal indicating that processing at the position on the object corresponding to the pixel has reached the end point.

According to a seventh aspect of the present invention, in the end point detecting method according to the sixth aspect, in the end point determining step, each of the signal intensities corresponding to the plurality of specific wavelengths in the predetermined pixel is determined. The magnitude of a difference vector between the first signal strength vector as each element and the second signal strength vector having each of the signal intensities corresponding to the plurality of specific wavelengths at the time of reaching the end point determined in advance is predetermined. When the values fall within the permissible error range, all the elements of both signal strength vectors are regarded as coincident, and the end point signal is output.

The end point detecting method according to an eighth aspect of the present invention is the end point detecting method according to the sixth aspect, based on the spectral reflectance before the start of processing and the spectral reflectance at the end of the processing, which are obtained in advance for the object. And a wavelength selecting step of selecting, as the plurality of specific wavelengths, a set of wavelengths in which the difference between the two reflectances satisfies a predetermined criterion.

An end point detection method according to a ninth aspect is the end point detection method according to the sixth aspect, wherein an end point signal distribution displaying step of displaying a distribution of an occurrence state of the end point signal in a predetermined area of the captured image, Is further included.

According to a tenth aspect of the present invention, in the end point detecting method according to the sixth aspect, the processing is performed at the time when the end point signals are output for all the pixels included in the predetermined area of the captured image. It is characterized by terminating.

[0017]

DESCRIPTION OF THE PREFERRED EMBODIMENTS <A. Outline of Apparatus> FIG. 1 is a schematic diagram showing a configuration of an end point detection apparatus 1 according to an embodiment of the present invention. The end point detection device 1 is an end point detection device that detects an end point of the processing when performing processing involving a change in film thickness on a thin film of the sample 9 which is an object having a thin film formed on a predetermined base material surface. .

The end point detecting device 1 includes a light source section 11 for emitting illumination light 11L irradiated on the sample 9, an illumination optical system 20a for guiding the illumination light 11L from the light source section 11 to the sample 9, and a reflection from the sample 9. An imaging optical system 20b that guides the light 31L to a predetermined light receiving position, and an image acquisition unit 31 that receives the reflected light 31L and generates an image signal of the sample 9 are provided.

The illumination optical system 20a has a lens 21, a wavelength selection mechanism 22, a half mirror 23, and an objective lens 24. Illumination light 11L emitted from the light source unit 11 passes through the lens 21 and the wavelength selection mechanism 22. Illumination light 11L incident on the half mirror 23 and reflected by the half mirror 23 is irradiated onto the sample 9 via the objective lens 24. The wavelength selection mechanism 22 includes a rotating disk 22a having a plurality of bandpass filters 22b,
And a drive motor M for rotating the rotating disk 22a. FIG. 2 is a plan view of the rotating disk 22a. As shown in FIG. 2, the rotating disk 22a has a plurality (4 in FIG. 2).
), And each of these filters 22b passes only light of a different specific wavelength. Then, the rotation angle of the rotating disk 22a is changed by the drive motor M to move the desired filter 22b into the optical path, and the position filter 22b selectively illuminates only light of a specific wavelength to illuminate the target object. It is possible. Therefore, it is possible to illuminate the target object with light having a plurality of different specific wavelengths sequentially or alternately by time division.

The imaging optical system 20b has an objective lens 24, a half mirror 23, and a lens 25, and its optical axis Lb and the optical axis La of the illumination optical system 20a intersect at the half mirror 23. That is, in the end point detection device 1, the illumination optical system 20 a and the imaging optical system 20
The form of coaxial epi-illumination in which b shares the optical axis Lb is adopted. The reflected light 31L from the sample 9 is
The light is guided to the image acquisition unit 31 via the half mirror 23 and the lens 25 in this order.

The image acquisition section 31 has a CCD camera 32 having two-dimensionally arranged light receiving elements, and receives reflected light 31L.
Is received to generate an image signal of the sample 9. The generated image signal is transferred to the image storage unit 33 and stored as a captured image.

The end point detecting device 1 further includes a control unit 40. The control unit 40 detects the end point of the processing based on the captured image obtained by the image acquisition unit 31.

FIG. 3 is a diagram showing a schematic configuration of the control unit 40. The control unit 40 includes an end point determination unit 42 for each pixel, an end point distribution output unit 44, and a processing end determination unit 46.

The end point determination unit 42 for each pixel includes the image acquisition unit 31
(For example, images A1 and A2)
End point of processing is determined for each pixel included in. Then, as a result of the determination, an end point signal is output for each pixel that has reached the end point. The end point distribution output unit 44 displays the distribution of the occurrence state of the end point signal in the captured image on a display of the input / output unit 50 (described later). Further, the processing end determination unit 46
It is determined that the processing should be terminated when the end point signals are output for all the pixels included in the captured image.

The end point detecting device 1 also includes an input / output unit 50 which is in charge of a human interface such as a keyboard for inputting various conditions as keys and a display for outputting images.

In this embodiment, the control unit 40 is configured using a computer system (hereinafter, referred to as a "computer"), and includes an end point determination unit 42 for each pixel and an end point distribution output unit 44. , Machining end determination unit 46
Are configured to operate by executing a program. The image acquisition unit 31 is configured as an electric circuit provided in a computer. All of these configurations may be configured as software, or all may be configured as hardware. Further, it may be constructed only partially as software.

<B. Outline of Operation> Next, in the end point detection device 1, when processing a thin film formed on the surface of the sample 9 which is a predetermined object with a change in film thickness, an operation for detecting the end point of the processing is performed. Will be described.
In the following, when forming the damascene wiring, CM
A case where the end point in the polishing process by the P technique is detected will be described.

FIG. 4 is a diagram showing an example of a procedure for forming a damascene wiring. The lower layer structure F0 as shown in FIG.
RIE (Reactive)
By performing ion etching or the like, a groove G is formed on the surface of the oxide film F1, and a pattern as shown in FIG. 4B is generated. Then, a metal wiring layer F2 is formed so as to be buried in the trench G (FIG. 4C). Then, the excess deposited portion H is polished by the CMP technique, and the state shown in FIG. Thus, the metal wiring F2 (G) is formed.

In the following, the base material (layer structure F0 + oxide film F
An object (sample 9) on which a thin film (metal wiring layer F2) is formed on the surface of 1) is polished with a decrease in the thickness of the thin film (metal wiring layer F2), and is polished as shown in FIG. In transitioning from the state to the state shown in FIG. 4D, it is assumed that the end point of the polishing process is detected. In the end point detection operation described below, polishing is performed in a separate polishing device arranged near the end point detection device 1, and each time the polishing process proceeds to a predetermined extent, the processing is temporarily interrupted to remove the sample 9. This is performed by moving to the end point detection device 1. Then, the metal residual film F
If 2 (I) is present (see FIG. 5), that is, if the polishing has not reached the end point, the sample 9 is returned to the polishing apparatus again and polishing is continued. When it is confirmed that the excess deposited portion H has been completely removed (FIG. 4D), that is, the polishing process has reached the end point, the polishing process is terminated.

<Principle of End Point Detection> Before describing the actual operation, the principle of end point detection in the present invention will be described.

FIG. 5 is a diagram showing an intermediate state of polishing when the lower layer structure F0 has a structure in which a lower oxide film F0b and a lower metal wiring layer F0c are formed on a silicon substrate F0a. And, on this layer structure F0,
SiO2 oxide film F1 and aluminum metal wiring layer F2
Is formed.

FIG. 6 is a graph obtained by calculating the spectral reflectance of aluminum when the film thickness D2 (see FIG. 4) is sufficiently large. The horizontal axis represents the wavelength, and the vertical axis represents the relative reflectance with respect to the silicon substrate. Is plotted. This represents the spectral reflectance at the time when the polishing process is started (FIG. 4C). FIG. 7 shows the thickness D of the SiO2 oxide film F1.
1 is a graph obtained by calculating the spectral reflectance when 700 (see FIG. 4) is 700 nm, in which the horizontal axis plots the wavelength and the vertical axis plots the relative reflectance with respect to the silicon substrate. As shown in FIG. 4D, FIG. 7 shows the spectral reflectance when the thickness D2 of the metal wiring layer F2 at the position PA becomes 0 and the polishing reaches the end point. As described above, as the polishing progresses, the spectral reflectance changes from the state in FIG. 6 to the state in FIG.

FIG. 8 is a graph showing the spectral reflectance in each intermediate state of the polishing process, and shows a case where the film thickness D2 of the aluminum metal wiring layer F2 changes sequentially to 100, 20, 5, and 0 nm. It is. In the figure, the spectral reflectance corresponding to each film thickness D2 is represented by each curve (white circles, white triangles, x marks, and black circles are added in order for distinction).

FIG. 9 is a graph showing the change over time in the relative reflectance shown in FIG. 8 during polishing, for each wavelength of 475 nm and 600 nm. Note that the horizontal axis represents the processing time t, and the vertical axis represents the difference from the relative reflectance at the end of polishing, that is, the relative reflectance at the end of polishing (film thickness D2 = 0) from the relative reflectance at the actual processing time t. The value after subtracting the rate is plotted. Two curves corresponding to each wavelength are marked with a white circle (475 nm) and a black circle (600 nm) to distinguish each other.

This difference in relative reflectance changes with the passage of time and becomes zero at the end of polishing (t = 100). However, at the end of the original polishing (t = 10
Even before 0), there is a point in time when the difference in relative reflectance becomes zero. For example, for light having a wavelength of 600 nm, the difference in relative reflectance decreases over time, and once becomes zero at time t = 80. However,
This is, as shown in FIG. 8, for a wavelength of 600 nm.
This corresponds to the fact that the relative reflectance when D2 is 20 nm matches the relative reflectance when D2 is 0 nm. At this point (t = 80), the actual film thickness is not 0 nm but 2 nm.
0 nm. Then, when the time further elapses (as the polishing process proceeds), the difference in the relative reflectance increases after becoming a negative value and becomes zero again. This zero point is the original end point. Similarly, for light having a wavelength of 475 nm, there is also a point where the difference in relative reflectance becomes zero even though it is not the original end point.

Therefore, in such a case, if the end point is determined only by the relative reflectance for a single wavelength, the end point may be determined even though the end point has not yet been reached.

Therefore, in this embodiment, a plurality (2
9) are selected, and as shown in FIG. 9, the point in time when the relative reflectances for those wavelengths become zero at the same time is determined as the end point. This makes it possible to reliably detect the end point. Therefore, the following equation 1 is defined as “residual error Er”.

[0038]

(Equation 1)

Here, λ1 and λ2 represent the specific wavelengths described above, Rm (*) is the current value of the relative reflectance, Rem
(*) Is the value of the relative reflectance at the end of polishing.

FIG. 10 is a graph in which the time t is plotted on the horizontal axis and the value of the residual error Er is plotted on the vertical axis. FIG.
The time on the horizontal axis of 0 corresponds to the time on the horizontal axis in FIG.
As shown in FIG. 10, the value decreases over time,
It takes a local minimum around t = 80, but does not reach zero, then increases once and then decreases to zero at t = 100. By detecting the time point at which the residual error Er becomes zero as the end point, the determination can be made reliably.

However, in practice, the residual error Er often does not become completely zero due to errors due to sampling intervals and the like and calculation errors. Therefore, it is also possible to provide a threshold value ε as an allowable error range, and detect a point in time that falls below the threshold value ε as an end point. The following equation 2 shows the condition for detecting the end point in that case.

[0042]

(Equation 2)

It should be noted that the threshold ε can be set to a value small enough not to cause erroneous detection.

As described above, in the present embodiment, the end point is detected by using light having two wavelengths of 475 nm and 600 nm as a plurality of specific wavelengths. It is preferable to use wavelengths apart from each other as the specific wavelength. This is based on the fact that the present principle utilizes the fact that the refractive index and the absorption coefficient of the film differ depending on the wavelength.

As the specific wavelength, it is preferable to use a wavelength having a large difference in relative reflectance with the progress of processing. This will be described with reference to FIG. FIG. 11 is a graph showing a difference (amount of change) in relative reflectance between a start point (FIG. 6) and an end point (FIG. 7) of polishing processing for each wavelength. As shown in FIG.
The graph has peaks near m and 600 nm, and has the largest variation at these two wavelengths. By using such a wavelength,
A large contrast can be obtained in a captured image described later. For this reason, it is very preferable to select 475 nm and 600 nm as specific wavelengths.

In the above description, the case where two wavelengths are used has been described. However, in general, the case where N wavelengths (N is a natural number of 2 or more) is used can be extended. In this case, the residual error Er may be expressed by the following equation 3 instead of the equation 1.

In general, as a criterion for selecting N wavelengths, for example, a wavelength near each peak in a distribution of a difference in relative reflectance as shown in FIG. 11 or a range within a half value width of a peak is selected. The selection of a plurality of wavelengths within the range can be adopted.

[0048]

(Equation 3)

Here, the vector Rm is a vector in which each of the relative reflectances corresponding to a plurality of specific wavelengths is an element, and the vector Re is a vector corresponding to a plurality of specific wavelengths at the time of reaching a predetermined end point. Is a vector having each of the relative reflectances as elements. When the magnitude of the difference vector between the two vectors falls within a threshold value ε which is a predetermined allowable error, it can be determined that all the elements of the two vectors have coincided, and that the end point has been reached.

<Operation> FIG. 12 is a flowchart showing the procedure for detecting the end point. The actual end point detection operation in the end point detection device 1 will be described with reference to FIG.

First, in step SP10, a plurality of specific wavelengths are determined. As the plurality of specific wavelengths, it is preferable to use wavelengths that satisfy the above conditions. In the present embodiment, two wavelengths of 475 nm (= λ1) and 600 nm (= λ2) are used.

Then, the sample 9 which has been polished to a predetermined degree is placed on the end point detecting device 1 to obtain a picked-up image for the above two specific wavelengths (λ1, λ2) (step SP20). Specifically, first, a bandpass filter 22b that allows only light having a wavelength of 475 nm to pass is selected, and the sample 9 is illuminated with light having a wavelength that passes through the bandpass filter 22b, and an image A1 of the sample 9 is captured. I do. Next, a band-pass filter 22b that allows only light having a wavelength of 600 nm to pass is selected by the wavelength selection mechanism 22,
The sample 9 is illuminated with light having a wavelength passing through the bandpass filter 22b, and an image A2 of the sample 9 is captured. As described above, by performing the imaging operation back and forth in time, a plurality of (two) images A1 and A2 (see FIG. 3) are captured. The imaging of the two images A1 and A2 is performed while the position of the sample 9 is fixed, and the same region on the sample 9 is imaged. This makes it easy to associate the pixels between the images A1 and A2.

Hereinafter, whether or not the processing has reached the end point will be considered based on the captured images A1 and A2. First, it is determined whether or not the processing at a predetermined position on the sample 9 has reached the end point. (Step SP3
0). Therefore, a pixel corresponding to a predetermined position, for example, a position PA (see FIGS. 3 and 4) is selected from each of the captured images A1 and A2. Let these corresponding pixels be P
E1 and PE2 (not shown). These corresponding pixels P
The pixels E1 and PE2 have signals reflecting the relative reflectance of the position PA on the sample 9 for each wavelength. Therefore, based on the strength of these signals, the position PA (FIG. 4)
) Can be determined.

More specifically, using the above equation (2), the end point determination unit 42 (FIG. 3) for each pixel determines the corresponding position P on the sample 9.
It is determined whether the processing of A has reached the end point, and FIG.
As shown in FIG. 0, when the residual error Er falls within a predetermined allowable error range (t = 100), Rm (λ1)
And Re (λ1) match, and Rm (λ2) and Re (λ2)
Assuming that (λ2) matches, an end point signal is output. In addition, in response to the output of the end point signal, an end point detection flag described later is set to “1”.

As described above, it is possible to determine whether the processing has reached the end point at the position PA. Then, the same operation is repeated for all the pixels in the image A1 (A2), whereby an end point distribution map can be created. The end point distribution map indicates whether an end point signal is output at each pixel in the image A1 (A2) based on the end point detection flag. Then, the end point distribution output section 44 outputs the end point distribution map to the input / output section 50.
(Step SP40). However, since the position PB (see FIG. 4) is different from the layer structure at the position PA, it is possible to detect that the processing has progressed to a predetermined film thickness, that is, the end point, based on a different reference from the above. For example, in Equation 2, Re
This can be handled by preparing signal intensity values for the wavelengths λ1 and λ2 at the end point at the position PB as (λ1) and Re (λ2).

FIG. 13 is a diagram for explaining the end point distribution map. FIG. 13A shows an image A1 (A2), and FIG. 13B shows an end point distribution map.
The end point distribution map is the original image A1 (A2) in FIG.
Is displayed in combination with. For example, in the original image A1 (A2) represented by the black-and-white gradation value, the pixel (for example, a white background pixel in the figure) to which the end point signal is output can be represented in blue. Thereby, the distribution of the portion reaching the end point can be easily visually confirmed. Conversely, the visibility can be further enhanced by expressing the pixels (hatched pixels in the figure) corresponding to the portion that has not reached the end point in another color (for example, red). is there.

FIG. 14 shows a part of the end point distribution map (FIG. 13).
(R1) of FIG. The end point distribution map is created based on the captured image obtained as described above each time the processing advances to a predetermined degree. FIG.
(A) shows the state at the start of processing. At the start of processing, since the end point signal has not been output for any pixel, the end point detection flags are all reset. This state is shown as a white background in the figure.

When the above-described detection operation is performed on all pixels of the captured image after further processing, an end point signal is output at positions corresponding to some of the pixels. This state is shown in FIG.
This is shown in FIG. In FIG. 14, the pixels PE1 (x, y) and PE1 ′ (x + 2,
y + 2), the end point detection flag “1” is displayed. As described above, the portion corresponding to the end point detection flag “1” can be displayed in color.
It is also possible to display a portion where the end point detection flag “1” is not set in a different color.

However, in the state of FIG. 14B, there are pixels for which the end point detection flag "1" has not yet been set. That is, sample 9 which has not yet reached the end point
This is a state where the upper processing point exists and the metal residual film exists. Therefore, it is determined that the polishing process should not be completed, and the polishing process is continued (step SP50).

Thereafter, after further processing, the operations from step SP20 to step SP40 are repeated.
An end point distribution map is similarly obtained for the image captured at this point. This state is shown in FIG.
In FIG. 14C, the number of pixels for which the end point detection flag “1” is set is further increased. This is shown in FIG.
Pixels PE1 and PE reaching the end point in (b)
This is because, in addition to 1 ′, further pixels also reach the end point.

However, even in the state as shown in FIG. 14 (c), since there is a processing point on the sample 9 which has not yet reached the end point, the polishing is further continued. After performing a predetermined degree of polishing again, the operations from step SP20 to step SP40 are repeated.

After the processing has progressed to a certain extent, an end point detection signal is output for all the pixels of the captured image A1 (A2). At this point, the processing ends (step SP50). Thereby, the image A
In the region of the sample 9 corresponding to No. 1, processing can be performed to a state corresponding to FIG. 4D so that the defect due to the metal residual film does not surely exist.

<C. Others> In the above embodiment,
In selecting a plurality of specific wavelengths, the spectral reflectance before processing (FIG. 6) and the spectral reflectance after processing are completed (FIG. 7)
Was obtained by calculation, but can also be obtained by actual measurement. For example, an ideal reference sample having no metal residual film is prepared in advance, and the spectral reflectance corresponding to FIG. 7 is obtained by actual measurement for the ideal reference sample, and a plurality of appropriate wavelengths are selected. It is possible to Thereafter, similarly to the above embodiment, it is possible to perform an end point detection operation on the sample 9 to be measured. In both the calculation and the actual measurement, a plurality of wavelengths are determined from the spectral reflectance of a model corresponding to the actual structure of the sample 9.

In the above embodiment, the wavelength selection mechanism 2
By selecting the two specific bandpass filters 22b and illuminating the target object with light having a specific wavelength passing through the specific bandpass filter 22b and taking an image of the target object in a time-division manner, a plurality of operations are performed. Has been captured, and the signal intensities corresponding to the respective wavelengths have been obtained based on the pixels at the same position in the plurality of images. However, the present invention is not limited to this. For example, a wavelength selection mechanism may be provided on the image sensor side, or a plurality of image sensors are arranged side by side and images of a plurality of wavelengths are simultaneously obtained by a dichroic mirror or the like. For example, signal strengths corresponding to respective wavelengths may be obtained based on pixels at the same position in a plurality of images.

In the above embodiment, the captured images A1,
Although the end point distribution map is obtained for all the pixels of A2, the end point distribution map can be obtained only for the pixels in a predetermined area of the images A1 and A2. The same applies to the determination of the end time of the processing.
It is also possible to end the processing when all the pixels included in the second predetermined area have reached the end point.

In the above embodiment, the point in time when the end point has been reached for all the pixels of the captured image in the predetermined area on the sample 9 is set as the processing end point. However, the present invention is not limited to this. For each of the regions,
By applying the present invention, the point in time at which the end point is reached in all the pixels in all the regions may be set as the processing end point.

In the above embodiment, the polishing step has been described as an example. However, the present invention is not limited to this, and the present invention can be applied to the detection of an end point in another processing step such as an etching step. In addition, the present invention is not limited to a processing step in which the film thickness is simply reduced such as a polishing step and an etching step, and is also applicable to a process in which the film thickness is simply increased such as a film forming step. It is possible to detect the end point. In this case, an increase in the film thickness to a predetermined value is detected as an end point. Thus, the present invention can be applied to a case where the film thickness is simply increased and simply reduced, that is, simply changed.

In the above embodiment, the end point detecting device 1
Is provided separately from other polishing apparatuses and the like, but may be configured as an apparatus having both functions as the same apparatus.

[0069]

As described above, according to the end point detecting device according to the first and second aspects and the end point detecting method according to the sixth and seventh aspects, an image is captured by illuminating with light having a plurality of specific wavelengths. When each of the signal intensities corresponding to a plurality of specific wavelengths in a predetermined pixel in a captured image of the target object coincides with each of the signal intensities corresponding to a plurality of specific wavelengths at the time of reaching an end point obtained in advance. Then, an end point signal indicating that the processing at the position on the object corresponding to the predetermined pixel position has reached the end point is output. Therefore, since end detection is performed using each of the signal intensities corresponding to a plurality of wavelengths, erroneous detection due to performing the end point detection for the signal intensity corresponding to only a single wavelength is prevented, and the end point is reliably detected. It is possible to do.

According to the end point detecting device according to the third aspect and the end point detecting method according to the eighth aspect, based on the spectral reflectance before the start of processing and the spectral reflectance at the end of the processing, the object is obtained in advance. Then, a set of wavelengths whose difference between the two reflectances satisfies a predetermined standard is selected as a plurality of specific wavelengths. Therefore, an image with high contrast can be obtained, and the end point can be detected with high accuracy.

According to the end point detecting device according to the fourth aspect and the end point detecting method according to the ninth aspect, the distribution of the occurrence state of the end point signal in a predetermined area of the captured image is displayed. Therefore, whether or not the end point has been reached at each pixel position can be easily visually recognized.

According to the end point detecting device according to the fifth aspect and the end point detecting method according to the tenth aspect, the processing is performed when the end point signal is output for all the pixels included in the predetermined area of the captured image. Since the process is completed, the end point can be reliably detected, and defects in the thin film processing can be prevented.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a schematic configuration of an end point detection device 1 according to an embodiment of the present invention.

FIG. 2 is a plan view of a rotating disk 22a.

FIG. 3 is a diagram showing a schematic configuration of a control unit 40.

FIG. 4 is a diagram showing an example of a procedure for forming a damascene wiring.

FIG. 5 is a diagram showing a layer structure in which a lower layer structure F0 of FIG. 4 is displayed in detail.

FIG. 6 is a graph showing the spectral reflectance at the start of polishing.

FIG. 7 is a graph showing the spectral reflectance at the end of polishing.

FIG. 8 is a diagram showing a spectral reflectance in each intermediate state of polishing.

FIG. 9 is a graph showing a change in relative reflectance with time during polishing.
5 is a graph showing a specific wavelength.

FIG. 10 is a graph showing a temporal change of a residual error Er.

FIG. 11 is a graph showing a difference (change amount) in relative reflectance between a start point and an end point of polishing.

FIG. 12 is a flowchart illustrating a procedure for detecting an end point.

FIG. 13 is a diagram for explaining an end point distribution map.

FIG. 14 is an enlarged view of a region R1 in the end point distribution map.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 End-point detection apparatus 9 Sample 11 Light source part 22 Wavelength selection mechanism 31 Image acquisition part 40 Control part A1, A2 Captured image

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Noriyuki Kondo 4-chome Tenjin Kitamachi 1-chome, Horikawa-dori-Terauchi, Kamigyo-ku, Kyoto F-term (reference) 2F065 AA30 AA53 BB02 BB03 BB17 CC03 CC18 CC19 CC31 CC32 DD09 FF01 GG23 JJ26 LL04 LL22 LL62 NN06 PP02 PP13 PP18 QQ31 3C058 AA07 AC02 BA01 CB02 DA17 4M106 AA01 AA09 BA04 CA19 CA21 CA48 DA15 DH03 DH12 DH31 DJ17 DJ20 DJ23

Claims (10)

[Claims] 1. An end point detection device for detecting an end point of a process involving a change in the thickness of a thin film of an object on which a thin film is formed on a surface of a predetermined base material, wherein the light has a plurality of specific wavelengths. Illuminating means for illuminating the object with; imaging means for receiving reflected light from the object to obtain a captured image of the object; and a plurality of specific wavelengths at predetermined pixels in the captured image. When each of the corresponding signal intensities coincides with each of the signal intensities corresponding to the plurality of specific wavelengths at the time of reaching the end point determined in advance, at a position on the object corresponding to the predetermined pixel. An end point detection device, comprising: an end point determination unit that outputs an end point signal indicating that the processing has reached the end point. 2. The end point detecting device according to claim 1, wherein the end point determining means includes a first signal intensity vector having each of the signal intensities corresponding to the plurality of specific wavelengths in the predetermined pixel as an element. And when the magnitude of the difference vector between the second signal strength vector having each of the signal strengths corresponding to the plurality of specific wavelengths when reaching the end point determined in advance as the respective elements falls within a predetermined allowable error range. And outputting the end point signal. 3. The end point detection device according to claim 1, wherein the illumination unit is configured to determine the two reflections based on a spectral reflectance before the processing and a spectral reflectance at the end of the processing, which are obtained in advance for the object. An end point detecting device, wherein the device is illuminated with light having a plurality of specific wavelengths whose difference in rate satisfies a predetermined criterion. 4. The end point detection device according to claim 1, further comprising end point signal distribution display means for displaying a distribution of an occurrence state of an end point signal in a predetermined area of the captured image. Detection device. 5. The end point detection apparatus according to claim 1, wherein the processing is terminated when an end point signal is output for all pixels included in a predetermined area of the captured image. apparatus. 6. An end point detection method for detecting an end point of a process involving a change in the thickness of a thin film on an object on which a thin film is formed on a surface of a predetermined base material, the method comprising: An imaging step of illuminating the object with light, receiving reflected light from the object to obtain a captured image of the object, and a signal corresponding to the plurality of specific wavelengths at predetermined pixels in the captured image When each of the intensities coincides with each of the signal intensities corresponding to the plurality of specific wavelengths at the time of reaching the end point determined in advance, the processing at the position on the object corresponding to the predetermined pixel ends. An end point determining step of outputting an end point signal indicating that the end point has been reached. 7. The end point detection method according to claim 6, wherein in the end point determination step, a first signal intensity vector having each of the signal intensities corresponding to the plurality of specific wavelengths in the predetermined pixel as an element. And when the magnitude of the difference vector between the second signal strength vector having each of the signal strengths corresponding to the plurality of specific wavelengths when reaching the end point determined in advance as the respective elements falls within a predetermined allowable error range. And outputting the end point signal assuming that all the elements of both signal strength vectors match. 8. The end point detecting method according to claim 6, wherein a difference between the two reflectances is predetermined based on a spectral reflectance before the processing start and a spectral reflectance at the end of the processing, which are obtained in advance for the object. Selecting a set of wavelengths satisfying the above criteria as the plurality of specific wavelengths. 9. The end point detecting method according to claim 6, further comprising: an end point signal distribution displaying step of displaying a distribution of an occurrence state of the end point signal in a predetermined area of the captured image. Detection method. 10. The end point detection method according to claim 6, wherein the processing is terminated when an end point signal is output for all pixels included in a predetermined area of the captured image. Method.