Reflectance spectrum acquisition method, end point detection method and apparatus, and polishing apparatus

Abstract

The present invention relates to a reflectance spectrum acquisition method, an end point detection method and apparatus, and a polishing apparatus.SOLUTION: The reflectance spectrum obtaining method includes obtaining an influence factor distribution curve characterizing a light intensity measurement value variability in different light intensity light source conditions, determining a critical light intensity value corresponding to a case of being dominated by a light source factor based on the influence factor distribution curve, splitting a source beam according to a predetermined proportion β to obtain a required measurement beam and a reference beam, and forming a reflected beam after the measurement beam is incident on a sample. The reflected light intensity is calibrated based on the reference light intensity to generate a target reflected light intensity. A reflectance spectrum of the sample is calculated based on the target reflected light intensity. The measurement beam and the reference beam have the same light intensity variability, and when the reflected light intensity is calibrated based on the reference light intensity, the accuracy and stability of the reflectance spectrum measurement are effectively improved.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to the technical field of chemical mechanical polishing, and in particular to a reflectance spectrum acquisition method, an endpoint detection method and apparatus, and a polishing apparatus. [Background technology]

[0002] Chemical mechanical polishing (CMP) is a key process step in the semiconductor integrated circuit chip manufacturing process. The CMP process uses a polishing pad and a polishing solution to grind wafers or thin films, achieving a planarized surface profile through a combination of mechanical and chemical means. A key challenge in CMP is determining the end point of the polishing process, i.e., whether the polishing process is complete and the target layer of the wafer or thin film has been planarized to the required flatness and thickness. Therefore, an end point detection (EPD) system capable of determining the film thickness during the polishing process is crucial. Spectral detection-based endpoint detection primarily utilizes the principle of optical interference. A broadband light source covering the visible wavelength range is used as the measurement light for irradiating the thin film surface. The light reflected from the upper interface of the thin film interferes with the light reflected from the lower interface to form a reflectance spectrum. Finally, a spectral detector is used to obtain reflectance spectral information related to the film thickness. Given the dispersion model and thin film layer structure of a thin film material with a given dielectric constant, the reflectance spectrum is related only to wavelength and thin film thickness. Therefore, for a given broadband spectral wavelength range, the measured reflectance spectrum can be back-calculated to determine the film thickness. In practical applications, the spectral characteristics of the light source itself have a significant impact on the shape of the reflectance spectrum. Therefore, the reflectance spectrum is divided by the light source spectrum to remove the influence of the light source spectral characteristics and obtain a normalized reflectance spectrum. During the measurement process, different thin film thicknesses correspond to different reflectance spectra. Therefore, the obtained reflectance spectrum can be used to perform a back-calculation algorithm to determine the film thickness, or the reflectance spectrum can be used directly to determine whether wafer or thin film grinding has reached its endpoint during chemical mechanical polishing.

[0003] Currently, in the chemical mechanical polishing endpoint detection process based on spectral detection technology, to meet the demands of system life and intensive sampling, the system light source often adopts a broadband light source capable of high-frequency pulse scintillation. To ensure the stability of the scintillation broadband light source output light, a reference system for detecting light intensity fluctuations is used for endpoint detection. The reference light intensity fluctuations detected by the reference system are mainly used to feed back to a controller, which then uses the controller to adjust the light source voltage to ensure the stability of the measured light intensity. However, because the measured light intensity is inconsistent with the reference light intensity fluctuations, the accuracy of light intensity calibration is poor, which affects the accuracy of endpoint detection. Summary of the Invention

[0004] To overcome the problems existing in the prior art, the present invention provides a reflectance spectrum acquisition method, an endpoint detection method and apparatus, and a polishing apparatus.

[0005] According to a first aspect of an embodiment of the present invention, there is provided a method for acquiring a reflectance spectrum, the method comprising: obtaining influence factor distribution curves characterizing light intensity measurement variability under different light intensity light source conditions, the influence factor distribution curves including a detector factor distribution curve and a light source factor distribution curve; Based on the influence factor distribution curve, the corresponding critical light intensity value I critical determining (λ), where λ is any wavelength value within the light source band range; performing a homogenization focusing process on the light source to obtain a source beam with consistent light intensity fluctuations at each position on the light-emitting surface; The source beam is divided according to a predetermined ratio β to obtain the required measurement and reference beams, and the measurement beam forms a reflected beam after being incident on a sample, and the reference light intensity I of the reference beam is reference (t, λ) and the reflected light intensity I of the reflected beam reflecton (t,λ) is I reflecton (t,λ)≧I reference(t,λ)≧I critical (λ), where t is the light intensity collection time; The reference light intensity I reference (t, λ) based on the reflected light intensity I reflecton (t, λ) is calibrated to obtain the target reflected light intensity I ’ reflecton (t,λ) and The target reflected light intensity I ’ reflecton Based on (t, λ), the reflectance spectrum R of the sample is calculated. means and computing (t, λ).

[0006] In some embodiments, obtaining the detector factor distribution curve comprises: providing a reference test beam that is a broadband beam; collecting output light intensities of the reference test beam at different wavelengths multiple times with a detector; Obtaining output light intensity noise corresponding to the reference test beam and an output light intensity average value corresponding to the reference test beam at different wavelengths, or obtaining output light intensity signal-to-noise ratio corresponding to the reference test beam and an output light intensity average value corresponding to the reference test beam at different wavelengths; establishing a first relationship curve between an output light intensity noise corresponding to the reference test beam and an output light intensity average value corresponding to the reference test beam, or establishing a second relationship curve between an output light intensity signal-to-noise ratio corresponding to the reference test beam and an output light intensity average value corresponding to the reference test beam; Obtaining the illuminant factor distribution curve involves: collecting, with a detector, the output light intensity of the source beam at different wavelengths within the bandpass range multiple times; Obtaining output light intensity noise corresponding to the light source and an output light intensity average value corresponding to the light source at different wavelengths, or obtaining output light intensity signal-to-noise ratio corresponding to the light source and an output light intensity average value corresponding to the light source at different wavelengths; establishing a third relationship curve between the output light intensity noise corresponding to the light source and the output light intensity average value corresponding to the light source, or establishing a fourth relationship curve between the output light intensity signal-to-noise ratio corresponding to the light source and the output light intensity average value corresponding to the light source.

[0007] In some embodiments, based on the influence factor distribution curve, a corresponding critical light intensity value I critical Determining (λ) is A first intersection point between the first relationship curve and the third relationship curve is obtained, a first light intensity value corresponding to the first intersection point is determined, and any light intensity value equal to or greater than the first light intensity value is set as the critical light intensity value I critical (λ), or A second intersection point between the second relationship curve and the fourth relationship curve is obtained, a second light intensity value corresponding to the second intersection point is determined, and any light intensity value equal to or greater than the second light intensity value is set as the critical light intensity value I critical (λ).

[0008] In some embodiments, splitting the source beam according to a predetermined proportion β comprises: and splitting the source beam according to the predetermined proportion β for each location on the source beam emitting surface using a fiber optic beam splitter or a beam splitting lens.

[0009] In some embodiments, the predetermined proportionality β is determined by the measured light intensity I measure (t,λ) and the reference light intensity I reference is the ratio of (t, λ), The predetermined proportionality β satisfies the following:

number

number

[0010] In some embodiments, the reference light intensity I reference (t, λ) based on the reflected light intensity I reflecton (t, λ) is calibrated to obtain the target reflected light intensity I ’ reflecton Generating (t,λ) is Select a time t0 within the collection time period, and calculate the reference light intensity I reference Setting (t0, λ) as the reference initial value; According to the following formula, I reflecton (t, λ) is corrected to obtain the target reflected light intensity I ’ reflecton and obtaining (t, λ).

number

[0011] In some embodiments, the target reflected light intensity I ’ reflecton Based on (t, λ), the reflectance spectrum R of the sample is calculated. means Calculating (t,λ) is The reflectance spectrum R is calculated based on the following formula: means This involves obtaining (t,λ)

number

[0012] According to a second aspect of an embodiment of the present invention, there is provided an endpoint detection method, comprising: measuring a reflectance spectrum R of a sample during a chemical mechanical polishing process; means(t, λ) is collected in real time, and the reflectance spectrum R means (t, λ) is obtained by the reflectance spectrum acquisition method according to any one of the first aspects described above, The reflectance spectrum R means Calculate the film thickness d of the sample in real time based on (t, λ), The end point of the chemical mechanical polishing is determined based on the film thickness d.

[0013] According to a third aspect of an embodiment of the present invention, there is provided an endpoint detection apparatus, the apparatus comprising: a light source for emitting a source beam; a beam homogenization and focusing processor for processing the source beam to obtain a source beam with consistent light intensity fluctuations at each position on the light-emitting surface; a beam splitter for splitting the source beam according to a predetermined ratio β to obtain a measurement beam and a reference beam, the measurement beam forming a reflected beam after being incident on a sample, and a reference light intensity I of the reference beam; reference (t, λ) and the reflected light intensity I of the reflected beam reflecton (t,λ) is I reflecton (t,λ)≧I reference (t,λ)≧I critical (λ) and I critical (λ) is the critical light intensity value corresponding to the case governed by the light source factor, beam splitter; The reflected light intensity I of the reflected beam reflecton (t, λ) and the reference light intensity I of the reference beam reference a detector module for acquiring (t, λ) a control processing unit electrically connected to the detector module, The control processing unit further The reference light intensity I reference (t, λ) based on the reflected light intensity I reflecton (t, λ) is calibrated to obtain the target reflected light intensity I ’ reflecton Generate (t,λ), The target reflected light intensity I ’ reflectonBased on (t, λ), the reflectance spectrum R of the sample is means Calculate (t,λ) and The reflectance spectrum R collected in real time means Based on (t, λ), the thickness d of the sample at the corresponding time t is calculated, and is used to determine the end point of chemical mechanical polishing based on the thickness d.

[0014] According to a fourth aspect of the present invention, there is provided a polishing apparatus, the polishing apparatus including the endpoint detection apparatus according to the third aspect described above.

[0015] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: the reflectance spectrum acquisition method of the embodiments of the present invention performs a homogenization convergence process on the light source, thereby making the light intensity fluctuations at different positions on the light source light-emitting surface homogenized and distributed at the convergence point, i.e., different fluctuations at different positions or regions on the original light-emitting surface of the light source are uniformly distributed at each position or region on the light-emitting surface after convergence, thereby obtaining a source beam with consistent light intensity fluctuations and fluctuation components at each position or region on the light-emitting surface after convergence, and by re-dividing the source beam to obtain a measurement beam and a reference beam, the measurement beam and the reference beam have the same light intensity variability, i.e., solving the problem of inconsistency in the light intensity fluctuations of the measurement beam and the reference beam caused by the light intensity distribution factor of the light source.

[0016] By dividing the reference beam by setting the division ratio β, the reference beam intensity I reference (t, λ), the reflected light intensity of the reflected beam I reflecton (t, λ), and the critical light intensity value I critical (λ) is I reflecton (t,λ)≧I reference (t,λ)≧I critical (λ) relationship is satisfied, and the reflected light intensity I reflecton (t, λ) and reference light intensity I reference The dominant influence factor of (t, λ) can be the light source factor, and the reflected light intensity I reflecton (t, λ) and the reference light intensity I referenceThe influence of the detector factor on (t, λ) is weakened, and at the same time, the variability of the measurement beam and the reference beam after the homogenization process is consistent, so that the reflected light intensity I due to the light source factor is reduced. reflecton (t, λ) and reference light intensity I reference The variability of (t,λ) is also consistent.

[0017] Based on this, the reflectance spectrum acquisition method of this embodiment is to obtain the reflected light intensity I reflecton (t, λ) and reference light intensity I reference (t, λ) can be obtained, and the reference light intensity I reference (t, λ) is used to calculate the reflected light intensity I reflecton By accurately correcting and calibrating the light intensity fluctuations of each wavelength in (t, λ), a highly stable reflectance spectrum can be obtained, improving the accuracy of spectrum detection, and ultimately improving the stability and accuracy of endpoint detection in chemical mechanical grinding. [Brief explanation of the drawings]

[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments in accordance with the present invention and, together with the description, serve to explain the principles of the invention.

[0019] [Figure 1] 1 is a flowchart of a method for obtaining a reflectance spectrum, according to an example embodiment.

[0020] [Figure 2a] 10 is an influence factor distribution curve of light intensity measurement variability illustrated by an example embodiment;

[0021] [Figure 2b] 10 is an influence factor distribution curve of light intensity measurement variability illustrated by another illustrative embodiment;

[0022] [Figure 3a] 1A and 1B are schematic diagrams illustrating variations in the intensity of reflected light and variations in the intensity of reference light in the related art.

[0023] [Figure 3b] 4 is a schematic diagram of reflected light intensity fluctuations and reference light intensity fluctuations illustrated by an example embodiment;

[0024] [Figure 4] FIG. 10 is a comparative diagram of stability before and after correction of reflected light intensity shown by an exemplary embodiment.

[0025] [Figure 5] 1 is a schematic diagram of an endpoint detection apparatus according to an exemplary embodiment;

[0026] [Figure 6] 1 is a schematic diagram of a polishing apparatus according to an exemplary embodiment; DETAILED DESCRIPTION OF THE INVENTION

[0027] Illustrative examples will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numerals in different accompanying drawings refer to the same or similar elements, unless otherwise specified. The embodiments described in the following illustrative examples do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present invention as detailed in the appended claims.

[0028] In order to solve the above technical problems, according to a first aspect of an embodiment of the present invention, there is provided a reflectance spectrum acquisition method, and as shown in FIG. 1, the reflectance spectrum acquisition method includes the following steps: Influence factor distribution curves are obtained that characterize the light intensity measurement variability under different light intensity light source conditions, the influence factor distribution curves including a detector factor distribution curve and a light source factor distribution curve.

[0029] Based on the influence factor distribution curve, the corresponding critical light intensity value I is determined when the light source factor is dominant. critical (λ), where λ is any wavelength value within the light source band range.

[0030] A homogenization focusing process is performed on the light source to obtain a source beam with consistent light intensity fluctuations at each location on the light-emitting surface.

[0031] The source beam is divided according to a predetermined ratio β to obtain the required measurement beam and reference beam, and the measurement beam forms a reflected beam after being incident on the sample, and the reference beam has a reference light intensity I reference (t, λ) and the reflected light intensity I of the reflected beam reflecton (t,λ) is I reflecton (t,λ)≧I reference (t,λ)≧I critical (λ), where t is the light intensity collection time.

[0032] Reference light intensity I reference Based on (t, λ), the reflected light intensity I reflecton (t, λ) is calibrated to obtain the target reflected light intensity I ’ reflecton Generate (t,λ).

[0033] Target reflected light intensity I ’ reflecton (t, λ) based on the reflectance spectrum R of the sample means Calculate (t,λ).

[0034] In the reflectance spectrum acquisition method of the embodiment of the present invention, by performing a homogenization convergence process on the light source, the light intensity fluctuations at different positions on the light source light-emitting surface can be homogenized and distributed at the convergence point, i.e., the different fluctuations at different positions or regions on the original light-emitting surface of the light source are uniformly distributed at each position or region on the light-emitting surface after convergence, so that a source beam with consistent light intensity fluctuations and fluctuation components can be obtained at each position or region on the light-emitting surface after convergence, and by re-dividing the source beam to obtain a measurement beam and a reference beam, the measurement beam and the reference beam have the same light intensity variability, i.e., the problem of inconsistency in the light intensity fluctuations of the measurement beam and the reference beam caused by the light intensity distribution factor of the light source is solved.

[0035] By dividing it according to a predetermined proportionality β, the reference intensity I of the reference beamreference (t, λ), the reflected light intensity of the reflected beam I reflecton (t, λ), and the critical light intensity value I critical (λ) is I reflecton (t,λ)≧I reference (t,λ)≧I critical (λ) relationship is satisfied, and the reflected light intensity I reflecton (t, λ) and reference light intensity I reference The dominant influence factor of (t, λ) can be the light source factor, and the reflected light intensity I reflecton (t,λ), reference light intensity I reference The influence of the detector factor on (t, λ) is weakened, and at the same time, the variability of the measurement beam and the reference beam after the homogenization process is consistent, so that the reflected light intensity I due to the light source factor is reduced. reflecton (t, λ) and reference light intensity I reference The variability of (t,λ) is also consistent.

[0036] Based on this, the reflectance spectrum acquisition method of this embodiment is to obtain the reflected light intensity I reflecton (t, λ) and reference light intensity I reference (t, λ) can be obtained, and the reference light intensity I reference (t, λ) is used to calculate the reflected light intensity I reflecton By accurately correcting and calibrating the light intensity fluctuations at each wavelength in (t, λ), a highly stable reflectance spectrum can be obtained, improving the accuracy of spectrum detection, and ultimately improving the stability and accuracy of EPD in CMP.

[0037] It should be understood that the variability of light intensity is assessed by light intensity measurements collected by a detector, which are affected by two factors: the variability of the light source itself and the inherent characteristics of the detector. In this example, a detector factor distribution curve and a light source factor distribution curve are obtained to characterize the influence of the light source and the detector on the light intensity variability, respectively.

[0038] Illustratively, a detector factor distribution curve is used to characterize the effect of detector background signal fluctuations on light intensity measurements under different light intensities, and the detector factor distribution curve is obtained by the following method. A reference test beam is provided, the reference test beam being a broadband beam.

[0039] A detector is used to collect the output light intensity of the reference test beam multiple times at different wavelengths.

[0040] Obtain output light intensity noise corresponding to the reference test beam and an average output light intensity corresponding to the reference test beam at different wavelengths, or obtain output light intensity signal-to-noise ratio corresponding to the reference test beam and an average output light intensity corresponding to the reference test beam at different wavelengths.

[0041] A first relationship curve is established between the output light intensity noise corresponding to the reference test beam and the output light intensity average value corresponding to the reference test beam, or a second relationship curve is established between the output light intensity signal-to-noise ratio corresponding to the reference test beam and the output light intensity average value corresponding to the reference test beam, and the first relationship curve or the second relationship curve is a detector factor distribution curve.

[0042] Preferably, the reference test beam refers to a beam emitted from a highly stable light source such as a tungsten lamp, and the stability of the self-generated emission light intensity of the reference test beam is higher than the stability of the detector background signal collection value, so that in the output light intensity detected by the detector, the self-generated fluctuations of the emission light intensity are covered by the fluctuations of the detector background signal, and thereby the fluctuations of the output light intensity are the fluctuations of the detector background signal.

[0043] The detector is a spectrometer with a spectral detection function, and can detect the optical intensity value of the reference test beam corresponding to a certain wavelength. Using the detector to collect the output optical intensity of the reference test beam at different wavelengths multiple times specifically involves first adjusting the emission optical intensity of the reference test beam to reach the maximum optical intensity value detectable by the detector, and then using the detector to collect the output optical intensity of the reference test beam at different wavelengths multiple times. Because the reference test beam is a broadband light source with different emission optical intensities at different wavelengths, the output optical intensity collected by the detector is actually not a single optical intensity value but an optical intensity range. The reason for first adjusting the emission optical intensity of the reference test beam to the maximum optical intensity value detectable by the detector is to ensure that the emission optical intensity range of the reference test beam covers the optical intensity measurement range of the detector, and then to obtain background signal fluctuations within the entire optical intensity measurement range of the detector, thereby achieving the critical optical intensity value I critical This helps to obtain (λ) more accurately.

[0044] Obtaining the output light intensity noise corresponding to the reference test beam and the output light intensity average value corresponding to the reference test beam at different wavelengths specifically means that after a detector acquires multiple output light intensity values ​​at a certain wavelength, the standard deviation of the multiple output light intensity values ​​is taken as the output light intensity noise corresponding to the reference test beam, and the average value of the multiple output light intensity values ​​is taken as the output light intensity average value corresponding to the reference test beam, and then similarly sequentially calculating the output light intensity noise and the output light intensity average value corresponding to different wavelengths.

[0045] Establishing the first relationship curve between the output light intensity noise corresponding to the reference test beam and the output light intensity average value corresponding to the reference test beam specifically involves obtaining the output light intensity noise and the output light intensity average value corresponding to different wavelengths, and then fitting the output light intensity noise and the output light intensity average value corresponding to each wavelength into a plane Cartesian coordinate system, respectively, to obtain the first relationship curve.

[0046] Obtaining the output light intensity signal-to-noise ratios and the output light intensity average values ​​corresponding to the reference test beam at different wavelengths specifically means that after a detector obtains multiple output light intensity values ​​at a certain wavelength, the average value of the multiple output light intensity values ​​is taken as the output light intensity average value corresponding to the reference test beam, and the ratio of the output light intensity average value and the multiple output light intensity values ​​as the output light intensity signal-to-noise ratio corresponding to the reference test beam, and then similarly sequentially calculating the output light intensity signal-to-noise ratios and the output light intensity average values ​​corresponding to different wavelengths.

[0047] Establishing the second relationship curve between the output light intensity signal-to-noise ratio corresponding to the reference test beam and the output light intensity average value corresponding to the reference test beam specifically involves obtaining the output light intensity signal-to-noise ratios and output light intensity average values ​​corresponding to different wavelengths, and then fitting the output light intensity signal-to-noise ratios and output light intensity average values ​​corresponding to each wavelength into a plane Cartesian coordinate system, respectively, to obtain the second relationship curve.

[0048] For example, the light source factor distribution curve is used to characterize the influence of light source fluctuations on light intensity measurements under different light intensity light sources, and the light source factor distribution curve is obtained by the following method. A detector is used to collect the output light intensity of the light source multiple times at different wavelengths within the light source band.

[0049] Obtain output light intensity noise corresponding to the light source and an output light intensity average value corresponding to the light source at different wavelengths, or obtain output light intensity signal-to-noise ratio corresponding to the light source and an output light intensity average value corresponding to the light source at different wavelengths.

[0050] A third relationship curve is established between the output light intensity noise corresponding to the light source and the output light intensity average value corresponding to the light source, or a fourth relationship curve is established between the output light intensity signal-to-noise ratio corresponding to the light source and the output light intensity average value corresponding to the light source.

[0051] Specifically, obtaining the output light intensity noise corresponding to the light source and the output light intensity average value corresponding to the light source at different wavelengths means that after the detector obtains multiple output light intensity values ​​at a certain wavelength, the standard deviation of the multiple output light intensity values ​​is taken as the output light intensity noise corresponding to the light source, and the average value of the multiple output light intensity values ​​is taken as the output light intensity average value corresponding to the light source, and then similarly sequentially calculating the output light intensity noise and the output light intensity average value corresponding to different wavelengths.

[0052] Establishing the third relationship curve between the output light intensity noise corresponding to the light source and the output light intensity average value corresponding to the light source specifically involves obtaining the output light intensity noise and the output light intensity average value corresponding to different wavelengths, and then substituting the output light intensity noise and the output light intensity average value corresponding to each wavelength into a plane Cartesian coordinate system for fitting to obtain the third relationship curve.

[0053] Obtaining the output light intensity signal-to-noise ratios corresponding to the reference test beam corresponding to the light source at different wavelengths and the output light intensity average values ​​corresponding to the light source specifically means that after the detector obtains multiple output light intensity values ​​at a certain wavelength, the average value of the multiple output light intensity values ​​is taken as the output light intensity average value corresponding to the light source, and the ratio of the standard deviation of the output light intensity average value and the multiple output light intensity values ​​is taken as the output light intensity signal-to-noise ratio corresponding to the light source, and then similarly sequentially calculating the output light intensity signal-to-noise ratios and output light intensity average values ​​corresponding to different wavelengths.

[0054] Establishing the fourth relationship curve between the output light intensity signal-to-noise ratio corresponding to the light source and the output light intensity average value corresponding to the light source specifically involves obtaining the output light intensity signal-to-noise ratios and output light intensity average values ​​corresponding to different wavelengths, and then fitting the output light intensity signal-to-noise ratios and output light intensity average values ​​corresponding to each wavelength into a plane Cartesian coordinate system, respectively, to obtain the fourth relationship curve.

[0055] In this embodiment, the critical light intensity value I critical (λ) can be obtained by the following method. A first intersection point between the first relationship curve and the third relationship curve is obtained, a first light intensity value corresponding to the first intersection point is determined, and any light intensity value equal to or greater than the first light intensity value is set as a critical light intensity value I critical (λ) is chosen.

[0056] As shown in FIG. 2a, the first light intensity value corresponding to the first intersection point between the first relationship curve and the third relationship curve is 3000. In this embodiment, the critical light intensity value I critical (λ) is defined as 3000. The critical light intensity value I critical Defining (λ) as 3000 is just an example, and any light intensity value greater than 3000 is considered to be the critical light intensity value I critical It should be understood that the λ can also be chosen as

[0057] Or, the critical light intensity value I critical (λ) may be obtained as follows: A second intersection point between the second relationship curve and the fourth relationship curve is obtained, a second light intensity value corresponding to the second intersection point is determined, and any light intensity value equal to or greater than the first light intensity value is set as a critical light intensity value I critical (λ) is chosen.

[0058] As shown in FIG. 2b, the second light intensity value corresponding to the second intersection point between the second relationship curve and the fourth relationship curve is 3000. In this embodiment, the critical light intensity value I critical (λ) is defined as 3000, and of course, any light intensity value greater than 3000 is considered to be the critical light intensity value I critical (λ) may be selected.

[0059] The critical light intensity value I critical The significance of determining (λ) is as follows: measure (t, λ) and reference light intensity I reference (t, λ) are both critical light intensity values ​​I critical When (λ) is larger than λ, the light source factor is dominant in both the light intensity variability influence factors of the test beam and the reference beam, which makes the light intensity variability of the test beam and the light intensity variability of the reference beam more consistent.

[0060] In some embodiments, performing a homogenization focusing process on the light source to obtain a source beam with consistent light intensity variability at each position on the light-emitting surface comprises performing a homogenization focusing process on the light source using a beam homogenization focusing processor.

[0061] The beam homogenization convergence processing unit can homogenize and distribute the light intensity fluctuations at different positions on the light source light-emitting surface at the convergence point, so that the different fluctuations at different positions or regions on the light source light-emitting surface are uniformly distributed at each position or region on the light-emitting surface after convergence, i.e., ensure that the light intensity fluctuations and fluctuation components at each position or region on the light-emitting surface after convergence are both consistent.

[0062] In some embodiments, when the source beam is transmitted through a spatial optical path, a beam splitter is used to split the entire source beam according to a predetermined proportion β, and the beam splitter comprises an optical splitter device such as a beam splitting lens.

[0063] Splitting the entire source beam according to a predetermined proportion β here means using a beam splitter to split the energy of each point on the propagation plane of the source beam according to the predetermined proportion β, which avoids splitting based on geometric division on the propagation plane of the source beam, i.e., avoiding splitting the optical energy of some areas into the measurement beam and the optical energy of other areas into the reference beam, and helps maintain consistency in the optical intensity fluctuations of the measurement beam and the reference beam.

[0064] In some embodiments, when the source beam is transmitted through a fiber optic cable, the beam splitter is a fiber optic beam splitter, and can split the light transmitted through each optical fiber into a measurement beam and a reference beam according to a predetermined ratio β. Common fiber optic beam splitters include a 1:2 fiber optic coupler. When the transmission optical path is multiple optical fibers, multiple 1:2 fiber optic couplers with the same splitting ratio should be used to uniformly split the light transmitted through each optical fiber into a measurement beam and a reference beam.

[0065] In some embodiments, the predetermined proportionality β is determined by the measured light intensity I measure (t, λ) and the reference beam intensity I reference (t,λ), i.e.,

number

number

[0066] The loss coefficient α is affected by the grinding fluid, the grinding pad, and the interconversion coupling of the beam from the optical fiber to the space transmission. The loss coefficient α is determined by the measured light intensity I measure (t, λ) and reflected light intensity I reflecton (t, λ) based on the following formula:

number

[0067] Among these, the predetermined proportionality β is set by adjusting the wavelength value λ within the light source band range. Reflected light intensity I reflecton (t, λ) and the reference light intensity I reference If the light intensity of (t, λ) is below the critical light intensity value I critical (λ), so that the reflected light intensity I at each wavelength value λ reflecton (t, λ) and reference light intensity I reference It is important to ensure the consistency of the (t, λ) fluctuations to avoid miscalibration of the reference system due to inconsistent light intensity fluctuations between the reflected and reference beams.

[0068] FIG. 3b is a schematic diagram of the reflected light intensity fluctuation and the reference light intensity fluctuation after dividing according to the predetermined proportion β. Compared with the schematic diagram of the reflected light intensity fluctuation and the reference light intensity fluctuation in the related art shown in FIG. 3a, the reflected light intensity I reflecton (t, λ) and reference light intensity I reference It can be seen that the consistency of the (t, λ) fluctuations is better than that of conventional related techniques.

[0069] In this embodiment, the reference light intensity I reference (t, λ) and the reflected light intensity I of the reflected beam reflecton (t, λ) can be collected using the same detector module, or can each be collected using a detector module with the same performance parameters.

[0070] In this embodiment, the reference light intensity I reference Based on (t, λ), the reflected light intensity I reflecton To calibrate (t, λ), select a time t0 and calculate the reference light intensity I at t0.reference This involves setting (t0, λ) as the reference initial value. reflecton Correct (t, λ) to obtain the target reflected light intensity I ’ reflecton (t,λ) is obtained.

number

number

[0071] FIG. 4 is a comparison diagram of the reflected light stability before and after correction shown by an exemplary embodiment. As shown in FIG. 4, after calibration using the above method, the fluctuation of reflected light intensity (standard deviation of light intensity value) is effectively reduced, and the stability of the target reflected light intensity is improved.

[0072] In an embodiment of the present invention, the target reflected light intensity I is calculated according to the following formula: ’ reflecton (t, λ) based on the reflectance spectrum R of the sample means Calculate (t,λ).

number

[0073] The step of eliminating the system stray light intensity is performed by illuminating a black sample having a reflectance of 0 under the same measurement light intensity to obtain the system stray light intensity I black This involves obtaining (λ).

[0074] Embodiments of the present invention further provide an endpoint detection method, comprising measuring the reflectance spectrum R of a sample during a chemical mechanical polishing process. means (t, λ) is collected in real time, and the reflectance spectrum R means (t, λ) is obtained using any of the reflectance spectrum acquisition methods described above. The reflectance spectrum R means The film thickness d of the sample is calculated in real time based on (t, λ), and the end point of chemical mechanical polishing is determined based on the film thickness d.

[0075] The film thickness d can be calculated as follows: If the complex refractive index of each layer in the sample thin film structure is known, the reflectance spectrum R means The relationship between (t, λ) and the sample thickness d shows a transcendental function, which can be expressed as follows:

number

[0076] In the end point detection method according to the embodiment of the present invention, the reflectance spectrum R obtained by any of the above-described reflectance spectrum acquisition methods is means By using (t, λ) and performing a homogenization convergence process on the light source, the light intensity fluctuations at different positions on the light source light-emitting surface can be homogenized and distributed at the convergence point, thereby obtaining a source beam with consistent light intensity fluctuations and fluctuation components at each position or region on the light-emitting surface after convergence.By subdividing the source beam to obtain a measurement beam and a reference beam, the measurement beam and the reference beam have the same light intensity fluctuations, thereby solving the problem of inconsistent light intensity fluctuations between the measurement beam and the reference beam caused by the light intensity distribution factor of the light source.

[0077] After the measurement beam irradiates the sample, it is reflected by the sample to obtain a reflected beam, which is then divided according to a predetermined proportion β to obtain the reference beam intensity I reference (t, λ), the reflected light intensity of the reflected beam I reflecton (t, λ), and the critical light intensity value I critical (λ) is I reflecton (t,λ)≧I reference (t,λ)≧I critical (λ) relationship is satisfied, and the reflected light intensity I reflecton (t, λ) and reference light intensity I reference The dominant influence factor of (t, λ) can be the light source factor, and the reflected light intensity I reflecton (t, λ) and the reference light intensity I reference The influence of the detector factor on (t, λ) is weakened, and at the same time, the variability of the measurement beam and the reference beam after the homogenization process is matched, so that the reflected light intensity I due to the light source factor is reduced. reflecton (t, λ) and reference light intensity I reference The variability of (t, λ) also matches. reference (t, λ) is used to calculate the reflected light intensity I reflecton By accurately correcting and calibrating the light intensity fluctuations at each wavelength in (t, λ), a highly stable reflectance spectrum R means (t, λ) can be obtained, and the detection accuracy of the film thickness d of the sample in the end point detection step can be effectively improved.

[0078] An embodiment of the present invention further provides an endpoint detection apparatus 100, which, as shown in FIG. 5 , includes a light source 110 for emitting a source beam. A beam homogenization and convergence processor 170 is used to process the source beam to obtain a source beam with consistent light intensity fluctuation at each position on the light-emitting surface. A beam splitter 120 is used to split the source beam according to a predetermined proportion β to obtain a measurement beam and a reference beam, and the measurement beam forms a reflected beam after being incident on a sample, and the reference light intensity I of the reference beam is reference (t, λ) and the reflected light intensity I of the reflected beam reflecton (t,λ) is I reflecton (t,λ)≧I reference (t,λ)≧I critical (λ) and I critical (λ) is the critical light intensity value that corresponds to the case governed by the light source factor. The detector module measures the reflected light intensity I of the reflected beam. reflecton (t, λ) and the reference beam intensity I reference (t, λ). The control processing unit 160 is electrically connected to the detector module. The control processing unit further controls the reference light intensity I reference Based on (t, λ), the reflected light intensity I reflecton (t, λ) is calibrated to obtain the target reflected light intensity I ’ reflection (t, λ) is used to generate the target reflected light intensity I ’ reflection Based on (t, λ), the reflectance spectrum R of the sample is means (t, λ) is calculated by collecting the reflectance spectrum R means Based on (t, λ), the film thickness d of the sample at the corresponding time t is calculated, and the end point of the chemical mechanical polishing is determined based on the film thickness d.

[0079] In some embodiments, the light source 110 uses a high-frequency scintillation broadband light source to provide a broadband spectrum beam covering the visible wavelength range. The stability of the output light intensity of the high-frequency scintillation broadband light source does not depend on the magnitude of the light intensity, and the light intensity distribution on the light source light-emitting surface is also non-uniform, with the light intensity at the center of the arc being the strongest, and the light intensity fluctuations at different positions on the light-emitting surface also being non-uniform and randomly distributed. Therefore, a beam homogenization and convergence processing unit 170 performs homogenization and convergence processing on the broadband spectrum beam to homogenize and distribute the light intensity fluctuations at different positions on the light source light-emitting surface at the convergence point, so that the different fluctuations at different positions or regions on the light source light-emitting surface are uniformly distributed at each position or region on the converged light-emitting surface, i.e., the light intensity fluctuations and fluctuation components at each position or region on the converged light-emitting surface are both consistent.

[0080] In some embodiments, the beam splitter 120 splits the entire source beam according to a predetermined proportion β. Specifically, the beam splitter 120 splits the energy of each point on the propagation plane of the source beam according to a predetermined proportion β, which avoids splitting the source beam based on a geometric division on the propagation plane, i.e., avoiding splitting the optical energy of some areas into the measurement beam and the optical energy of other areas into the reference beam. This helps maintain the consistency of the optical intensity fluctuations of the measurement beam and the reference beam.

[0081] In some embodiments, the source beam is transmitted through an optical fiber, and the beam splitter 120 is an optical fiber beam splitter with the same splitting ratio, which can uniformly split the source beam in each optical fiber into a measurement beam and a reference beam according to a predetermined splitting ratio, thereby avoiding using multiple optical fibers in the optical fiber cable to transmit the measurement beam and the reference beam separately.

[0082] In some embodiments, the source beam is transmitted through a spatial optical path, and the beam splitter 120 includes a beam splitting lens that splits the entire source beam according to a predetermined split ratio.

[0083] In some embodiments, the detector module comprises a measurement detector 140 and a reference detector 150, which respectively measure the reflected light intensity I of the reflected beam. reflecton (t, λ) and the reference beam intensity I reference is used to obtain (t,λ).

[0084] In the end-point detection device 100 according to the embodiment of the present invention, a beam homogenization and convergence processing unit 170 is used to obtain a source beam with consistent light intensity fluctuations and fluctuation components at each position or region on the light-emitting surface of the light source 110, and the source beam is then subdivided to obtain a measurement beam and a reference beam, so that the measurement beam and the reference beam have the same light intensity fluctuations, i.e., the problem of inconsistency in the light intensity fluctuations between the measurement beam and the reference beam caused by the light intensity distribution factor of the light source is resolved. By dividing according to a predetermined proportionality β, the reference light intensity I of the reference beam is reference (t, λ), the reflected light intensity of the reflected beam I reflecton (t, λ), and the critical light intensity value I critical (λ) is I reflecton (t,λ)≧I reference (t,λ)≧I critical (λ) relationship is satisfied, and the reflected light intensity I reflecton (t, λ) and reference light intensity I reference The dominant influence factor of (t, λ) can be the light source factor, and the reflected light intensity I reflecton (t, λ) and the reference light intensity I reference The influence of the detector module factor on (t, λ) is weakened, and at the same time, the variability of the measurement beam and the reference beam after the homogenization process is matched, so that the reflected light intensity I due to the light source factor is reduced. reflecton (t, λ) and reference light intensity I reference The variability of (t, λ) also matches. Based on this, the endpoint detection device 100 of this embodiment detects the reflected light intensity I reflecton (t, λ) and reference light intensity I reference (t, λ) can be obtained, and the control processing unit 160 calculates the reference light intensity I reference (t, λ) is used to calculate the reflected light intensity I reflectonThe light intensity fluctuations of each wavelength at (t, λ) can be accurately corrected and calibrated, a highly stable reflectance spectrum can be obtained, the accuracy of spectrum detection can be improved, and the measurement accuracy of film thickness d can be improved, thereby improving the stability and accuracy of endpoint detection during chemical mechanical polishing by the endpoint detection device 100 of this embodiment of the present invention.

[0085] An embodiment of the present invention further provides a polishing apparatus, which includes any of the endpoint detection devices 100 described above, as shown in FIG.

[0086] In some embodiments, the apparatus further includes a grinding table 200, the grinding table 200 having a chamber therein, the endpoint detection device 100 being disposed in the chamber of the grinding table 200, and a grinding pad 210 being provided on the upper surface of the grinding table 200, the grinding pad 210 being used to grind the sample 500. The grinding table 200 and the grinding pad 210 are provided with a through-hole that can vertically pass through the grinding table 200 and the grinding pad 210. A transmission window 220 is used to close the through-hole, and the upper surface of the transmission window 220 is flush with the grinding pad 210.

[0087] The probe 130 of the endpoint detection device 100 faces the transmission window 220 and is used to irradiate the sample 500 with measurement light through the transmission window 220 and receive light reflected from the sample 500 through the transmission window 220. The clamping unit 300 is used to clamp the sample 500 so that the sample 500 can be ground on the grinding pad 210.

[0088] In some embodiments, the grinding table 200 is provided with a first drive device for rotating the grinding table 200, and the rotation axis of the first drive device is perpendicular to the plane on which the side of the grinding pad 210 used for grinding is located. When the grinding table 200 rotates, the endpoint detection device 100 in its chamber also rotates together with the grinding table 200, and when the transmission window 220 rotates to the sample 500, the endpoint detection device 100 measures the thickness of the sample 500 and determines whether the liquid of the sample to be measured has been ground to the endpoint.

[0089] In some embodiments, the apparatus further includes a grinding fluid sprayer 400, the nozzle of which faces the upper surface of the grinding pad 210 used for grinding, and the nozzle is used to spray the grinding fluid onto the grinding pad 210.

[0090] In some embodiments, the clamping unit 300 includes a carrier head 310, which is used to clamp the sample 500. The second drive unit 330 is connected to the carrier head 310 via a rotation shaft and is used to rotate the carrier head 310. The support 320 is used to support the second drive unit 330, the rotation shaft, and the carrier head 310, and is disposed on the side of the grinding pad 210 that is used for grinding.

[0091] In some embodiments, the support 320 is provided with a third drive device, which is used to drive the linear movement of the support 320, and the motion plane of the support 320 is parallel to the plane on which the side of the grinding pad 210 used for grinding is located, i.e., the sample 500 is moved on the grinding pad 210 for grinding.

[0092] In the grinding process of the polishing apparatus according to the embodiment of the present invention, every time the sample 500 passes through the transmission window 220, the endpoint detection device 100 measures the current reflectance spectrum of the sample 500. The endpoint detection device 100 detects the reflected light intensity I reflecton (t, λ) and reference light intensity I reference (t, λ) is used to calculate the reference light intensity I reference Based on (t, λ), the reflected light intensity I reflecton By accurately correcting and calibrating the light intensity fluctuations of each wavelength at (t, λ), the stability and accuracy of the current reflectance spectrum of the sample 500 is improved, the measurement accuracy of the current film thickness d of the sample 500 is improved, and the accuracy of detecting the end point of the sample 500 during the grinding process by the polishing device is improved.

[0093] It should be understood that "plurality" in the present invention means two or more, and similarly applies to other quantifiers. "And / or" indicates a relation between related objects, and indicates that three types of relation may exist, for example, A and / or B can represent A occurring alone, A and B occurring simultaneously, and B occurring alone. The character " / " generally indicates that the related objects before and after are in an "or" relation. The singular forms "a," "said," and "the" are intended to include the plural unless the context clearly indicates otherwise.

[0094] Furthermore, while terms such as "first" and "second" are used to describe various pieces of information, it should be understood that such information should not be limited to these terms. These terms are used only to distinguish between pieces of information of the same type and do not indicate a particular order or importance. In fact, terms such as "first" and "second" can be used completely interchangeably. For example, first information can also be referred to as second information, and similarly, second information can also be referred to as first information, without departing from the scope of the present invention.

[0095] Furthermore, it should be understood that the orientations or positional relationships indicated by terms such as "center," "longitudinal," "lateral," "front," "rear," "upper," "lower," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer" are based on those shown in the accompanying drawings, are intended only to facilitate and simplify the description of the present embodiments, and are not intended to indicate or imply that the referred-to devices or elements must have a particular orientation, be configured, or operate in a particular orientation.

[0096] Furthermore, unless otherwise specified, "connection" should be understood to include not only a direct connection between two elements without other elements involved, but also an indirect connection between two elements involving other elements.

[0097] Additionally, although embodiments of the present invention describe operations in a particular order in the accompanying figures, it should be understood that this should not be construed as requiring the operations to be performed in the particular order or serial order shown, or to perform all of the operations shown, to achieve desirable results. Multitasking or parallel processing may be advantageous in certain environments.

[0098] Other embodiments of the present invention will be readily apparent to those skilled in the art from a consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present invention in accordance with the general principles of the present invention, including means common knowledge or customary in the art that are not invented by the present invention. The specification and examples are considered exemplary only, with the true scope and spirit of the invention being indicated by the following claims.

[0099] It should be understood that the present invention is not limited to the exact construction described above and illustrated in the accompanying drawings, but that various modifications and changes are possible without departing from the scope of the present invention, which is limited only by the appended claims.

Claims

1. obtaining influence factor distribution curves characterizing light intensity measurement variability under different light intensity light source conditions, the influence factor distribution curves including a detector factor distribution curve and a light source factor distribution curve; Based on the influence factor distribution curve, the corresponding critical light intensity value I ｃｒｉｔｉｃａｌ (λ), where λ is any wavelength value within the light source band range; performing a homogenization focusing process on the light source to obtain a source beam with consistent light intensity fluctuations at each position on the light-emitting surface; The source beam is divided according to a predetermined ratio β to obtain the required measurement and reference beams, the measurement beam forming a reflected beam after being incident on a sample, and the reference light intensity I of the reference beam ｒｅｆｅｒｅｎｃｅ (t, λ) and the reflected light intensity I of the reflected beam ｒｅｆｌｅｃｔｏｎ (t, λ) is I ｒｅｆｌｅｃｔｏｎ (t, λ) ≧ I ｒｅｆｅｒｅｎｃｅ (t, λ) ≧ I ｃｒｉｔｉｃａｌ (λ), where t is the light intensity collection time; The reference light intensity I ｒｅｆｅｒｅｎｃｅ (t, λ) based on the reflected light intensity I ｒｅｆｌｅｃｔｏｎ (t, λ) is calibrated to obtain the target reflected light intensity I ’ ｒｅｆｌｅｃｔｏｎ (t, λ); The target reflected light intensity I ’ ｒｅｆｌｅｃｔｏｎ (t, λ) based on the reflectance spectrum R of the sample ｍｅａｓ (t, λ), A reflectance spectrum acquisition method characterized by:

2. obtaining the detector factor distribution curve providing a reference test beam that is a broadband beam; collecting output light intensities of the reference test beam at different wavelengths multiple times with a detector; Obtaining output light intensity noise corresponding to the reference test beam and an output light intensity average value corresponding to the reference test beam at different wavelengths, or obtaining output light intensity signal-to-noise ratio corresponding to the reference test beam and an output light intensity average value corresponding to the reference test beam at different wavelengths; establishing a first relationship curve between an output light intensity noise corresponding to the reference test beam and an output light intensity average value corresponding to the reference test beam, or establishing a second relationship curve between an output light intensity signal-to-noise ratio corresponding to the reference test beam and an output light intensity average value corresponding to the reference test beam; Obtaining the illuminant factor distribution curve involves: collecting, with a detector, the output light intensity of the source beam at different wavelengths within the bandpass range multiple times; Obtaining output light intensity noise corresponding to the light source and an output light intensity average value corresponding to the light source at different wavelengths, or obtaining output light intensity signal-to-noise ratio corresponding to the light source and an output light intensity average value corresponding to the light source at different wavelengths; establishing a third relationship curve between an output light intensity noise corresponding to the light source and an output light intensity average value corresponding to the light source, or establishing a fourth relationship curve between an output light intensity signal-to-noise ratio corresponding to the light source and an output light intensity average value corresponding to the light source. The reflectance spectrum acquisition method according to claim 1 .

3. Based on the influence factor distribution curve, the corresponding critical light intensity value I ｃｒｉｔｉｃａｌ Determining (λ) is A first intersection point between the first relationship curve and the third relationship curve is obtained, a first light intensity value corresponding to the first intersection point is determined, and any light intensity value equal to or greater than the first light intensity value is set as the critical light intensity value I ｃｒｉｔｉｃａｌ (λ), or A second intersection point between the second relationship curve and the fourth relationship curve is obtained, a second light intensity value corresponding to the second intersection point is determined, and any light intensity value equal to or greater than the second light intensity value is set as the critical light intensity value I ｃｒｉｔｉｃａｌ (λ) as The reflectance spectrum acquisition method according to claim 2 .

4. Dividing the source beam according to a predetermined proportion β comprises: splitting the source beam according to the predetermined proportion β for each location on the source beam emitting surface using a fiber optic beam splitter or a beam splitting lens; The reflectance spectrum acquisition method according to claim 1 .

5. The predetermined proportionality β is the measured light intensity I of the measurement beam ｍｅａｓｕｒｅ (t, λ) and the reference light intensity I ｒｅｆｅｒｅｎｃｅ (t, λ), The predetermined proportionality β satisfies the following: [Equation 1] where S is the light source intensity, α is the loss coefficient during transmission of the measurement beam, and I ｃｒｉｔｉｃａｌ (λ) is the critical light intensity value, The loss coefficient α is the measured light intensity I ｍｅａｓｕｒｅ (t, λ) and the reflected light intensity I ｒｅｆｌｅｃｔｏｎ (t, λ) is obtained based on the following formula: [Equation 2] The reflectance spectrum acquisition method according to claim 4 .

6. The reference light intensity I ｒｅｆｅｒｅｎｃｅ (t, λ) based on the reflected light intensity I ｒｅｆｌｅｃｔｏｎ (t, λ) is calibrated to obtain the target reflected light intensity I ’ ｒｅｆｌｅｃｔｏｎ Generating (t, λ) is A certain time t0 is selected within the collection time period, and the reference light intensity I at t0 is ｒｅｆｅｒｅｎｃｅ (t ０ , λ) as a reference initial value; According to the following formula, I ｒｅｆｌｅｃｔｏｎ (t, λ) is corrected to obtain the target reflected light intensity I ’ ｒｅｆｌｅｃｔｏｎ (t, λ), [Equation 3] The reflectance spectrum acquisition method according to claim 1 .

7. The target reflected light intensity I ’ ｒｅｆｌｅｃｔｏｎ (t, λ) based on the reflectance spectrum R of the sample ｍｅａｓ Calculating (t, λ) is The reflectance spectrum R is calculated based on the following formula: ｍｅａｓ (t, λ), [Equation 4] R ｓｉ (λ) is the known reflectance spectrum of a reference bare silicon sample, and I Ｓｉ (λ) is the reflected light intensity of the reference bare silicon sample, and I ｂｌａｃｋ (λ) is the measurement system stray light intensity, The reflectance spectrum acquisition method according to claim 6 .

8. During the chemical mechanical polishing process, the reflectance spectrum R ｍｅａｓ (t, λ) is collected in real time, and the reflectance spectrum R ｍｅａｓ (t, λ) is obtained by the reflectance spectrum acquisition method according to any one of claims 1 to 7, The reflectance spectrum R ｍｅａｓ Calculating the film thickness d of the sample in real time based on (t, λ); determining an end point of the chemical mechanical polishing based on the film thickness d; An endpoint detection method comprising:

9. a light source for emitting a source beam; a beam homogenization and focusing processor for processing the source beam to obtain a source beam with consistent light intensity fluctuations at each position on the light-emitting surface; a beam splitter for splitting the source beam according to a predetermined ratio β to obtain a measurement beam and a reference beam, the measurement beam forming a reflected beam after being incident on a sample, and a reference light intensity I of the reference beam; ｒｅｆｅｒｅｎｃｅ (t, λ) and the reflected light intensity I of the reflected beam ｒｅｆｌｅｃｔｏｎ (t, λ) is I ｒｅｆｌｅｃｔｏｎ (t, λ) ≧ I ｒｅｆｅｒｅｎｃｅ (t, λ) ≧ I ｃｒｉｔｉｃａｌ (λ) relationship is satisfied, and I ｃｒｉｔｉｃａｌ (λ) is the critical light intensity value corresponding to the case governed by the light source factor, the beam splitter; The reflected light intensity I of the reflected beam ｒｅｆｌｅｃｔｏｎ (t, λ) and the reference light intensity I of the reference beam ｒｅｆｅｒｅｎｃｅ a detector module for acquiring (t, λ); a control processing unit electrically connected to the detector module, The control processing unit further The reference light intensity I ｒｅｆｅｒｅｎｃｅ (t, λ) based on the reflected light intensity I ｒｅｆｌｅｃｔｏｎ (t, λ) is calibrated to obtain the target reflected light intensity I ’ ｒｅｆｌｅｃｔｏｎ (t, λ), The target reflected light intensity I ’ ｒｅｆｌｅｃｔｏｎ Based on (t, λ), the reflectance spectrum R of the sample is calculated. ｍｅａｓ Calculate (t, λ), The reflectance spectrum R collected in real time ｍｅａｓ (t, λ) to calculate a film thickness d of the sample at a corresponding time t, which is used to determine an endpoint of chemical mechanical polishing based on the film thickness d. An endpoint detection device characterized by:

10. 10. An end point detection device comprising: A polishing apparatus characterized by:

Images