Spectroscopic analysis method, method for fabricating semiconductor device using the same, and substrate process system using the same

Abstract

A spectroscopic analysis method with improved precision and correlation, a method for manufacturing a semiconductor device using the same, and a substrate processing system using the same are provided. The spectroscopic analysis method includes receiving plasma light emitted from a plasma to generate an emission spectrum, detecting n peak wavelengths (wherein is a natural number greater than or equal to 2) from the emission spectrum, generating a plurality of correlation factor time series from correlation factors between the peak wavelengths, filtering the plurality of correlation factor time series, and analyzing the plasma using the filtered correlation factor time series.

Description

Technology Field

[0001] The present invention relates to a spectroscopic analysis method, a method for manufacturing a semiconductor device using the same, and a substrate processing system using the same. More specifically, the present invention relates to a spectroscopic analysis method using optical emission spectrometry (OES), a method for manufacturing a semiconductor device using the same, and a substrate processing system using the same. Background Technology

[0002] Semiconductor devices can be formed by various semiconductor manufacturing processes, such as etching, deposition, ashing, and cleaning. In particular, plasma processes that utilize plasma to accelerate desired chemical reactions (e.g., deposition or etching) are being widely used recently.

[0003] Optical Emission Spectrometry (OES) can be used to optically diagnose abnormalities in the plasma process. OES has the advantage of being able to analyze the plasma generated inside the process chamber from outside the chamber without the need to insert probes. The problem to be solved

[0004] The technical problem that the present invention aims to solve is to provide a spectroscopic analysis method with improved precision and correlation.

[0005] Another technical problem that the present invention aims to solve is to provide a method for manufacturing a semiconductor device using a spectroscopic analysis method with improved precision and correlation.

[0006] Another technical problem that the present invention aims to solve is to provide a substrate processing system using a spectroscopic analysis method with improved precision and correlation.

[0007] The technical problems of the present invention are not limited to those mentioned above, and other unmentioned technical problems will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0008] A method for manufacturing a semiconductor device according to some embodiments for achieving the above technical problems includes loading a substrate into a process chamber, processing the substrate using plasma generated in the process chamber, receiving plasma light emitted from the plasma to generate an emission spectrum, detecting n peak wavelengths (wherein n is a natural number greater than or equal to 2) from the emission spectrum, generating a plurality of correlation factor time series from correlation factors between the peak wavelengths, filtering the plurality of correlation factor time series, and analyzing the substrate using the filtered correlation factor time series.

[0009] A method for manufacturing a semiconductor device according to some embodiments for achieving the above technical problems includes loading a substrate into a process chamber, performing an etching process on the substrate using plasma generated in the process chamber, detecting an etching endpoint of the etching process using plasma light emitted from the plasma, and terminating the etching process after detecting the etching endpoint, wherein detecting the etching endpoint includes receiving plasma light to generate an emission spectrum, detecting n peak wavelengths (wherein is a natural number greater than or equal to 2) from the emission spectrum, generating a plurality of correlation factor time series from correlation factors between the peak wavelengths, and detecting the inflection point of the plurality of correlation factor time series.

[0010] A spectroscopic analysis method according to some embodiments for achieving the above technical problems includes receiving plasma light emitted from a plasma to generate an emission spectrum, detecting n peak wavelengths (wherein is a natural number greater than or equal to 2) from the emission spectrum, generating a plurality of correlation factor time series from correlation factors between the peak wavelengths, filtering the plurality of correlation factor time series, and analyzing the plasma using the filtered correlation factor time series.

[0011] Specific details of other embodiments are included in the detailed description and drawings. Brief explanation of the drawing

[0012] FIG. 1 is an exemplary drawing for illustrating a substrate processing system according to some embodiments. Figure 2 is an exemplary drawing for explaining the spectrometer of Figure 1. FIGS. 3 to 7 are drawings for illustrating a spectroscopic analysis method according to some embodiments. FIGS. 8 and 9 are drawings for illustrating a spectroscopic analysis method according to some embodiments. FIGS. 10 to 12 are drawings for explaining a method of manufacturing a semiconductor device according to some embodiments. Specific details for implementing the invention

[0013] Hereinafter, a substrate processing system according to exemplary embodiments is described with reference to FIGS. 1 and FIGS. 2.

[0014] A substrate processing system according to some embodiments may analyze the plasma state within a process chamber using Optical Emission Spectrometry (OES), which is an optical diagnostic technique for plasma processes. The substrate processed by the substrate processing system according to some embodiments may be, for example, a semiconductor substrate for manufacturing semiconductor devices, a glass substrate for manufacturing flat panel display devices, etc., but is not limited thereto. The processing process performed by the substrate processing system according to some embodiments may be, for example, an etching process, a deposition process, an ashing process, and a cleaning process, but is not limited thereto.

[0015] FIG. 1 is an exemplary drawing for illustrating a substrate processing system according to some embodiments. FIG. 2 is an exemplary drawing for illustrating a spectrometer of FIG. 1.

[0016] Referring to FIGS. 1 and 2, a substrate processing system according to some embodiments includes a process chamber (100), a stage (120), a shower head (130), a first power supply (210), a second power supply (220), a matching unit (300), a spectrometer (500), and an analyzer (600).

[0017] A process chamber (100) may provide an internal space (105) for performing a processing process on a substrate (W; e.g., a wafer). Plasma (P) may be generated within the internal space (105) to perform a plasma process on the substrate (W). The process chamber (100) may be provided with a sealed structure to maintain the internal space (105) at a vacuum (e.g., about 1 Torr or less). The process chamber (100) may have various shapes, such as a hollow hexahedron or a hollow cylinder, but is not limited thereto. The process chamber (100) may include, for example, at least one of a metal and an insulator, but is not limited thereto.

[0018] The process chamber (100) may include a window (110). The number and arrangement of the windows (110) are merely exemplary and are not limited to those illustrated. The window (110) may receive plasma light generated from the plasma (P). The window (110) may include a material capable of transmitting the plasma light, for example, glass or quartz, but is not limited thereto.

[0019] A stage (120) may be positioned within the internal space (105) of a process chamber (100) to support a substrate (W). For example, the stage (120) may be positioned at the bottom of the internal space (105). A substrate (W) may be loaded onto the stage (120). For example, the stage (120) may be an electrostatic chuck that holds the substrate (W) by electrostatic force. In some embodiments, the stage (120) may include a heater (not shown) for heating the substrate (W) to a temperature suitable for a plasma process.

[0020] A shower head (130) may be positioned within the internal space (105) of a process chamber (100) to supply process gas into the process chamber (100). For example, the shower head (130) may be positioned at the top of the internal space (105). The shower head (130) may be positioned opposite the stage (120). In this way, the shower head (130) may supply process gas onto a substrate (W) loaded on the stage (120).

[0021] The first power supply unit (210) and the second power supply unit (220) can each be connected to a shower head (130) or a stage (120). The first power supply unit (210) and the second power supply unit (220) can each apply radio frequency (RF) power to the shower head (130) or the stage (120) for generating or controlling plasma (P). For example, the first power supply unit (210) can provide first RF power to the shower head (130), and the second power supply unit (220) can provide second RF power to the stage (120). In the following description, the shower head (130) connected to the first power supply unit (210) may also be referred to as the upper electrode, and the stage (120) connected to the second power supply unit (220) may also be referred to as the lower electrode.

[0022] The first power supply unit (210) may be an RF source power supply that applies RF source power. That is, the first power supply unit (210) can form plasma (P) within the process chamber (100). For example, when the first RF power is applied from the first power supply unit (210) to an upper electrode (e.g., a shower head (130)), an electric field may be formed by the potential difference between the lower electrode (e.g., a stage (120)) and the upper electrode (e.g., a shower head (130)). Through this, plasma (P) may be generated from the process gas supplied into the process chamber (100) by the shower head (130). In some embodiments, the first power supply unit (210) may be provided in multiple units.

[0023] The second power supply (220) may be an RF bias power supply that applies RF bias power. For example, when the second RF power is applied from the second power supply (220) to a lower electrode (e.g., stage (120)), positive ions generated from the plasma (P) may be incident toward the substrate (W). In some embodiments, the frequency of the second RF power may be lower than the frequency of the first RF power. For example, the frequency of the first RF power may be about 13 MHz to about 14 MHz, and the frequency of the second RF power may be about 12 MHz to about 13 MHz.

[0024] The matching unit (300) can minimize losses of RF power generated from the first power supply unit (210) or the second power supply unit (220). For example, the matching unit (300) can be interposed between the stage (120) and the second power supply unit (220) to increase the transfer efficiency of the second RF power applied to the stage (120). In some embodiments, the matching unit (300) may be arranged to correspond to the number of the first power supply unit (210) and the second power supply unit (220) and connected to each of them, but is omitted for convenience of explanation.

[0025] The spectrometer (500) can receive the plasma light that has passed through the window (110). For example, an optical fiber (410) placed adjacent to the window (110) may be provided. The optical fiber (410) can receive the plasma light that has passed through the window (110) and transmit it to the spectrometer (500).

[0026] In some embodiments, a filter (420) may be provided between the window (110) and the optical fiber (410). The filter (420) may filter at least a portion of the plasma light received by the optical fiber (410). For example, the filter (420) may remove noise components generated inside the process chamber (100) (e.g., noise amplified by reflection between the shower head (130) and the substrate (W)) by filtering at least a portion of the TE (Transverse Electric) mode components from the plasma light received by the optical fiber (410).

[0027] In some embodiments, a collimator (430) may be provided between the window (110) and the optical fiber (410). The collimator (430) may be positioned outside the window (110). The collimator (430) may be a collimator lens, but is not limited thereto. The collimator (430) may collimate the plasma light received by the optical fiber (410).

[0028] The spectrometer (500) can spectrometer the plasma light transmitted from the optical fiber (410). For example, as shown in FIG. 2, the spectrometer (500) may include an inlet slit (510), an imaging mirror (522, 524), a diffraction grating (530), and an array detector (540).

[0029] The plasma light transmitted from the optical fiber (410) can be incident on the spectrometer (500) through the entrance slit (510). The imaging mirrors (522, 524) can collect the plasma light spreading out from the entrance slit (510) and form an image on the array detector (540). The number and arrangement of the imaging mirrors (522, 524) are merely exemplary and are not limited to those depicted. The array detector (540) may be, for example, a Charged Coupled Device (CCD), but is not limited thereto. A diffraction grating (530) is placed in the middle of the optical path from the entrance slit (510) toward the array detector (540) to spectrally analyze the plasma light according to wavelength. The diffraction grating (530) may be, for example, a reflective diffraction grating, but is not limited thereto. Through this, the light spectrally separated by wavelength from the plasma light can form images at different locations on the array detector (540). The spectrometer (500) shown in FIG. 2 is merely exemplary, and it goes without saying that the spectrometer (500) can be provided in various other structures to perform optical emission spectroscopy.

[0030] The analysis system (600) can receive data regarding the plasma light from the spectrometer (500) and analyze the state of the plasma (P). For example, data regarding the plasma light provided through the array detector (540) can be transmitted to the analysis system (600). The spectroscopic analysis method by the analysis system (600) will be described in more detail later in the description regarding FIGS. 3 to 9.

[0031] The analysis system (600) may include software (e.g., an analysis program) for analyzing the state of the plasma (P) from data transmitted from the spectrometer (500). The analysis system (600) may include, for example, at least one of a personal computer (PC), a desktop computer, a laptop computer, a computer workstation, a tablet PC, a server, a mobile computing device, and a combination thereof, but is not limited thereto. The above mobile computing device may be implemented, for example, as a mobile phone, a smartphone, an enterprise digital assistant (EDA), a digital still camera, a digital video camera, a portable multimedia player (PMP), a personal navigation device or portable navigation device (PND), a mobile internet device (MID), a wearable computer, an Internet of Things (IOT) device, an Internet of Everything (IOE) device, or an e-book, but is not limited thereto.

[0032] Hereinafter, a spectroscopic analysis method according to exemplary embodiments will be described with reference to FIGS. 1 to 9.

[0033] FIGS. 3 to 7 are drawings illustrating a spectroscopic analysis method according to some embodiments. For convenience of explanation, parts that overlap with those described above using FIGS. 1 and 2 are briefly described or omitted. The spectroscopic analysis method according to some embodiments may be performed, for example, by the analysis system (600) of FIG. 1, but is not limited thereto.

[0034] Referring to FIGS. 3 and FIGS. 4, a plurality of peak wavelengths of the emission spectrum generated from the plasma light are detected (S10). For reference, FIG. 3 is a flowchart illustrating a spectroscopic analysis method according to some embodiments.

[0035] For example, the emission spectrum of FIG. 4 may be provided from the spectrometer (500) of FIG. 1 and FIG. 2. As shown in FIG. 4, the emission spectrum may be provided as a graph by measuring the intensity according to wavelength from the spectrally separated plasma light. In some embodiments, the emission spectrum may include a full spectrum of a predetermined wavelength range. For example, as shown in FIG. 4, a full spectrum in which intensity is measured over a continuous wavelength range from 200 nm to 800 nm may be provided.

[0036] Subsequently, a plurality of peak wavelengths can be detected from the emission spectrum. For example, as shown in FIG. 4, the analyzer (600) of FIG. 1 detects n peak wavelengths (where n is a natural number greater than or equal to 2) (λ1~λ) from the emission spectrum of FIG. 4. n It can detect ).

[0037] Referring to FIGS. 3 and FIGS. 5, a plurality of correlation factor time series are generated from correlation factors between detected peak wavelengths (S20).

[0038] The above correlation factor may include, for example, at least one of the Pearson correlation coefficient, Spearman correlation coefficient, Cronbach's alpha coefficient, and combinations thereof, but is not limited thereto.

[0039] In some embodiments, generating each of the above-mentioned correlation factor time series involves the detected n peak wavelengths (λ1~λ n It may include deriving a time integration correlation factor between two peak wavelengths among ). For example, when using the Pearson correlation coefficient as the correlation factor, the peak wavelengths (λ1~λ n The time cumulative correlation factor (r) between two peak wavelengths selected from ) (e.g., first peak wavelength (x) and second peak wavelength (y)) xy ) can be calculated as shown in the following mathematical formula 1.

[0040]

[0041] That is, in the above mathematical formula 1, σ x is intensity data (x1~x) of the first peak wavelength (x) measured as process time elapses i Refers to the standard deviation for ), and σ y is intensity data of the second peak wavelength (y) measured as process time elapses (y1~y i It refers to the standard deviation for ). Also, in the above mathematical formula 1, COV(x, y) is the intensity data (x1~x) of the first peak wavelength (x). i ) and intensity data of the second peak wavelength (y) (y1~y i It refers to the covariance of ).

[0042] Next, the time cumulative correlation factor (r) between the first peak wavelength (x) and the second peak wavelength (y) xy ) can be derived chronologically as process time elapses. Through this, a correlation factor time series can be generated from the first peak wavelength (x) and the second peak wavelength (y). In the same way, peak wavelengths (λ1~λ nOther correlation factor time series can be generated from the correlation factors between two other peak wavelengths selected from ). For example, as shown in FIG. 5, the analysis system (600) of FIG. 1 can provide multiple correlation factor time series as a graph by calculating the correlation factors between peak wavelengths over time.

[0043] In some embodiments, n peak wavelengths (λ1~λ n The number of multiple correlation factor time series generated from ) is total n There can be C2 (i.e., n(n-1) / 2). That is, n peak wavelengths (λ1~λ n The above correlation factor time series can be generated for all combinations of two peak wavelengths selected from ).

[0044] Referring to FIGS. 3, FIGS. 6 and FIGS. 7, a plurality of the above-mentioned correlation factor time series are filtered (S30).

[0045] For example, generated n Among the C2 correlation factor time series, only some of the correlation factor time series satisfying a predetermined criterion may be selected. In some embodiments, filtering the plurality of correlation factor time series may include performing clustering on the generated plurality of correlation factor time series to generate a plurality of groups, and selecting at least one group satisfying a predetermined criterion among the generated plurality of groups.

[0046] For example, as illustrated in FIG. 6, the analysis system (600) of FIG. 1 can generate first to twelf groups (1 to 12) each comprising classified correlation factor time series by performing clustering on the multiple correlation factor time series of FIG. 5. The number of groups generated is merely exemplary and is not limited thereto.

[0047] Subsequently, at least one group satisfying a predetermined criterion among the first to twelfth groups (1 to 12) of FIG. 6 may be selected. For example, as illustrated in FIG. 7, the analysis system (600) of FIG. 1 may select a second group (2) from the first to twelfth groups (1 to 12) of FIG. 6. Through this, the analysis system (600) of FIG. 1 may provide filtered correlation factor time series.

[0048] In some embodiments, selecting at least one group satisfying the predetermined criteria may include selecting at least one group satisfying a predetermined range of standard deviations among a plurality of generated groups. For example, among the first to twelfth groups (1 to 12), a group having a standard deviation greater than or equal to a predetermined standard deviation may be selected, or among the first to twelfth groups (1 to 12), a group having a standard deviation less than or equal to a predetermined standard deviation may be selected, and filtered correlation factor time series may be provided.

[0049] Referring to FIGS. 3 and FIGS. 7, a substrate (e.g., substrate (W) of FIG. 1) is analyzed using filtered correlation factor time series (S40).

[0050] Filtered correlation factor time series can reflect the state of the plasma (P) occurring within the process chamber (100) of FIG. 1. For example, as illustrated in FIG. 7, the analyzer (600) of FIG. 1 can detect the inflection point (EP) of the filtered correlation factor time series. Through the inflection point (EP), changes in the substrate (W) as the plasma process proceeds can be inferred. For example, through the inflection point (EP), the etching endpoint of the etching process using the plasma (P) can be detected. Detecting the inflection point (EP) can be performed by various known methods, such as a feature point detection algorithm.

[0051] FIGS. 8 and 9 are drawings illustrating a spectroscopic analysis method according to some embodiments. For convenience of explanation, parts that overlap with those described above using FIGS. 1 to 7 are briefly explained or omitted.

[0052] Referring to FIGS. 8 and 9, a spectroscopic analysis method according to some embodiments includes matching filtered correlation factor time series to chemical species (S35) and analyzing a substrate (e.g., substrate (W) of FIG. 1) using the matched correlation factor time series (S45). For reference, FIG. 9 is an exemplary graph illustrating the result of matching the filtered correlation factor time series of FIG. 7 to corresponding chemical species.

[0053] Specifically, after filtering the plurality of the above-mentioned correlation factor time series (S30), each correlation factor time series has peak wavelengths (λ1~λ n It can be matched to chemical species corresponding to ) (S35). The chemical species may refer to active species or byproducts generated according to the plasma process. For example, detected peak wavelengths (λ1~λ n Among them, the peak wavelength having a wavelength of approximately 206.0 nm (e.g., λ1) can correspond to Si, N2, CO, SiF, etc. As another example, the detected peak wavelengths (λ1~λ n Among them, the peak wavelength having a wavelength of approximately 218.5 nm (e.g., λ2) can correspond to CN, C2H2, CO, CO+, etc. As described above, the detected peak wavelengths (λ1~λ n Two selected peak wavelengths can generate a single correlation factor time series. For example, as shown in FIG. 9, the analyzer (600) of FIG. 1 can match each filtered correlation factor time series to a pair of chemical species (e.g., NO-O, CO-O, or NO-F, etc.) corresponding to the two selected peak wavelengths.

[0054] Next, the substrate (W) can be analyzed using the matched correlation factor time series. The matched correlation factor time series can reflect the state of the plasma (P) occurring within the process chamber (100) of FIG. 1. For example, as shown in FIG. 9, the analysis system (600) of FIG. 1 can detect the inflection point (EP) of the matched correlation factor time series. Since detecting the inflection point (EP) is similar to that described above using FIG. 7, a detailed explanation is omitted below.

[0055] As semiconductor devices become increasingly highly integrated, limitations are emerging in existing spectroscopic analysis methods for plasma processes. For example, in the case of spectroscopic analysis methods that detect inflection points in an etching process using specific wavelengths of emission spectra generated from plasma light, there is a problem in that it is difficult to detect inflection points due to the low signal-to-noise ratio (SNR) in semiconductor devices with fine patterns or composite films. However, as described above, the spectroscopic analysis method according to some embodiments utilizes multiple peak wavelengths detected within a predetermined wavelength range and can analyze the plasma process by calculating correlation factors between the peak wavelengths, thereby enabling a high signal-to-noise ratio. Through this, precise detection of inflection points is possible, and a spectroscopic analysis method with improved precision can be provided.

[0056] Furthermore, in the case of spectroscopic analysis methods that detect inflection points using full-spectrum-based data statistical analysis regarding emission spectra, there is a problem in that it is difficult to interpret the correlation with plasma process results because the derived components lack specific physical meaning. However, as described above, spectroscopic analysis methods according to some embodiments utilize correlation factor time series that can be matched to corresponding chemical species, and thus can have physical meaning regarding active species or byproducts generated according to the plasma process. Through this, a spectroscopic analysis method with improved correlation with plasma process results can be provided.

[0057] Hereinafter, a method for manufacturing a semiconductor device according to exemplary embodiments is described with reference to FIGS. 1 to 12.

[0058] FIGS. 10 to 12 are drawings for explaining a method for manufacturing a semiconductor device according to some embodiments. For convenience of explanation, parts that overlap with those described above using FIGS. 1 and 2 are briefly explained or omitted. For reference, FIG. 10 is a flowchart for explaining a method for manufacturing a semiconductor device according to some embodiments. The method for manufacturing a semiconductor device described with respect to FIGS. 10 to 12 is described with a focus on the etching process, but this is merely illustrative. As another example, the method for manufacturing a semiconductor device according to some embodiments may, of course, be applied to various other plasma processes such as deposition, ashing, and cleaning processes.

[0059] Referring to FIGS. 1, FIGS. 10 and FIGS. 11, a substrate (W) is loaded into a process chamber (100) (S100).

[0060] The substrate (W) may, for example, include a semiconductor substrate (10), a pattern (20), an anti-reflection film (30), and an etching target film (40). The pattern (20), the anti-reflection film (30), and the etching target film (40) may be formed on the semiconductor substrate (10).

[0061] The semiconductor substrate (10) may be bulk silicon or SOI (silicon-on-insulator). Alternatively, the semiconductor substrate (10) may be a silicon substrate or may include other materials, such as silicon germanium, SGOI (silicon germanium on insulator), indium antimonide, lead telluride compound, indium arsenide, indium phosphide, gallium arsenide, or gallium antimonide, but is not limited thereto. For convenience of explanation, the semiconductor substrate (10) is described below as being a silicon substrate.

[0062] The pattern (20) may be a fine pattern formed on a semiconductor substrate (10) to implement a semiconductor device. The anti-reflection film (30) may be formed on the pattern (20). The etching target film (40) may be formed on the pattern (20) and the anti-reflection film (30). The semiconductor substrate (10), the pattern (20), and the anti-reflection film (30), excluding the etching target film (40), may correspond to non-etching target films.

[0063] The etching target film (40) may include at least one of a semiconductor material, a conductive material, an insulating material, and a combination thereof. For example, the etching target film (40) may include a semiconductor material such as polysilicon and / or an epitaxial layer. As another example, the etching target film (40) may include a conductive material such as doped polysilicon, a metal, a metal silicide and / or a metal nitride. As yet another example, the etching target film (40) may include an insulating material such as silicon oxide, silicon nitride, silicon oxynitride, a low-dielectric (low-k) material having a dielectric constant lower than silicon oxide and / or a high-dielectric (high-k) material having a dielectric constant higher than silicon oxide.

[0064] Although the etching target film (40) is illustrated as being formed as a single film, this is merely illustrative, and it goes without saying that the etching target film (40) may be formed as a multi-film in which a plurality of films are stacked. For example, the etching target film (40) may include a plurality of stacked insulating films, or may include a conductive film or a semiconductor film interposed between the stacked insulating films.

[0065] Referring to FIGS. 1, FIGS. 10 and FIGS. 12, plasma (P) is generated in a process chamber (100) to perform an etching process (S200), the process chamber (100) is monitored to detect the etching endpoint of the etching process (S300), and the etching process is terminated (S400).

[0066] Performing an etching process (S200) can be performed, for example, by a first power supply unit (210) and a second power supply unit (220). For example, plasma (P) can be formed within the process chamber (100) by the first power supply unit (210), and positive ions generated from the plasma (P) can be incident toward the substrate (W) by the second power supply unit (220). Through this, an etching process can be performed on the etching target film (40).

[0067] Monitoring the process chamber (100) and detecting the etching endpoint of the etching process (S300) can be performed, for example, by a spectrometer (500) and an analyzer (600). For example, the spectrometer (500) can spectrometer plasma light generated from the plasma (P) in the process chamber (100) and transmit it to the analyzer (600). The analyzer (600) can monitor the process chamber (100) and detect the etching endpoint using the spectroscopic analysis method described above with reference to FIGS. 3 to 9. For example, as described above in the description regarding FIG. 7, the analyzer (600) can detect the etching endpoint by detecting an inflection point (EP).

[0068] When an etching endpoint is detected, the etching process can be terminated (S400). For example, as shown in FIG. 12, as the etching process is performed on the etching target film (40), the upper surface of the semiconductor substrate (10) may be exposed and the state of the plasma (P) in the process chamber (100) may change. Through this, the analyzer (600) can detect the point in time when the upper surface of the semiconductor substrate (10) is exposed as an etching endpoint by using the spectroscopic analysis method described above with FIGS. 3 to 9. For example, the analyzer (600) can detect the inflection point (EP) of FIG. 7 as an etching endpoint.

[0069] Next, referring to FIG. 10, a subsequent process is performed (S500).

[0070] The post-processing may include various semiconductor processes for the substrate (W). For example, the semiconductor processes may include, but are not limited to, deposition processes, etching processes, ion processes, and cleaning processes. The semiconductor processes for the substrate (W) may also include a test process for a wafer-level semiconductor device. As the post-processing for the substrate (W) is performed, various integrated circuits and wirings required for the semiconductor device may be formed.

[0071] When semiconductor chips are completed within a substrate (W; e.g., a wafer) through the semiconductor process on the substrate (W), the substrate (W) can be individualized into individual semiconductor chips. Individualization into individual semiconductor chips can be achieved through a sawing process using a blade or a laser. Subsequently, a packaging process for each semiconductor chip can be performed. The packaging process may refer to a process of mounting each semiconductor chip onto a circuit board (e.g., a printed circuit board (PCB)) and sealing it with a sealant. Additionally, the packaging process may include forming a stack package by stacking multiple semiconductor chips in multiple layers on a circuit board, or forming a Package On Package (POP) structure by stacking stack packages on top of each stack package. A semiconductor package can be formed through the packaging process for each semiconductor chip. The semiconductor process on the substrate (W) may also include a test process for semiconductor devices at the package level.

[0072] Although embodiments of the present invention have been described above with reference to the attached drawings, the present invention is not limited to the above embodiments and can be manufactured in various different forms, and those skilled in the art will understand that the present invention can be implemented in other specific forms without changing the technical concept or essential features of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. Explanation of the symbols

[0073] 10: Semiconductor substrate 20: Pattern 30: Anti-reflective film 40: Etching target film 100: Process chamber 105: Internal space 110: Window 120: Stage 130: Shower head 210: First power supply unit 220: Second power supply unit 300: Matching unit 410: Optical fiber 420: Filter 430: Collimator 500: Spectrometer 510: Entrance slit 522, 524: Imaging mirror 530: Diffraction grating 540: Array detector 600: Analysis system P: Plasma W: Substrate

Claims

Claim 1 A method for manufacturing a semiconductor device, comprising loading a substrate into a process chamber, processing the substrate using plasma generated within the process chamber, receiving plasma light emitted from the plasma to generate an emission spectrum, detecting n peak wavelengths (wherein n is a natural number greater than or equal to 2) from the emission spectrum, generating a plurality of correlation factor time series from correlation factors between the peak wavelengths, filtering the plurality of correlation factor time series, and analyzing the substrate using the filtered correlation factor time series, wherein generating each of the correlation factor time series includes deriving a time integration correlation factor between two peak wavelengths among the n peak wavelengths in a time series. Claim 2 A method for manufacturing a semiconductor device according to claim 1, wherein the emission spectrum comprises a full spectrum of a predetermined wavelength range. Claim 3 delete Claim 4 In claim 1, the number of the generated plurality of correlation factor time series is a total n Method for manufacturing a semiconductor device with C2. Claim 5 A method for manufacturing a semiconductor device according to claim 1, wherein filtering a plurality of the correlation factor time series comprises performing clustering on the plurality of the correlation factor time series to generate a plurality of groups, and selecting at least one group satisfying a predetermined criterion among the generated plurality of groups. Claim 6 A method for manufacturing a semiconductor device according to claim 5, wherein selecting at least one group satisfying the above-mentioned predetermined criteria includes selecting at least one group satisfying a predetermined range of standard deviations among the generated plurality of groups. Claim 7 A method for manufacturing a semiconductor device according to claim 1, wherein analyzing the substrate includes detecting the inflection point of the filtered correlation factor time series. Claim 8 A method for manufacturing a semiconductor device according to claim 1, further comprising, after filtering a plurality of the above-mentioned correlation factor time series, matching each of the above-mentioned correlation factor time series to chemical species corresponding to the peak wavelengths. Claim 9 A method for manufacturing a semiconductor device, comprising loading a substrate into a process chamber, performing an etching process on the substrate using plasma generated within the process chamber, detecting an etching endpoint of the etching process using plasma light emitted from the plasma, and terminating the etching process after detecting the etching endpoint, wherein detecting the etching endpoint includes receiving the plasma light to generate an emission spectrum, detecting n peak wavelengths (where n is a natural number greater than or equal to 2) from the emission spectrum, generating a plurality of correlation factor time series from correlation factors between the peak wavelengths, and detecting the inflection point of the plurality of correlation factor time series. Claim 10 A spectroscopic analysis method comprising receiving plasma light emitted from a plasma to generate an emission spectrum, detecting n peak wavelengths (wherein n is a natural number greater than or equal to 2) from the emission spectrum, generating a plurality of correlation factor time series from correlation factors between the peak wavelengths, filtering the plurality of correlation factor time series, and analyzing the plasma using the filtered correlation factor time series, wherein generating each of the correlation factor time series includes deriving a time integration correlation factor between two peak wavelengths among the n peak wavelengths in a time series.

Images