DEFECT CLASSIFICATION EQUIPMENT FOR SILICON CARBIDE SUBSTRATE USING SINGLE INCIDENT LIGHT-BASED PHOTOLUMINESCENCE AND DEFECT CLASSIFICATION METHOD USING THE SAME

Abstract

Stack fault inspection apparatus and method are disclosed. The apparatus includes a sample stage fixing the silicon carbide substrate and allow the incident light to scan the substrate surface; an incident light source configured to irradiate a vertical illumination light of a wavelength corresponding to an energy greater than a band gap energy of the substrate to at least a portion of a surface of the substrate in a direction substantially perpendicular to the surface of the substrate; a photomultiplier tube (PMT) configured to obtain a photoluminescence mapping image having a wavelength corresponding to the band gap energy of the substrate from the surface of the substrate; and a controller configured to process the mapping image and identify stacking faults.

Claims

1. A method of inspecting a stacking fault of a silicon carbide substrate using single incident light-based photoluminescence, the method comprising: irradiating a vertical illumination light of a wavelength corresponding to an energy greater than a band gap energy of the substrate to at least a portion of a surface of the substrate, in a direction substantially perpendicular to the surface of the substrate; obtaining a photoluminescence mapping image having a wavelength corresponding to the band gap energy of the substrate from the surface of the substrate using a photomultiplier tube (PMT); classifying identifiable defects having different wavelengths from the mapping image having the wavelength corresponding to the band gap energy of the substrate obtained from the PMT into shapes and sizes and securing location data of the identifiable defects; classifying stacking faults from the classified defects and assigning coordinates to each central position of the stacking faults using the location data; sequentially irradiating the vertical illumination light for the stacking faults to which the coordinates are assigned; acquiring a photoluminescence spectrum emitted from each stacking fault to which the vertical illumination light is irradiated with a spectrometer; and comparing a peak wavelength in a region of 400 nm or more of the photoluminescence spectrum obtained from each stacking fault with a central wavelength in stacking fault database to classify characteristics of each stacking fault.

2. The method of claim 1, wherein classifying the stacking faults from the classified defects comprises comparing the defects classified by shape and size with the stacking fault database comprising shapes and sizes of stacking faults of 1SSF, 2SSF, 3SSF, 4SSF and 3C.

3. The method of claim 1, wherein the wavelength corresponding to the band gap energy of the substrate is 390 nm, and the wavelength of the vertical illumination light is 355 nm.

4. The method of claim 1, wherein the center wavelengths of 1S SF, 2SSF, 3SSF, 4SSF and 3C of the stacking fault database are 420 nm, 500 nm, 480 nm, 455 nm and 540 nm, respectively, and wherein in the range of 400 nm or more of the photoluminescence spectrum, the peak wavelength 5 nm above and below the center wavelength of each stacking fault database is classified as a stacking fault.

5. An apparatus for inspecting a stacking fault of a silicon carbide substrate using single incident light-based photoluminescence, the apparatus comprising: a sample stage configured to fix the silicon carbide substrate and allow the incident light to scan the substrate surface; an incident light source configured to irradiate a vertical illumination light of a wavelength corresponding to an energy greater than a band gap energy of the substrate to at least a portion of a surface of the substrate in a direction substantially perpendicular to the surface of the substrate; a photomultiplier tube (PMT) configured to obtain a photoluminescence mapping image having a wavelength corresponding to the band gap energy of the substrate from the surface of the substrate; at least one controller configured to: classify identifiable defects displayed on the photoluminescence mapping image of the substrate obtained from the PMT into shapes and sizes and secure position data, classify stacking faults from the classified defects and assigning coordinates to the respective central positions of the stacking faults using position data, and adjust the stage and the incident light source to sequentially irradiate the vertical illumination light to the coordinate assigned to each of the stacking faults; a spectrometer configured to obtain a photoluminescence spectrum emitted from each stacking fault to which the vertical illumination light is irradiated; and the at least one controller configured to compare a peak wavelength in a region of 400 nm or more of the photoluminescence spectrum obtained from each stacking fault with a central wavelength in the stacking fault database classify characteristics for each stacking fault.

6. The apparatus of claim 5, wherein when the at least one controller classifies the stacking faults from the classified defects, the at least one controller is further configured to compare the identifiable defects classified by shape and size with the stacking fault database comprising shapes and sizes of stacking faults of 1SSF, 2SSF, 3SSF, 4SSF and 3C.

7. The apparatus of claim 5, the wavelength corresponding to the band gap energy of the substrate is 390 nm, and the wavelength of vertical illumination light having a wavelength corresponding to energy greater than the band gap energy of the substrate is 355 nm.

8. The apparatus of claim 5, wherein the center wavelengths of 1SSF, 2SSF, 3SSF, 4SSF and 3C of the stacking fault database are 420 nm, 500 nm, 480 nm, 455 nm and 540 nm respectively, and wherein, in the range of 400 nm or more of the photoluminescence spectrum, the peak wavelength is between 5 nm above and below the center wavelength of each stacking fault database is to be classified as a corresponding stacking fault.

9. A non-transitory computer-readable storage medium storing at least one instruction therein, wherein the at least one instruction, when executed by a processor of an inspection apparatus, causes the processor to perform a method of inspecting a stacking fault of a silicon carbide substrate using single incident light-based photoluminescence, the method comprising: irradiating a vertical illumination light of a wavelength corresponding to an energy greater than a band gap energy of the substrate to at least a portion of a surface of the substrate, in a direction substantially perpendicular to the surface of the substrate; obtaining a photoluminescence mapping image having a wavelength corresponding to the band gap energy of the substrate from the surface of the substrate using a photomultiplier tube (PMT); classifying identifiable defects having different wavelengths from the mapping image having the wavelength corresponding to the band gap energy of the substrate obtained from the PMT into shapes and sizes and securing location data of the identifiable defects; classifying stacking faults from the classified defects and assigning coordinates to each central position of the stacking faults using the location data; sequentially irradiating the vertical illumination light for the stacking faults to which the coordinates are assigned; acquiring a photoluminescence spectrum emitted from each stacking fault to which the vertical illumination light is irradiated with a spectrometer; and comparing a peak wavelength in a region of 400 nm or more of the photoluminescence spectrum obtained from each stacking fault with a central wavelength in stacking fault database to classify characteristics of each stacking fault.

Description

DESCRIPTION OF THE DRAWINGS

[0015] FIG. 1 is a conceptual diagram of a device for classifying stacking faults of a silicon carbide substrate using single incident light-based photoluminescence, according to an embodiment of the present disclosure.

[0016] FIG. 2 is a line profile regarding the shape (left) and photoluminescence intensity at a specific location of stacked bonding through mapping by a photomultiplier tube (PMT) according to an embodiment of the present disclosure.

[0017] FIG. 3 shows a mapping image of a photoluminescence wavelength (390 nm) corresponding to the band gap energy of a SiC substrate and a median range of photoluminescence wavelengths for each surface stacking fault, according to an embodiment of the present disclosure.

[0018] FIG. 4 is a view showing the classification of stacking fault types based on the coordinates of each surface stacking flaw shown in the mapping image of FIG. 3 and the spectroscopically analyzed central wavelength, according to an embodiment of the present disclosure.

[0019] FIGS. 5A-5C are graphs showing the spectral data of the stacking fault of FIG. 4, according to an embodiment of the present disclosure.

[0020] FIG. 6 is a conceptual diagram illustrating a sequence of a method for classifying stacking faults of a silicon carbide substrate using single incident light-based photoluminescence, according to an embodiment of the present disclosure.

DETAILED DESCRIPTION OF EMBODIMENTS

[0021] Prior to the detailed description of the present disclosure, terms or words used in the present specification and claims described below should not be construed as being limited to ordinary or dictionary meaning. Therefore, embodiments described in the present specification and the configurations illustrated in the drawings are merely examples and are not intended to represent all of the technical spirit of the present disclosure, such that it should be understood that various equivalents and deformed examples capable of replacing these at the time of filing the present disclosure can be present.

[0022] When there is a defect in the sample material such as impurities, an energy level can be formed in the band gap. When an excited electron transition happens, it may come down to an energy level above the valence band created by the impurity. On the other hand, the energy level located below the conduction band formed by the impurity may come down to the energy level above the valence band formed by the impurity. Therefore, if photoluminescence analysis is performed, photoluminescence corresponding to energy corresponding to various electron transitions occurring between various energy levels formed by raw materials and impurities can be observed.

[0023] Stacking faults occurring on the substrate surface also generate photoluminescence having different wavelengths for each type. However, in a classification method using photoluminescence, a photomultiplier tube (PMT) is used to classify or inspect the stacking faults by installing a band pass filter for every wavelength at which the stacking faults appear. It is also possible to obtain a luminescence image with a time delay and integration (TDI) camera using a line beam, etc., pass it through a bandpass filter for each wavelength, and then separate the types of stacking faults. Alternatively, defects had to be identified by performing analysis using a spectrometer at all points on the front side of the substrate. However, in the classification of the bandpass filter by wavelength, accurate classification is difficult due to overlapping wavelengths of actual defects, and since simultaneous measurement of several pass paths (ports) may need to be performed for classification by wavelength, the data becomes too large. When spectroscopic analysis is performed at all points on the substrate surface, the minimum exposure time for signal classification of the spectrometer is several milliseconds (msec) or more for each point, so the measurement time is also excessively long.

[0024] When there is a defect in the sample material such as impurities, an energy level can be formed in the band gap. When an excited electron transition happens, it may come down to an energy level above the valence band created by the impurity. On the other hand, the energy level located below the conduction band formed by the impurity may come down to the energy level above the valence band formed by the impurity. Therefore, if photoluminescence analysis is performed, photoluminescence corresponding to energy corresponding to various electron transitions occurring between various energy levels formed by raw materials and impurities can be observed.

[0025] Stacking faults occurring on the substrate surface also generate photoluminescence having different wavelengths for each type. However, in the classification method using photoluminescence, a photomultiplier tube (PMT) is used to classify the stacking faults by installing a band pass filter for every wavelength at which the stacking faults appear. It is also possible to obtain a luminescence image with a time delay and integration (TDI) camera using a line beam, etc., pass it through a bandpass filter for each wavelength, and then separate the types of stacking faults. Alternatively, defects had to be identified by performing analysis using a spectrometer at all points on the front side of the substrate. However, in the classification of the bandpass filter by wavelength, accurate classification is difficult due to overlapping wavelengths of actual defects, and since simultaneous measurement of several pass paths (ports) may need to be performed for classification by wavelength, the data becomes too large. When spectroscopic analysis is performed at all points on the substrate surface, the minimum exposure time for signal classification of the spectrometer is several milliseconds (msec) or more for each point, so the measurement time is also excessively long.

[0026] Hereinafter, embodiments of the present disclosure are further discussed will be described in detail below with reference to the accompanying drawings so that those skilled in the art to which the present disclosure pertains can easily practice them. FIG. 1 is a conceptual diagram of a device for classifying stacking faults of a silicon carbide substrate using single incident light-based photoluminescence, according to an embodiment of the present disclosure.

[0027] In an embodiment of the present disclosure, an apparatus for classifying or inspecting a stacking fault of a silicon carbide substrate using single incident light-based photoluminescence, the apparatus comprising: a sample stage assembly 20 capable of fixing the silicon carbide substrate and allowing the incident light to scan the substrate surface; an incident light source unit 10 irradiating a vertical illumination light of a wavelength corresponding to an energy greater than the band gap energy of the substrate to every portion of the surface of the substrate in a direction substantially perpendicular to the surface of the substrate; a photomultiplier tube (PMT) 30 obtaining a photoluminescence mapping image having a wavelength corresponding to the band gap energy of the substrate from the substrate surface; a control unit comprising a shape classification unit classifying or identifying the identifiable defects displayed on the photoluminescence mapping image of the substrate obtained from the PMT into shapes and sizes and secures position data, a coordinate assignment unit classifying stacking faults from the classified defects and assigning coordinates to the respective central positions of the selected stacking faults using position data, and an adjustment unit adjusting the stage assembly and the incident light source to sequentially irradiate the vertical illumination light to the coordinate assigned stacking faults; a spectrometer 50 for obtaining a photoluminescence spectrum emitted from each stacking fault to which the vertical illumination light is irradiated; and a characteristic classification unit for classifying characteristics for each stacking fault by comparing a peak wavelength in a region of 400 nm or more of the photoluminescence spectrum obtained from each stacking fault with a central wavelength in the stacking fault database. In an embodiment of the present disclosure, the stacking fault database includes various related information such as the size and shape of the stacking faults.

[0028] In an embodiment of the present disclosure, the control unit and the characteristic classification unit may include a computer represented by a central processing unit, an arithmetic unit, and the like. In an embodiment of the present disclosure, the condenser 25 may be positioned in the incident light incident path on the sample stage assembly and in the substrate surface emission photoluminescence passing path by the incident light. In an embodiment of the present disclosure, a total reflection or partial reflection mirror may be located on the path of the incident light 100 from the incident light source and the path of the photoluminescence 200 emitted from the substrate to the photoelectric amplifier tube and the spectrometer.

[0029] In an embodiment of the present disclosure, the incident light 100 is reflected by the mirror 15 toward the sample stage assembly 20, and the emitted photoluminescence 200 passes through the mirror 15. Then, it is reflected from the mirror 35 that reflects the photoluminescence 300 of the wavelength (390 nm) corresponding to the band gap energy and may be directed toward the photoelectric amplifier 30. In an embodiment of the present disclosure, the photoluminescence 200 emitted from the stacking defect may pass through the mirror 15 and then be reflected 500 from the mirror 55 to face the spectrometer 50. In an embodiment of the present disclosure, the classifying of the stacking fault can be executed by a computer having a program of comparing the identifiable defects classified by shape and size with the stacking fault database comprising shapes and sizes of stacking faults of 1SSF, 2SSF, 3SSF, 4SSF and 3C. In the present disclosure, SSF is an abbreviation of Shockley stacking fault as described in Sun et al., “Shockley-Frank stacking faults in 6H-SiC,” J. Appl. Phys. 111, 113527 (2012), the disclosure of which is incorporated by reference herein. 3C stands for 3C-like faulted regions in the 4H-SiC matrix.

[0030] In addition, a wavelength corresponding to the band gap energy of the substrate may be 390 nm, and a wavelength of vertical illumination light having a wavelength corresponding to an energy greater than the band gap energy of the substrate may be selected as 355 nm. In one embodiment of the present disclosure, the equipment may further install an extra detector 40 for detecting the photoluminescence 400 passing through the mirror 35. In an embodiment of the present disclosure, any known method may be used for the machine learning.

[0031] FIG. 2 is a line profile regarding the shape (left) and photoluminescence intensity at a specific location of stacked bonding through mapping by a photomultiplier tube (PMT) according to an embodiment of the present disclosure. Most of the Shockley type stacking faults in 4H-SiC have triangular or quadrangular plate-shaped defects, which are clearly observed in photoluminescence images with a wavelength (390 nm) corresponding to the energy band gap of the substrate. As can be seen in FIG. 2, it can be confirmed that the intensity of photoluminescence is lowered at the location of the surface lamination defect. Through this, it is possible to easily identify the location of the stacking defect through photoluminescence mapping.

[0032] FIG. 3 shows a mapping image of a photoluminescence wavelength (390 nm) corresponding to the band gap energy of a SiC substrate and a median range of photoluminescence wavelengths for each surface stacking fault, according to an embodiment of the present disclosure. FIG. 4 is a view showing the classification of stacking fault types based on the coordinates of each surface stacking flaw shown in the mapping image of FIG. 3 and the spectroscopically analyzed central wavelength, according to an embodiment of the present disclosure. The wavelength ranges of the stack faults, 1SSF, 2SSF, 3SSF, 4SSF and 3C are as follows: [0033] 1SSF: 420±5 nm, [0034] 2SSF: 500±5 nm, [0035] 3SSF: 480±5 nm, [0036] 4SSF: 455±5 nm, and [0037] 3C Type: 540±5 nm.

[0038] In an embodiment of the present disclosure, the center wavelengths of 1SSF, 2SSF, 3SSF, 4SSF and 3C of the stacking fault database are 420 nm, 500 nm, 480 nm, 455 nm and 540 nm respectively, and in the range of 400 nm or more of the photoluminescence spectrum, at the step of comparing the peak wavelength of the stacking fault to the center wavelength of the stacking fault database, the peak wavelengths within 5 nm above and below the center wavelength of each stacking fault database may be classified as corresponding stacking fault.

[0039] FIGS. 5A-5C are graphs showing the spectral data of the stacking fault of FIG. 4, according to an embodiment of the present disclosure. FIG. 5A is spectral data for stacking defects Nos. 1 to 3, and it can be seen that stacking defects 1 to 3 share a wavelength of 390 nm emitted from the substrate and have peak wavelengths of 541 nm, 483 nm and 483 nm, respectively. Therefore, it can be seen that stacking defect 1 is a 3C-type stacking defect, and stacking defects 2 and 3 are 3SSF stacking defects. FIG. 5B is spectral data for stacking faults 4 to 6 and 11, and it can be seen that stacking fault 4 is a 3C type, stacking fault 5 is a 3SSF, and stacking fault 11 is a 4SSF stacking fault. However, the peak wavelength of stacking fault 6 is 463 nm, which is a value not found in the stacking fault database identified so far. Such a value is classifying as unknown and can be added to the stacking fault database when characteristics are classifying according to the results of further research. Classification and additional functions through comparison of such peak wavelengths may be included in the machine learning function. FIG. 5C is spectral data for stacking faults 7 to 10, and it can be understood that stacking fault 7 is a 3SSF, stacking fault 8 is a 4SSF, and stacking faults 9 and 10 are 3C-type stacking faults.

[0040] FIG. 6 is a conceptual diagram illustrating a sequence of a method for classifying stacking faults of a silicon carbide substrate using single incident light-based photoluminescence, according to an embodiment of the present disclosure. In an embodiment of the present disclosure, a method 600 for classifying a stacking fault of a silicon carbide substrate using single incident light-based photoluminescence is comprising the steps of: irradiating 601 a vertical illumination light of a wavelength corresponding to an energy greater than the band gap energy of the substrate to every portion of the surface of the substrate, in a direction substantially perpendicular to the surface of the substrate; obtaining 602 a photoluminescence mapping image having a wavelength corresponding to the band gap energy of the substrate from the surface of the substrate using a photomultiplier tube (PMT); classifying 603 identifiable defect images having different wavelengths from the mapping image having a wavelength corresponding to the band gap energy of the substrate obtained from the PMT into shapes and sizes and securing location data; classifying 604 stacking faults from the classifying defects and assigning coordinates to each central position of the selected stacking faults using the location data; sequentially irradiating 605 the vertical illumination light for the stacking faults to which the coordinates are assigned; acquiring 606 a photoluminescence spectrum emitted from each stacking fault to which the vertical illumination light is irradiated with a spectrometer; and classifying or determining 607 the characteristics of each stacking fault by comparing the peak wavelength in the region of 400 nm or more of the photoluminescence spectrum obtained from each stacking fault with the central wavelength in the stacking fault database.

[0041] In an embodiment of the present disclosure, the classifying of the stacking fault can be executed by a computer having a program of comparing the identifiable defects classified by shape and size with the stacking fault database comprising shapes and sizes of stacking faults of 1SSF, 2SSF, 3SSF, 4SSF and 3C. The wavelength corresponding to the band gap energy of the substrate may be 390 nm, and a wavelength of vertical illumination light having a wavelength corresponding to an energy greater than the band gap energy of the substrate may be selected as 355 nm. In an embodiment of the present disclosure, the center wavelengths of 1SSF, 2SSF, 3SSF, 4SSF and 3C of the stacking fault database are 420 nm, 500 nm, 480 nm, 455 nm and 540 nm respectively, and in the range of 400 nm or more of the photoluminescence spectrum, at the step of comparing the peak wavelength of the stacking fault to the center wavelength of the stacking fault database, the peak wavelengths within 5 nm above and below the center wavelength of each stacking fault database may be classified as corresponding stacking fault.

[0042] Embodiments disclosed herein can be implemented or performed by a computing device having at least one processor, at least one memory and at least one communication interface. The elements of a method, process, or algorithm described in connection with embodiments disclosed herein can be embodied directly in hardware, in a software module executed by at least one processor, or in a combination of the two. Computer-executable instructions for implementing a method, process, or algorithm described in connection with embodiments disclosed herein can be stored in a non-transitory computer readable storage medium.

[0043] Although embodiments of the present application have been described in detail above, the scope of the present application is not limited thereto, and various modifications and improvements by those skilled in the art using the basic concept defined in the following claims.

[0044] All technical terms used in the present disclosure, unless otherwise defined, have the meaning as commonly understood by one of ordinary skill in the art of The present disclosure. The contents of all publications herein incorporated by reference are incorporated herein by reference.