A method for rapid non-destructive testing of silicon carbide crystal quality

Abstract

The application discloses a method for rapidly and nondestructively detecting the quality of silicon carbide crystals, which comprises the following steps: placing the silicon carbide crystals on a computer tomography (CT) carrier table, and rapidly and nondestructively detecting the typical defects in the silicon carbide crystals by means of computer tomography under specific parameters, image contrast and the identification of the topographic features, so as to directly obtain the defect topography, density distribution and the evolution of the defects with the growth process of the measured ingot at different sections, and significantly improve the visualization degree of the defects in the silicon carbide crystals.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of semiconductor material test characterization, and particularly relates to a rapid nondestructive testing method for silicon carbide crystal quality. BACKGROUND

[0002] Silicon carbide (SiC) is known as the third generation semiconductor material, which has high breakdown field strength, high thermal conductivity, high electron saturation velocity and super strong anti-radiation ability, and is very suitable for making high-temperature, high-frequency and high-power electronic devices, as well as blue-green, ultraviolet light emitting devices and photoelectric detection devices. At present, in the microelectronic field, there are various types of SiC-based microelectronic devices, such as high electron mobility transistors, thyristors, metal-oxide-semiconductor field effect transistors, heterojunction bipolar transistors, etc., and have been widely used in aerospace, military equipment, communication radar and automobile electrification fields.

[0003] However, the quality of silicon carbide material and the cost of devices are still not comparable with silicon-based devices, which limits large-scale application. Due to the characteristics of the commonly used PVT method for preparing silicon carbide crystals, it is very difficult to effectively and timely control the defects in the actual crystal growth process, resulting in low crystalline quality of the obtained crystals and low utilization rate of the crystals.

[0004] Generally, the defects in SiC crystals mainly include microtubes, hexagonal voids, polytype inclusion defects, stacking fault defects, and carbon inclusions, etc. The existence of these structural defects causes the performance of the devices to deteriorate sharply. Therefore, reducing or eliminating the defects in the crystal is crucial for promoting the growth of high-quality silicon carbide material and realizing the preparation of high-performance devices, and how to effectively characterize the defect density and distribution of silicon carbide crystals quickly and nondestructively has become a research focus.

[0005] The traditional method for detecting the quality of silicon carbide crystals is wet etching after slicing, and then detecting by optical microscopy. However, this method is destructive to the sample, and can only realize small area observation within the microscopic field of view. With the increase of crystal size and cost, it is more difficult to characterize the quality of complete silicon carbide crystals. Therefore, a method for quickly and nondestructively detecting the quality of silicon carbide crystals is needed.

[0006] In addition, in the field of semiconductor substrate or epitaxial wafer detection, although the XRT (X-ray topography) detection technology based on similar imaging principles is very mature and can obtain fine features such as defect morphology of wafer surface [Defect Inspection Techniques in SiC, Chen et al. Nanoscale Research Letters (2022) 17:30], the related detection method is usually mainly based on reflection fine scanning, which cannot perform internal tomographic scanning on large-size, whole-piece SiC crystal ingots, thereby limiting the application of the technology in the nondestructive detection of silicon carbide crystal ingots. SUMMARY

[0007] The present application provides a rapid nondestructive detection method for the quality of silicon carbide crystal, which performs full wafer scanning on the treated silicon carbide crystal through a specific X-ray source and scanning mode, and can quickly obtain the defect density and intuitive image of the distribution of the whole wafer or local defects of the silicon carbide crystal on the computer tomographic imaging according to the local difference in X-ray absorption or scattering of different defects of the crystal material.

[0008] The technical scheme of the present application is as follows:

[0009] A rapid nondestructive detection method for the quality of silicon carbide crystal, comprising:

[0010] S1. Surface cleaning treatment is performed on the silicon carbide crystal.

[0011] S2. The silicon carbide crystal is placed on the object table of the computer tomography, the X-rays emitted by the ray generator pass through the collimator and the silicon carbide crystal in turn and then enter the detector, the detector receives the transmitted X-rays of the scanning layer to generate a signal, and the computer reconstructs the image to obtain the tomographic image of the whole wafer or local defects of the silicon carbide crystal.

[0012] S3. According to the difference in X-ray absorption / scattering of different defect sites in the silicon carbide crystal compared with the ideal crystal, the contrast and topographic features in the tomographic image are different, thereby realizing the identification of the defect type.

[0013] Optionally, the surface cleaning treatment is to clean the surface of the silicon carbide crystal with toluene, acetone, ethanol or isopropanol, and then dry it with nitrogen.

[0014] Optionally, the silicon carbide crystal is a silicon carbide crystal ingot, a silicon carbide substrate or an epitaxial wafer. The silicon carbide crystal is prepared by physical vapor phase transmission method or liquid phase method; the silicon carbide substrate is a single crystal substrate after cutting, grinding and polishing; and the epitaxial wafer is prepared by chemical vapor deposition method.

[0015] Optionally, the ray generator is a micro-focus high-energy X-ray tube, the X-ray source has an accelerating voltage range of 200-350 kV, a light source focal point size of ≤10 microns, an exposure time of ≥1000 ms, and a spatial resolution of ≤80 microns.

[0016] Optionally, when the silicon carbide crystal is placed on the object table, the Si surface of the silicon carbide crystal faces upward, and the growth direction is perpendicular to the object table plane.

[0017] Optionally, the X-ray is scanned in a segmented scanning manner.

[0018] Optionally, the defect types include microtubule defects, carbon inclusions, polytype or crystal type inclusion defects, hexagonal void defects, and stacking fault defects.

[0019] Optionally, the identification of the defect types includes:

[0020] The microtubule defects appear as bright white spots in the image and have an extension along the crystal axis direction; the carbon inclusions appear as black spots distributed in specific positions inside the crystal and do not change with the fault growth direction; the polycrystalline region appears as a regional, shallow white block that evolves with the crystal growth direction; the single crystal, defect-free position of the normal crystal region appears as a uniform gray color.

[0021] Optionally, it further includes extracting image features of the defect types for defect distribution, color presentation, and distribution density calculation.

[0022] Optionally, the silicon carbide crystal is a 6-inch 4H-SiC single crystal, the ray generator is a micro-focus 240 kV high-energy X-ray tube, the light source focal point is 5 microns, the exposure time is 1000 ms, and the number of projections is 2400.

[0023] The semiconductor crystal belongs to a homogenous material with very high crystallinity, and has very small changes in density, so conventional CT imaging and material analysis methods cannot be applied to the crystal material. The crystal material has very high penetration consistency to X-rays, and thus the scattering or absorption enhancement of the micro-crystal defects in the crystal material to X-rays is easy to produce differentiated features in imaging. According to the generation mechanism of specific defects and the difference characteristics of the effects on X-rays, the identification and distribution information extraction of typical defect types in the silicon carbide single crystal can be realized.

[0024] The signal generated by the detector receiving the transmitted X-rays of the scanning layer is analog / digital converted and input into a computer for reconstruction imaging to obtain a full or partial tomographic scan of the silicon carbide crystal. According to the contrast and distribution characteristics of different defects in the imaging graph, the identification of the defect types is performed.

[0025] When the X-rays formed by the collimator hit the silicon carbide ingot, the lattice arrangement is not neat at the position with growth defects, which leads to enhanced X-ray scattering and weakened transmission, thereby forming a white phase with strong contrast on the computer imaging; at the defect-free position, i.e. the single crystal with good crystallinity, the lattice arrangement is neat, the sample has relatively less scattering / absorption of X-rays, and the transmission of X-rays is relatively more, thereby forming a uniform gray phase; and in the carbon inclusions, the transmission of the sample is enhanced due to the small density of carbon, the received X-ray signal on the detector is enhanced, and a black spot with strong contrast is shown on the computer imaging; in addition, the polycrystalline or polymorphic region of the crystal has regional scattering enhancement on the fault plane due to the destruction of the crystal periodicity, which leads to increased scattering of X-rays and weakened transmission, and thus a light region with obvious characteristics is shown on the computer imaging.

[0026] In addition to the typical microtubule defects, polymorphic / inclusion defects, and carbon inclusion defects, hexagonal void defects, macroscopic stacking fault defects, and crystal cracks can also be identified by specific feature patterns.

[0027] The present application has the following beneficial effects:

[0028] Compared with the traditional wet etching method, the present application can realize rapid and non-destructive detection of large-size silicon carbide crystals by computer tomography combined with image contrast and defect identification based on topographic features. It has been verified that the complete scanning process of a 6-inch silicon carbide crystal only takes 40 minutes, which is simple to operate and easy to popularize.

[0029] By obtaining the defect distribution map of the whole or part of the silicon carbide crystal, the present application can realize rapid and non-destructive detection of the silicon carbide ingot, directly obtain the intuitive image of the defects existing in different fault planes of the measured ingot, and the evolution of the defects with the spatial position, which can significantly improve the visualization of the silicon carbide crystal defects, and analyze the quality of the crystal more simply and quickly without special sample preparation, thereby improving the production efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 The schematic diagram of the non-destructive detection process of the silicon carbide crystal of the embodiment is shown in the figure, wherein 1 is an X-ray tube light source, 2 is a collimator, 3 is a fan-shaped light source after collimation, 4 is a silicon carbide crystal, 5 is a sample stage, 6 is an X-ray detector, and 7 is a tomographic image.

[0031] Figure 2 The principle diagram of identifying the defects of the silicon carbide crystal based on the computer tomographic image features is shown in the figure.

[0032] Figure 3Fig. 6 is a tomographic image of four typical positions of the computer tomography scanned 6-inch 4H-SiC crystal of the embodiment, respectively containing a microtubular defect with a characteristic feature, a single crystal with good crystallinity, a carbon inclusion defect, and a polycrystal / polymorph defect;

[0033] Figure 4 Fig. 7 is a tomographic image of the embodiment and the corresponding position Figure 3 Fig. 8 is an optical microscope result image of the position 1 in the middle, which proves that the defect of the position belongs to a typical microtubular defect;

[0034] Figure 5 Fig. 9 is a three-dimensional image of the distribution of the typical microtubular defect of the crystal extracted based on image processing after computer tomography. DETAILED DESCRIPTION

[0035] The present application is further explained with reference to the accompanying drawings and specific embodiments. The drawings of the present application are only schematic and the specific proportions can be varied as design requirements dictate.

[0036] A rapid nondestructive testing method of silicon carbide crystal, which is to segmentally scan and splice and reconstruct the silicon carbide wafer after surface treatment, so as to quickly obtain the typical defect density and distribution in the complete silicon carbide crystal. Figure 1 The testing method mainly includes the following steps:

[0037] S101: Surface treatment is performed on the silicon carbide crystal 4, the surface is cleaned with toluene, acetone, ethanol or isopropanol, and the surface of the crystal is then dried with nitrogen;

[0038] S102: The silicon carbide crystal 4 is placed on the object table 5 of the computer tomography (CT);

[0039] S103: The rays emitted by the X-ray tube light source 1 form a fan-shaped light source 3 after passing through the collimator 2, then pass through the silicon carbide crystal 4 and are received by the detector 6;

[0040] S104: The diffraction / scattering X-rays of the scanned layer received by the detector 6 generate a signal, which is then analog / digital converted and input into a computer for reconstruction imaging to obtain a tomographic image of the whole or part of the silicon carbide crystal;

[0041] S105: According to the difference in X-ray absorption / scattering of different defect parts in the crystal compared with the ideal crystal, the contrast and topographic features in the imaging image are different, and the defect type is identified.

[0042] In order to realize clear scanning of the crystal micro-defects, the X-ray tube light source 1 uses a micro-focus high-energy X-ray tube, the acceleration voltage range is 200-350kV, the light source focal point size is less than or equal to 10 microns, the exposure time is greater than or equal to 1000ms, and the spatial resolution is less than or equal to 80 microns. The detector can be a detector array.

[0043] When the X-rays formed by the collimator hit the silicon carbide crystal ingot, at the position with growth defects, due to the irregular arrangement of the crystal lattice, the X-ray scattering is enhanced, and the transmission ability is weakened, thereby forming a white phase with relatively strong contrast on the computer imaging; at the defect-free position, that is, the single crystal with good crystallinity, the crystal lattice is arranged in an orderly manner, the sample has relatively less X-ray scattering / absorption, and relatively more X-ray transmission is performed, thereby forming a uniform gray phase; and in the carbon inclusions defect, due to the small density of carbon, the transmission of the sample is enhanced, the received X-ray signal on the detector is enhanced, and a black spot with relatively strong contrast is shown on the computer imaging; in addition, the polycrystalline or polymorphic region of the crystal has regional scattering enhancement and poor transmission on the fault plane due to the destruction of the periodicity of the crystal, and thus a light region with obvious characteristics is shown on the computer imaging.

[0044] The X-ray scanning imaging mode of the crystal can also be a reflection mode imaging, and the absorption characteristics of the described defects on the X-rays are different from those of the ideal crystal, which are consistent with the above, that is, the identification principle is the same.

[0045] The contrast change of the defect type on the computer imaging is for the conventional computer tomography scanning imaging, and according to different graphic processing, the intensity change characteristics of the above-mentioned defects on the X-ray absorption or reflection are presented in the form of negative color or color image, but are not limited to this.

[0046] The computer tomography scanning can be further processed through subsequent computer image processing to extract the image characteristics of the above-mentioned defect types, such as defect distribution, color presentation, and distribution density calculation, so as to present the characteristic information of different defects in the crystal in a more intuitive form. Through extraction of the imaging spatial distribution and topographic characteristics corresponding to specific defects, a defect distribution map of the whole or part of the silicon carbide crystal can be obtained. The characteristics formed by the above-mentioned different defect types can be verified by a plurality of optical detection means, such as a polarization microscope, a laser confocal microscope, an ultraviolet fluorescence image, and a Raman spectrum plane scanning.

[0047] Compared with the traditional wet etching method, the method of computer tomography scanning combined with image contrast and defect identification of topographic characteristics can realize rapid and non-destructive detection of large-size silicon carbide crystals. It has been verified that the complete scanning process of a 6-inch silicon carbide crystal only takes 40 minutes, and the operation is simple and easy to popularize. The embodiment of the present application can quickly and non-destructively evaluate the silicon carbide crystal by obtaining the defect distribution, thereby greatly improving the production efficiency.

[0048] By acquiring the defect distribution map of the whole or partial silicon carbide crystal, the silicon carbide crystal ingot can be quickly and nondestructively detected, and the intuitive image of the defects existing in different sections of the measured ingot and the evolution of the defects with the spatial position can be directly acquired, so that the visualization degree of the silicon carbide crystal defects can be significantly improved, and the quality of the crystal can be analyzed more simply and quickly without special sample preparation.

[0049] Based on the above principle, the method provided by the present application will be described in more detail in combination with specific embodiments.

[0050] Embodiment 1

[0051] The 6-inch 4H-SiC single crystal is grown by the physical vapor transport method, and the temperature is controlled at 2100-2300 ℃ during the growth process, and the growth pressure is 5-30 mbar. The conductive N-type 4H-SiC single crystal is realized by introducing nitrogen into the growth atmosphere. A 6-inch silicon carbide crystal ingot is obtained after growth.

[0052] The surface of the silicon carbide crystal is cleaned with ethanol and then dried with nitrogen. The treated silicon carbide crystal is placed on the sample stage of a computer tomography (industrial CT) device, the model of which is GE phoenix V. The X-ray source is a micro-focus 240 kV high-energy X-ray tube, the light source focal point is 5 microns, the exposure time is 1000 ms, and the number of projections is 2400. The X-rays pass through the collimator, the crystal and the detector in turn. After scanning and computer image reconstruction, the tomographic image of any region or direction of the crystal can be obtained. According to the above defect identification principle, the defect distribution of the whole or partial silicon carbide crystal can be acquired. After further processing of the computer image, detailed information such as the distribution and density of specific defects and the evolution process can be obtained.

[0053] Figure 2 The principle diagram for identifying the defects of the silicon carbide crystal by the computer tomography image features is shown in the figure. ① is the imaging mechanism of the microtubule defect. For the microtubule defect, the lattice distortion (periodic destruction) caused by the microtubule defect outside the microtubule leads to the enhancement of X-ray scattering, which shows the weakening of the ability of the X-ray to penetrate the sample; ② is the imaging mechanism of the ideal single crystal. For the single crystal with good crystallinity, the crystal penetration and scattering characteristics are consistent when the X-ray scans in this region; ③ is the imaging mechanism of the carbon inclusions defect. Because the density of carbon is lower than that of silicon carbide, the X-ray scanning in the defect shows strong penetration; ④ is the imaging mechanism of the polycrystal / polymorphic region. Since the periodicity of the crystal is destroyed, the scattering is enhanced in the region on the tomographic plane, which shows the increase of the absorption of X-rays. According to the energy conservation law, the energy of the scattered photons hitting the electrons is reduced, and the ability to penetrate the material is poor.

[0054] Figure 3 The results of the computer tomography 6 inch 4H-SiC crystal quality detection provided by the embodiment of the present application show one of the faults along the growth direction. It can be seen that the typical microtubular defects are less in the central region of the crystal ingot and more in the edge region. Since the edge expansion region of the silicon carbide crystal is prone to polycrystalline, the crystal quality is relatively poor, the crystal form is unstable, and the scanning results also show a regional white flaky distribution with high contrast.

[0055] In order to more clearly compare the computer tomography (CT) results of different types of defects, we compare the computer tomography (CT) results of four different positions, which are microtubular defects, single crystals with good crystallinity, carbon inclusions, and polycrystalline / multi-type regions. It can be seen that Figure 3 The dotted bright image in the middle ① position corresponds to the microtubular defect. The lattice distortion (periodic destruction) caused by the microtubular defect leads to enhanced X-ray diffraction / scattering and reduced X-ray penetration ability. Figure 3 The contrast uniform gray in the middle ② position corresponds to the single crystal. For single crystals with good crystallinity, the lattice arrangement is neat, the sample scatters less X-rays, and more X-rays are transmitted, resulting in a uniform gray. Figure 3 The strong contrast black dots in the middle ③ position correspond to carbon inclusion defects, which are due to the small density of carbon, which has less X-ray absorption and scattering in this area. Figure 3 The regional, shallow white area in the middle ④ position that evolves with the growth direction of the crystal corresponds to the polycrystalline / multi-type region. The polycrystalline or multi-type region of the crystal has a regional scattering enhancement on the fault plane due to the periodic destruction of the crystal, and the X-rays have reduced penetration into the sample.

[0056] In order to verify the reliability of the detection method provided by the embodiment for observing the defects of the silicon carbide crystal, we simultaneously observed the microstructure of the microtubular defect at the ① position in Figure 3 Figure 4 b is the result observed in a 10x optical microscope, Figure 4 c is the result obtained by three-dimensional scanning of a 150x laser confocal microscope, which can observe the typical stepped crystal surface inside the microtubule, which corresponds to the morphology characteristics of the microtubule in the literature. Proves the feasibility of computer tomography (CT) detection of defects in cubic silicon carbide crystals.

[0057] Correspondingly, through subsequent processing of computer graphics, the above typical defects can be extracted and presented in the three-dimensional reconstruction of the crystal, and the distribution density and other information can be calculated. In order to verify this function, the embodiment extracts the distribution of microtubular defects in the crystal and performs color processing in the three-dimensional image to present the typical microtubular defect distribution characteristics and the evolution with the growth of the crystal, as shown in Figure 5 .​

[0058] As described above, according to the present application, the quality of silicon carbide crystal can be detected quickly and non-destructively. By means of computer tomography combined with image contrast and defect identification of topography, the silicon carbide crystal ingot can be detected quickly and non-destructively, and the intuitive image of the defects existing in different sections of the measured ingot and the evolution of the defects with the spatial position can be directly obtained, which significantly improves the visualization degree of the defects of silicon carbide crystal and greatly improves the production efficiency.

[0059] The application of computer tomography technology in the detection of conventional materials such as amorphous and polycrystalline materials mainly relies on the difference in the distribution of the internal composition (density) of the materials to produce spatial differences in X-ray absorption, thereby realizing three-dimensional reconstruction, material structure analysis or substance identification. Unlike the above-mentioned detection methods of conventional metal parts, ceramics and rocks, semiconductor crystals are homogeneous and have high crystallinity, and the density changes very little. Therefore, the conventional CT imaging and material analysis method cannot be applied to crystal materials. Because the crystal material has very high penetration consistency to X-ray, in principle, the X-ray scattering or absorption enhancement caused by the micro-crystal defects in the crystal can be used to produce differentiated features in imaging to non-destructively detect the defects in the ingot. To implement this detection technology, the computer tomography equipment needs to improve the X-ray accelerating voltage, reduce the focal point size, prolong the exposure time, and use a specific segmented scanning method to improve the distinguishability of the micro-defects in the crystal.

[0060] To illustrate the difference in principle between the detection method of the present application and the computer tomography detection of conventional amorphous materials, and the difference in settings, Table 1 lists the system settings for detecting typical SiC crystals using the method of the present application and the system settings for detecting conventional amorphous materials.

[0061] Table 1

[0062] Voltage Current Focus size Magnification Exposure time Number of projections Resolution Scanning mode SiC 210 kV 180 uA 5 um 4x 1000 ms 2400 79 um Segmented scan Amorphous sample 150 kV 120 uA 400 um 2x 500 ms 1200 140 um Full area scan

[0063] The above examples are only used to further illustrate the present application, but the present application is not limited to the examples. Any simple modification, equivalent change and modification made according to the technical essence of the present application to the above examples all fall within the protection scope of the technical solution of the present application.

Claims

1. A rapid, non-destructive testing method for the quality of silicon carbide crystals, characterized in that, The method includes: S1. Perform surface cleaning treatment on silicon carbide single crystals; S2. A silicon carbide single crystal is placed on a stage for computed tomography scanning. X-rays emitted by the X-ray generator pass sequentially through a collimator and the silicon carbide single crystal before entering a detector. The detector receives the transmitted X-rays from the scanning layer and generates signals. The images are then reconstructed by a computer to obtain a tomographic image of the entire or partial section of the silicon carbide single crystal. The X-ray generator is a micro-focus high-energy X-ray tube with an accelerating voltage range of 200~350kV, a focal spot size of ≤10 micrometers, an exposure time of ≥1000ms, and a spatial resolution of ≤80 micrometers. S3. Based on the differences in X-ray absorption / scattering at different defect locations within a silicon carbide single crystal compared to an ideal crystal, differences in contrast and morphological features are generated in the tomographic images, thereby enabling the identification of defect types. These defect types include microtubule defects, carbon inclusion defects, polymorphic or crystalline inclusion defects, hexagonal void defects, and stacking fault defects. Microtubule defects appear as bright white spots extending along the crystal axis in the image; carbon inclusion defects appear as black spots distributed at specific locations within the crystal and not changing with the direction of tomographic growth; polycrystalline regions appear as regional light-colored blocks that evolve with the crystal growth direction; and single-crystal and defect-free locations in normal crystal regions appear as a uniform gray phase.

2. The rapid non-destructive testing method for silicon carbide crystal quality according to claim 1, characterized in that: The surface cleaning process involves cleaning the silicon carbide single crystal surface with toluene, acetone, ethanol, or isopropanol, followed by drying with nitrogen gas.

3. The rapid non-destructive testing method for silicon carbide crystal quality according to claim 1, characterized in that: The silicon carbide single crystal is a silicon carbide ingot, a silicon carbide substrate, or an epitaxial wafer.

4. The rapid non-destructive testing method for silicon carbide crystal quality according to claim 1, characterized in that: When the silicon carbide single crystal is placed on the stage, the Si surface of the silicon carbide single crystal faces upward, and the growth direction is perpendicular to the plane of the stage.

5. The rapid non-destructive testing method for silicon carbide crystal quality according to claim 1, characterized in that: The X-rays are scanned using a segmented scanning method.

6. The rapid non-destructive testing method for silicon carbide crystal quality according to claim 1, characterized in that: It also includes extracting image features of the defect type to calculate defect distribution, color representation, and distribution density.

7. The rapid non-destructive testing method for silicon carbide crystal quality according to claim 1, characterized in that: The silicon carbide crystal is a 6-inch 4H-SiC single crystal, and the X-ray generator is a micro-jog 240 kV high-energy X-ray tube with a focal point of 5 micrometers, an exposure time of 1000ms, and a projection number of 2400.

Images