EVALUATION METHODS AND MANUFACTURING METHODS OF SiC EPITAXIAL WAFERS

Abstract

Evaluation method for a SiC epitaxial wafer, comprising: a first observation step of producing a SiC epitaxial wafer with a high-concentration epitaxial layer with an impurity concentration of 1 × 10 18 cm -3 or more, irradiating a surface of the high-concentration epitaxial layer with an impurity concentration of 1 × 10 18 cm -3 or more with excitation light and observing a surface irradiated with the excitation light through a bandpass filter with a wavelength band of 430 nm or less, a second observation step of observing the surface irradiated with the excitation light through a bandpass filter with a wavelength band of more than 430 nm; and a first determination step of comparing an observation result of the first observation step with an evaluation result of the second observation step.

Description

BACKGROUND OF THE INVENTIONField of the Invention

[0001] The present invention relates to an evaluation method and a manufacturing method for a SiC epitaxial wafer. Description of the state of the art

[0002] Silicon carbide (SiC) exhibits an order of magnitude larger dielectric breakdown field, a three times larger band gap, and approximately three times higher thermal conductivity than silicon (Si). Therefore, silicon carbide (SiC) is expected to be applied in power supplies, high-frequency devices, high-temperature operating devices, and the like.

[0003] To promote the practical application of SiC devices, the establishment of high-quality SiC epitaxial wafers and high-quality epitaxial growth techniques is necessary.

[0004] A SiC device is formed on a SiC epitaxial wafer comprising a SiC substrate and an epitaxial layer stacked on the substrate. The SiC substrate can be obtained by processing a bulk single crystal of SiC grown by a sublimation recrystallization method or the like. The epitaxial layer is formed by a CVD (chemical vapor deposition) method and becomes an active region of the device.

[0005] Specifically, the epitaxial layer is formed on the SiC substrate, which has a plane with a cutout in the <11-20> direction from a (0001) plane as its growth surface. The epitaxial layer undergoes step flow growth (growth in the lateral direction from atomic steps) on the SiC substrate and becomes 4H-SiC.

[0006] Basal plane dislocation (BPD) is known as one of the device-killer defects in SiC devices. Due to the recombination energy of minority carriers flowing when a current is applied to a bipolar device, the BPD is extended through the epitaxial layer of the SiC substrate and becomes a high-resistance stacking fault. Furthermore, if a high resistance fraction is formed in the device, the reliability of the device is reduced. Therefore, a reduction in the BPD achieved by the epitaxial layer has been pursued.

[0007] Most BPDs in the SiC substrate can be converted into threading edge dislocations (TEDs), which do not cause defect strain during epitaxial layer formation. Currently, a conversion ratio of 99.9% or more can be realized, and device defects caused by BPDs in the epitaxial layer can be almost ignored. On the other hand, it has been found that BPDs in the SiC substrate that have been converted into TEDs will form stacking faults in the epitaxial layer when a large forward current flows. Without suppressing this, even if the conversion ratio to TEDs is increased, it cannot be said that the device defects caused by BPDs are completely eliminated. Therefore, preventing minority carriers near BPDs in the substrate from recombination is considered effective for this suppression.

[0008] Patent Documents 1 and 2 describe that by stacking an epitaxial layer doped with impurities at a high concentration on a SiC substrate, the probability of a minority carrier reaching BPDs in the substrate when a forward current is applied to a device can be suppressed and the formation of stacking faults that are high-resistance due to elongation can be prevented.

[0009] JP 2010-153464 A discloses an evaluation method for a SiC epitaxial wafer, comprising a first observation step of preparing a SiC epitaxial wafer having a high-concentration epitaxial layer, irradiating a surface of the high-concentration epitaxial layer with excitation light, and observing a surface irradiated with the excitation light through a bandpass filter having different wavelengths.

[0010] US 2013 / 0 320 357 A1 discloses a measurement method for defect detection on highly doped SiC wafers.

[0011] DE 11 2016 004 194 T5 discloses photoluminescence measurements on highly doped SiC wafers. PRIOR ART DOCUMENTSPatent documents Patent Document 1: PCT International Publication No. WO 2017 / 094 764 A1 Patent Document 2: PCT International Publication No. WO 2017 / 104 751 A1 SUMMARY OF THE INVENTIONProblems to be solved by the invention

[0012] BPDs in an epitaxial layer are generally identified using a photoluminescence (PL) method. For example, a SiC epitaxial wafer is irradiated with excitation light at 313 nm, and the irradiated surface is observed through a bandpass filter with a wavelength band of 660 nm or more. If a BPD is present in the irradiated surface, the position emits light.

[0013] On the other hand, in a case where a high-concentration epitaxial layer is used, BPDs in the high-concentration epitaxial layer are invisible because the light emission from the environment becomes stronger than the light emission from the BPDs in the wavelength band.

[0014] This means that BPDs present in a highly concentrated epitaxial layer could not be identified.

[0015] By using a wafer in which a high-concentration epitaxial layer is stacked on the SiC substrate as described above, it is possible to prevent BPDs from expanding and becoming high-resistance stacking faults when a forward current is applied to a device where conduction is caused by minority carriers, such as a bipolar device. However, from a quality control perspective, there is a problem that BPDs that have not been converted into TEDs cannot be identified due to the high-concentration epitaxial layer.

[0016] The present invention has been made in consideration of the above problems, and an object is to obtain an evaluation method and a manufacturing method for a SiC epitaxial wafer capable of evaluating a basal plane dislocation (BPD) in a high-concentration epitaxial layer. Means of solving the problem

[0017] As a result of intensive research, the inventors found that when a surface irradiated with excitation light is observed through a bandpass filter with a wavelength band of 430 nm or less, BPDs appear dark compared to other regions. In the prior art method, BPDs are identified by light emission, whereas in the present invention, BPDs are identified as dark lines. The detection process for such BPDs is different and cannot be achieved by simply changing the wavelength band of the bandpass filter.

[0018] The present invention is defined in the appended claims. Effects of the invention

[0019] With the aspect evaluation method of a SiC epitaxial wafer, it is possible to evaluate a basal plane dislocation (BPD) in the high-concentration epitaxial layer. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a view of a high-concentration epitaxial layer observed at the same position, wherein (a) is an image of a surface of the high-concentration epitaxial layer taken by a confocal microscope, (b) shows a result of observing the surface irradiated with the excitation light through a band-pass filter having a wavelength band of 660 nm, and (c) shows a result of observing the surface irradiated with the excitation light through a band-pass filter having a wavelength band of 420 nm. Fig. Figure 2 is a view of an epitaxial layer with an impurity concentration of about 1 × 10 16 cm -3observed at the same position, wherein (a) is an image of a surface of the high concentration epitaxial layer taken by a confocal microscope, (b) is a result of observation of the surface irradiated with the excitation light at the position through the bandpass filter having a wavelength band of 660 nm, and (c) is a result of observation of the surface irradiated with the excitation light through the bandpass filter having a wavelength band of 420 nm. Fig. 3 is a comparison view between a result of observation of the surface irradiated with the excitation light through the bandpass filter with a wavelength band of 420 nm and a result of surface observation by the confocal microscope after the surface of the high concentration epitaxial layer at the same location was etched with KOH. Fig. 4 is a view summarizing the results of observation of the surface irradiated with the excitation light at the same position by changing the wavelength band of the bandpass filter. Fig. 5 is a view showing a result of observing the surface irradiated with the excitation light through the bandpass filter having a wavelength band of 420 nm resulting from the change in the impurity concentration of the high concentration epitaxial layer. Fig. Figure 6 is a view showing the measurement results of a BPD in the case of conversion to a TED in the epitaxial layer at two points in the first observation step. Fig. 7 is a view of a prismatic stacking fault observed at the same position, wherein (a) is an image of the surface of the high concentration epitaxial layer of the confocal microscope, (b) shows a result of observation of the surface irradiated with the excitation light through the bandpass filter with a wavelength band of 660 nm, and (c) shows a result of observation of the surface irradiated with the excitation light through the bandpass filter with a wavelength band of 420 nm. Fig. 8 is a view of basal plane stacking faults observed at the same position, where (a) is an image of the surface of the high concentration epitaxial layer taken by the confocal microscope, (b) shows a result of observing the surface irradiated with the excitation light through the bandpass filter with a wavelength band of 660 nm, and (c) shows a result of observing the surface irradiated with the excitation light through the bandpass filter with a wavelength band of 420 nm. DETAILED DESCRIPTION OF THE INVENTION

[0020] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. In the drawings used in the following description, in order to better understand the features of the present invention, there are cases where characteristic portions are enlarged for convenience, and the dimensions, ratio, and the like of each component part may be the same or different from reality. The materials, dimensions, and the like shown in the following description are merely examples, and the present invention is not limited thereto and can be embodied in an appropriately modified manner within a range that does not change the gist. “Manufacturing process for SiC epitaxial wafers”.

[0021] A manufacturing method of a SiC epitaxial wafer according to the embodiment includes a step of stacking a high-concentration epitaxial layer having an impurity concentration of 1 × 10 18 cm -3 or more on a SiC substrate, a step of evaluating the high concentration epitaxial layer using a predetermined evaluation method of a SiC epitaxial wafer, and a step of stacking a drift layer on the high concentration epitaxial layer. (Step of stacking high-concentration epitaxial layers)

[0022] First, the SiC substrate is prepared. A method for manufacturing the SiC substrate is not particularly limited. For example, the SiC substrate can be obtained by slicing a SiC ingot obtained by a sublimation method or the like. In this application, the SiC epitaxial wafer refers to a wafer after the formation of an epitaxial film, and the SiC substrate refers to a wafer before the formation of the epitaxial layer.

[0023] In the SiC substrate, BPDs are present along a (0001) plane (c-plane). The number of BPDs exposed to a growth surface of the SiC substrate is preferably as small as possible, but not particularly limited. According to the current state of the art, the number of BPDs on the surface (growth surface) of a 6-inch SiC substrate is approximately 500 to 5000 per 1 cm 2 .

[0024] Subsequently, the high-concentration epitaxial layer is epitaxially grown on the SiC substrate. Nitrogen, boron, titanium, vanadium, aluminum, gallium, phosphorus, or the like can be used as doping impurities. The high-concentration epitaxial layer is a layer formed by epitaxial growth and a layer with an impurity concentration of 1 × 10 18 cm -3 or more. In a case where a forward current flows through a bipolar device with BPDs, the stack layer prevents minority carriers from reaching the BPDs. This suppresses the formation of Shockley stacking noise and the propagation of noise. This prevents deterioration of the device's forward-bias characteristics.

[0025] The thickness of the high concentration epitaxial layer is preferably 0.1 µm or more, preferably 1 µm or more, and more preferably 3 µm or more.

[0026] In addition, a buffer layer with an impurity concentration equal to or lower than that of the high-concentration epitaxial layer can be stacked between the SiC substrate and the high-concentration epitaxial layer. The buffer layer is a layer for reducing the difference in carrier concentration between the high-concentration epitaxial layer and the SiC substrate. (Step for evaluating the high-concentration epitaxial layer: First observation step)

[0027] Subsequently, as a step for evaluating the high-concentration epitaxial layer (first observation step), it is evaluated whether BPDs are present in the high-concentration epitaxial layer or not. In the first observation step, the surface of the high-concentration epitaxial layer is impurity-coated with an impurity concentration of 1 × 10 18 cm -3 or more is irradiated with excitation light and the surface irradiated with the excitation light is observed through a bandpass filter with a wavelength band of 430 nm or less.

[0028] A mercury lamp is preferably used as the excitation light source. The irradiation time can be selected as needed, but is preferably 10 msec or more and 100 sec or less, and more preferably 200 msec or more and 10 sec or less. When the excitation light is sufficiently irradiated, the contrast between the BPD and the other regions becomes clear. However, the "enhanced background PL" is generated by the excitation light that has been present in the meantime, which simultaneously leads to a reduction in detection sensitivity. Therefore, it is preferable to reduce the intensity of the irradiated excitation light to a low level. In particular, the intensity of the excitation light is preferably 1 Wcm. -2 or less and more preferably 500 mWcm -2or less. The emitted excitation light preferably has a wavelength of 280 nm or more and 375 nm or less. When a mercury lamp is used instead of a laser, the intensity of the emitted excitation light can be reduced to a low level. In the observation examples described below, a mercury lamp was used for irradiation.

[0029] Fig. Figure 1 is a view of a high-concentration epitaxial layer with an impurity concentration of 1 × 10 18 cm -3 or more in the observed SiC epitaxial wafer. (a) in Fig. 1 shows an image taken by a confocal microscope of a surface of the high concentration epitaxial layer at a predetermined position, (b) in Fig. 1 shows a result of observing the surface irradiated with the excitation light at the position through a bandpass filter with a wavelength band, and (c) in Fig. Figure 1 shows a result of observing the surface irradiated with the excitation light at the position through a bandpass filter with a wavelength band of 420 nm.

[0030] As in Fig. As shown in Figure 1, no BPDs were identified for the high-concentration epitaxial layer even when the surface irradiated with excitation light was observed through a bandpass filter with a wavelength band of 660 nm. Furthermore, BPDs were not identified even when the surface was observed with a confocal microscope. In contrast, BPDs were observed as dark when the surface irradiated with excitation light was observed through a bandpass filter with a wavelength band of 420 nm.

[0031] On the other hand, Fig. 2 a view of an epitaxial layer with an impurity concentration of 1 × 10 16 cm -3 in the observed SiC epitaxial wafer. (a) in Fig. 2 shows an image taken by a confocal microscope of a surface of the high concentration epitaxial layer at a predetermined position, (b) in Fig. 2 shows a result of observation of the surface irradiated with the excitation light at the position through the bandpass filter with a wavelength band of 660 nm, and (c) in Fig. Figure 2 shows a result of observation of the surface irradiated with the excitation light at the position through the bandpass filter with a wavelength band of 420 nm.

[0032] As in Fig. 2, the epitaxial layer was treated with an impurity concentration corresponding to that of the drift layer (approx. 1 × 10 16 cm -3), when observing the surface irradiated with the excitation light through the bandpass filter with a wavelength band of 660 nm, the light emission by BPDs (photoluminescence light) was observed.

[0033] As in Fig. 2, was also present in the epitaxial layer with an impurity concentration of about 1 × 10 16 cm -3 The surface irradiated with the excitation light is observed through the bandpass filter with a wavelength band of 420 nm.

[0034] In this case, too, it was observed that BPDs were dark. That is, although the observation state of BPDs differs between light emission and absorption, it was confirmed that BPDs were identified by observation through the bandpass filter with a wavelength band of 420 nm. Furthermore, BPDs could not be identified in the film even with the confocal microscope.

[0035] Fig. Figure 3 is a comparison view between the result of observation of the surface irradiated with the excitation light through the bandpass filter with a wavelength band of 420 nm and the result of surface observation through the confocal microscope after the surface of the high-concentration epitaxial layer at the same location was etched with KOH. When the epitaxial layer containing BPDs is etched with KOH, pits are created at the locations where BPDs are present. As shown in Fig. As shown in Figure 3, the result of observing the surface irradiated with excitation light through a bandpass filter with a wavelength band of 420 nm and the result of observing the surface of the high-concentration epitaxial layer etched with KOH and observed with a confocal microscope have a correspondence relationship. Therefore, from this point of view, it was also confirmed that BPDs were identified by observing through a bandpass filter with a wavelength band of 420 nm.

[0036] In Fig. 1, BPDs were identified using the bandpass filter with a wavelength band of 420 nm. However, the bandpass filter used in the first observation step is not limited to the wavelength band, and a bandpass filter with a wavelength band of 430 nm or less can be used.

[0037] Fig. 4 is a view summarizing the results of observation of the surface irradiated with the excitation light in the SiC epitaxial wafer having the high-concentration epitaxial layer with an impurity concentration of 1 × 10 18 cm -3 or more by changing the wavelength band of the bandpass filter. As in Fig. As shown in Figure 4, BPDs could be identified when using the bandpass filter with a wavelength band of 420 nm or less. However, when using the bandpass filter with a wavelength band greater than 420 nm, BPDs could not be identified. Here, a bandpass filter with a specific wavelength can transmit light of a wavelength band of about ±10 nm of a specific wavelength. That is, if BPDs can be identified by the bandpass filter with a wavelength band of 420 nm, the BPDs can also be identified by the bandpass filter with a wavelength band of 430 nm. Therefore, the bandpass filter with a wavelength band of 430 nm or less can be used in the first observation step.

[0038] The impurity concentration of the high-concentration epitaxial layer is preferably 1 × 10 18 cm -3 or more and 2 × 10 19 cm -3or less and more preferably 1 × 10 19 cm -3 or less. If the impurity concentration of the high-concentration epitaxial layer is too high, the probability of defects other than BPDs occurring increases.

[0039] Fig. Figure 5 shows a result of observing the surfaces irradiated with the excitation light through the bandpass filter with a wavelength band of 420 nm in the three SiC epitaxial wafers with the high-concentration epitaxial layer with an impurity concentration of 1 × 10 18 cm -3 or more. The results of the evaluation of SiC epitaxial wafers with the high-concentration epitaxial layer with different impurity concentrations are presented. As in Fig. As shown in Figure 5, BPDs can be identified by repeating the first observation step at each impurity concentration of 1 × 10 18 cm -3 up to 1 × 10 19 cm -3is carried out.

[0040] In the method, there may be cases where defects other than BPDs are identified. This can be achieved if the presence or absence of BPDs can be determined. However, it is more preferable to identify the positions of the BPDs. Therefore, it is more preferable that BPDs can be distinguished from other defects.

[0041] One of the discrimination methods is discrimination using the shape of the defects to be identified (first discrimination method).

[0042] A BPD is a dislocation that exists on a (0001) plane (c-plane), which is the basal plane of a SiC substrate. Generally, the SiC substrate has a surface with an offset angle in the <11-20> direction of (0001) as its growth surface. Therefore, BPDs extending in the offset direction are observed. That is, a BPD is a defect with a major axis with an inclination angle relative to the <11-20> direction of 45° or less.

[0043] Therefore, it is preferable to distinguish a defect by determining a defect that extends in the offset direction and has an aspect ratio greater than 1 as BPD. Depending on the accuracy of distinguishing defects, the aspect ratio may be 1.5 or more or 2.0 or more. Here, the aspect ratio is a value obtained by dividing the length of the major axis of a flaw to be measured by the length of the minor axis.

[0044] As in Fig. As shown in Figure 1, when an epitaxial layer following a BPD extends from the SiC substrate to the surface of the epitaxial layer, the aspect ratio is 2.0 or more. On the other hand, when the BPD in the epitaxial layer is converted to a TED, the aspect ratio is less than 2.0 in some cases. Fig. Figure 6 shows the measurement results of a BPD in the case of conversion to a TED in the epitaxial layer in the first observation step. The aspect ratio of the layers in (a) is Fig. 6 is 1.67, and the aspect ratio of the image shown in (b) is Fig. The BPD shown in Figure 6 is 1.85.

[0045] As another method for distinguishing a BPD from other defects, there is a method for comparing the other evaluation results. For example, a second discrimination method further includes a second observation step of observing the surface irradiated with the excitation light through a bandpass filter with a wavelength band of 430 nm or more, and a first determination step of comparing the observation result of the first observation step with an evaluation result of the second observation step.In addition, for example, a third discrimination method further includes a third observation step of performing surface observation on a surface corresponding to that irradiated with the excitation light with a microscope or the like, and a second determination step of comparing the observation result of the first observation step with an evaluation result of the third observation step.

[0046] Defects with a similar shape to a BPD include prismatic stacking faults (carrot-shaped faults), basal plane stacking faults, and the like. Fig. Figure 7 is a view of an observed prismatic stacking fault. (a) in Fig. 7 is an image taken by a confocal microscope of the surface of the high concentration epitaxial layer at a predetermined position, (b) in Fig. 7 is a result of observing the surface irradiated with the excitation light at the position through the bandpass filter with a wavelength band of 660 nm, and (c) in Fig. 7 is a result of the observation of the surface irradiated with the excitation light at the position through the bandpass filter with a wavelength band of 420 nm.

[0047] In addition, Fig. 8 a view of the observed basal plane stacking faults. (a) in Fig. Fig. 8 is an image taken by a confocal microscope of the surface of the high concentration epitaxial layer at a predetermined position, (b) in Fig. Figure 8 shows a result of observation of the surface irradiated with the excitation light at the position through the bandpass filter with a wavelength band of 660 nm, and (c) in Fig. Figure 8 shows a result of observation of the surface irradiated with the excitation light at the position through the bandpass filter with a wavelength band of 420 nm.

[0048] As in Fig. As shown in Figure 7, a prismatic stacking fault can be identified using a confocal microscope. Furthermore, when the surface irradiated with excitation light is observed through a bandpass filter with a wavelength band of 660 nm, the prismatic stacking fault does not emit light and appears dark. That is, by performing the second observation step and / or the third observation step and performing the first determination step and / or the second determination step, a prismatic stacking fault and a BPD can be distinguished from each other. That is, a BPD is not the one that can be confirmed by surface observation before irradiation without a bandpass filter or by using a bandpass filter with a wavelength band of 660 nm.

[0049] As in Fig. As shown in Figure 8, a basal plane stacking fault cannot be identified with the confocal microscope.

[0050] On the other hand, when the surface irradiated with the excitation light is observed through a bandpass filter with a wavelength band of 660 nm, the basal plane stacking fault does not emit light and appears dark. Therefore, a basal plane stacking fault can be distinguished by performing the second observation step and the first determination step.

[0051] A third determination step for comparing the evaluation results of the first, second, and third observation steps may be performed to determine a defect observed as a basal plane dislocation only in the observation result of the first observation step, to determine a defect observed as a prismatic stacking fault in all of the first to third observation steps, and to determine a defect observed as a basal plane stacking fault only in the observation results of the first and second observation steps.

[0052] As described above, when using the evaluation method of a SiC epitaxial wafer according to the embodiment, basal plane dislocations (BPDs) can be evaluated even when the high-concentration epitaxial layer is stacked.

[0053] In addition, the BPDs measured using the evaluation method can be automatically detected or quantitatively counted using image analysis software. (Step of stacking the drift layer)

[0054] Finally, it is preferable that the drift layer be stacked on the high-concentration epitaxial layer. The drift layer is stacked using a known method. The impurity concentration of the drift layer is lower than that of the high-concentration epitaxial layer and is below 1 × 10 17 cm -3The drift layer is a layer on which a SiC device is formed. When BPDs are included in the drift layer, they cause degradation of the forward conduction characteristics of the SiC device. In this context, by converting BPDs into TEDs at the interface between the substrate and the epitaxial layer, the number of BPDs contained in the drift layer can be reduced. Furthermore, since the drift layer is stacked on the high-concentration epitaxial layer, the degradation of characteristics caused by BPDs in the substrate is suppressed.

[0055] Furthermore, in the high-concentration epitaxial layer evaluation step, predetermined acceptance criteria for the presence or absence of BPDs, the positions of BPDs, and / or the number of BPDs may be specified with respect to the BPDs present in the high-concentration epitaxial layer. For example, a step for determining whether the result of the first observation step meets a predetermined acceptance criterion may be provided to stack the drift layer only on the high-concentration epitaxial layer that meets the acceptance criterion. Alternatively, in a case where the predetermined acceptance criterion is not met, an additional step may be performed to compensate for the non-met criterion, and then the drift layer may be stacked.

[0056] As described above, according to the manufacturing method of a SiC epitaxial wafer according to the embodiment, it is possible to prevent BPDs from causing deterioration of device characteristics. Furthermore, despite the high-concentration epitaxial layer, the presence or absence and positions of BPDs can be identified, so that quality control of the SiC epitaxial wafer can be easily performed.

[0057] According to the present invention, it is possible to obtain an excellent evaluation method and a manufacturing method for a SiC epitaxial wafer capable of evaluating a basal plane dislocation (BPDs) in a high concentration epitaxial layer.

Claims

[1] Evaluation method for a SiC epitaxial wafer, comprising: a first observation step of producing a SiC epitaxial wafer with a high-concentration epitaxial layer with an impurity concentration of 1 × 10 18 cm -3 or more, irradiating a surface of the high-concentration epitaxial layer with an impurity concentration of 1 × 10 18 cm -3 or more with excitation light and observing a surface irradiated with the excitation light through a bandpass filter with a wavelength band of 430 nm or less, a second observation step of observing the surface irradiated with the excitation light through a bandpass filter with a wavelength band of more than 430 nm; and a first determination step of comparing an observation result of the first observation step with an evaluation result of the second observation step. [2] Evaluation method for a SiC epitaxial wafer, comprising: a first observation step of producing a SiC epitaxial wafer with a high-concentration epitaxial layer with an impurity concentration of 1 × 10 18 cm -3 or more, irradiating a surface of the high-concentration epitaxial layer with an impurity concentration of 1 × 10 18 cm -3 or more with excitation light and observing a surface irradiated with the excitation light through a bandpass filter with a wavelength band of 430 nm or less, a third observation step of performing a surface observation on a surface which is the same as the surface irradiated with the excitation light, before its irradiation with the excitation light; and a second determination step of comparing an observation result of the first observation step with an evaluation result of the third observation step. [3] Evaluation method of a SiC epitaxial wafer according to claim 1 or 2, wherein in the first observation step a defect is observed which extends in an offset direction and has an aspect ratio of more than 1. [4] Evaluation method of a SiC epitaxial wafer according to claim 1 or 2, wherein the SiC epitaxial wafer includes a SiC substrate and the epitaxial layer stacked on the substrate, and Presence or absence and a position of a basal plane dislocation of the wafer are detected in the first observation step. [5] The evaluation method of a SiC epitaxial wafer according to claim 1, further comprising: a third observation step of performing a surface observation on a surface which is the same as the surface irradiated with the excitation light, before its irradiation with the excitation light; and a third determination step of comparing the evaluation results of the first, second and third observation steps, wherein in the third determination step, a defect that is only observed in the observation result of the first observation step is determined as a basal plane dislocation. [6] Evaluation method of a SiC epitaxial wafer according to claim 5, wherein in the third determination step, a defect observed in all of the first to third observation steps is determined as a prismatic stacking fault, and a defect observed only in the observation results of the first and second observation steps is determined as a basal plane stacking fault. [7] A manufacturing method for a SiC epitaxial wafer, comprising the steps of: Stacking a high-concentration epitaxial layer with an impurity concentration of 1 × 10 18 cm -3 or more on a surface of a SiC substrate; Evaluating the high-concentration epitaxial layer using the evaluation method of a SiC epitaxial wafer according to any one of claims 1 to 6; and Stacking a drift layer on the high concentration epitaxial layer. [8] A manufacturing method of a SiC epitaxial wafer according to claim 7, further comprising: a step of stacking a buffer layer of a lower impurity concentration than that of the high concentration epitaxial layer between the SiC substrate and the high concentration epitaxial layer. [9] A manufacturing method of a SiC epitaxial wafer according to claim 7, further comprising: a step of determining whether the obtained result of the first observation step meets an acceptance criterion or not, after the step of evaluating the high-concentration epitaxial layer, wherein in the step of stacking the drift layer, the drift layer is stacked on the high concentration epitaxial layer satisfying the acceptance criterion.

Images