SiC EPITAXIAL WAFER

Abstract

To provide a SiC epitaxial wafer with high surface uniformity of doping concentration in a high concentration layer.SOLUTION: A SiC epitaxial wafer 100 includes a SiC single crystal substrate 10 and a high concentration layer 21 (20) provided on the SiC single crystal substrate, having an average value of n-type doping concentration of 1.03×1018 / cm3 or more and 1×1019 / cm3 or less, and having a doping concentration uniformity on the surface of 30% or less. Nitrogen is doped in the high concentration layer 21 (20). The uniformity within the surface is the absolute value of (the maximum value of the doping concentration within the surface-the minimum value of the doping concentration within the surface) / the average value of the doping concentration within the surface.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a SiC epitaxial wafer. [Background technology]

[0002] Silicon carbide (SiC) has an electric breakdown field one order of magnitude larger than that of silicon (Si), a band gap three times larger, and a thermal conductivity three times higher. For this reason, silicon carbide (SiC) is expected to be used in power devices, high-frequency devices, high-temperature operating devices, and other applications.

[0003] To promote the practical use of SiC devices, it is necessary to develop high-quality, low-cost SiC epitaxial wafers and establish epitaxial growth technology.

[0004] SiC devices are formed on a SiC epitaxial wafer that includes a SiC substrate and an epitaxial layer stacked on the substrate. The SiC substrate is obtained by processing a bulk single crystal of SiC grown by a sublimation recrystallization method or the like. The epitaxial layer is fabricated by a chemical vapor deposition (CVD) method or the like, and serves as the region that maintains the device's voltage resistance.

[0005] More specifically, the epitaxial layer is formed on a SiC substrate with a growth plane that is off-angled in the <11-20> direction from the (0001) plane. The epitaxial layer grows on the SiC substrate by step-flow growth (lateral growth from atomic steps) to become 4H-SiC.

[0006] In SiC epitaxial wafers, basal plane dislocations (BPDs) are known to be one of the device killer defects that cause fatal defects in SiC devices. For example, when a forward current is applied to a bipolar device, the recombination energy of the flowing carriers causes partial dislocations of basal plane dislocations inherited from the SiC substrate to the epitaxial layer to move and expand, forming high-resistance stacking faults. If high-resistance areas are generated in the device, this causes a decrease in device reliability (forward degradation). For this reason, efforts have been made to reduce basal plane dislocations inherited by the epitaxial layer.

[0007] Many of the basal plane dislocations in a SiC substrate can be converted into threading edge dislocations (TEDs), which do not cause defect expansion when an epitaxial layer is formed (Patent Document 1). However, it has become clear in recent years that when a large current flows in the forward direction, basal plane dislocations that are converted to threading edge dislocations at the interface between the SiC substrate and the epitaxial layer also expand into stacking faults (SFs) in the epitaxial layer. Therefore, for high-current power devices, which are expected to see market expansion in the future, the formation of stacking faults cannot be sufficiently suppressed by simply converting basal plane dislocations to threading edge dislocations, and there is always the concern that device reliability may deteriorate.

[0008] Patent Document 2 discloses that the conversion efficiency of basal plane dislocations to threading edge dislocations at the interface between the SiC single crystal substrate and the epitaxial layer is increased by forming an epitaxial layer with a higher impurity concentration in addition to the normal epitaxial layer in the SiC epitaxial wafer. By increasing the conversion efficiency to basal plane dislocations, the extension and expansion of basal plane dislocations can be suppressed. The extension and expansion of basal plane dislocations are the cause of forward degradation of devices. Therefore, the formation of an epitaxial layer with a high impurity concentration is considered to be a promising solution for suppressing the forward degradation of SiC devices using SiC epitaxial wafers.

[0009] Patent Document 3 discloses a manufacturing method for improving the in-plane uniformity of the doping concentration of a low-concentration layer, but does not mention the in-plane uniformity of the doping concentration of a high-concentration layer. [Prior art documents] [Patent documents]

[0010] [Patent Document 1] JP 2009-88223 A [Patent Document 2] International Publication No. 2017 / 094764 [Patent Document 3] Patent No. 6386706 Summary of the Invention [Problem to be solved by the invention]

[0011] The present inventors found a problem that, during the production of a SiC epitaxial wafer having a SiC single crystal substrate, a normal epitaxial layer, and an epitaxial layer with a high n-type doping concentration therebetween, the in-plane uniformity of the n-type doping concentration of the epitaxial layer with a high n-type doping concentration (herein, in this specification, the in-plane uniformity of the doping concentration refers to "the absolute value of the maximum doping concentration - the minimum doping concentration) / the average doping concentration") deteriorates, and as a result of extensive investigations, they came up with the present invention, which solves this problem.

[0012] The present invention has been made in view of the above circumstances, and has an object to provide a SiC epitaxial wafer in which the high concentration layer has high in-plane uniformity of the n-type doping concentration. [Means for solving the problem]

[0013] In order to solve the above problems, the present invention provides the following means.

[0014] (1) Aspect 1 of the present invention is a method for manufacturing a SiC single crystal substrate, and a method for manufacturing a SiC single crystal substrate, comprising:18 / cm 3 That's it, 1×10 19 / cm 3 and a high concentration layer having an in-plane uniformity of doping concentration of 30% or less, wherein the high concentration layer is doped with nitrogen; here, the in-plane uniformity is the absolute value of (maximum in-plane doping concentration - minimum in-plane doping concentration) / average in-plane doping concentration.

[0015] (2) Aspect 2 of the present invention is a SiC single crystal substrate, and a SiC single crystal substrate having an n-type doping concentration of 3×10 18 / cm 3 That's it, 1×10 19 / cm 3 and a high concentration layer having an in-plane uniformity of doping concentration of 30% or less, wherein the high concentration layer is doped with nitrogen; here, the in-plane uniformity is the absolute value of (maximum in-plane doping concentration - minimum in-plane doping concentration) / average in-plane doping concentration.

[0016] (3) Aspect 3 of the present invention is a SiC single crystal substrate, and a SiC single crystal substrate having an n-type doping concentration of 5×10 18 / cm 3 That's it, 1×10 19 / cm 3 and a high concentration layer having an in-plane uniformity of doping concentration of 30% or less, wherein the high concentration layer is doped with nitrogen; here, the in-plane uniformity is the absolute value of (maximum in-plane doping concentration - minimum in-plane doping concentration) / average in-plane doping concentration.

[0017] (4) A fourth aspect of the present invention is a method for manufacturing a semiconductor device comprising the steps of: forming a SiC single crystal substrate; and forming a semiconductor device having an n-type doping concentration of 1×10 on the SiC single crystal substrate. 18 / cm 3 That's it, 1×10 19 / cm 3a high-concentration layer having a doping concentration of 30% or less and an in-plane uniformity of 30% or less, and a buffer layer between the SiC single crystal substrate and the high-concentration layer, which converts basal plane dislocations into threading edge dislocations, wherein the high-concentration layer is doped with nitrogen; here, the in-plane uniformity is defined as the absolute value of (maximum in-plane doping concentration - minimum in-plane doping concentration) / average in-plane doping concentration.

[0018] (5) A fifth aspect of the present invention is the SiC epitaxial wafer of any one of the first to fourth aspects, wherein the high concentration layer is free of basal plane dislocations.

[0019] (6) A sixth aspect of the present invention is a SiC epitaxial wafer according to any one of the first to fifth aspects, wherein the high concentration layer is a buffer layer, and a drift layer is provided on the buffer layer, the drift layer having an average doping concentration lower than an average doping concentration of the buffer layer.

[0020] (7) A seventh aspect of the present invention is a SiC epitaxial wafer according to any one of the first to sixth aspects, wherein the in-plane uniformity is 20% or less.

[0021] (8) According to an eighth aspect of the present invention, in the SiC epitaxial wafer of any one of the first to sixth aspects, the in-plane uniformity is 10% or less.

[0022] (9) A ninth aspect of the present invention is a SiC epitaxial wafer according to any one of the first to eighth aspects, the diameter of which is 150 mm or more. Effect of the Invention

[0023] According to the SiC epitaxial wafer of the present invention, it is possible to provide a SiC epitaxial wafer in which the high concentration layer has high in-plane uniformity of the n-type doping concentration. [Brief description of the drawings]

[0024] [Figure 1]1 is a schematic cross-sectional view showing a SiC epitaxial wafer according to one embodiment of the present invention. [Diagram 2] FIG. 2 is a schematic cross-sectional view showing a SiC epitaxial wafer according to another embodiment of the present invention. [Diagram 3] 1 is a schematic cross-sectional view showing an example of a configuration capable of independently controlling the in-plane distribution of a C-based gas supply and a Si-based gas supply. FIG. [Figure 4] 1 is a graph showing the relationship between the input nitrogen flow rate and the obtained doping concentration (average value) for three samples shown in Table 1 when the C / Si ratio was set to 1.15. [Diagram 5] This is the result of investigating the relationship between the average doping concentration and the growth rate. [Figure 6] (a) is an image of nitrogen (N) doping. Compared to (a), (b) is an image of a case where the C / Si is low, (c) is an image of a case where the C / Si is high, and (d) is an image of a case where the doping flow rate is high. [Figure 7] FIG. 13 is a conceptual diagram for explaining how to estimate a substantially insufficient amount of C-based gas input from a calibration curve and a growth rate decrease amount. [Figure 8] 1 is a graph showing the relationship between the average value of the doping concentration, the in-plane uniformity of the doping concentration, and C / Si. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0025] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the following embodiments, the same or equivalent parts may be denoted by the same reference numerals in the drawings. In addition, the drawings used in the following description may show characteristic parts in an enlarged scale for the sake of convenience in order to make the features easier to understand, and the dimensional ratios of each component may not necessarily be the same as in reality. In addition, the materials, dimensions, etc. exemplified in the following description are merely examples, and the present invention is not limited thereto. The present invention can be implemented by appropriately changing the materials, dimensions, etc. within the scope of the effects of the present invention. The configuration shown in one embodiment can also be applied to other embodiments.

[0026] (SiC Epitaxial Wafer) FIG. 1 is a schematic cross-sectional view showing an SiC epitaxial wafer according to an embodiment of the present invention, and FIG. 2 is a schematic cross-sectional view showing an SiC epitaxial wafer according to another embodiment.

[0027] The SiC epitaxial wafer 100 shown in FIG. 1 includes an SiC single crystal substrate 10 and an SiC epitaxial layer 20 formed on a main surface 10a of the SiC single crystal substrate 10. The SiC epitaxial layer 20 included in the SiC epitaxial wafer 100 has an average doping concentration of 1 × 10 18 / cm 3 or more and 1 × 10 19 / cm 3 or less, and is composed of a high-concentration layer 21 having an in-plane uniformity of doping concentration of 30% or less.

[0028] In the SiC epitaxial wafer 200 shown in FIG. 2, the SiC epitaxial layer 20 has a high-concentration layer 21 as a buffer layer, and a drift layer 22 having an average doping concentration lower than the average doping concentration of the buffer layer is provided on the buffer layer.

[0029] <SiC Single Crystal Substrate> As the SiC single crystal substrate 10, one obtained by slicing an SiC ingot obtained by a sublimation method or the like can be used. In this specification, the SiC epitaxial wafer means a wafer after forming an epitaxial layer, and the SiC single crystal substrate means a wafer before forming an epitaxial layer.

[0030] The SiC single crystal substrate 10 is not limited in size, but is preferably 100 mm, and more preferably 150 mm or more.

[0031] As the SiC single crystal substrate 10, one having a plane with an offset angle in the <11-20> direction from (0001) as a growth plane can be used.

[0032] Basal plane dislocations exist along the (0001) plane (c-plane) in the SiC single crystal substrate 10. The number of basal plane dislocations exposed on the growth surface of the SiC single crystal substrate is preferably small, but is not particularly limited.

[0033] When the growth plane of the SiC single crystal substrate 10 is a plane that has an offset angle from (0001) in the <11-20> direction, the basal plane dislocations are inclined with respect to the growth plane.

[0034] The SiC single crystal substrate 10 is doped with, for example, nitrogen. There are no particular limitations on the doping concentration of the SiC single crystal substrate 10, and it may be a normal one for a SiC substrate for a power semiconductor.

[0035] <High concentration layer> The high-concentration layer 21 has an average n-type doping concentration of 1×10 18 / cm 3 That's it, 1×10 19 / cm 3 or less, and the in-plane uniformity of the n-type doping concentration is 30% or less. The in-plane uniformity of the n-type doping concentration of the high concentration layer 21 is preferably within 20%, and more preferably 10% or less. Note that the lower the in-plane uniformity of the n-type doping concentration, the better the quality of the SiC epitaxial wafer, but from the viewpoint of yield, the lower limit can be set to, for example, 1%.

[0036] At the interface between the epitaxial layer and the SiC single crystal substrate, positive carriers (holes) and negative carriers (electrons) recombine with each other at the basal plane dislocations of the SiC single crystal substrate, expanding the basal plane dislocations in the epitaxial layer. High concentration layer 21, which has carriers at a high concentration, prevents carriers in the epitaxial layer from reaching the SiC single crystal substrate.

[0037] When the high-concentration layer 21 is a buffer layer and the drift layer 22 is provided thereon, the high-concentration buffer layer 21 and the drift layer 22 can be clearly distinguished from each other by the difference in doping concentration.

[0038] When the inventors fabricated a SiC epitaxial wafer having a high-concentration buffer layer and a drift layer on a SiC single crystal substrate, the in-plane uniformity of the doping concentration of the high-concentration buffer layer was 50% or more. By using the manufacturing method described later, it is possible to manufacture a high concentration layer with an in-plane uniformity of the doping concentration of 30% or less. If the doping concentration uniformity is poor, there will be areas on the wafer where the concentration is lower than the target, reducing the carrier recombination effect, or there will be areas on the wafer where the concentration is higher than the target, causing defects due to high concentration. These problems can be prevented by maintaining a good uniformity of 30% or less on the wafer.

[0039] The high concentration layer 21 is of n-type, and nitrogen is used as the doping impurity.

[0040] The average n-type doping concentration of the high concentration layer 21 is 1×10 18 / cm 3 That's it, 1×10 19 / cm 3 The following is the result.

[0041] The thickness of the high concentration layer 21 is not particularly limited, but may be, for example, about 1 μm to 10 μm. If it is too thin, the effect of suppressing carriers from reaching the SiC single crystal substrate is reduced, and if it is too thick, the cost becomes high.

[0042] The thickness distribution of the high concentration layer 21 is preferably 10% or less. When the film thickness distribution of the high-concentration layer 21 is 10% or less, it can be said that at least one of the in-plane distribution of the C-based gas supply on the substrate surface and the in-plane distribution of the Si-based gas on the substrate surface is 10% or less, and the surface of the high-concentration layer 21 is likely to become a mirror surface.

[0043] <Drift layer> The high concentration layer 21 is a buffer layer, on which a drift layer 22 may be provided. The drift layer 22 is a layer through which a drift current flows and functions as a device. The drift current is a current generated by the flow of carriers when a voltage is applied to a semiconductor. The doping concentration of the drift layer 22 is, for example, 1×10 14 cm -3 That's it, 1×10 17 cm -3 is less than or equal to 1×10 16 cm -3 That's about it.

[0044] <Method of measuring doping concentration> The n-type doping concentration of the high concentration layer can be measured by a mercury probe (Hg-CV) method or secondary ion mass spectrometry (SIMS). In the Hg-CV method, the n-type doping concentration is N d -N a is measured, where N d is the donor concentration and N a is the acceptor concentration. N d Compared to N a If it is confirmed that is sufficiently small, N d -N a ≒N d It can be safely assumed that. By using secondary ion mass spectrometry (SIMS), the doping concentration of a high-concentration layer of a SiC epitaxial wafer can be measured by cutting the high-concentration layer in the depth direction while performing measurements. The same can be said for a SiC epitaxial wafer with a drift layer on the high-concentration layer.

[0045] The measurement points may be any points that can reflect the distribution within the wafer surface, but they include the wafer center and the measurement point 5 mm from the wafer edge as the measurement point farthest from the wafer center, and do not include measurement points less than 5 mm from the edge. As a specific procedure for measuring the n-type doping concentration of a high concentration layer, for example, in the case of a 6-inch wafer, the n-type doping concentration is measured at multiple points, for example, 21 points, in the direction of the cross with the center of the wafer as the origin. The average value of the n-type doping concentration is calculated using the n-type doping concentration obtained at each point, and the absolute value of the difference between the maximum and minimum values ​​of the n-type doping concentration is divided by the calculated average value of the n-type doping concentration to obtain the in-plane uniformity. One of the directions of the cross can be parallel to the orientation flat.

[0046] <Other layers> The SiC epitaxial wafer of the present invention may include other layers as long as the effects of the present invention are achieved. For example, another buffer layer (hereinafter referred to as a first buffer layer) made of an n-type or p-type semiconductor having a lower impurity concentration (doping concentration) than the SiC single crystal substrate 10 may be provided between the SiC single crystal substrate 10 and the high concentration layer 21. The first buffer layer can be provided to convert basal plane dislocations into threading edge dislocations. From this viewpoint, the first buffer layer is a BPD conversion layer. The impurity concentration of the first buffer layer is preferably lower than that of the SiC single crystal substrate 10, and is preferably equal to or lower than the impurity concentration of the high concentration layer 21. The impurity concentration of the first buffer layer is preferably 1×10 17 cm -3 The impurity concentration of the first buffer layer is preferably 1×10 19 cm -3 The impurity concentration of the first buffer layer can be set to be intermediate between the impurity concentrations of the SiC single crystal substrate 10 and the high concentration layer 21 in order to alleviate the lattice mismatch between the two.

[0047] (SiC epitaxial wafer manufacturing method) The method for manufacturing the SiC epitaxial wafer 100 or the SiC epitaxial wafer 200 according to this embodiment involves, for example, growing the epitaxial layer 20 on a SiC single crystal substrate 10 whose main surface has an off angle of 0.4° to 5° with respect to the (0001) plane.

[0048] First, prepare a SiC single crystal substrate 10. There is no particular limitation on the method for producing the SiC single crystal substrate 10. For example, it can be obtained by slicing a SiC ingot obtained by a sublimation method or the like.

[0049] Next, the SiC epitaxial layer 20 is epitaxially grown on the SiC single crystal substrate 10 to produce the SiC epitaxial wafer 100. The SiC epitaxial layer 20 can be formed on the growth surface 10a of the SiC single crystal substrate 10 by step-flow growth (lateral growth from atomic steps) using a chemical vapor deposition (CVD) method.

[0050] The step of forming the SiC epitaxial layer 20 is performed by flowing source gas and dopant gas onto a SiC single crystal substrate maintained at high temperature.

[0051] The source gas is the gas used as the raw material when forming the SiC epitaxial layer. Generally, source gases are divided into Si-based source gases that contain Si in the molecule and C-based source gases that contain C in the molecule.

[0052] The silicon-based source gas may be a known one, for example, silane (SiH 4 ) In addition, dichlorosilane (SiH 2 Cl 2 ), trichlorosilane (SiHCl 3 ), tetrachlorosilane (SiCl 4 It is also possible to use a chlorine-based Si source-containing gas (chloride-based source) containing Cl, which has an etching effect, such as propane (C 3 H 8 ), ethylene (C 2 H 4 ) etc. can be used.

[0053] A dopant gas is a gas that contains an element that acts as a donor or acceptor (carrier). Nitrogen or ammonia is used as a dopant gas to grow an n-type epitaxial layer, and trimethylaluminum (TMA) or triethylaluminum (TEA) is used to grow a p-type epitaxial layer.

[0054] In addition, a gas for transporting these gases into the reactor may be used at the same time, for example, hydrogen, which is inactive with respect to SiC.

[0055] When manufacturing SiC epitaxial wafer 100, the process of forming SiC epitaxial layer 20 is a high-concentration layer process of forming high-concentration layer 21 on SiC single crystal substrate 10, and when manufacturing SiC epitaxial wafer 200, the process of forming SiC epitaxial layer 20 is divided into a high-concentration layer process of forming high-concentration layer 21 on SiC single crystal substrate 10 and a drift layer process of forming a drift layer on high-concentration layer 21.

[0056] <High concentration layer process> The growth temperature can be, for example, 1400 to 1800° C., and more preferably 1500 to 1700° C. If the temperature is too low, polytypes other than 4H are likely to occur, and if the temperature is too high, surface roughness is likely to occur.

[0057] As described later, the higher the C / Si ratio of the source gas, the higher the in-plane uniformity of the doping concentration can be, but increasing the doping concentration lowers the effective C / Si. Here, the C / Si ratio is the molar ratio of C atoms in the C-based source gas to Si atoms in the Si-based source gas. Therefore, the C / Si of the source gas input must be further increased to compensate for this. In other words, if the average target doping concentration is 1×10 18 cm -3 ~5×10 18 cm -3 In this case, C / Si is preferably 1.1 to 1.5, more preferably 1.2 to 1.4. 18cm -3 ~1×10 19 cm -3 In this case, it is preferably 1.3 or more and 1.7 or less, and more preferably 1.4 or more and 1.6 or less.

[0058] When forming a high-concentration layer, it is necessary to dope a large amount of impurities to achieve a high concentration. In this case, the C / Si ratio is usually lowered. This is because if one tries to dope at a high concentration while keeping the C / Si ratio the same, it becomes necessary to introduce a large amount of doping gas. By lowering the C / Si ratio, it is possible to prevent the introduction of too much doping gas. In response to this, the inventors have conducted extensive research and found that lowering the C / Si ratio during deposition of the high-concentration layer leads to a decrease in the in-plane uniformity of the n-type doping concentration. It has also been found that the decrease in the in-plane uniformity of the n-type doping concentration can be suppressed by depositing the high-concentration layer using a C / Si ratio higher than usual. Furthermore, it has been found that it is important that the in-plane distribution of the C-based gas and the Si-based gas on the substrate surface is good before deposition of the high-concentration layer.

[0059] <<Confirmation of C-based gas supply and Si-based gas distribution on the substrate surface>> As described later, when forming a high-concentration layer, a high-concentration layer having a mirror-like surface is formed on a SiC single crystal substrate using a predetermined C / Si ratio higher than usual (SiC epitaxial layer formation), but a non-mirror region may occur in part of the surface. The non-mirror region is often in the range of 5% to 50% in size. By carefully studying the phenomenon of the non-mirror region occurrence, it was found that the cause is poor in-plane distribution of C-based gas supply and Si-based gas on the substrate surface, and it was found that in order to improve this, it is effective to configure the gas supply to the film formation device so that the in-plane distribution of C-based gas and Si-based gas supply can be controlled independently.

[0060] FIG. 3 shows an example of a configuration capable of independently controlling the in-plane distribution of the C-based gas supply and the Si-based gas supply. In a vertical film forming apparatus 30 as shown in Fig. 3, gas supply units 32a, 32b, and 32c are provided to supply gas from the top to the bottom of the apparatus toward the substrate 10 placed thereon, and the gas introduction unit includes a supply unit 32a for C-based gas only and a supply unit 32b for Si-based gas only, and the supply unit 32a for C-based gas only and the supply unit 32b for Si-based gas only are structured so that their positions can be adjusted independently in the horizontal direction (in-plane direction of the substrate). For example, a carrier gas is supplied to the gas supply unit 32c. In the example shown in Fig. 3, the C-based gas supply unit 32a includes gas supply pipes 32aa, 32ab, and 32ac, and the Si-based gas supply unit 32b includes gas supply pipes 32ba and 32bb.

[0061] Here, the in-plane distribution of the C-based gas supply on the substrate surface can be measured by measuring the film thickness distribution (growth rate distribution) of an epitaxial wafer formed under C supply rate-limiting conditions. In this specification, the in-plane distribution of the gas supply on the substrate surface refers to "the absolute value of the maximum film thickness of the epitaxial wafer - the minimum film thickness of the epitaxial wafer) / the average film thickness of the epitaxial wafer". The film thickness of the SiC epitaxial wafer can be measured by a known method, for example, FT-IR (Fourier Transform Infrared Spectroscopy). The C supply rate-limiting condition is a state in which the supply of the C-based gas is insufficient compared to the Si-based gas, and C / Si is preferably within the range of 0.6 to 0.9. However, in order to prevent the supply of the N-based gas from changing the effective C / Si, the carrier concentration of the epitaxial wafer is set to 1×10 17 cm -3 The film is formed under conditions where On the other hand, the in-plane distribution of the Si-based gas supply on the substrate surface can be measured by measuring the film thickness distribution (growth rate distribution) of an epitaxial wafer formed under Si supply rate-limiting conditions. The Si supply rate-limiting conditions are a state in which the supply of Si-based gas is insufficient compared to the C-based gas, and preferably the C / Si is in the range of 1.1 to 1.2. However, in order to prevent the supply of N-based gas from changing the effective C / Si, the carrier concentration of the epitaxial wafer is set to 1×10 17 cm -3 The film is formed under conditions where

[0062] In this way, the in-plane distribution of C-based gas supply (film thickness distribution of epitaxial wafers formed under C supply rate-limiting conditions) and the in-plane distribution of Si-based gas supply (film thickness distribution of epitaxial wafers formed under Si supply rate-limiting conditions) are measured, and if the in-plane distribution of C-based gas supply is not 10% or less and the in-plane distribution of Si-based gas is not 10% or less, the supply positions of the C-based gas supply unit and the Si-based gas supply unit are adjusted. For example, in a vertical film formation apparatus, if the supply of C-based gas is small at the center and large at the periphery, the position of the C-based gas supply unit in the gas introduction unit is moved to the center. When the in-plane distribution of the C-based gas supply on the substrate surface was 10% or less and the in-plane distribution of the Si-based gas supply on the substrate surface was 10% or less, no non-mirror region was generated.

[0063] Before preparing the samples described below, the supply positions of the C-based gas supply unit and the Si-based gas supply unit were adjusted, and the in-plane distribution of the C-rate-limited and Si-rate-limited growth rates was measured. (i)C rate limiting At C / Si=0.8, the doping concentration is 8×10 15 cm -3 The amount of nitrogen was adjusted so that the film was formed. The position of the supply portion for only the C-based gas was adjusted so that the in-plane distribution of the C-based gas supply was 5.6%. (ii) Silicon-Limited At C / Si=1.1, the doping concentration is 1.3×10 16 cm -3 The amount of nitrogen was adjusted so that the film was formed. The position of the supply part for only the Si-based gas was adjusted so that the in-plane distribution of the Si-based gas supply was 3.4%. The in-plane distribution of C-based gas supply and the in-plane distribution of Si-based gas supply were measured at 21 points on the wafer, in the same manner as the in-plane uniformity of the n-type doping concentration described below.

[0064] Table 1 shows the results of investigating the in-plane uniformity of the n-type doping concentration of the high-concentration epitaxial layer of a sample fabricated by using a 4H-SiC single crystal substrate with a diameter of 150 mm and a main surface with an off-angle of 4° and forming a high-concentration epitaxial layer on the Si surface using nitrogen as an n-type dopant. Specifically, the C / Si ratio of the source gas was set to one of 1.05, 1.15, and 1.35, 21 points on the wafer were measured, and the doping gas was introduced to aim for a specified value as the average doping concentration, and the in-plane uniformity of the n-type doping concentration was measured. The samples were fabricated under the same conditions except for the C / Si ratio and the flow rate of the doping gas. The measurement points, in mm, were the center of the wafer as (0,0) and the orientation flat in the Y direction, and were 21 points: (X,Y) = (0,70), (0,60), (0,45), (0,30), (0,15), (0,0), (0,-15), (0,-30), (0,-45), (0,-60), (0,-67), (-70,0), (-60,0), (-45,0), (-30,0), (-15,0), (15,0), (30,0), (45,0), (60,0), (70,0).

[0065] A higher doping concentration (average value) can be achieved by increasing the flow rate of the nitrogen introduced. z shows the relationship between the nitrogen flow rate introduced and the obtained doping concentration (average value) for the three samples shown in Table 1 when the C / Si ratio was set to 1.15. In Fig. 4, the horizontal axis indicates the doping concentration (average value) of 1.03 x 10 18 cm -3 The vertical axis indicates the obtained doping concentration (average value), and the vertical axis indicates the relative value of the input nitrogen flow rate when the input nitrogen flow rate at the time is set to 1.

[0066] [Table 1]

[0067] As shown in Table 1, when the commonly used C / Si ratio is 1.05, the average n-type doping concentration is 2 × 10 18 cm -3It can be seen that even at this level, the in-plane uniformity of the n-type doping concentration exceeds 30%, and that if the average value of the n-type doping concentration is increased, the in-plane uniformity of the n-type doping concentration further deteriorates. On the other hand, when the C / Si ratio is 1.15, which is higher than the usual ratio, the average n-type doping concentration is 1×10 18 cm -3 The uniformity of the n-type doping concentration was as high as 12.2% at about 100 nm, and the average n-type doping concentration was 2.6×10 18 cm -3 When the concentration is increased to about 17.8%, the uniformity of the n-type doping concentration on the surface is somewhat decreased, but remains good at 17.8%. Furthermore, when the C / Si ratio was 1.35, which is much higher than the commonly used ratio, the average n-type doping concentration was 2.2 × 10 18 cm -3 An extremely high in-plane uniformity of n-type doping concentration of 6.1% was obtained. From the above results, it was found that the in-plane uniformity of the n-type doping concentration is improved by using a C / Si ratio higher than the commonly used C / Si ratio of 1.05, and that the higher the C / Si ratio, the better the in-plane uniformity of the n-type doping concentration. It was also found that regardless of the C / Si ratio used, increasing the average value of the n-type doping concentration deteriorates the in-plane uniformity of the n-type doping concentration.

[0068] Figure 5 shows the results of investigating the relationship between the average n-type doping concentration and the growth rate for a sample of a 150 mm diameter SiC epitaxial wafer. The horizontal axis is the average n-type doping concentration, and the vertical axis is the average n-type doping concentration of 1×10 16 cm -3 This is the standard value of the growth rate when the growth rate in the case where epitaxial growth is performed so that

[0069] From Figure 5, we can see that the higher the average n-type doping concentration is, the slower the growth rate tends to be. This tendency is presumed to be due to nitrogen (N) inhibiting the growth of SiC (site-competition effect). This point will be explained below.

[0070] Figures 6(a)-(d) are conceptual diagrams of nitrogen (N) doping, where (b) shows a case where C / Si is low, (c) shows a case where C / Si is high, and (d) shows a case where the flow rate of the doping gas is high (aiming for a high average doping concentration) compared to (a). Figures 6(a)-(c) have been conventionally known as the site-competition effect, while Figure 6(d) is a new finding made by the present inventor.

[0071] As shown in FIG. 6(a), Si is incorporated into the Si site, C into the C site, and N into the C site. In contrast, in the case of low C / Si, as shown in FIG. 6(b), the ratio of N to C increases, and therefore the probability of N occupying the C site increases. Furthermore, in the case of a high C / Si ratio, as shown in FIG. 6(c), the ratio of N to C decreases, and the probability that N will occupy the C site decreases. In addition, when the N concentration is as high as that of the source gas, the ratio of N to C increases as shown in Fig. 6(d). Therefore, C is less likely to be incorporated into the C site, which results in a decrease in the effective C / Si and a decrease in the growth rate.

[0072] Here, the inventors investigated the relationship between the amount of C-based gas input and the growth rate in the CVD apparatus in which the above-mentioned samples were produced, and obtained a calibration curve showing a positive correlation in which the growth rate increases as the amount of C-based gas input increases. From this calibration curve and the amount of growth rate reduction caused by increasing the n-type doping concentration, the actual shortage of C-based gas input can be estimated. This is explained using the conceptual diagram in Figure 6. 7, the horizontal axis represents the amount of C-based gas input, and the vertical axis represents the growth rate of the SiC epitaxial layer. Note that the amount of C-based gas input corresponds to C / Si when the amount of Si-based gas input is fixed. When fabricating a SiC epitaxial wafer with an average n-type doping concentration higher than the average n-type doping concentration of the SiC epitaxial wafer obtained under the condition P1 on the calibration curve (C-based gas input amount C1, growth rate R1), if the growth rate is R3 when epitaxial growth is performed with a C-based gas input amount C1, then based on the calibration curve in FIG. 7, the effective C-based gas input amount is C3. In this case, the ratio of N atoms to C atoms has increased, making it difficult for C atoms to be incorporated into the C site, and as a result, the effective C-based gas input amount is C3. By inputting an amount of C-based gas that compensates for this shortage, it becomes possible to perform epitaxial growth at the growth rate R1.

[0073] As shown in Table 1, by increasing the C / Si to compensate for the shortage of C-based gas, the in-plane uniformity of the n-type doping concentration was significantly improved, reaching 6.1% at C / Si=1.35. C / Si=1.35, not high doping concentration, but normal concentration (e.g. 1×10 16 / cm 3 When a SiC epitaxial layer with a high doping concentration is formed, the surface becomes non-mirror, but the entire surface of this highly doped SiC epitaxial layer was mirror-like. This is thought to support the idea that nitrogen (N) atoms inhibit the growth of SiC through the site compensation effect described above. For the same C / Si, it is believed that a SiC epitaxial layer with a high doping concentration is more affected by the site compensation effect than a SiC epitaxial layer with a low doping concentration, which inhibits the uniform growth of SiC on the growth surface and impairs the in-plane uniformity of the doping concentration. Therefore, in a SiC epitaxial layer with a high doping concentration, it is believed that by increasing the C / Si, the influence of the site compensation effect can be reduced, inhibition of the uniform growth of SiC on the growth surface can be suppressed, and deterioration of the in-plane uniformity of the doping concentration can be suppressed.

[0074] In this way, the amount of inhibition of SiC epitaxial growth by the dopant can be estimated from the correlation (e.g., a calibration curve) between the amount of C-based gas input and the growth rate obtained in advance, and the source gas can be input with a C / Si ratio that compensates for that amount, thereby manufacturing SiC epitaxial wafers with high in-plane uniformity.

[0075] Based on the above experimental results and considerations, and taking into account the results shown in Table 1, Figure 8 is a graph showing the relationship between the average n-type doping concentration, the in-plane uniformity of the n-type doping concentration, and the C / Si ratio for a sample SiC epitaxial wafer with a diameter of 150 mm. Based on the relationship shown in Fig. 8, the growth conditions of C / Si and doping gas flow rate were estimated, and the average n-type doping concentration was 1 × 10 18 / cm 3 That's it, 1×10 19 / cm 3 It is possible to manufacture a SiC epitaxial wafer having a high concentration layer with an n-type doping concentration of 30% or less within the surface uniformity.

[0076] 8, in the graph where C / Si is 1.35, the data is only one point, but it was obtained based on the mechanism based on the site compensation effect described above. This will be explained. In Fig. 8, the graphs for C / Si of 1.05 and C / Si of 1.15 are obtained from data at three points. These graphs have a positive slope, and the slope of the graph for C / Si of 1.15 is lower than that of the graph for C / Si of 1.05, which indicates that the mechanism based on the site compensation effect is well applied. In addition, the doping concentration of 2 × 10 18 / cm 3 In the nearby SiC epitaxial layer, the in-plane uniformity of the doping concentration is best when C / Si is 1.35, followed by C / Si 1.15 and finally C / Si 1.05, which also indicates that the mechanism based on the site compensation effect applies well. Then, when C / Si is 1.35, 1×10 18 / cm 3 That's it, 1×10 19 / cm 3 It can be assumed that the mechanism based on the site compensation effect is well established within the doping concentration range below 1000 nm. If the mechanism based on the site compensation effect is well suited, the graph when C / Si is 1.35 will have a positive slope, and the slope will be lower than the slope of the graph when C / Si is 1.15. As described above, the graph when C / Si is 1.35 is obtained based on the mechanism based on the site compensation effect. [Explanation of symbols]

[0077] 10 SiC single crystal substrate 20 SiC epitaxial layer 21 High concentration layer 22 Drift Layer

Claims

1. A SiC single crystal substrate; On the SiC single crystal substrate, the average n-type doping concentration is 1.03 × 10 １８ / cm ３ That's it, 1 x 10 １９ / cm ３ and a high concentration layer having an in-plane uniformity of doping concentration of 30% or less, a SiC epitaxial wafer, wherein the high concentration layer is doped with nitrogen; Here, the in-plane uniformity is defined as the absolute value of (maximum value of doping concentration in the plane-minimum value of doping concentration in the plane) / average value of doping concentration in the plane.

2. A SiC single crystal substrate; On the SiC single crystal substrate, the average n-type doping concentration is 1.91 × 10 １８ / cm ３ That's it, 1 x 10 １９ / cm ３ and a high concentration layer having a doping concentration uniformity of 30% or less; a SiC epitaxial wafer, wherein the high concentration layer is doped with nitrogen; Here, the in-plane uniformity is defined as the absolute value of (maximum value of doping concentration in the plane-minimum value of doping concentration in the plane) / average value of doping concentration in the plane.

3. A SiC single crystal substrate; On the SiC single crystal substrate, the average n-type doping concentration is 2.20 × 10 １８ / cm ３ That's it, 1 x 10 １９ / cm ３ and a high concentration layer having a doping concentration uniformity of 30% or less; a SiC epitaxial wafer, wherein the high concentration layer is doped with nitrogen; Here, the in-plane uniformity is defined as the absolute value of (maximum value of doping concentration in the plane-minimum value of doping concentration in the plane) / average value of doping concentration in the plane.

4. A SiC single crystal substrate; On the SiC single crystal substrate, the average n-type doping concentration is 2.59 × 10 １８ / cm ３ That's it, 1 x 10 １９ / cm ３ and a high concentration layer having a doping concentration uniformity of 30% or less; a SiC epitaxial wafer, wherein the high concentration layer is doped with nitrogen; Here, the in-plane uniformity is defined as the absolute value of (maximum value of doping concentration in the plane-minimum value of doping concentration in the plane) / average value of doping concentration in the plane.

5. A SiC single crystal substrate; On the SiC single crystal substrate, the average n-type doping concentration is 1.03 × 10 １８ / cm ３ That's it, 2.59 x 10 １８ / cm ３ and a high concentration layer having a doping concentration uniformity of 30% or less; a SiC epitaxial wafer, wherein the high concentration layer is doped with nitrogen; Here, the in-plane uniformity is defined as the absolute value of (maximum value of doping concentration in the plane-minimum value of doping concentration in the plane) / average value of doping concentration in the plane.

6. A SiC single crystal substrate; On the SiC single crystal substrate, the average n-type doping concentration is 1.91 × 10 １８ / cm ３ That's it, 2.59 x 10 １８ / cm ３ and a high concentration layer having an in-plane uniformity of doping concentration of 30% or less, a SiC epitaxial wafer, wherein the high concentration layer is doped with nitrogen; Here, the in-plane uniformity is defined as the absolute value of (maximum value of doping concentration in the plane-minimum value of doping concentration in the plane) / average value of doping concentration in the plane.

7. A SiC single crystal substrate; On the SiC single crystal substrate, the average n-type doping concentration is 2.20 × 10 １８ / cm ３ That's it, 2.59 x 10 １８ / cm ３ and a high concentration layer having a doping concentration uniformity of 30% or less; a SiC epitaxial wafer, wherein the high concentration layer is doped with nitrogen; Here, the in-plane uniformity is defined as the absolute value of (maximum value of doping concentration in the plane-minimum value of doping concentration in the plane) / average value of doping concentration in the plane.

8. A SiC single crystal substrate; On the SiC single crystal substrate, the average n-type doping concentration is 1.03 × 10 １８ / cm ３ That's it, 2.20 x 10 １８ / cm ３ and a high concentration layer having a doping concentration uniformity of 30% or less; a SiC epitaxial wafer, wherein the high concentration layer is doped with nitrogen; Here, the in-plane uniformity is defined as the absolute value of (maximum value of doping concentration in the plane-minimum value of doping concentration in the plane) / average value of doping concentration in the plane.

9. A SiC single crystal substrate; On the SiC single crystal substrate, the average n-type doping concentration is 1.03 × 10 １８ / cm ３ That's it, 1.91 x 10 １８ / cm ３ and a high concentration layer having a doping concentration uniformity of 30% or less; a SiC epitaxial wafer, wherein the high concentration layer is doped with nitrogen; Here, the in-plane uniformity is defined as the absolute value of (maximum value of doping concentration in the plane-minimum value of doping concentration in the plane) / average value of doping concentration in the plane.

10. A SiC single crystal substrate; On the SiC single crystal substrate, the average n-type doping concentration is 1.91 × 10 １８ / cm ３ That's it, 2.20 x 10 １８ / cm ３ and a high concentration layer having a doping concentration uniformity of 30% or less; a SiC epitaxial wafer, wherein the high concentration layer is doped with nitrogen; Here, the in-plane uniformity is defined as the absolute value of (maximum value of doping concentration in the plane-minimum value of doping concentration in the plane) / average value of doping concentration in the plane.

11. The SiC epitaxial wafer according to any one of claims 1 to 10, wherein the in-plane uniformity of the doping concentration is 20% or less.

12. The SiC epitaxial wafer according to any one of claims 1 to 10, wherein the in-plane uniformity of the doping concentration is 10% or less.

13. The SiC epitaxial wafer according to any one of claims 1 to 10, wherein the high concentration layer is a buffer layer, and a drift layer having an average doping concentration lower than the average doping concentration of the buffer layer is provided on the buffer layer.

14. The SiC epitaxial wafer according to any one of claims 1 to 10, further comprising a buffer layer between the SiC single crystal substrate and the high concentration layer, the buffer layer converting basal plane dislocations into threading edge dislocations.

15. The SiC epitaxial wafer according to any one of claims 1 to 10, having a diameter of 150 mm or more.

16. The SiC epitaxial wafer of claim 11 having a diameter of 150 mm or more.

17. The SiC epitaxial wafer of claim 12 having a diameter of 150 mm or more.

18. The SiC epitaxial wafer of claim 13 having a diameter of 150 mm or more.

19. 15. The SiC epitaxial wafer of claim 14, having a diameter of 150 mm or more.

Images