Multilayer substrate

Abstract

To provide a multilayered substrate in which in-plane non-uniformity of an n-type carrier concentration resulting from in-plane non-uniformity of a Si concentration distribution in a GaN layer is suppressed.SOLUTION: A multilayered substrate includes an underlying substrate, and a gallium nitride layer epitaxially grown above the underlying substrate and comprising gallium nitride doped with silicon. The gallium nitride layer has a top surface with a radius of 50 mm or more, and the gallium nitride layer has a thickness of 4 μm or more. A silicon concentration on the top surface of the gallium nitride layer has a distribution in which an outer circumferential silicon concentration at a radial position 10 mm from an edge of the top surface is higher than a central silicon concentration at a center of the top surface, the central silicon concentration is 4×1015 cm-3 or more and less than 2×1016 cm-3, and a carbon concentration on the top surface of the gallium nitride layer has a distribution in which an outer circumferential carbon concentration at a radial position 10 mm from an edge of the top surface is higher than a central carbon concentration at a center of the top surface.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present invention relates to a laminated substrate.

Background Art

[0002] A laminated substrate in which a gallium nitride (GaN) layer is epitaxially grown on various base substrates is used as a material for manufacturing semiconductor devices (see, for example, Patent Document 1). In order to improve the production efficiency of semiconductor devices, the diameter of the laminated substrate is being increased. Silicon (Si) is used as an n-type impurity added to the GaN layer.

[0003] As will be described in detail later, the inventor of the present application studied a laminated substrate having a large diameter, a low added Si concentration, and a thick GaN layer, and found that in-plane non-uniformity of the Si concentration distribution in the GaN layer (specifically, non-uniformity in which the Si concentration increases on the outer peripheral side of the GaN layer) is likely to occur.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] One object of the present invention is to provide a laminated substrate in which in-plane non-uniformity of the n-type carrier concentration caused by in-plane non-uniformity of the Si concentration distribution in the GaN layer is suppressed.

Means for Solving the Problems

[0006] According to one aspect of the present invention a base substrate, a gallium nitride layer that is epitaxially grown above the base substrate and is composed of gallium nitride to which silicon is added, and The gallium nitride layer has an upper surface with a radius of 50 mm or more. The thickness of the gallium nitride layer is 4 μm or more. The silicon concentration on the upper surface of the gallium nitride layer has a distribution where the outer peripheral silicon concentration at a radial position 10 mm from the edge of the upper surface is higher than the central silicon concentration at the center of the upper surface. The aforementioned central silicon concentration is 4 × 10 15 cm -3 The above 2 x 10 16 cm -3 It is less than, The carbon concentration on the upper surface of the gallium nitride layer has a distribution where the outer peripheral carbon concentration at a radial position 10 mm from the edge of the upper surface is higher than the central carbon concentration at the center of the upper surface. Multilayer substrate It will be provided. [Effects of the Invention]

[0007] A laminated substrate is provided in which in-plane non-uniformity of n-type carrier concentration caused by in-plane non-uniformity of Si concentration distribution in the GaN layer is suppressed. [Brief explanation of the drawing]

[0008] [Figure 1] Figure 1 is a schematic cross-sectional view illustrating a laminated substrate according to an embodiment. [Figure 2] Figure 2 is a top view of the GaN layer, illustrating how to measure the Si and C concentrations in the GaN layer. [Figure 3] Figure 3 is a schematic side view showing a GaN layer manufacturing apparatus according to an embodiment. [Figure 4] Figure 4 is a graph showing an example of the C concentration distribution on the upper surface of the GaN layer according to the embodiment. [Figure 5] Figure 5 is a timing chart schematically showing the supply pattern of C-doping raw material gas at representative positions on the outer perimeter. [Figure 6] Figure 6 is a schematic cross-sectional view illustrating a GaN layer according to an embodiment. [Figure 7] Figure 7 is a schematic side view showing a GaN layer manufacturing apparatus in a comparative configuration. [Figure 8] Figure 8 is a graph showing an example of the Si concentration distribution on the upper surface of the GaN layer according to the embodiment or comparative configuration. [Modes for carrying out the invention]

[0009] A laminated substrate 100 according to an embodiment of the present invention will be described. Figure 1 is a schematic cross-sectional view illustrating the laminated substrate 100. The laminated substrate 100 comprises at least a base structure 10 and a gallium nitride (GaN) layer 20 formed on the base structure 10.

[0010] The substrate structure 10 includes at least a substrate which is a crystalline substrate for epitaxial growth of the GaN layer 20. The substrate structure 10 may optionally have a buffer layer between the substrate and the GaN layer 20. The structure of the buffer layer may be appropriately selected as needed.

[0011] As the base substrate, a GaN substrate, sapphire substrate, silicon (Si) substrate, silicon carbide (SiC) substrate, etc., may be used as appropriate. For example, if the base substrate is a GaN substrate, the base structure 10 may be a GaN substrate. Also, for example, if the base substrate is a Si substrate, the base structure 10 may have a Si substrate and an aluminum nitride (AlN) layer provided as a nucleation layer.

[0012] The laminated substrate 100 may have an upper layer 30 formed on the GaN layer 20, if necessary. The structure of the upper layer 30 may be appropriately selected depending on the operation of the semiconductor device to be fabricated using the laminated substrate 100. The upper layer 30 may be, for example, a group III nitride layer epitaxially grown on the GaN layer 20.

[0013] The GaN layer 20 according to the embodiment has the following characteristics. The GaN layer 20 has an upper surface 25 with a radius of 50 mm or more (a diameter of 100 mm or more). That is, the laminated substrate 100 has a circular shape with a diameter of 100 mm or more. Note that the laminated substrate 100 may have an orientation flat or a notch as necessary.

[0014] The GaN layer 20 contains silicon (Si) as an n-type impurity. The Si concentration on the upper surface 25 of the GaN layer 20 has a distribution in which the outer peripheral Si concentration CSi1 at the outer periphery of the upper surface 25 is higher than the central Si concentration CSi0 at the center of the upper surface 25.

[0015] The central Si concentration CSi0 is a low concentration at which concentration fluctuations on the order of 10 15 cm -3 have an impact. Specifically, it is 4×10 15 cm -3 or more and less than 2×10 16 cm -3 . In an aspect where the central Si concentration CSi0 is less than 1×10 16 cm -3 (that is, on the order of 10 15 cm -3 ), such an impact is more significant. The thickness of the GaN layer 20 is 4 μm or more. Also, the thickness of the GaN layer 20 is, for example, 40 μm or less.

[0016] The GaN layer 20 is a layer that functions as at least a part of an operating layer in a semiconductor device including the laminated substrate 100. Therefore, it is preferable that the GaN layer 20 has both breakdown voltage characteristics and resistance characteristics in a balanced manner. From the viewpoint of improving the breakdown voltage, the central Si concentration CSi0 of the GaN layer 20 is preferably less than 2×10 16 cm -3 . Also, from the viewpoint of improving the breakdown voltage, the thickness of the GaN layer 20 is preferably 4 μm or more. From the viewpoint of reducing resistance, the central Si concentration CSi0 of the GaN layer 20 is 4×10 15 cm -3The above is preferable. Furthermore, from the viewpoint of reducing resistance, the thickness of the GaN layer 20 is preferably 40 μm or less.

[0017] As will be explained in detail later, the peripheral Si contamination concentration (i.e., CSi1-CSi0), which is the excess of the peripheral Si concentration CSi1 from the central Si concentration CSi0, is the concentration of Si that has been mixed into the outer periphery of the GaN layer 20 due to the Si-containing deposition 230. The central Si concentration CSi0 is approximately equal to the target addition concentration, which is the Si concentration that is intentionally added.

[0018] The inclusion of Si results in an excess of n-type carrier concentration at the outer periphery compared to the center. The GaN layer 20 contains carbon (C) as an impurity to compensate for the excess n-type carrier concentration at the outer periphery. The C concentration at the top surface 25 of the GaN layer 20 has a higher distribution at the outer periphery C concentration CC1 compared to the central C concentration CC0 at the center of the top surface 25. The outer periphery C addition concentration (i.e., CC1-CC0), which is the excess of the outer periphery C concentration CC1 from the central C concentration CC0, is the concentration of C added to compensate for the excess n-type carrier concentration at the outer periphery.

[0019] In the following description, the radius of the GaN layer 20 is 50 mm, the thickness of the GaN layer 20 is 4 μm, and the target Si doping concentration in the GaN layer 20 is 9 × 10⁻¹⁶. 15 cm -3 Examples of such embodiments are given.

[0020] This section describes how to measure the Si and C concentrations in the GaN layer 20. To avoid complexity in the explanation, the term "concentration" may be used for common aspects of both Si and C concentrations. Figure 2 is a top view of the GaN layer 20, illustrating how to measure the concentrations. An XY Cartesian coordinate system is set up on the top surface 25 of the GaN layer 20, with the center of the top surface 25 as the origin.

[0021] First, let's explain the measurement positions within the plane. At least five measurement points are set on the top surface 25 to measure the concentration: a central measurement point and four outer circumference measurement points. The central measurement point is set at the center of the top surface 25. The four outer circumference measurement points are set at radial positions 10 mm from the edge of the top surface 25 for each of the X-positive, X-negative, Y-positive, and Y-negative directions (i.e., every 90° in the circumferential direction). Here, the radial positions 10 mm from the edge of the top surface 25 are set as representative positions on the outer circumference when defining the concentration on the outer circumference of the top surface 25.

[0022] The Si concentration at the central measurement point is the central Si concentration CSi0. The Si concentration obtained by averaging the Si concentrations at the outer circumference measurement points in the positive X direction (CSiX+), the outer circumference measurement points in the negative X direction (CSiX-), the outer circumference measurement points in the positive Y direction (CSiY+), and the outer circumference measurement points in the negative Y direction (CSiY-) is the outer circumference Si concentration CSi1. In other words, the outer circumference Si concentration CSi1 represents the circumferentially averaged Si concentration at the outer circumference. The excess of the outer circumference Si concentration CSi1 from the central Si concentration CSi0, i.e., CSi1-CSi0, is the outer circumference Si contamination concentration.

[0023] The C concentration at the central measurement point is the central C concentration CC0. The average C concentration obtained by averaging the C concentrations at the outer circumference measurement points in the positive X direction (CCX+), the outer circumference measurement points in the negative X direction (CCX-), the outer circumference measurement points in the positive Y direction (CCY+), and the outer circumference measurement points in the negative Y direction (CCY-) is the outer circumference C concentration CC1. In other words, the outer circumference C concentration CC1 represents the circumferentially averaged C concentration at the outer circumference. The excess of the outer circumference C concentration CC1 from the central C concentration CC0, i.e., CC1-CC0, is the outer circumference C added concentration.

[0024] In this example, the radius of the top surface 25 is 50 mm. The four outer circumference measurement points are set at radial positions 40 mm from the center in the positive X direction, negative X direction, positive Y direction, and negative Y direction, respectively.

[0025] In addition to the central measurement point and the four outer circumference measurement points, intermediate measurement points may be set at radial positions midway between the central measurement point and the outer circumference measurement points. In this example, intermediate measurement points are set at radial positions of 10 mm, 20 mm, and 30 mm from the center for each of the positive X, negative X, positive Y, and negative Y directions. In Figure 2, for ease of illustration, the central measurement point and outer circumference measurement points are shown as large black circles, and the intermediate measurement points are shown as small black circles.

[0026] Furthermore, even in embodiments where the radius of the upper surface 25 of the GaN layer 20 exceeds 50 mm, the four outer circumference measurement points are set at radial positions 10 mm from the edge of the upper surface 25. Also, even in embodiments where the radius of the upper surface 25 of the GaN layer 20 exceeds 50 mm, the intermediate measurement points may be set at radial positions 10 mm apart from the center in each of the positive X direction, negative X direction, positive Y direction, and negative Y direction.

[0027] The average concentration obtained by averaging the concentrations at four measurement points set at equal radial positions from the center of the upper surface 25 in the positive X direction, negative X direction, positive Y direction, and negative Y direction, that is, the circumferentially averaged concentration with respect to each radial position, is called the average Si concentration and average C concentration, respectively, for Si and C. The outer periphery Si concentration CSi1 and outer periphery C concentration CC1 are the average Si concentration and average C concentration at the outer periphery, respectively.

[0028] As described later, the GaN layer 20 is grown in a manner that ensures uniformity in the circumferential direction. However, in reality, variations in concentration occur even between four measurement points set at equal radial positions. To reduce the effects of such variations, the radial concentration distribution in the GaN layer 20 is evaluated using the average Si concentration, which is the circumferentially averaged Si concentration, and the average C concentration, which is the circumferentially averaged C concentration, for Si and C respectively. The concentration of added Si is the excess of the average Si concentration from the central Si concentration CSi0 at each radial position. The concentration of added C is the excess of the average C concentration from the central C concentration CC0 at each radial position.

[0029] Furthermore, when defining "the peripheral Si concentration CSi1 is higher than the central Si concentration CSi0," it is not simply that the peripheral Si concentration CSi1, that is, the average value of the Si concentrations at the four peripheral measurement points, is higher than the central Si concentration CSi0; rather, it is also assumed that the Si concentrations at at least three (preferably four) of the four peripheral measurement points are higher than the central Si concentration CSi0.

[0030] Furthermore, when defining "the peripheral C concentration CC1 is higher than the central C concentration CC0," it is not simply that the peripheral C concentration CC1, that is, the average value of the C concentrations at the four peripheral measurement points, is higher than the central C concentration CC0; it is also that the C concentrations at at least three (preferably four) of the four peripheral measurement points are higher than the central C concentration CC0.

[0031] Next, the measurement position in the height direction will be described. The concentration in the GaN layer 20 (Si concentration and C concentration, respectively) is measured at the top surface 25 by secondary ion mass spectrometry (SIMS). However, "concentration at the top surface 25" does not mean the concentration at the exact height of the top surface 25, but rather the concentration in the vicinity of the top surface 25, as will be explained below. In the embodiment where the upper layer 30 is not laminated, the top surface 25 is the uppermost surface of the laminated substrate 100, and in the embodiment where the upper layer 30 is laminated, it is the interface between the upper layer 30 and the GaN layer 20.

[0032] Generally, the reliability of SIMS measurements at the surface is not high. For this reason, it is preferable to perform SIMS measurements at a height slightly towards the substrate structure 10 from the top surface 25, that is, slightly inside the GaN layer 20. It is also preferable to use averaged measurements over a certain height range (depth range, thickness range). Here, for each measurement point in the plane, the average value of SIMS measurements in a depth range of 100 nm to 200 nm from the top surface 25 is defined as the concentration at the top surface 25. For example, in an embodiment where the thickness of the GaN layer 20 is 4 μm, the average value of the concentration obtained by SIMS measurements in the range of 3.8 μm to 3.9 μm from the bottom surface of the GaN layer 20 is defined as the concentration at the top surface 25.

[0033] Furthermore, the concentration at a height midway between the upper surface 25 of the GaN layer 20 and the lower surface of the GaN layer 20 (the concentration at a predetermined depth position from the upper surface 25 of the GaN layer 20) is defined in accordance with the concentration at the upper surface 25. Specifically, the concentration at the predetermined depth position is defined as the average value of SIMS measurements in a depth range of 100 nm to 200 nm from that predetermined depth position.

[0034] Before further describing the embodiments, a comparative configuration will be explained. Figure 7 is a schematic side view showing a GaN layer 20 manufacturing apparatus 200 according to the comparative configuration. The manufacturing apparatus 200 is an organometallic vapor deposition (MOVPE) apparatus and has a susceptor 210 on which a wafer 40 is placed. The wafer 40 is the object to be processed (an intermediate of the laminated substrate 100) at the stage in which the GaN layer 20 growth process is carried out by the manufacturing apparatus 200. Specifically, at the start of GaN layer 20 growth, it is the base structure 10, and after the start of GaN layer 20 growth, it is a laminate of the base structure 10 and the GaN layer 20.

[0035] The manufacturing apparatus 200 is a lateral flow type MOVPE apparatus in which the raw material gas 220 is supplied in a direction parallel to the upper surface of the wafer 40. The raw material gas 220 includes a GaN raw material gas for growing GaN constituting the GaN layer 20, and a Si doping raw material gas for supplying Si to be added to the GaN layer 20. The GaN raw material gas includes a gallium (Ga) raw material gas and a nitrogen (N) raw material gas. For example, trimethylgallium (TMG) gas is used as the Ga raw material gas, and for example, ammonia (NH3) gas is used as the N raw material gas. For example, silane (SiH4) gas is used as the Si doping raw material gas.

[0036] The growth temperature can be selected, for example, in the range of 900°C to 1400°C, and the V / III ratio, which is the flow rate ratio of Group V raw material gas to Group III raw material gas, can be selected, for example, in the range of 10 to 5000. The target doping concentration of the Si doping raw material gas in the GaN layer 20 is 4 × 10⁻¹⁰. 15 cm -3 The above 2 x 10 16 cm -3 A predetermined value within a range less than 9 × 10 15 cm -3 It is supplied in such a manner.

[0037] The susceptor 210 holds the wafer 40 so as to rotate the wafer 40 with its center as the center of rotation. By growing the GaN layer 20 while performing this rotation, the growth conditions of the GaN layer 20 can be made uniform in the circumferential direction of the wafer 40.

[0038] The inner peripheral edge 211 of the susceptor is the position on the susceptor 210 where the edge of the wafer 40 is located, in a plan view. The upper surface 213 of the susceptor edge is the upper surface of the susceptor 210 between the inner peripheral edge 211 and the outer peripheral edge 212 of the susceptor.

[0039] During the growth process of the GaN layer 20, a Si-containing deposition 230 adheres to the upper surface 213 of the susceptor edge, causing Si to be mixed into the GaN layer 20 from the Si-containing deposition 230. The inclusion of Si in the GaN layer 20 due to the Si-containing deposition 230 is called "improper inclusion," and is distinguished from "addition," which is the intentional inclusion of Si in the GaN layer 20 by using a Si-doping raw material gas. The Si concentration on the upper surface 25 of the GaN layer 20 is the sum of the concentration of Si added by the Si-doping raw material gas and the concentration of Si mixed in due to the Si-containing deposition 230.

[0040] The concentration of Si added by the Si doping source gas is, schematically speaking, constant across the entire upper surface 25 of the GaN layer 20. The concentration of Si introduced due to Si-containing deposition 230 is higher closer to the inner peripheral edge 211 of the susceptor, that is, closer to the outer edge of the wafer 40.

[0041] Figure 8 is a graph showing an example of the Si concentration distribution on the upper surface 25 of the GaN layer 20 under comparative configuration. The horizontal axis shows the in-plane measurement position in mm, and the vertical axis shows the Si concentration in 10°C. 15 cm -3 The values ​​are shown in units. The Si concentration at each measurement point in the X direction is shown by a black circle, and the Si concentration at each measurement point in the Y direction is shown by a white square. The average Si concentration is shown by an "X".

[0042] Here, we will describe a comparative configuration in which carbon (C) is not added, as described in the embodiments below. Therefore, except for the absence of carbon addition, the comparative configuration is the same as the embodiments, and the Si concentration distribution shown in Figure 8 is the same for both the comparative configuration and the embodiments.

[0043] The central Si concentration CSi0 is the target addition concentration of 9 × 10⁻⁶. 15 cm -3 It is approximately equal to 9.01 × 10 15 cm -3 Furthermore, the Si concentration remains almost constant up to about 10 mm from the center. From this, it can be said that the contaminated Si does not reach the center of wafer 40.

[0044] At radial positions approximately 20 mm or more away from the center, the concentration of contaminated Si, which is the excess from the central Si concentration CSi0, is significantly higher. Furthermore, the concentration of contaminated Si increases towards the outer periphery. The outer periphery Si concentration CSi1 is 10.65 × 10⁻⁶ 15 cm -3 Therefore, the peripheral Si contamination concentration, which is the excess of the peripheral Si concentration CSi1 from the central Si concentration CSi0, is 1.64 × 10⁻⁶. 15 cm -3 That is the case.

[0045] The Si resulting from the Si-containing deposition 230 is mixed in radially from the edge of the wafer 40 to a width of approximately 40 mm. In a large-diameter wafer 40, such as the one in this example with a radius of 50 mm, Si mixing does not occur in the center. In contrast, if a smaller-diameter wafer, such as one with a radius of 25 mm, is used, Si mixing will occur even in the center of the wafer.

[0046] In small-diameter configurations where Si inclusion occurs at the center, the Si inclusions from both edges on either side of the center are superimposed at the center, resulting in a higher Si concentration at the center due to the inclusions. As a result, the Si concentration distribution due to the inclusions becomes more uniform across the plane.

[0047] In contrast, in the case of a large-diameter wafer, as in this example, where Si contamination does not occur in the center, the Si concentration distribution due to the contaminated Si becomes non-uniform, with a large difference between the center and the outer edge. Thus, in large-diameter wafers, specifically those with a radius of 50 mm or more, the in-plane non-uniformity of the Si concentration distribution due to the contaminated Si becomes pronounced.

[0048] Although the example shows a wafer with a radius of 50 mm, even for wafers 40 with a radius greater than 50 mm, the concentration of Si at the outer edge (10 mm radially from the edge of the wafer 40) is about the same as in the case of a wafer with a radius of 50 mm.

[0049] The Si concentration at the outer edge, which is the difference in Si concentration between the center and the outer edge, is 1.64 × 10⁻⁶ 15 cm-3 On the other hand, the target Si addition concentration, that is, the central Si concentration CSi0, is 10 15 cm -3 The concentration is so low that even a concentration fluctuation of the order of magnitude can have an effect. Therefore, the influence of the Si contamination concentration on the outer periphery on the in-plane Si concentration distribution becomes significant.

[0050] Thus, the radius of the upper surface 25 is 50 mm or more, and the central Si concentration CSi0 is 4 × 10 15 cm -3 The above 2 x 10 16 cm -3 Less than (and even 1 x 10) 16 cm -3 In a GaN layer 20 (less than ), a significant difference in Si concentration between the center and the outer periphery occurs due to the inclusion of Si. In other words, a significant difference in n-type carrier concentration between the center and the outer periphery occurs due to this difference in Si concentration. As described below, the embodiment proposes a technique that can suppress such a difference in n-type carrier concentration between the center and the outer periphery compared to the comparative form.

[0051] The embodiment will be described further. As mentioned above, Figure 8 is also a graph showing the Si concentration distribution on the upper surface 25 of the GaN layer 20 according to the embodiment. The Si concentration distribution will be described further.

[0052] The lower limit of the target additive concentration, that is, the lower limit of the central Si concentration CSi0, is 4 × 10⁻⁶. 15 cm -3 A Si concentration of 0.4 × 10⁻⁶ is 10% higher than the limit. 15 cm -3 This is defined as the contamination threshold value ΔSi. The region where the concentration of contaminated Si is less than or equal to the contamination threshold value ΔSi is called the low-concentration contaminated Si region 23. The region where the concentration of contaminated Si exceeds the contamination threshold value ΔSi is called the high-concentration contaminated Si region 24.

[0053] The high-concentration Si region 24 is an annular region on the edge of the upper surface 25 of the GaN layer 20. The entire circular region inside the annular region that is the high-concentration Si region 24 becomes the low-concentration Si region 23. The width of the annular region that is the high-concentration Si region 24 from the edge of the upper surface 25 is approximately 30 mm.

[0054] The Si-high concentration region 24 is a region where the excess n-type carrier concentration due to the Si contamination is large. In this embodiment, carbon is added to the Si-high concentration region 24 to compensate for at least a portion of the excess n-type carrier concentration, thereby suppressing the difference in n-type carrier concentration between the center and the outer periphery.

[0055] Figure 3 is a schematic side view showing a manufacturing apparatus 200 for the GaN layer 20 according to an embodiment. The manufacturing apparatus 200 according to this embodiment is a lateral flow type MOVPE apparatus, similar to the comparative embodiment, and the main raw material gases 220 (GaN raw material gas and Si doping raw material gas) are supplied in a direction parallel to the upper surface of the wafer 40.

[0056] The manufacturing apparatus 200 according to this embodiment further includes a mechanism for supplying a compensating raw material gas 240. The compensating raw material gas 240 includes a carbon-doping raw material gas for supplying carbon to be added to the GaN layer 20. For example, methane (CH4) gas is used as the carbon-doping raw material gas.

[0057] Furthermore, since the Ga raw material gas (e.g., TMG gas) contained in the main raw material gas 220 is an organic raw material gas, carbon (C) originating from the Ga raw material gas will also be contained in the GaN layer 20. The inclusion of C in the GaN layer 20 due to the Ga raw material gas is called "contamination," and is distinguished from "addition," which is the intentional inclusion of C in the GaN layer 20 by using a C-doping raw material gas. The C concentration on the upper surface of the GaN layer 20 is the sum of the C concentration added by the C-doping raw material gas and the C concentration contaminated due to the Ga raw material gas.

[0058] The concentration of carbon (C) introduced due to the Ga raw material gas is, schematically speaking, constant across the entire upper surface 25 of the GaN layer 20. The compensating raw material gas 240, i.e., the C doping raw material gas, is selectively supplied to the high-concentration Si region 24.

[0059] In this embodiment, in order to preferentially add C to the high-concentration Si region 24, the C-doping raw material gas is selectively supplied to the high-concentration Si region 24, that is, to a predetermined range on the outer periphery with respect to the radial direction (specifically, in this example, a range with a width of approximately 30 mm from the edge of the top surface 25). In addition, since the Si concentration is slightly higher in the outer periphery portion of the low-concentration Si region 23, the C-doping raw material gas may also be supplied to this portion as an auxiliary measure.

[0060] The C-doping raw material gas is supplied from a direction perpendicular to the upper surface of the wafer 40. Specifically, the C-doping raw material gas is supplied from a direction perpendicular to the high-concentration Si region 24 downstream of the main raw material gas 220 with respect to the rotation center of the wafer 40. This enables selective supply to the high-concentration Si region 24.

[0061] In the high-concentration Si region 24, the concentration of Si is higher towards the outer periphery. Therefore, it is preferable that the C-doping raw material gas be supplied to multiple radial positions in the high-concentration Si region 24, and that the amount of C-doping raw material gas supplied to these multiple radial positions be controlled to be higher towards the outer periphery.

[0062] The growth conditions for the GaN layer 20 in this embodiment are the same as those in the comparative embodiment, except for the supply conditions for the carbon-doping raw material gas. The supply conditions for the carbon-doping raw material gas (supply position, supply amount, supply timing, etc.) may be set appropriately through experimental studies so as to compensate for any excess n-type carrier concentration caused by the mixed-in Si. A rough guideline is to set the supply conditions for the carbon-doping raw material gas so that the concentration of mixed-in Si and the concentration of added carbon are equal at each position in the radial direction. However, since it is not easy to determine how much of the added carbon compensates for the n-type carriers, the concentration of added carbon may be adjusted as appropriate.

[0063] Furthermore, adding C compensates for at least a portion of the excess n-type carrier concentration compared to the comparative configuration without C. Therefore, it is not essential to add C to the extent that the excess n-type carrier concentration is completely compensated.

[0064] For example, it is not essential that the outer perimeter carbon (C) concentration be as high as (or higher than) the outer perimeter silicon (Si) concentration. However, it is preferable that the outer perimeter carbon (C) concentration be 0.5 times or more the outer perimeter silicon concentration. Furthermore, to prevent the outer perimeter carbon (C) concentration from becoming excessive, it is preferable that the outer perimeter carbon (C) concentration be 2 times or less the outer perimeter silicon concentration.

[0065] The concentration of carbon (C) introduced due to the Ga raw material gas, i.e., the central C concentration CC0, is, for example, 2 × 10⁻⁶ as described later. 15 cm -3 While this is a general guideline, other values ​​are also acceptable. However, it is preferable to keep the central C concentration CC0 as low as possible so that it is below the lower limit of the central Si concentration CSi0. The concentration of added C is the excess from the central C concentration CC0, and this excess C concentration contributes to compensating for the excess n-type carrier concentration in the embodiment.

[0066] Figure 4 is a graph showing an example of the C concentration distribution on the upper surface 25 of the GaN layer 20 according to the embodiment. The horizontal axis shows the in-plane measurement position in mm, and the vertical axis shows the C concentration in 10 degrees. 15 cm -3 The values ​​are shown in units. The carbon concentration at each measurement point in the X direction is shown by a black circle, and the carbon concentration at each measurement point in the Y direction is shown by a white square. The average carbon concentration is shown by an "X".

[0067] The central C concentration CC0 is 2.01 × 10⁻⁶. 15 cm -3 In the region 24 with a high concentration of mixed Si, the C concentration is higher than the central C concentration CC0, meaning that C is added. Corresponding to the higher concentration of mixed Si towards the outer edge, C is added so that the concentration of added C is higher towards the outer edge. The outer edge C concentration CC1 is 3.26 × 10⁻⁶ 15 cm -3 Therefore, the peripheral carbon concentration, which is the excess of the peripheral carbon concentration CC1 from the central carbon concentration CC0, is 1.25 × 10⁻⁶. 15 cm -3 The concentration of carbon added to the outer periphery is 1.25 × 10⁻⁶. 15 cm -3 The outer Si content concentration is 1.64 × 10⁻⁶. 15 cm -3 It is 0.76 times that amount.

[0068] The following describes the characteristics of the C doping in the height direction (depth direction, thickness direction) according to the embodiment. Generally, the concentration of mixed Si in the GaN layer 20 increases as the GaN layer 20 becomes thicker. This is because the amount of Si-containing deposition 230 adhering to the upper surface 213 of the susceptor edge increases as the growth time of the GaN layer 20 increases. Therefore, the supply of the C doping raw material gas may be started at the timing when the growth thickness reaches a level where the mixed Si concentration exceeds a predetermined level.

[0069] Figure 5 is a timing chart schematically showing the supply pattern of the C-doping raw material gas at a representative position on the outer periphery (10 mm radially from the edge). The horizontal axis represents the growth thickness of the GaN layer 20, and the vertical axis represents the supply amount of the C-doping raw material gas. The Si contamination concentration on the outer periphery is a predetermined concentration, for example, the above-mentioned contamination standard value ΔSi(0.4 × 10) 15 cm -3 The supply of C-doping raw material gas is started when the growth thickness T1 reaches the target thickness. Subsequently, the Si content on the outer edge increases until the final growth thickness T2 of the GaN layer 20 is reached. For this reason, it is preferable to increase the supply amount of C-doping raw material gas as the growth thickness increases, in accordance with the increase in the Si content on the outer edge.

[0070] Furthermore, the amount of Si mixed in is less towards the center than at the representative position on the outer circumference. For this reason, the start timing of supplying the C-doping raw material gas may be delayed at the center compared to the representative position on the outer circumference. Also, the amount of Si mixed in is more at the outer edge than at the representative position on the outer circumference. For this reason, the start timing of supplying the C-doping raw material gas may be advanced at the outer edge compared to the representative position on the outer circumference. At the center or outer edge than the representative position on the outer circumference, the amount of C-doping raw material gas supplied may be increased as the growth thickness increases, in accordance with the increasing concentration of Si mixed in.

[0071] Figure 6 is a schematic cross-sectional view illustrating a GaN layer 20 according to an embodiment. The region where C-doping raw material gas is not supplied is the C-free region 21, and the region where C-doping raw material gas is supplied is the C-doped region 22. Hereinafter, a representative position on the outer periphery (a radial position 10 mm from the edge) will be used as an example. The range of thickness T2 downward from the upper surface 25 of the GaN layer 20 is the C-doped region 22, and the range of thickness T1 downward from the lower end of the C-doped region 22 is the C-free region 21.

[0072] In the C-doped region 22, the C concentration is highest at the top surface 25, and decreases downwards from the top surface 25. Also, the C concentration in the non-doped region 21 is lower than the C concentration at the top surface 25. In other words, in the GaN layer 20, at a representative position on the outer periphery (10 mm radially from the edge), there exists a region at a predetermined depth from the top surface 25 where the C concentration is lower than the outer periphery C concentration at the top surface 25.

[0073] Furthermore, as mentioned above, the concentration of Si in the GaN layer 20 generally increases as the GaN layer 20 becomes thicker. In other words, the concentration of Si in the GaN layer 20 generally decreases towards the lower side of the GaN layer 20. That is, in the GaN layer 20, at a representative position on the outer periphery (10 mm radially from the edge), there exists a region at a predetermined depth from the top surface 25 where the Si concentration is lower than that of the outer periphery Si concentration on the top surface 25.

[0074] The characteristics of the GaN layer 20 according to this embodiment will be described below. The concentration of mixed Si decreases schematically monotonically from the outer periphery to the center of the GaN layer 20. Therefore, the GaN layer 20 of this embodiment has the characteristic that the intermediate Si concentration CSi2 at a radial position midway between the center and the outer periphery is higher than the central Si concentration CSi0 and lower than the outer periphery Si concentration CSi1.

[0075] The intermediate Si concentration CSi2 is defined, for example, as the average Si concentration at a radial position 30 mm from the center. In the example shown in Figure 8, the intermediate Si concentration CSi2 (9.94 × 10) 15 cm -3 ) is the central Si concentration CSi0(9.01 × 10⁻¹⁰ 15 cm -3 ) is higher than the peripheral Si concentration CSi1(10.65 × 10 15 cm -3 It is lower than ).

[0076] In accordance with the concentration distribution shape of the mixed Si, the concentration of added C is preferably monotonically decreased from the outer periphery to the center of the GaN layer 20. In other words, the GaN layer 20 of the embodiment preferably has the characteristic that the intermediate C concentration CC2 at a radial position midway between the center and the outer periphery is higher than the central C concentration CC0 and lower than the outer periphery C concentration CC1.

[0077] The intermediate C concentration CC2 is defined, for example, as the average C concentration at a radial position 30 mm from the center. In the example shown in Figure 4, the intermediate C concentration CC2 is 2.73 × 10⁻¹⁰. 15 cm -3 ) is the central C concentration CC0 (2.01 × 10⁻¹). 15 cm -3 Higher than ), peripheral C concentration CC1 (3.26 × 10 15 cm -3 It is lower than ).

[0078] <Preferred Embodiments of the Invention> Preferred embodiments of the present invention are described below.

[0079] (Note 1) Substrate and A gallium nitride layer, epitaxially grown above the aforementioned substrate and composed of silicon-doped gallium nitride, Equipped with, The gallium nitride layer has an upper surface with a radius of 50 mm or more. The thickness of the gallium nitride layer is 4 μm or more. The silicon concentration on the upper surface of the gallium nitride layer has a distribution where the outer peripheral silicon concentration at a radial position 10 mm from the edge of the upper surface is higher than the central silicon concentration at the center of the upper surface. The aforementioned central silicon concentration is 4 × 10 15 cm -3 The above 2 x 10 16 cm -3 It is less than, The carbon concentration on the upper surface of the gallium nitride layer has a distribution where the outer peripheral carbon concentration at a radial position 10 mm from the edge of the upper surface is higher than the central carbon concentration at the center of the upper surface. Multilayer substrate.

[0080] (Note 2) In the gallium nitride layer, at a radial position 10 mm from the edge of the upper surface, and at a predetermined depth from the upper surface, there exists a region with a lower carbon concentration than the outer peripheral carbon concentration on the upper surface. The laminated substrate described in Appendix 1.

[0081] (Note 3) In the gallium nitride layer, at a radial position 10 mm from the edge of the upper surface, and at a predetermined depth from the upper surface, there exists a region with a lower silicon concentration than the silicon concentration at the outer periphery of the upper surface. A laminated substrate as described in Appendix 1 or 2.

[0082] (Note 4) The silicon concentration on the upper surface of the gallium nitride layer has a distribution in which the intermediate silicon concentration at a radial position 30 mm from the center of the upper surface is higher than the central silicon concentration and lower than the outer peripheral silicon concentration. A multilayer substrate as described in any one of the appendices 1 to 3.

[0083] (Note 5) The carbon concentration on the upper surface of the gallium nitride layer has a distribution in which the intermediate carbon concentration at a radial position 30 mm from the center of the upper surface is higher than the central carbon concentration and lower than the outer peripheral carbon concentration. A laminated substrate as described in any one of the appendices 1 to 4. [Explanation of symbols]

[0084] 10…Underlayment structure, 20…GaN layer, 21…C-free region, 22…C-doped region, 23…Low Si concentration region, 24…High Si concentration region, 25…(GaN layer) top surface, 30…Upper layer, 40…Wafer, 100…Laminated substrate, 200…Manufacturing equipment, 210…Susceptor, 211…Inner edge of susceptor, 212…Outer edge of susceptor, 213…Upper surface of susceptor edge, 220…(Main) raw material gas, 230…Si-containing deposition, 240…Compensation raw material gas

Claims

1. Substrate and A gallium nitride layer, epitaxially grown above the aforementioned substrate and composed of silicon-doped gallium nitride, Equipped with, The gallium nitride layer has an upper surface with a radius of 50 mm or more. The thickness of the gallium nitride layer is 4 μm or more. The silicon concentration on the upper surface of the gallium nitride layer has a distribution in which the outer peripheral silicon concentration at a radial position 10 mm from the edge of the upper surface is higher than the central silicon concentration at the center of the upper surface. The central silicon concentration is 4 × 10 １５ cm －３ The above 2 x 10 １６ cm －３ It is less than, The carbon concentration on the upper surface of the gallium nitride layer has a distribution where the outer peripheral carbon concentration at a radial position 10 mm from the edge of the upper surface is higher than the central carbon concentration at the center of the upper surface. Multilayer substrate.

2. In the gallium nitride layer, at a radial position 10 mm from the edge of the upper surface, and at a predetermined depth from the upper surface, there exists a region with a lower carbon concentration than the outer peripheral carbon concentration on the upper surface. The laminated substrate according to claim 1.

3. In the gallium nitride layer, at a radial position 10 mm from the edge of the upper surface, and at a predetermined depth from the upper surface, there exists a region with a lower silicon concentration than the silicon concentration at the outer periphery of the upper surface. A laminated substrate according to claim 1 or 2.

4. The silicon concentration on the upper surface of the gallium nitride layer has a distribution in which the intermediate silicon concentration at a radial position 30 mm from the center of the upper surface is higher than the central silicon concentration and lower than the outer peripheral silicon concentration. A laminated substrate according to claim 1 or 2.

5. The carbon concentration on the upper surface of the gallium nitride layer has a distribution in which the intermediate carbon concentration at a radial position 30 mm from the center of the upper surface is higher than the central carbon concentration and lower than the outer peripheral carbon concentration. A laminated substrate according to claim 1 or 2.

Images