Nitride semiconductor device

By incorporating a carbon layer with a graphene layer between a silicon carbide layer and nitride semiconductor layer, the substrate diameter of nitride semiconductor devices can be increased, improving reliability and reducing costs while enhancing performance.

JP2024166957AInactive Publication Date: 2024-11-29PANASONIC HOLDINGS CORP
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
JP2023083402
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-19
Publication Date
2024-11-29
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Conventional nitride semiconductor devices face limitations in increasing the diameter of the substrate, which affects reliability and cost due to the use of AlN buffer layers that are not suitable for vertical transistors and lead to cracking and high resistance.

Method used

The use of a carbon layer, including a graphene layer, between a silicon carbide layer and a nitride semiconductor layer, which weakens the bond and allows for the formation of a nitride semiconductor device with a larger substrate diameter, using Si or SiC substrates, and includes features like openings to enhance adhesion and reduce dislocations.

Benefits of technology

This configuration enables the use of larger substrates, improves crystallinity, reduces manufacturing costs, and enhances the reliability and performance of nitride semiconductor devices by suppressing cracking and leakage currents.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024166957000001_ABST
    Figure 2024166957000001_ABST
Patent Text Reader

Abstract

To provide a nitride semiconductor device capable of achieving increase in a diameter of a substrate.SOLUTION: A nitride semiconductor device 1 includes: a substrate 10; a drift layer 14; a first foundation layer 16; a second foundation layer 18; a gate opening 20 that penetrates through the second foundation layer 18 and the first foundation layer 16 to reach the drift layer 14; an electron transit layer 22 and an electron supply layer 24 that are provided, in order from the bottom, so as to cover the top surface of the second foundation layer 18 and the gate opening 20; a gate electrode 32 that is provided above the electron supply layer 24 so as to cover the gate opening 20; a source electrode 34 that is provided at a position away from the gate electrode 32; and a drain electrode 36 that is provided below the substrate 10. The substrate 10 includes a silicon carbide layer 12 and a carbon layer 13 provided above the silicon carbide layer 12, the carbon layer 13 including at least one of a graphene layer and a graphene buffer layer.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to nitride semiconductor devices. [Background technology]

[0002] Nitride semiconductors, such as GaN, are wide-gap semiconductors with a large band gap, and have the advantage of a large dielectric breakdown field and a high electron saturation drift velocity compared to compound semiconductors such as GaAs or Si semiconductors. For example, the band gaps of GaN and AlN are 3.4 eV and 6.2 eV, respectively, at room temperature. For this reason, research and development of power transistors using nitride semiconductors, which are advantageous for achieving high output and / or high voltage resistance, is currently being actively conducted.

[0003] In addition, in the AlGaN / GaN heterostructure, spontaneous and piezoelectric polarization on the (0001) surface generate a high concentration of two-dimensional electron gas (2DEG) at the heterointerface, and the density of 2DEG is as high as 1×10 13 cm -2 This has the advantage that a sheet carrier concentration of the above can be obtained. For example, Patent Document 1 discloses a GaN-based vertical transistor in which a channel made of two-dimensional electron gas generated at an AlGaN / GaN heterointerface can be opened and closed by a gate voltage. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 6511645 [Patent Document 2] JP 2019-102767 A [Non-patent literature]

[0005] [Non-Patent Document 1] Ponce et al., “Microstructure of GaN epitaxy on SiC using AlN buffer layers”, Appl. Phys. Lett., 67, 410-412, (1995) Summary of the Invention [Problem to be solved by the invention]

[0006] For conventional nitride semiconductor devices, there is room for improvement in increasing the diameter of the substrate.

[0007] In view of this, the present disclosure provides a nitride semiconductor device that enables a substrate with a large diameter to be used. [Means for solving the problem]

[0008] a first nitride semiconductor layer provided above the substrate; a first p-type nitride semiconductor layer provided above the first nitride semiconductor layer; a second nitride semiconductor layer provided above the first p-type nitride semiconductor layer; a first opening penetrating the second nitride semiconductor layer and the first p-type nitride semiconductor layer and reaching the first nitride semiconductor layer; an electron transit layer and an electron supply layer provided in this order from below to cover an upper surface of the second nitride semiconductor layer and the first opening; a gate electrode provided above the electron supply layer to cover the first opening; a source electrode provided at a position spaced from the gate electrode; and a drain electrode provided below the substrate, wherein the substrate includes a silicon carbide layer and a carbon layer provided above the silicon carbide layer, and the carbon layer includes at least one of a graphene layer and a graphene buffer layer.

[0009] A nitride semiconductor device according to another embodiment of the present disclosure includes a substrate, a first silicon carbide layer provided above the substrate, a p-type silicon carbide layer provided above the first silicon carbide layer so as to expose at least a portion of an upper surface of the first silicon carbide layer, and at least one of a graphene layer or a graphene buffer layer, and includes a carbon layer provided above the p-type silicon carbide layer, an electron transit layer provided above the first silicon carbide layer and the carbon layer, an electron supply layer provided above the electron transit layer, a gate electrode provided above the electron supply layer, a source electrode provided at a position spaced from the gate electrode, and a drain electrode provided below the substrate, and the carbon layer has an opening exposing a portion of the first silicon carbide layer. Effect of the Invention

[0010] According to the present disclosure, it is possible to provide a nitride semiconductor device that enables a substrate with a large diameter to be formed. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a cross-sectional view of a nitride semiconductor device according to the first embodiment. [Diagram 2] FIG. 2 shows a graphene buffer layer and an epitaxial graphene layer on SiC. [Diagram 3] FIG. 3 is a diagram showing the relationship between the number of epitaxial graphene layers and the crystallinity of the GaN film. [Figure 4] FIG. 4 is a cross-sectional view of a nitride semiconductor device according to the second embodiment. [Figure 5A] FIG. 5A is a plan view showing an example of the shape of openings in a carbon layer. [Figure 5B] FIG. 5B is a plan view showing another example of the shape of the openings in the carbon layer. [Figure 6] FIG. 6 is a cross-sectional view of a nitride semiconductor device according to a modification of the second embodiment. [Figure 7A] FIG. 7A is a plan view showing an example of a nitride semiconductor device according to a modification of the second embodiment. FIG. [Figure 7B] FIG. 7B is a plan view of a carbon layer included in the nitride semiconductor device shown in FIG. 7A. [Figure 8A] FIG. 8A is a plan view showing another example of the nitride semiconductor device according to the second embodiment. [Figure 8B] FIG. 8B is a plan view of a carbon layer included in the nitride semiconductor device shown in FIG. 8A. [Figure 9] FIG. 9 is a cross-sectional view of a nitride semiconductor device according to the third embodiment. [Figure 10] FIG. 10 is a cross-sectional view of a nitride semiconductor device according to a modification of the third embodiment. [Figure 11] FIG. 11 is a cross-sectional view of a nitride semiconductor device according to the fourth embodiment. [Figure 12] FIG. 12 is a cross-sectional view of a nitride semiconductor device according to a modification of the fourth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] (Findings on which this disclosure is based) The present inventors have found that the following problems occur with conventional nitride semiconductor devices.

[0013] Conventional nitride semiconductor devices use n-type epitaxially grown GaN layers on conductive GaN substrates. - - A transistor structure was formed using nitride semiconductor films such as GaN drift layers. However, it is difficult to increase the diameter of a GaN substrate, so the size of the substrate is limited. This is because, for example, when the diameter of a GaN substrate is increased, the reliability decreases due to the effect of the variation in the off-angle.

[0014] As substrates that can be made larger in diameter, Si substrates or SiC substrates are known. A technique for epitaxially growing GaN on a Si substrate is disclosed in Patent Document 2. Specifically, in the technique disclosed in Patent Document 2, an AlN buffer layer is formed on a 3C-SiC thin film on a Si substrate. Also, as disclosed in Non-Patent Document 1, epitaxial growth of GaN is possible on a 6H-SiC substrate by forming an AlN buffer layer in the same manner. The reason for using AlN as a buffer layer is due to the relationship between the lattice constant and the thermal expansion coefficient of semiconductor materials shown in Table 1.

[0015] [Table 1]

[0016] As shown in Table 1, SiC has a smaller thermal expansion coefficient than GaN, so when the temperature is lowered from the growth temperature to room temperature, a large tensile stress is applied to the GaN epitaxial film, causing cracks. By using AlN, which has a larger thermal expansion coefficient than GaN and a small lattice mismatch with SiC, as the buffer layer, compressive stress can be generated in the GaN, preventing cracks.

[0017] However, since the AlN buffer layer has high resistance, it cannot be applied to vertical transistors in which current flows perpendicular to the substrate. Therefore, there is a demand for technology that realizes nitride semiconductor devices using large-diameter substrates without using an AlN buffer layer. Note that, as a large-diameter substrate, for example, a Si substrate or a SiC substrate with a diameter of more than 6 inches is known, but the diameter of the substrate is not limited to this. Even for substrates with a diameter of less than 6 inches, it is expected to reduce costs and improve reliability by using a Si substrate instead of a GaN substrate.

[0018] a first nitride semiconductor layer provided above the substrate; a first p-type nitride semiconductor layer provided above the first nitride semiconductor layer; a second nitride semiconductor layer provided above the first p-type nitride semiconductor layer; a first opening penetrating the second nitride semiconductor layer and the first p-type nitride semiconductor layer and reaching the first nitride semiconductor layer; an electron transit layer and an electron supply layer provided in this order from below to cover an upper surface of the second nitride semiconductor layer and the first opening; a gate electrode provided above the electron supply layer to cover the first opening; a source electrode provided at a position spaced from the gate electrode; and a drain electrode provided below the substrate, wherein the substrate includes a silicon carbide layer and a carbon layer provided above the silicon carbide layer, and the carbon layer includes at least one of a graphene layer and a graphene buffer layer.

[0019] In this way, by providing the carbon layer, the bond between the silicon carbide layer and the first nitride semiconductor layer can be weakened. Therefore, a first nitride semiconductor layer with good crystallinity can be formed without providing AlN. Therefore, existing large-diameter substrates such as Si substrates or SiC substrates can be applied. In this way, according to this embodiment, a nitride semiconductor device that allows the substrate diameter to be increased can be realized.

[0020] In addition, in a nitride semiconductor device according to a second aspect of the present disclosure, in the nitride semiconductor device according to the first aspect, the carbon layer has an opening that exposes a portion of the silicon carbide layer.

[0021] This strengthens the bond between the silicon carbide layer and the first nitride semiconductor layer in the portion where the opening is provided, thereby suppressing peeling of the substrate and realizing a highly reliable nitride semiconductor device.

[0022] In addition, in the nitride semiconductor device according to a third aspect of the present disclosure, in the nitride semiconductor device according to the second aspect, the opening has a quadrangular shape in plan view.

[0023] In addition, in the nitride semiconductor device according to a fourth aspect of the present disclosure, in the nitride semiconductor device according to the second aspect, the opening has a hexagonal shape in plan view.

[0024] In addition, in the nitride semiconductor device according to a fifth aspect of the present disclosure, in the nitride semiconductor device according to the second aspect, the opening has a circular shape in a plan view.

[0025] Thus, the planar shape of the openings provided in the carbon layer is not particularly limited. The shape and arrangement can be adjusted to a desired design value by patterning or the like, or lattice defects randomly generated by laser irradiation or the like can be used as the openings. Since the openings can be formed by a method suitable for the required specifications, a highly reliable or easily manufactured nitride semiconductor device can be realized.

[0026] Furthermore, a nitride semiconductor device according to a sixth aspect of the present disclosure is the nitride semiconductor device according to any one of the first to fifth aspects, further comprising a second p-type nitride semiconductor layer provided between the electron supply layer and the gate electrode.

[0027] This makes it possible to reduce the carrier concentration directly below the gate electrode and shift the threshold voltage of the transistor to the positive side, which makes it possible to easily realize the nitride semiconductor device according to this embodiment as a normally-off type FET.

[0028] Furthermore, in a nitride semiconductor device according to a seventh aspect of the present disclosure, in the nitride semiconductor device according to the sixth aspect, the carbon layer is provided only at a position overlapping the second p-type nitride semiconductor layer in a plan view.

[0029] This makes it possible to suppress dislocations occurring in the portion of the first nitride semiconductor layer between the gate electrode and the drain electrode, thereby suppressing leakage current that can flow between the gate and drain when the device is off.

[0030] Furthermore, a nitride semiconductor device according to an eighth aspect of the present disclosure is the nitride semiconductor device according to any one of the first to seventh aspects, further comprising a second opening that penetrates the electron supply layer and the electron transit layer at a position away from the gate electrode and reaches the first p-type nitride semiconductor layer, and the source electrode is provided so as to cover the second opening.

[0031] This allows the two-dimensional electron gas and the source electrode to be directly connected to each other on the side surface of the second opening, thereby reducing the source contact resistance.

[0032] In addition, in a nitride semiconductor device according to a ninth aspect of the present disclosure, in the nitride semiconductor device according to any one of the first to eighth aspects, the substrate includes a silicon substrate layer located below the silicon carbide layer, and the drain electrode is provided on a lower surface of the silicon substrate layer.

[0033] This makes it possible to use a Si substrate, which is inexpensive and can be easily made large in diameter.

[0034] In addition, in a nitride semiconductor device according to a tenth aspect of the present disclosure, in the nitride semiconductor device according to any one of the first to eighth aspects, the silicon carbide layer is a hexagonal silicon carbide substrate layer, and the drain electrode is provided on a lower surface of the silicon carbide layer.

[0035] As a result, since hexagonal silicon carbide has a higher mobility than cubic silicon carbide, it is possible to realize a lower on-resistance and a larger drain current.

[0036] In addition, in a nitride semiconductor device according to an eleventh aspect of the present disclosure, in the nitride semiconductor device according to any one of the first to eighth aspects, the silicon carbide layer is a cubic silicon carbide substrate layer, and the drain electrode is provided on a lower surface of the silicon carbide layer.

[0037] As a result, cubic silicon carbide has a higher thermal conductivity than hexagonal silicon carbide, and can quickly dissipate heat generated during operation, thereby making it possible to suppress malfunctions caused by heat and realizing a highly reliable nitride semiconductor device.

[0038] In addition, in a nitride semiconductor device according to a twelfth aspect of the present disclosure, in the nitride semiconductor device according to any one of the first to eighth aspects, the substrate includes a hexagonal silicon carbide substrate layer or a cubic silicon carbide substrate layer located below the silicon carbide layer, and the impurity concentration of the silicon carbide layer is lower than the impurity concentration of the hexagonal silicon carbide substrate layer or the cubic silicon carbide substrate layer.

[0039] This allows the carbon layer to be used for the purpose of diffusing current in a direction parallel to the main surface of the substrate, thereby realizing a large drain current.

[0040] A nitride semiconductor device according to a thirteenth aspect of the present disclosure includes a substrate, a first silicon carbide layer provided above the substrate, a p-type silicon carbide layer provided above the first silicon carbide layer so as to expose at least a portion of an upper surface of the first silicon carbide layer, and at least one of a graphene layer or a graphene buffer layer, and is equipped with a carbon layer provided above the p-type silicon carbide layer, an electron transit layer provided above the first silicon carbide layer and the carbon layer, an electron supply layer provided above the electron transit layer, a gate electrode provided above the electron supply layer, a source electrode provided at a position spaced from the gate electrode, and a drain electrode provided below the substrate, and the carbon layer has an opening exposing a portion of the first silicon carbide layer.

[0041] As a result, the carbon layer is provided, and thus the bond between the silicon carbide layer and the first nitride semiconductor layer can be weakened. Therefore, a first nitride semiconductor layer having good crystallinity can be formed without providing AlN. Therefore, an existing large-diameter substrate such as a Si substrate or a SiC substrate can be applied. Thus, according to this embodiment, a nitride semiconductor device that allows the substrate to have a large diameter can be realized. In addition, since it is not necessary to provide a gate opening, the manufacturing process can be simplified and the reliability of the device can be improved.

[0042] In addition, in a nitride semiconductor device according to a fourteenth aspect of the present disclosure, in the nitride semiconductor device according to the thirteenth aspect, the gate electrode is provided at a position overlapping the opening in a plan view of the substrate.

[0043] In addition, in a nitride semiconductor device according to a fifteenth aspect of the present disclosure, in the nitride semiconductor device according to the thirteenth aspect, the gate electrode is provided at a position overlapping the p-type silicon carbide layer in a planar view of the substrate.

[0044] As a result, a p-type silicon carbide layer is disposed between the gate electrode and the drain electrode, which reduces the parasitic capacitance between the gate and drain, thereby enabling the transistor to operate at a higher speed.

[0045] Furthermore, a nitride semiconductor device according to a sixteenth aspect of the present disclosure is the nitride semiconductor device according to any one of the thirteenth to fifteenth aspects, further comprising a p-type nitride semiconductor layer provided between the electron supply layer and the gate electrode.

[0046] This makes it possible to reduce the carrier concentration directly below the gate electrode and shift the threshold voltage of the transistor to the positive side, which makes it possible to easily realize the nitride semiconductor device according to this embodiment as a normally-off type FET.

[0047] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0048] The embodiments described below are all comprehensive or specific examples. The numerical values, shapes, materials, arrangement and connection of components, order of steps, etc. shown in the following embodiments are merely examples and are not intended to limit the present disclosure. Furthermore, among the components in the following embodiments, components that are not described in the independent claims are described as optional components.

[0049] In addition, each figure is a schematic diagram and is not necessarily illustrated precisely. Therefore, for example, the scales in each figure do not necessarily match. In addition, in each figure, substantially the same configurations are given the same reference numerals, and duplicated explanations are omitted or simplified.

[0050] Furthermore, in this specification, terms indicating the relationship between elements, such as parallel or perpendicular, terms indicating the shape of elements, such as rectangle or trapezoid, and numerical ranges are not expressions that only express a strict meaning, but are expressions that include a substantially equivalent range, for example, a difference of about a few percent.

[0051] In this specification and the drawings, the x-axis, y-axis, and z-axis indicate the three axes of a three-dimensional Cartesian coordinate system. When the planar shape of the substrate is rectangular, the x-axis and y-axis are directions parallel to a first side of the rectangle and a second side perpendicular to the first side, respectively. The z-axis is the thickness direction of the substrate. In this specification, the "thickness direction" of the substrate refers to a direction perpendicular to the main surface of the substrate. The thickness direction is the same as the stacking direction of the semiconductor layers, and is also referred to as the "vertical direction". The direction parallel to the main surface of the substrate may be referred to as the "horizontal direction".

[0052] In addition, the side of the substrate on which the gate electrode and source electrode are provided (the positive side of the z-axis) is regarded as the "upper" or "upper side", and the side of the substrate on which the drain electrode is provided (the negative side of the z-axis) is regarded as the "lower" or "lower side".

[0053] In addition, in this specification, the terms "upper" and "lower" do not refer to the upward (vertically upward) and downward (vertically downward) directions in absolute spatial recognition, but are used as terms defined by the relative positional relationship based on the stacking order in the stacked structure. Further, the terms "upper" and "lower" are applicable not only when two components are arranged with a space therebetween and another component exists between the two components, but also when the two components are arranged in close contact with each other and the two components are in contact with each other.

[0054] In addition, in this specification, "plan view" means, unless otherwise specified, when viewed from a direction perpendicular to the main surface of the substrate of the nitride semiconductor device, that is, when the main surface of the substrate is viewed from the front.

[0055] In addition, in this specification, ordinal numbers such as "first" and "second" do not mean the number or order of components, unless otherwise specified, but are used for the purpose of avoiding confusion and distinguishing between the same kind of components.

[0056] In addition, in this specification, AlGaN represents ternary mixed crystal Al x Ga 1-x N (0 < x < 1). Hereinafter, the multi-component mixed crystal is abbreviated by the arrangement of the respective component symbols, for example, AlInN, GaInN, etc. For example, Al x Ga 1-x-y In y N (0 < x < 1, 0 < y < 1, and 0 < x + y < 1) is abbreviated as AlGaInN.

[0057] (Embodiment 1) [Configuration] First, the configuration of the nitride semiconductor device according to Embodiment 1 will be described with reference to FIG. 1.

[0058] Fig. 1 is a cross-sectional view of a nitride semiconductor device 1 according to the present embodiment. In Fig. 1, each component, such as a semiconductor layer and electrodes, is shaded with diagonal lines to indicate a cross section. Note that the electron transit layer 22 is not shaded with diagonal lines to make it easier to see the dashed lines that typically represent the two-dimensional electron gas 26 in the electron transit layer 22.

[0059] 1, the nitride semiconductor device 1 according to the present embodiment is a so-called vertical field effect transistor (FET). Specifically, in the nitride semiconductor device 1, a current flows between a drain electrode 36 and a source electrode 34 in a direction perpendicular to a main surface of a substrate 10.

[0060] The nitride semiconductor device 1 is a device having a laminated structure of nitride semiconductor layers containing, as a main component, nitride semiconductors such as GaN and AlGaN. Note that "A contains B as a main component" means that the content of B in A is 50% or more.

[0061] The nitride semiconductor device 1 according to the present embodiment is a normally-off type FET. In the nitride semiconductor device 1, for example, the source electrode 34 is grounded (i.e., the potential is 0 V), and a positive potential is applied to the drain electrode 36. The potential applied to the drain electrode 36 is, for example, not limited to, 100 V or more and 1200 V or less. When the nitride semiconductor device 1 is in an off state, 0 V or a negative potential (for example, −5 V) is applied to the gate electrode 32. When the nitride semiconductor device 1 is in an on state, a positive potential (for example, +5 V) is applied to the gate electrode 32. The nitride semiconductor device 1 may be a normally-on type FET.

[0062] 1, the nitride semiconductor device 1 includes a substrate 10, a drift layer 14, a first underlayer 16, a second underlayer 18, a gate opening 20, an electron transit layer 22, an electron supply layer 24, a threshold adjustment layer 28, a source opening 30, a gate electrode 32, a source electrode 34, and a drain electrode 36. At the interface between the electron transit layer 22 and the electron supply layer 24, a two-dimensional electron gas (2DEG) 26 that functions as a channel is generated.

[0063] Each of the components of the nitride semiconductor device 1 will be described in detail below.

[0064] The substrate 10 includes a silicon substrate layer 11, a silicon carbide layer 12, and a carbon layer 13. The shape of the substrate 10 in a plan view is, for example, rectangular, but is not limited to this.

[0065] The silicon substrate layer 11 is a substrate body located below the silicon carbide layer 12. The silicon substrate layer 11 is, for example, + The crystal plane orientation of the upper surface of the silicon substrate layer 11 is, for example, (111). Note that n-type and p-type indicate the conductivity types of semiconductors. + The term "type" refers to a state in which a semiconductor is heavily doped with n-type dopants. - The term "type" refers to a state in which a semiconductor is doped with a low concentration of n-type dopants, i.e. lightly doped. + Type and n - Both types are examples of n-type, and may be referred to as n-type without distinction. + Type and p - The same is true for types.

[0066] The silicon carbide layer 12 is provided above the silicon substrate layer 11. The silicon carbide layer 12 is, for example, a cubic silicon carbide layer, specifically, a layer made of 3C-SiC. The crystal plane orientation of the upper surface of the silicon carbide layer 12 is, for example, (111). The silicon carbide layer 12 has a thickness of, for example, 100 nm or less. For example, the silicon carbide layer 12 is undoped, but may be doped to an n-type. The donor concentration of the silicon carbide layer 12 is, for example, 1×10 15 cm -3 That's it, 1×10 17 cm -3 These include, but are not limited to, the following:

[0067] The silicon carbide layer 12 is formed by epitaxial growth of 3C-SiC crystals on the silicon substrate layer 11. For the epitaxial growth, a growth method such as MBE (Molecular Beam Epitaxy) method or thermal CVD (Chemical Vapor Deposition) method is mainly used.

[0068] The crystal growth of 3C-SiC requires a temperature of 1000°C or higher. In order to suppress the sublimation of Si from the surface of the silicon substrate layer 11 during the crystal growth, a thin carbonized film is formed on the upper surface of the silicon substrate layer 11 at a low temperature in the early stage using a hydrocarbon gas such as ethylene (C2H4) or acetylene (C2H2). Then, the epitaxial growth process is generally performed using a mixed gas of a hydrocarbon and a silane-based gas. As the silane-based gas, for example, dichlorosilane (SiH2Cl2) or silane (SiH4) is used. Instead of the mixed gas of a hydrocarbon and a silane-based gas, monomethylsilane (CH3SiH3) may be used. Nitrogen gas or the like may be mixed during the crystal growth of 3C-SiC to form n-type 3C-SiC.

[0069] The carbon layer 13 is provided above the silicon carbide layer 12. The carbon layer 13 is located on the uppermost layer of the substrate 10. The carbon layer 13 includes at least one of a graphene layer and a graphene buffer layer. In this embodiment, the carbon layer 13 is provided over the entire interface between the silicon carbide layer 12 and the drift layer 14. Details of the carbon layer 13 will be described later.

[0070] The drift layer 14 is an example of a first nitride semiconductor layer provided above the substrate 10. The drift layer 14 is, for example, a n-type nitride semiconductor layer having a thickness of 8 μm. - The drift layer 14 is a film made of GaN of the type. The donor concentration of the drift layer 14 is, for example, 1×10 15 cm -3 More than 1×10 17 cm -3 As an example, 16 cm -3 The carbon concentration (C concentration) of the drift layer 14 is, for example, 1×10 15 cm -3 More than 2×10 17 cm -3 The following is the result.

[0071] The drift layer 14 is provided, for example, in contact with an upper surface (main surface) of the substrate 10. The drift layer 14 is formed on the main surface of the substrate 10 by crystal growth, for example, by a metal organic vapor phase epitaxy (MOVPE) method, a hydride vapor phase epitaxy (HVPE) method, or the like.

[0072] The first underlayer 16 is an example of a first p-type nitride semiconductor layer provided above the drift layer 14. The first underlayer 16 has a thickness of, for example, 400 nm and a carrier concentration of 1×10 17 cm -3 The first underlayer 16 is a film made of p-type GaN, which is a p-type GaN film having a p-type conductivity. The first underlayer 16 is provided in contact with the upper surface of the drift layer 14. The first underlayer 16 is formed on the drift layer 14 by crystal growth, for example, by MOVPE, HVPE, or the like.

[0073] Although the first underlayer 16 is formed by crystal growth, it may be formed, for example, by injecting magnesium (Mg) into the formed i-GaN. Furthermore, the first underlayer 16 may be an insulating layer obtained by injecting iron (Fe) or boron (B) instead of a p-type nitride semiconductor layer.

[0074] The first underlayer 16 is also called a block layer, and suppresses leakage current between the source electrode 34 and the drain electrode 36. For example, when a reverse voltage is applied to the pn junction formed by the first underlayer 16 and the drift layer 14, specifically, when the drain electrode 36 has a higher potential than the source electrode 34, a depletion layer extends to the drift layer 14. This allows the nitride semiconductor device 1 to have a high breakdown voltage. In this embodiment, the drain electrode 36 has a higher potential than the source electrode 34 in both the off state and the on state, except in the case of reverse conduction. This allows the nitride semiconductor device 1 to have a high breakdown voltage.

[0075] The second underlayer 18 is an example of a second nitride semiconductor layer provided above the first underlayer 16. The second underlayer 18 is a high-resistance layer having a higher resistance than the first underlayer 16. The second underlayer 18 is, for example, a film made of undoped GaN (i-GaN) having a thickness of 200 nm. The second underlayer 18 is provided in contact with the first underlayer 16. For example, the second underlayer 18 is formed on the first underlayer 16 by crystal growth such as MOVPE or HVPE.

[0076] The second underlayer 18 is assumed to be an undoped semiconductor layer, but may be an insulating layer or a semi-insulating layer. Here, "undoped" means that it is not doped with a dopant such as Si or Mg that changes the polarity of GaN to n-type or p-type. In this embodiment, the second underlayer 18 may be doped with carbon (C). For example, the carbon concentration of the second underlayer 18 is higher than the carbon concentration of the first underlayer 16.

[0077] For example, the carbon concentration of the second underlayer 18 is 3×10 17 cm -3 That's it, but 1×10 18 cm -3 In this case, the concentration of silicon (Si) or oxygen (O) serving as an n-type impurity is lower than the carbon concentration. For example, the silicon concentration or oxygen concentration of the second underlayer 18 is 5×10 16 cm -3 Below, 2×10 16 cm -3 The following may also be used: As for the type of ions implanted into the second underlayer 18 and the first underlayer 16, ion species other than those mentioned above can be used to obtain the same effect as long as they are capable of increasing the resistance of the semiconductor layer.

[0078] In addition, a layer for suppressing diffusion of p-type impurities such as Mg from the first underlayer 16 may be provided on the upper surface of the second underlayer 18. For example, an AlGaN layer having a thickness of 20 nm may be provided on the second underlayer 18.

[0079] The gate opening 20 is an example of a first opening that penetrates the second underlayer 18 and the first underlayer 16 to reach the drift layer 14. A bottom surface 20a of the gate opening 20 is a part of the upper surface of the drift layer 14. As shown in FIG. 1, the bottom surface 20a is located below the lower surface of the first underlayer 16. The lower surface of the first underlayer 16 corresponds to the interface between the first underlayer 16 and the drift layer 14. The bottom surface 20a is, for example, parallel to the main surface of the substrate 10. A drain current flows through the bottom surface 20a of the gate opening 20 when the nitride semiconductor device 1 is on.

[0080] In this embodiment, the gate opening 20 is formed so that the opening area increases as it gets farther from the substrate 10. Specifically, the side surface 20b of the gate opening 20 is inclined obliquely. As shown in FIG. 1, the cross-sectional shape of the gate opening 20 is an inverted trapezoid, more specifically, an inverted isosceles trapezoid.

[0081] The inclination angle of the side surface 20b with respect to the bottom surface 20a is, for example, 20° to 80°, but may be 30° to 45°. The smaller the inclination angle, the closer the side surface 20b is to the c-plane, and the better the film quality of the electron transit layer 22 and the like formed along the side surface 20b by crystal regrowth. On the other hand, the larger the inclination angle, the more the gate opening 20 is prevented from becoming too large, and the more compact the nitride semiconductor device 1 can be achieved.

[0082] The gate opening 20 is formed by forming the drift layer 14, the first underlayer 16, and the second underlayer 18 in this order by continuous film deposition on the main surface of the substrate 10, and then removing a portion of each of the second underlayer 18 and the first underlayer 16 so as to partially expose the drift layer 14. At this time, by removing a surface portion of the drift layer 14 by a predetermined thickness, for example, 300 nm, the bottom surface 20a of the gate opening 20 is formed below the lower surface of the first underlayer 16.

[0083] As a method for removing the second underlayer 18 and the first underlayer 16, dry etching such as inductively coupled plasma etching (ICP) is often used, and a chlorine-based gas is often used as the process gas.

[0084] The electron transit layer 22 is an example of a first regrown layer provided to cover the upper surface of the second underlayer 18 and the gate opening 20. Specifically, a part of the electron transit layer 22 is provided along the bottom surface 20a and the side surface 20b of the gate opening 20, and the other part of the electron transit layer 22 is provided on the upper surface of the second underlayer 18. The electron transit layer 22 is, for example, a film made of undoped GaN with a thickness of 150 nm. Note that although the electron transit layer 22 is assumed to be undoped, a part of it may be made n-type by Si doping or the like.

[0085] The electron transit layer 22 is in contact with the drift layer 14 at the bottom surface 20a and the side surface 20b of the gate opening 20. The electron transit layer 22 is in contact with each of the first underlying layer 16 and the second underlying layer 18 at the side surface 20b of the gate opening 20. Furthermore, the electron transit layer 22 is in contact with the upper surface of the second underlying layer 18.

[0086] The electron transit layer 22 has a channel region. Specifically, a two-dimensional electron gas 26 that serves as a channel is generated near the interface between the electron transit layer 22 and the electron supply layer 24. In Fig. 1, the two-dimensional electron gas 26 is diagrammatically illustrated by a dashed line. The two-dimensional electron gas 26 is bent along the interface between the electron transit layer 22 and the electron supply layer 24, i.e., along the inner surface of the gate opening 20.

[0087] 1, an AlN layer with a thickness of about 1 nm is provided as a second regrown layer between the electron transit layer 22 and the electron supply layer 24. This makes it possible to suppress alloy scattering, improve channel mobility, and reduce on-resistance. Note that the AlN layer is not necessarily required.

[0088] The electron supply layer 24 is an example of a third regrown layer provided so as to cover the upper surface of the second underlayer 18 and the gate opening 20. The electron transit layer 22 and the electron supply layer 24 are provided in this order from the substrate 10 side. The electron supply layer 24 is a film made of undoped AlGaN having a thickness of, for example, 20 nm. The electron supply layer 24 is formed in a shape that follows the upper surface of the electron transit layer 22 and has an approximately uniform thickness.

[0089] The electron supply layer 24 has a larger band gap than the electron transit layer 22. Therefore, an AlGaN / GaN heterointerface is formed between the electron supply layer 24 and the electron transit layer 22. The electron supply layer 24 supplies electrons to a channel region (two-dimensional electron gas 26) formed in the electron transit layer 22.

[0090] The threshold adjustment layer 28 is an example of a second p-type nitride semiconductor layer provided between the electron supply layer 24 and the gate electrode 32. The threshold adjustment layer 28 has a thickness of 100 nm and a carrier concentration of 1×10 17 cm -3 The film is made of p-type AlGaN.

[0091] The potential of the channel portion is increased by providing the threshold adjustment layer 28. This makes it possible to increase the threshold of the transistor and realize a normally-off state.

[0092] Alternatively, the threshold adjustment layer 28 may not be provided, and the gate electrode 32 may be provided in contact with the electron supply layer 24. Alternatively, instead of the threshold adjustment layer 28, an insulating layer such as SiN or SiO2 may be provided between the gate electrode 32 and the electron supply layer 24. This makes it possible to suppress the gate current and shift the threshold in the positive direction to achieve a normally-off operation.

[0093] The electron transit layer 22, the electron supply layer 24, and the threshold adjustment layer 28 are formed by forming the gate opening 20, and then successively depositing nitride semiconductor films by a crystal regrowth process and patterning them into a predetermined shape. Specifically, an undoped GaN film that is the base of the electron transit layer 22, an undoped AlGaN film that is the base of the electron supply layer 24, and a p-type AlGaN film that is the base of the threshold adjustment layer 28 are successively deposited by MOVPE, HVPE, or the like. After deposition, a part of the p-type AlGaN film is removed by etching until the undoped AlGaN film is exposed, thereby forming the threshold adjustment layer 28. Furthermore, a part of each of the undoped AlGaN film and the undoped GaN film and a part of the second underlayer 18 are continuously removed by etching until the first underlayer 16 is exposed. As a result, a source opening 30 that reaches the first underlayer 16 is formed, and the electron supply layer 24 and the electron transit layer 22 that are patterned into a predetermined shape are formed.

[0094] The source opening 30 is an example of a second opening that penetrates the electron supply layer 24 and the electron transit layer 22 at a position away from the gate electrode 32, and reaches the first underlayer 16. In the present embodiment, the source opening 30 is provided at a position away from both the gate opening 20 and the threshold adjustment layer 28 in a plan view of the substrate 10.

[0095] A bottom surface 30a of the source opening 30 is a part of the upper surface of the first underlayer 16. In FIG. 1, the bottom surface 30a is flush with the interface between the second underlayer 18 and the first underlayer 16, but is not limited to this. The bottom surface 30a may be located below the interface between the second underlayer 18 and the first underlayer 16. That is, the source opening 30 may be formed by removing a surface portion of the first underlayer 16. The bottom surface 30a is, for example, parallel to the main surface of the substrate 10.

[0096] 1, the source opening 30 is formed so that the opening area is constant regardless of the distance from the substrate 10. Specifically, the side surface 30b of the source opening 30 is perpendicular to the bottom surface 30a. That is, the cross-sectional shape of the source opening 30 is rectangular.

[0097] Alternatively, the source opening 30 may be formed so that the opening area increases as it is farther from the substrate 10. Specifically, the side surface 30b of the source opening 30 may be inclined obliquely. For example, the cross-sectional shape of the source opening 30 may be an inverted trapezoid, more specifically, an inverted isosceles trapezoid. In this case, the inclination angle of the side surface 30b with respect to the bottom surface 30a may be, for example, in the range of 30° to 60°. Since the side surface 30b is inclined obliquely, the contact area between the source electrode 34 and the electron transit layer 22 (the two-dimensional electron gas 26) increases, so that an ohmic connection is easily made. The two-dimensional electron gas 26 is exposed to the side surface 30b of the source opening 30, and is connected to the source electrode 34 at the exposed portion.

[0098] The source opening 30 is not necessarily required. The provision of the source opening 30 can reduce the ohmic contact resistance between the two-dimensional electron gas 26 functioning as a channel and the source electrode 34. In addition, the first underlayer 16 and the source electrode 34 can be electrically connected, so that the potential of the first underlayer 16 can be stabilized to obtain effects such as improved breakdown voltage.

[0099] The gate electrode 32 is provided above the electron supply layer 24 so as to cover the gate opening 20. Specifically, the gate electrode 32 is provided in contact with the upper surface of the threshold adjustment layer .

[0100] The gate electrode 32 is formed using a conductive material such as a metal. For example, the gate electrode 32 may be made of a material that is in ohmic contact with the p-type GaN layer. For example, palladium (Pd), nickel (Ni)-based material, tungsten silicide (WSi), gold (Au), etc. may be used.

[0101] The gate electrode 32 is formed by forming a conductive film by sputtering, vapor deposition, or the like after the formation of the threshold adjustment layer 28, after the formation of the source opening 30, or after the formation of the source electrode 34, and then patterning the formed conductive film.

[0102] The source electrode 34 is provided at a position away from the gate electrode 32. In this embodiment, the source electrode 34 is provided so as to cover the source opening 30. Specifically, the source electrode 34 is provided so as to be in contact with the bottom surface 30a and the side surface 30b of the source opening 30 so as to fill the source opening 30. The source electrode 34 is in direct contact with the two-dimensional electron gas 26 at the side surface 30b of the source opening 30. This makes it possible to reduce the contact resistance between the source electrode 34 and the two-dimensional electron gas 26. Note that a part of the source electrode 34 may be provided on the upper surface of the electron supply layer 24.

[0103] The source electrode 34 is formed using a conductive material such as a metal. For example, a material that can be ohmic-connected to the n-type GaN layer by heat treatment, such as Ti / Al (a laminated structure of a Ti layer and an Al layer), can be used as the material for the source electrode 34. The source electrode 34 is formed by patterning a conductive film formed by, for example, sputtering or vapor deposition.

[0104] The drain electrode 36 is provided below the substrate 10. Specifically, the drain electrode 36 is provided on the lower surface of the silicon substrate layer 11 of the substrate 10.

[0105] The drain electrode 36 is formed using a conductive material such as a metal. For example, a material that makes ohmic contact with n-type Si, such as AuSb / Au, can be used as the material of the drain electrode 36. The drain electrode 36 is formed by, for example, sputtering or vapor deposition.

[0106] [Main characteristic configuration] Next, a main characteristic configuration of the nitride semiconductor device 1 according to the present embodiment will be described.

[0107] As described above, in the nitride semiconductor device 1, the carbon layer 13 is provided on the uppermost layer of the substrate 10. As shown in Fig. 2, the carbon layer 13 includes a graphene buffer layer 13a and an epitaxial graphene layer 13b. Fig. 2 is a diagram showing the graphene buffer layer 13a and the epitaxial graphene layer 13b on SiC (silicon carbide layer 12).

[0108] The graphene buffer layer 13a is a carbon atom layer provided on the outermost surface of the silicon carbide layer 12. Four bonds of the carbon atom are bonded to three other carbon atoms, and the remaining bond is bonded to a silicon atom on the outermost surface of the silicon carbide layer 12.

[0109] The epitaxial graphene layer 13b is a carbon atom layer provided above the silicon carbide layer 12. A plurality of carbon atoms are arranged in a plane to form a hexagonal lattice structure. Although Fig. 2 shows an example in which one epitaxial graphene layer 13b is provided, a plurality of epitaxial graphene layers 13b may be stacked.

[0110] The carbon layer 13 is formed on the silicon carbide layer 12 by high-temperature annealing in a vacuum and in an inert gas atmosphere such as Ar. The temperature of the high-temperature annealing is equal to or lower than 1420°C, which is the melting point of Si, and is, for example, 1150°C to 1250°C. As shown in FIG. 2, Si is desorbed from the SiC surface during annealing, and the remaining C forms a graphene layer. Annealing at 1150°C forms a graphene buffer layer 13a on the SiC surface. Furthermore, annealing at 1250°C advances the thermal decomposition of SiC, and a new graphene buffer layer 13a is formed at the interface between the graphene buffer layer 13a and SiC. The previously formed graphene buffer layer 13a is then positioned at the top surface, becoming an epitaxial graphene layer 13b. At this time, only weak bonds such as van der Waals forces exist at the interface between the epitaxial graphene layer 13b and the graphene buffer layer 13a. On the other hand, at the interface between the graphene buffer layer 13a and SiC, Si atoms on the outermost surface of the SiC are covalently bonded to C atoms of the graphene buffer layer 13a.

[0111] In hetero growth on a hetero substrate, the difference in thermal expansion coefficient and lattice mismatch deteriorate the crystallinity of the grown film, causing cracks. This is due to the presence of strong bonds such as covalent bonds at the hetero interface. The presence of strong bonds distorts the original crystal structure, and distortion energy accumulates with increasing film thickness or when the growth temperature rises or falls. When the accumulated energy exceeds a critical value, a crystal defect called a dislocation occurs.

[0112] In contrast, as shown in FIG. 2, by forming a graphene buffer layer 13a or an epitaxial graphene layer 13b on SiC, the dangling bonds on the top surface of the SiC are terminated, and only weak bonds such as van der Waals forces exist between the epitaxial graphene layer 13b or the graphene buffer layer 13a and the GaN formed on the SiC. Therefore, even if the lattice constants and thermal expansion coefficients are different, the GaN can be epitaxially grown with its own crystal structure and lattice constant. This can improve the crystallinity of the nitride semiconductor film represented by GaN.

[0113] Note that when the epitaxial graphene layer 13b has three or more layers, the crystallinity of GaN deteriorates rapidly. This is because, as shown in Fig. 3, although dangling bonds are terminated on the top surface of the SiC, GaN on the epitaxial graphene layer 13b grows epitaxially by sensing the weak surface potential of SiC.

[0114] FIG. 3 is a diagram showing the relationship between the number of epitaxial graphene layers 13b and the crystallinity of the GaN film. In FIG. 3, the van der Waals forces occurring between the layers are represented by double-headed arrows. When the number of epitaxial graphene layers 13b as the underlayer of GaN becomes thinner in the order of two layers, one layer, and only the graphene buffer layer 13a, GaN itself can strongly sense the surface potential of SiC. This improves the crystallinity of GaN.

[0115] As described above, by providing the carbon layer 13 including at least one of the graphene buffer layer 13a and the epitaxial graphene layer 13b on the outermost surface of the substrate 10, a nitride semiconductor film with excellent crystallinity can be formed on the substrate 10, which is different from the nitride semiconductor film. Since a substrate made of Si or SiC, which is inexpensive and can be made large in diameter, can be used as the substrate 10, it becomes possible to make the substrate 10 of the nitride semiconductor device 1 large in diameter.

[0116] [Positions of the ends of the threshold adjustment layer and gate electrode] In the nitride semiconductor device 1, the threshold of the transistor can be adjusted according to the position of the end of the threshold adjustment layer 28. When the threshold adjustment layer 28 is not provided, the threshold of the transistor can be adjusted according to the position of the end of the gate electrode 32.

[0117] Specifically, when the end position of the threshold adjustment layer 28 is located inside the opening end of the gate opening 20 (specifically, the upper end of the side surface 20b), the threshold value of the transistor is determined only by the portion along the side surface 20b of the gate opening 20, i.e., the inclined portion of the electron transit layer 22. In this configuration, the carrier concentration in the flat portion of the electron transit layer 22, i.e., the portion located above the second underlayer 18, can be increased, thereby reducing the on-resistance.

[0118] On the other hand, when the end position of the threshold adjustment layer 28 is disposed outside the opening end of the gate opening 20, the threshold of the transistor is determined by the sloped portion or the flat portion of the electron transit layer 22, whichever has the larger threshold. For example, when the threshold of the transistor is determined by the flat portion, it is necessary to make only the threshold of the sloped portion smaller (specifically, negative). In this case, the film thickness of the electron transit layer 22 in the direction parallel to the main surface of the substrate 10 is made larger than the film thickness of the electron transit layer 22 in the direction perpendicular to the main surface of the substrate 10. This makes it possible to make the distance between the p-type first underlayer 16 and the two-dimensional electron gas 26 (channel) of the sloped portion larger than the distance between the first underlayer 16 and the two-dimensional electron gas 26 (channel) of the flat portion. Therefore, the influence of depletion from the first underlayer 16 on the channel of the sloped portion can be reduced compared to the flat portion, so that it is possible to selectively make the threshold of the channel of the sloped portion smaller. Therefore, it is possible to make the threshold of the transistor determined by the flat portion.

[0119] [Thickness of electron transport layer] In the nitride semiconductor device 1, the on-resistance can be reduced by varying the film thickness of the electron transit layer 22 depending on the location. Specifically, the film thickness of the electron transit layer 22 in a direction parallel to the main surface of the substrate 10 is greater than the film thickness of the electron transit layer 22 in a direction perpendicular to the main surface of the substrate 10. The film thickness of the electron transit layer 22 in the direction parallel to the main surface of the substrate 10 is the distance between the upper end of the gate opening 20 and the electron supply layer 24 in the direction parallel to the main surface of the substrate 10.

[0120] In this embodiment, as described above, the distance between the two-dimensional electron gas 26 in the inclined portion and the first underlayer 16 can be increased to suppress the decrease in carrier concentration. In general, vertical GaN crystal growth is performed so that the c-plane of the GaN crystal is parallel to the main surface of the substrate 10. In this case, the two-dimensional electron gas 26 inclined from the main surface of the substrate 10 has a smaller polarization than the two-dimensional electron gas 26 parallel to the main surface of the substrate 10, and therefore the carrier concentration is decreased. That is, the two-dimensional electron gas 26 in the inclined portion has a lower carrier concentration than the two-dimensional electron gas 26 in the flat portion. Therefore, the two-dimensional electron gas 26 in the inclined portion is more susceptible to the constriction effect of the depletion layer extending from the p-type first underlayer 16 than the two-dimensional electron gas 26 in the flat portion. Therefore, the on-resistance can be further reduced by moving the two-dimensional electron gas 26 in the inclined portion away from the first underlayer 16.

[0121] On the other hand, when the thickness of the electron transit layer 22 in the direction perpendicular to the main surface of the substrate 10 is large, the thickness of the electron transit layer 22 in the flat portion is thin, and thus the depth of the source opening 30 for contacting the source electrode 34 with the first underlayer 16 can be made shallower. The shallower the depth of the source opening 30, the shorter the process time and the better the coverage of the source electrode 34 to be formed later. This makes it possible to reduce the on-resistance while facilitating the process. In this way, the thickness of the electron transit layer 22 in the direction parallel to the main surface of the substrate 10 is greater than the thickness in the direction perpendicular to the main surface of the substrate 10, thereby making it possible to reduce the on-resistance while facilitating the process.

[0122] (Embodiment 2) Next, a second embodiment will be described.

[0123] The main difference between the second embodiment and the first embodiment is that an opening is provided in the carbon layer. The following description will focus on the differences from the first embodiment, and description of the commonalities will be omitted or simplified.

[0124] Fig. 4 is a cross-sectional view of a nitride semiconductor device 101 according to the present embodiment. As shown in Fig. 4, the nitride semiconductor device 101 according to the present embodiment differs from the nitride semiconductor device 1 shown in Fig. 1 in that it includes a substrate 110 instead of the substrate 10. The substrate 110 includes a carbon layer 113 instead of the carbon layer 13.

[0125] 4, the carbon layer 113 has an opening 113a exposing a part of the silicon carbide layer 12. In other words, the carbon layer 113 does not cover the entire upper surface of the silicon carbide layer 12, but is provided partially on the upper surface of the silicon carbide layer 12. A plurality of openings 113a are provided, but there may be only one opening 113a.

[0126] 5A and 5B are plan views showing an example of the shape of the openings 113a in the carbon layer 113. The shape of the openings 113a in plan view is, for example, a rectangle as shown in FIG. 5A. Alternatively, the shape of the openings 113a in plan view may be another polygon such as a hexagon, or a circle or an ellipse. Furthermore, as shown in FIG. 5B, the shapes of the openings 113a in plan view may be different from each other.

[0127] The opening 113a is formed, for example, by patterning with a resist and then selectively etching using O2 plasma ashing, thereby making it possible to provide the opening 113a of a desired shape at a desired position.

[0128] Alternatively, the openings 113a may be formed by randomly introducing lattice defects by low-energy rare gas ion beam irradiation and N2 plasma irradiation. When lattice defects are randomly introduced, multiple openings 113a with different shapes in plan view can be easily formed as shown in FIG. 5B.

[0129] The interface between the carbon layer 113 and the drift layer 14 made of GaN has only a weak bond such as van der Waals force, and therefore has a weak adhesion strength. Since the nitride semiconductor device 101 may be used as a power transistor for switching at a large current, if the adhesion strength between the carbon layer 113 and the drift layer 14 is weak, the reliability may be impaired.

[0130] Therefore, in the nitride semiconductor device 101, the carbon layer 113 is partially provided on the upper surface of the silicon carbide layer 12. That is, the carbon layer 113 has an opening 113a that exposes the upper surface of the silicon carbide layer 12. The surface of the silicon carbide layer 12 is exposed in the portion where the carbon layer 113 is not provided, that is, in the opening 113a. Therefore, in the opening 113a, a strong bond such as a covalent bond exists at the interface between the silicon carbide layer 12 and the drift layer 14. Therefore, the adhesive strength between the drift layer 14 and the substrate 110 can be increased, and the reliability of the nitride semiconductor device 101 can be improved.

[0131] Although there is a risk of dislocations (schematically shown by dotted lines in the drawing) being formed in the drift layer 14 within the openings 113a, good GaN is formed on the carbon layer 113. Therefore, it is possible to achieve both improved adhesion strength and good GaN crystallinity.

[0132] [Variations] Fig. 6 is a cross-sectional view of a nitride semiconductor device 102 according to a modification of the present embodiment. As shown in Fig. 6, the nitride semiconductor device 102 according to this modification is different from the nitride semiconductor device 101 shown in Fig. 4 in the position of the opening 113a in the carbon layer 113. Specifically, the opening 113a is provided so as not to overlap the threshold adjustment layer 28 in a planar view. In other words, the carbon layer 113 is provided only at a position that overlaps the threshold adjustment layer 28 in a planar view.

[0133] When a high voltage is applied to the drain electrode 36 during off-state, an electric field tends to concentrate under the threshold adjustment layer 28 and at the end of the first underlayer 16. Therefore, in the drift layer 14, a high-quality epitaxial film is required for the portion located below the threshold adjustment layer 28 and the end of the first underlayer 16 in order to suppress off-leakage current.

[0134] In the nitride semiconductor device 102 according to this modification, the carbon layer 113 is provided at a position overlapping the threshold adjustment layer 28 in a planar view. This makes it possible to improve the crystallinity of the portion of the drift layer 14 that overlaps the threshold adjustment layer 28 in a planar view. As a result, it is possible to sufficiently suppress the off-leak current.

[0135] Fig. 7A is a plan view showing an example of a nitride semiconductor device 102 according to this modification. Fig. 7B is a plan view of a carbon layer 113 included in the nitride semiconductor device 102 shown in Fig. 7A. The cross section taken along line VI-VI shown in Fig. 7A corresponds to the cross section shown in Fig. 6.

[0136] In the example shown in FIG. 7A, elongated source electrodes 34 extending along the y-axis direction are arranged side by side in each of the x-axis direction and the y-axis direction. The gate electrode 32 is provided so as to surround the multiple source electrodes 34 individually. Specifically, the gate electrode 32 is a plate-shaped electrode provided with multiple elongated openings corresponding to the multiple source electrodes 34. In a plan view, the gate electrode 32 and the multiple source electrodes 34 are provided at a distance from each other and do not overlap. The threshold adjustment layer 28 is a plate-shaped semiconductor layer provided with multiple elongated openings corresponding to the multiple source electrodes 34, similar to the gate electrode 32. In a plan view, the threshold adjustment layer 28 and the multiple source electrodes 34 are provided at a distance from each other and do not overlap. The elongated opening provided in the threshold adjustment layer 28 is one size smaller than the elongated opening provided in the gate electrode 32. Therefore, in a plan view, the open end of the threshold adjustment layer 28 is exposed from the elongated opening provided in the gate electrode 32.

[0137] The nitride semiconductor device 102 shown in Fig. 7A includes a carbon layer 113 shown in Fig. 7B. The carbon layer 113 shown in Fig. 7B has a plurality of openings 113a, and the planar shape of the plurality of openings 113a is the same as the planar shape of the elongated openings provided in the threshold adjustment layer 28 (or the gate electrode 32). The positions of the plurality of openings 113a are the same as the positions of the elongated openings provided in the threshold adjustment layer 28. This makes it possible to realize a configuration in which the openings 113a of the carbon layer 113 are located at positions overlapping the source electrode 34, and the carbon layer 113 is provided at positions overlapping the gate electrode 32 and the threshold adjustment layer 28 in a planar view.

[0138] It should be noted that the shape and arrangement of the multiple source electrodes 34 are not limited to this. Fig. 8A is a cross-sectional view of a nitride semiconductor device 103 according to another modified example. Fig. 8B is a plan view of a carbon layer 113 included in the nitride semiconductor device 103 shown in Fig. 8A. The cross section taken along line VI-VI shown in Fig. 8A corresponds to the cross section shown in Fig. 6.

[0139] In the example shown in FIG. 8A, the source electrodes 34, each having a regular hexagonal shape in plan view, are arranged side by side in the plane. Specifically, the source electrodes 34 are arranged so that the centers of the source electrodes 34 are located at the vertices of the equilateral triangles arranged in a packed arrangement. The gate electrode 32 is provided so as to surround the source electrodes 34 individually. Specifically, the gate electrode 32 is a plate-like electrode provided with a plurality of regular hexagonal openings corresponding to the source electrodes 34, respectively. In a plan view, the gate electrode 32 and the source electrodes 34 are provided at a distance from each other and do not overlap. The threshold adjustment layer 28 is a plate-like semiconductor layer provided with a plurality of regular hexagonal openings corresponding to the source electrodes 34, respectively, like the gate electrode 32. In a plan view, the threshold adjustment layer 28 and the source electrodes 34 are provided at a distance from each other and do not overlap. The regular hexagonal openings provided in the threshold adjustment layer 28 are one size smaller than the regular hexagonal openings provided in the gate electrode 32. Therefore, in plan view, the opening end of the threshold adjustment layer 28 is exposed from the regular hexagonal opening provided in the gate electrode 32 .

[0140] The nitride semiconductor device 103 shown in Fig. 8A includes a carbon layer 113 shown in Fig. 8B. The carbon layer 113 shown in Fig. 8B has a plurality of openings 113a, and the planar shape of the plurality of openings 113a is the same as the planar shape of the regular hexagonal openings provided in the threshold adjustment layer 28 (or the gate electrode 32). The positions of the plurality of openings 113a are the same as the positions of the regular hexagonal openings provided in the threshold adjustment layer 28. This makes it possible to realize a configuration in which the openings 113a of the carbon layer 113 are located at positions overlapping the source electrode 34, and the carbon layer 113 is provided at positions overlapping the gate electrode 32 and the threshold adjustment layer 28 in a planar view.

[0141] In this way, the patterning of the carbon layer 113 can be changed according to the layout of the vertical transistor to be fabricated. Note that the shape of the opening 113a provided in the carbon layer 113 in a plan view does not need to be the same as the shape of the opening 113a provided in the gate electrode 32 or the threshold adjustment layer 28 in a plan view.

[0142] For example, the carbon layer 113 may be provided at a position overlapping with a bottom portion of the threshold adjustment layer 28 located within the gate opening 20 in a plan view, and may not overlap with an inclined portion along the side surface 20b of the gate opening 20 and an upper end portion along the upper end of the gate opening 20. In other words, the opening 113a may be provided at a position overlapping with a part of the threshold adjustment layer 28, specifically, the inclined portion and the upper end portion of the threshold adjustment layer 28, in a plan view. Even in this case, the crystallinity of the portion of the drift layer 14 between the bottom portion of the threshold adjustment layer 28 and the drain electrode 36, where the electric field is likely to concentrate, is increased, and thus the leakage current can be suppressed.

[0143] (Embodiment 3) Next, a third embodiment will be described.

[0144] The main difference between the third embodiment and the first embodiment is that the substrate includes a silicon carbide layer instead of a silicon substrate layer. The following description will focus on the differences from the first embodiment, and the description of the commonalities will be omitted or simplified.

[0145] Fig. 9 is a cross-sectional view of a nitride semiconductor device 201 according to the present embodiment. As shown in Fig. 9, the nitride semiconductor device 201 of the present embodiment differs from the nitride semiconductor device 1 shown in Fig. 1 in that it includes a substrate 210 instead of the substrate 10. The substrate 210 includes a SiC substrate layer 211 and a carbon layer 13.

[0146] The SiC substrate layer 211 is an example of a silicon carbide layer. The SiC substrate layer 211 is a hexagonal silicon carbide substrate layer or a cubic silicon carbide substrate layer, and specifically, is a substrate body made of 4H-SiC, 6H-SiC, or 3C-SiC. A drain electrode 36 is provided on the lower surface of the SiC substrate layer 211. The SiC substrate layer 211 is, for example, an n + For example, the thickness is 100 μm and the donor concentration is 1×10 17 cm -3 More than 1×10 19 cm -3 The following is the result.

[0147] In the nitride semiconductor device 201 according to this embodiment, a hexagonal or cubic SiC substrate is used, thereby making it possible to obtain a GaN vertical transistor with improved heat dissipation characteristics. In general, the thermal conductivity of SiC is 4.9 W / cm·K, which is higher than the thermal conductivity of Si, 1.5 W / cm·K. Therefore, when applied to high-power equipment, heat can be efficiently dissipated, thereby reducing transistor loss.

[0148] [Variations] Fig. 10 is a cross-sectional view of a nitride semiconductor device 202 according to a modification of the present embodiment. As shown in Fig. 10, the nitride semiconductor device 202 according to this modification is different from the nitride semiconductor device 201 shown in Fig. 9 in that a substrate 210A is provided instead of the substrate 210. The substrate 210A includes a SiC substrate layer 211, a drift layer 212, and a carbon layer 13.

[0149] In this modification, the drift layer 212 is an example of a silicon carbide layer. The drift layer 212 is provided above the SiC substrate layer 211. The drift layer 212 is, for example, an n - For example, the donor concentration of the drift layer 212 is 1×10 15 cm -3 More than 1×10 17 cm -3 The thickness of the drift layer 212 is, for example, less than 7 μm, but is not limited to this. For example, the drift layer 212 can be formed by epitaxial growth on the upper surface of the SiC substrate layer 211. The thickness of the drift layer 212 is, for example, thinner than the thickness of the drift layer 14. This allows the carbon layer 13 to be disposed close to the bottom surface 20a of the gate opening 20.

[0150] The carbon layer 13 includes at least one of the graphene buffer layer 13a and the epitaxial graphene layer 13b, which have high mobility characteristics, and can function as a current diffusion layer. In this modification, the carbon layer 13 is close to the bottom surface 20a of the gate opening 20, so a wide range of the nitride semiconductor device 202 can be used as a current path. This makes it possible to realize low on-resistance characteristics of the nitride semiconductor device 202. In addition, like the SiC substrate layer 211, the drift layer 212 made of SiC also has high thermal conductivity, and therefore heat dissipation characteristics can be improved.

[0151] (Embodiment 4) Next, a fourth embodiment will be described.

[0152] The main difference between the fourth embodiment and the first embodiment is that there is no gate opening formed by etching. In the following, the differences from the first embodiment will be mainly described, and the description of the commonalities will be omitted or simplified.

[0153] Fig. 11 is a cross-sectional view of a nitride semiconductor device 301 according to this embodiment. As shown in Fig. 11, the nitride semiconductor device 301 according to this embodiment includes a substrate 310, a drift layer 312, a block layer 314, a carbon layer 313, an electron transit layer 22, an electron supply layer 24, a threshold adjustment layer 28, a source opening 30, a gate electrode 32, a source electrode 34, and a drain electrode 36. The electron transit layer 22, the electron supply layer 24, the threshold adjustment layer 28, the source opening 30, the gate electrode 32, the source electrode 34, and the drain electrode 36 are substantially the same as those in the first embodiment.

[0154] The substrate 310 is a hexagonal silicon carbide substrate or a cubic silicon carbide substrate, specifically, a substrate made of 4H-SiC, 6H-SiC, or 3C-SiC. + For example, the thickness is 100 μm and the donor concentration is 1×10 17 cm -3 More than 1×10 19 cm-3 The following is the result.

[0155] The drift layer 312 is an example of a first silicon carbide layer provided above the substrate 310. The drift layer 312 is, for example, - The drift layer 312 is a film made of SiC of the type. The donor concentration of the drift layer 312 is 1×10 15 cm -3 More than 1×10 17 cm -3 The drift layer 312 may be formed on the upper surface of the substrate 310 by epitaxial growth, for example.

[0156] The block layer 314 is a p-type silicon carbide layer provided above the drift layer 312 so as to expose at least a part of the upper surface of the drift layer 312. Note that above the drift layer 312 means above at least a part of the drift layer 312.

[0157] In this embodiment, the blocking layer 314 is an n-type GaN layer formed by epitaxial growth on the upper surface of the substrate 310. - It is formed by ion-implanting p-type impurities into a predetermined region of the surface layer portion of the SiC film of the mold. The portion where the p-type impurities are not implanted becomes the upper surface of drift layer 312. The upper surface of block layer 314 and the upper surface of drift layer 312 are flush with each other.

[0158] The carbon layer 313 is provided above the block layer 314. Like the carbon layer 13 in the first embodiment, the carbon layer 313 includes at least one of a graphene buffer layer or an epitaxial graphene layer.

[0159] In this embodiment, the carbon layer 313 has an opening 313a. The opening 313a is provided to prevent the carbon layer 313 from contacting the drift layer 312. This makes it possible to prevent the source electrodes 34 arranged on either side of the gate electrode 32 from being electrically short-circuited via the carbon layer 313. The gate electrode 32 is provided at a position overlapping the opening 313a in a plan view of the substrate 310.

[0160] Carbon layer 313 having opening 313a is formed by the same method as in embodiment 2. Specifically, drift layer 312 and block layer 314 are each subjected to high-temperature annealing treatment to form a graphene buffer layer and / or an epitaxial graphene layer on the outermost surfaces of drift layer 312 and block layer 314, respectively, and then the graphene buffer layer and / or epitaxial graphene layer are patterned using a resist and selectively etched by O2 plasma ashing, thereby forming carbon layer 313 having opening 313a.

[0161] After the carbon layer 313 is formed, GaN is regrown to sequentially form the electron transit layer 22 and the electron supply layer 24. The electron transit layer 22 and the electron supply layer 24 are provided so as to cover the drift layer 312, the block layer 314, and the carbon layer 313, respectively. Note that while the thickness of the carbon layer 313 is exaggerated in FIG. 11, the thickness of the carbon layer 313 is actually the thickness of several layers of graphene. Therefore, the electron transit layer 22 and the electron supply layer 24 are formed so as to have a substantially uniform thickness and a flat upper surface.

[0162] In this embodiment, the electron transit layer 22 made of GaN and the drift layer 312 made of SiC are in contact with each other at the openings 313a in the carbon layer 313. Therefore, dislocations due to SiC (drift layer 312) may occur in the electron transit layer 22. However, since the electron transit layer 22 has a thin thickness of only 150 nm, the occurrence of dislocations is suppressed.

[0163] In the present embodiment, the source opening 30 is provided so as to expose at least a part of the p-type block layer 314. For example, as shown in FIG. 11 , the bottom surface 30a of the source opening 30 is flush with the interface between the block layer 314 and the carbon layer 313, but is not limited to this. The bottom surface 30a of the source opening 30 may be located below the interface between the block layer 314 and the carbon layer 313. In this way, the source electrode 34 is electrically connected to the p-type block layer 314.

[0164] According to this embodiment, a SiC substrate having high thermal conductivity can be used, so that the heat dissipation of the nitride semiconductor device 301 can be improved. - Since the pn junction between the drift layer 312 made of p-type SiC and the block layer 314 made of p-type SiC can have a high avalanche resistance, the avalanche resistance of the pn junction is improved. In general, since SiC can be formed into a good quality p-type by ion implantation, the avalanche resistance of the pn junction is higher than that of GaN. Therefore, device destruction due to avalanche breakdown can be suppressed.

[0165] [Variations] Fig. 12 is a cross-sectional view of a nitride semiconductor device 302 according to a modification of the present embodiment. As shown in Fig. 12, the nitride semiconductor device 302 according to this modification is different from the nitride semiconductor device 301 shown in Fig. 11 in that it includes a threshold adjustment layer 328 and a gate electrode 332 instead of the threshold adjustment layer 28 and the gate electrode 32.

[0166] The threshold adjustment layer 328 and the gate electrode 332 are provided at positions different from those of the threshold adjustment layer 28 and the gate electrode 32. Specifically, as shown in Fig. 12, the threshold adjustment layer 328 and the gate electrode 332 are provided at positions overlapping the p-type block layer 314 in a planar view. The threshold adjustment layer 328 and the gate electrode 332 are provided at positions overlapping the carbon layer 313 in a planar view of the substrate 310, and do not overlap the opening 313a.

[0167] In this way, since the p-type block layer 314 is disposed between the gate electrode 332 and the drain electrode 36, the area where the gate electrode 332 and the drain electrode 36 face each other can be substantially eliminated. This allows the parasitic capacitance Cgd between the gate and the drain to be significantly reduced. The feedback capacitance Crss is the parasitic capacitance Cgd between the gate and the drain, and if the feedback capacitance Crss is large, the rise of the drain current is slow even when the gate is turned on, and the fall of the drain current is slow when the gate is turned off. In other words, by reducing the feedback capacitance Crss (parasitic capacitance Cgd), the switching speed can be improved, and high-speed switching can be achieved to reduce the size of power devices.

[0168] In the present embodiment and the modified examples, the threshold adjustment layer 28 or 328 is provided as an example of a p-type nitride semiconductor layer provided between the electron supply layer 24 and the gate electrode 32 or 332, but the present invention is not limited to this. The gate electrode 32 or 332 may be provided in contact with the electron supply layer 24 without providing the threshold adjustment layer 28 or 328. Alternatively, an insulating layer such as SiN or SiO2 may be provided between the electron supply layer 24 and the gate electrode 32 or 323.

[0169] (Other embodiments) Although the nitride semiconductor device according to one or more aspects has been described based on the embodiments, the present disclosure is not limited to these embodiments. As long as it does not deviate from the gist of the present disclosure, various modifications conceivable by a person skilled in the art to the present embodiment and forms constructed by combining components of different embodiments are also included within the scope of the present disclosure.

[0170] For example, the carbon layer 13 may include only the graphene buffer layer 13a. The high-temperature annealing treatment for SiC may be performed at 1150° C., and the high-temperature annealing treatment may be terminated when only the graphene buffer layer 13a is formed.

[0171] Alternatively, the carbon layer 13 may include only one or more epitaxial graphene layers 13b without including the graphene buffer layer 13a. For example, when the plane orientation of the top surface of the silicon carbide layer 12 included in the substrate 10 is (110) or (001), the epitaxial graphene layer 13b can be formed without forming the graphene buffer layer 13a by performing a high-temperature annealing process.

[0172] Also, for example, the drift layer 14, 212 or 312 may have a graded structure in which the impurity concentration (donor concentration) is gradually reduced from the substrate 10 side toward the upper side. The donor concentration may be controlled by Si serving as a donor, or by carbon serving as an acceptor that compensates for Si. Alternatively, the drift layer 14, 212 or 312 may have a stacked structure of multiple nitride semiconductor layers having different impurity concentrations.

[0173] Furthermore, each of the above embodiments can be modified, substituted, added, omitted, and the like in various ways within the scope of the claims or their equivalents. [Industrial Applicability]

[0174] Nitride semiconductor devices according to the present disclosure can be used, for example, in power transistors used in power supply circuits such as main inverters for xEVs, on-board chargers, and general-purpose inverters for power conditioners. [Explanation of symbols]

[0175] 1, 101, 102, 103, 201, 202, 301, 302 Nitride Semiconductor Devices 10, 110, 210, 210A, 310 board 11 Silicon substrate layer 12 Silicon carbide layer 13, 113, 313 Carbon layer 13a Graphene buffer layer 13b Epitaxial graphene layer 14, 212, 312 Drift layer 16 First Undercoat Layer 18 Second Undercoat 20 Gate opening 20a, 30a bottom 20b, 30b side 22 Electron transport layer 24 Electron supply layer 26 Two-dimensional electron gas 28, 328 Threshold adjustment layer 30 Source opening 32, 332 Gate electrode 34 Source electrode 36 Drain electrode 113a, 313a opening 211 SiC substrate layer 314 Block Layer

Claims

1. A substrate; a first nitride semiconductor layer provided above the substrate; a first p-type nitride semiconductor layer provided above the first nitride semiconductor layer; a second nitride semiconductor layer provided above the first p-type nitride semiconductor layer; a first opening that penetrates the second nitride semiconductor layer and the first p-type nitride semiconductor layer and reaches the first nitride semiconductor layer; an electron transit layer and an electron supply layer provided in this order from below so as to cover an upper surface of the second nitride semiconductor layer and the first opening; a gate electrode provided above the electron supply layer so as to cover the first opening; a source electrode provided at a position spaced from the gate electrode; a drain electrode provided below the substrate; The substrate is a silicon carbide layer; a carbon layer disposed above the silicon carbide layer; The carbon layer includes at least one of a graphene layer and a graphene buffer layer. Nitride semiconductor devices.

2. the carbon layer having an opening exposing a portion of the silicon carbide layer; The nitride semiconductor device of claim 1 .

3. The planar shape of the opening is rectangular. The nitride semiconductor device of claim 2 .

4. The planar shape of the opening is hexagonal. The nitride semiconductor device of claim 2 .

5. The planar shape of the opening is circular. The nitride semiconductor device of claim 2 .

6. a second p-type nitride semiconductor layer provided between the electron supply layer and the gate electrode; The nitride semiconductor device of claim 1 .

7. the carbon layer is provided only at a position overlapping the second p-type nitride semiconductor layer in a plan view; The nitride semiconductor device of claim 6.

8. a second opening that is disposed at a position away from the gate electrode, passes through the electron supply layer and the electron transit layer, and reaches the first p-type nitride semiconductor layer; The source electrode is provided to cover the second opening. The nitride semiconductor device of claim 1 .

9. the substrate includes a silicon substrate layer underlying the silicon carbide layer; The drain electrode is provided on a lower surface of the silicon substrate layer. The nitride semiconductor device according to any one of claims 1 to 8.

10. the silicon carbide layer is a hexagonal silicon carbide substrate layer; The drain electrode is provided on a lower surface of the silicon carbide layer. The nitride semiconductor device according to any one of claims 1 to 8.

11. the silicon carbide layer is a cubic silicon carbide substrate layer; The drain electrode is provided on a lower surface of the silicon carbide layer. The nitride semiconductor device according to any one of claims 1 to 8.

12. the substrate includes a hexagonal silicon carbide substrate layer or a cubic silicon carbide substrate layer underlying the silicon carbide layer; The impurity concentration of the silicon carbide layer is lower than the impurity concentration of the hexagonal silicon carbide substrate layer or the cubic silicon carbide substrate layer; The nitride semiconductor device according to any one of claims 1 to 8.

13. A substrate; a first silicon carbide layer disposed above the substrate; a p-type silicon carbide layer provided above the first silicon carbide layer so as to expose at least a portion of an upper surface of the first silicon carbide layer; a carbon layer including at least one of a graphene layer and a graphene buffer layer and disposed above the p-type silicon carbide layer; an electron transit layer provided above the first silicon carbide layer and the carbon layer; an electron supply layer provided above the electron transit layer; a gate electrode provided above the electron supply layer; a source electrode provided at a position spaced from the gate electrode; a drain electrode provided below the substrate; the carbon layer has an opening exposing a portion of the first silicon carbide layer. Nitride semiconductor devices.

14. the gate electrode is provided at a position overlapping the opening in a plan view of the substrate; The nitride semiconductor device of claim 13.

15. The gate electrode is provided at a position overlapping the p-type silicon carbide layer in a plan view of the substrate. The nitride semiconductor device of claim 13.

16. a p-type nitride semiconductor layer provided between the electron supply layer and the gate electrode; 16. The nitride semiconductor device according to any one of claims 13 to 15.

Citation Information

Patent Citations

  • Compound semiconductor substrate

    JP2019102767A

  • Nitride Semiconductor Devices

    JP6511645B2