High electron mobility transistor

Abstract

A high electron mobility transistor capable of reducing a sheet resistance is provided. The high electron mobility transistor has a channel layer having a first main surface, and a barrier layer having a second main surface opposite to the first main surface, the channel layer has a first nitride semiconductor layer containing gallium, the barrier layer has a second nitride semiconductor layer containing scandium and aluminum, and the channel layer includes a tensile strain parallel to the first main surface.

Description

Technical Field

[0001] This disclosure relates to high electron mobility transistors. Background Technology

[0002] A high electron mobility transistor (HEMT) is known, which has a channel layer formed of gallium nitride (GaN) and a barrier layer formed of aluminum gallium nitride (AlGaN).

[0003] Existing technical documents Patent documents Patent Document 1: Japanese Patent Publication No. 2021-524164 In recent years, the demand for reducing the resistance of thin films has been increasing. Summary of the Invention

[0004] The purpose of this disclosure is to provide a high electron mobility transistor that can reduce the resistance of a thin film.

[0005] The high electron mobility transistor of this disclosure has: a channel layer having a first main surface; and a barrier layer having a second main surface opposite to the first main surface, the channel layer having a first nitride semiconductor layer comprising gallium, the barrier layer having a second nitride semiconductor layer comprising scandium and aluminum, and the channel layer including tensile strain parallel to the first main surface.

[0006] Invention Effects According to this disclosure, the resistance of the thin film can be reduced. Attached Figure Description

[0007] Figure 1 This is a graph showing the relationship between strain along the a-axis and the effective mass of free electrons in a GaN crystal.

[0008] Figure 2 This is a cross-sectional view showing the high electron mobility transistor of the first embodiment.

[0009] Figure 3 This is a graph showing the relationship between composition and spontaneous polarization in Al-containing nitride semiconductors.

[0010] Figure 4 This is a graph showing the relationship between the composition of an Al-containing nitride semiconductor and the lattice constant along the a-axis.

[0011] Figure 5 This is a graph showing the relationship between the composition of the second nitride semiconductor layer and the thin-film resistance.

[0012] Figure 6 It means Sc 0.18（1－y） Al0.82（1－y） Ga y A graph showing the relationship between the composition of layer N and the arithmetic mean roughness Ra.

[0013] Figure 7 This is a diagram showing the composition of the barrier layer in the high electron mobility transistor of the second embodiment.

[0014] Figure 8 This is a diagram showing the composition of the barrier layer in the high electron mobility transistor of the third embodiment.

[0015] Figure 9 This is a cross-sectional view showing the high electron mobility transistor of the fourth embodiment.

[0016] Figure 10 This is a diagram showing the composition of the barrier layer in the high electron mobility transistor of the fourth embodiment.

[0017] Figure 11 This is a diagram showing the composition of the barrier layer in the high electron mobility transistor of the fifth embodiment.

[0018] Explanation of reference numerals in the attached figures 1: High electron mobility transistor; 44D: Drain electrode; 44S: Source electrode; 50: Gate electrode; 110: Substrate; 120: Nitride semiconductor stacked structure; 122: Buffer layer; 124: Channel layer; 124A: First nitride semiconductor layer; 126: Barrier layer; 126A: Second nitride semiconductor layer; 126B: Third nitride semiconductor layer; 128: Cap layer; 130: Insulating layer; 130D, 130G, 130S: Openings; 140D, 140S: Recesses; 142D, 142S: Regeneration layers; 155: Channel region; 161: First main surface; 162: Second main surface. Detailed Implementation

[0019] First, the characteristics of a conventional HEMT having a channel layer formed of GaN and a barrier layer formed of AlGaN will be explained. Figure 1 This graph shows the relationship between strain along the a-axis and the effective mass of free electrons in a GaN crystal. When the strain is zero, the crystal is in a bulk state. Hereinafter, negative strain will be called compressive strain, and positive strain will be called tensile strain. Compressive strain is the strain that compresses the lattice by decreasing the lattice constant relative to the bulk state, while tensile strain is the strain that stretches the lattice by increasing the lattice constant relative to the bulk state. Figure 1 In this context, the larger the positive value of the strain, the greater the tensile strain; the smaller the negative value of the strain, the greater the compressive strain. That is, the larger the absolute value, the greater the magnitude of the strain.

[0020] The lattice constant along the a-axis of GaN crystals is greater than that of AlN crystals. Therefore, in conventional HEMTs, near the interface between the channel layer and the barrier layer, the channel layer incorporates compressive strain along the a-axis by lattice matching with the barrier layer, while the barrier layer exhibits tensile strain along the a-axis by lattice matching with the channel layer.

[0021] In conventional HEMTs, the electron density in the channel region can be increased by increasing the proportion of Al in the barrier layer. On the other hand, in GaN, such as... Figure 1 As shown, the smaller the strain in the a-axis direction, i.e., the greater the compressive strain in the a-axis direction, the greater the effective mass. Therefore, in conventional HEMTs, increasing the proportion of Al in the barrier layer increases the effective mass of electrons in the channel region and decreases the electron mobility.

[0022] Therefore, in conventional HEMTs, even if the electron density in the channel region can be increased by increasing the proportion of Al in the barrier layer, the electron mobility in the channel region will decrease. Consequently, it is difficult to increase the conductivity of the channel region, thus making it difficult to reduce the sheet resistance.

[0023] Based on this insight, the inventors of this application conducted repeated and in-depth research to prevent a decrease in the mobility of electrons in the channel region, and as a result, conceived of the following implementation method.

[0024] [Description of embodiments of this disclosure] First, the implementation plan disclosed herein will be listed for illustration.

[0025] [1] A high electron mobility transistor of one embodiment of the present disclosure has: a channel layer having a first main surface; and a barrier layer having a second main surface opposite to the first main surface, the channel layer having a first nitride semiconductor layer comprising gallium, the barrier layer having a second nitride semiconductor layer comprising scandium and aluminum, and the channel layer including tensile strain parallel to the first main surface.

[0026] Scandium induces strong spontaneous polarization in the second nitride semiconductor layer, thereby increasing the electron density in the channel region. Furthermore, the tensile strain in the channel layer, parallel to the first principal plane, enhances electron mobility in the channel region. Therefore, the sheet resistance can be reduced.

[0027] [2] In [1], the number of scandium atoms contained in the second nitride semiconductor layer may also be greater than 0.22 times the number of aluminum atoms contained in the second nitride semiconductor layer. In this case, it is easier to generate tensile strain in the channel layer parallel to the first main surface.

[0028] [3] In [1] or [2], the second nitride semiconductor layer may also comprise gallium. In this case, excellent flatness is readily obtained on the upper surface of the barrier layer.

[0029] [4] In [3], the number of gallium atoms contained in the second nitride semiconductor layer may also be greater than or equal to 0.3 times the sum of the number of scandium atoms contained in the second nitride semiconductor layer, the number of aluminum atoms contained in the second nitride semiconductor layer, and the number of gallium atoms contained in the second nitride semiconductor layer. In this case, it is easy to obtain particularly excellent flatness on the upper surface of the barrier layer.

[0030] [5] In any of [1] to [4], the further away from the second main surface, the smaller the ratio of the number of scandium atoms in the second nitride semiconductor layer to the number of aluminum atoms in the second nitride semiconductor layer. In this case, the charge near the main surface of the barrier layer opposite to the second main surface can be reduced, thereby reducing gate leakage.

[0031] [6] In any of [1] to [5], the barrier layer may also have a third nitride semiconductor layer comprising aluminum, the second nitride semiconductor layer being located between the channel layer and the third nitride semiconductor layer, the third nitride semiconductor layer including tensile strain parallel to the first main surface. In this case, the charge near the main surface of the barrier layer opposite to the second main surface can also be reduced, thereby reducing gate leakage.

[0032] [7] In [6], the third nitride semiconductor layer may also contain scandium, wherein the first ratio of the number of scandium atoms in the second nitride semiconductor layer to the number of aluminum atoms in the second nitride semiconductor layer is greater than the second ratio of the number of scandium atoms in the third nitride semiconductor layer to the number of aluminum atoms in the third nitride semiconductor layer. By including scandium in the third nitride semiconductor layer, it is easier to reduce the sheet resistance compared to the case where the third nitride semiconductor layer does not contain scandium. Furthermore, compared to the case where the first and second ratios are equal, it is easier to stably perform crystal growth on the third nitride semiconductor layer.

[0033] [8] In any of [1] to [7], the second nitride semiconductor layer may also contain indium (In). By adding In, the relationship between the polarization charge and the lattice constant of the second nitride semiconductor layer can be adjusted.

[0034] [Details of the embodiments disclosed herein] The embodiments of this disclosure will now be described in detail, but this disclosure is not limited thereto. It should be noted that in this specification and accompanying drawings, components with substantially the same functional configuration may sometimes be omitted by using the same reference numerals to avoid repetitive descriptions. In this disclosure, "top view" refers to viewing an object from above.

[0035] (First Implementation) The first embodiment will be described. The first embodiment relates to a high electron mobility transistor (HEMT). Figure 2 This is a cross-sectional view showing the high electron mobility transistor of the first embodiment.

[0036] like Figure 2 As shown, the HEMT1 of the first embodiment has a substrate 110, a nitride semiconductor stacked structure 120, a regenerated layer 142S, a regenerated layer 142D, an insulating layer 130, a gate electrode 50, a source electrode 44S, and a drain electrode 44D.

[0037] The substrate 110 is, for example, a substrate for growing gallium nitride (GaN) semiconductor layers, or a semi-insulating silicon carbide (SiC) substrate. When the substrate 110 is a SiC substrate, its upper surface is a silicon (Si) polar surface. When the surface of the substrate 110 is a Si polar surface, the nitride semiconductor stack 120 grows crystals using the gallium (Ga) polar surface as the growth surface.

[0038] The nitride semiconductor stacked structure 120 has a buffer layer 122, a channel layer 124, a barrier layer 126, and a cap layer 128.

[0039] A buffer layer 122 is located on the substrate 110. The buffer layer 122 is, for example, an aluminum nitride (AlN) layer. The buffer layer 122 may also have an AlN layer, a GaN layer on top of an AlN layer, or an aluminum gallium nitride (AlGaN) layer.

[0040] The channel layer 124 is located above the buffer layer 122. The channel layer 124 has a first nitride semiconductor layer 124A comprising gallium (Ga). The first nitride semiconductor layer 124A is, for example, an undoped gallium nitride (GaN) layer. The entire channel layer 124 is the first nitride semiconductor layer 124A. The thickness of the channel layer 124 (first nitride semiconductor layer 124A) is, for example, greater than or equal to 20 nm and less than or equal to 800 nm. The channel layer 124 has an upper surface that serves as a first main surface 161.

[0041] A barrier layer 126 is located above the channel layer 124. The barrier layer 126 has a lower surface serving as a second principal surface 162. The second principal surface 162 is opposite to the first principal surface 161. The barrier layer 126 has a second nitride semiconductor layer 126A comprising scandium (Sc) and aluminum (Al). The second nitride semiconductor layer 126A is, for example, an n-type ScAlN layer or a ScAlGaN layer. The entire barrier layer 126 is the second nitride semiconductor layer 126A. The composition of the barrier layer 126 (second nitride semiconductor layer 126A) is fixed from the lower surface to the upper surface. The thickness of the barrier layer 126 (second nitride semiconductor layer 126A) is, for example, greater than or equal to 5 nm and less than or equal to 30 nm. The band gap of the barrier layer 126 is greater than the band gap of the channel layer 124.

[0042] The channel layer 124 includes tensile strain (hereinafter referred to as "tensile strain") parallel to the first principal surface 161, and the barrier layer 126 includes compressive strain (hereinafter referred to as "compressive strain") parallel to the first principal surface 161. The tensile and compressive strains are interrelated strains resulting from the lattice matching of the channel layer 124 and the barrier layer 126; when one increases, the other also increases, and when one decreases, the other also decreases. A channel region 155 comprising a two-dimensional electron gas (2DEG) exists near the upper surface (first principal surface 161) of the channel layer 124.

[0043] The cap layer 128 is located above the barrier layer 126. The cap layer 128 is, for example, an n-type GaN layer. The thickness of the cap layer 128 is, for example, greater than or equal to 3 nm and less than or equal to 10 nm.

[0044] A recess 140S for the active electrode and a recess 140D for the drain electrode are formed in a portion of the cap layer 128, the barrier layer 126, and the channel layer 124. Recesses 140S and 140D penetrate the cap layer 128 and the barrier layer 126 and extend into the channel layer 124. The channel layer 124 is exposed from the recesses 140S and 140D.

[0045] An insulating layer 130 is located above the cap layer 128. The insulating layer 130 is, for example, a silicon nitride (SiN) film. The thickness of the insulating layer 130 is, for example, greater than or equal to 1 nm and less than or equal to 20 nm. An opening 130S for the active electrode and an opening 130D for the drain electrode are formed in the insulating layer 130. The opening 130S is connected to the recess 140S, and the opening 130D is connected to the recess 140D.

[0046] The regenerated layer 142S lies above the channel layer 124 within the recess 140S and the opening 130S. The regenerated layer 142D lies above the channel layer 124 within the recess 140D and the opening 130D. The regenerated layers 142S and 142D are, for example, n-type GaN layers. The resistance of the regenerated layers 142S and 142D is lower than that of the channel region 155.

[0047] The source electrode 44S is located on the regenerated layer 142S, and the drain electrode 44D is located on the regenerated layer 142D. The source electrode 44S is in direct contact with the regenerated layer 142S, and the drain electrode 44D is in direct contact with the regenerated layer 142D. The source electrode 44S is in ohmic contact with the regenerated layer 142S, and the drain electrode 44D is in ohmic contact with the regenerated layer 142D.

[0048] An opening 130G for a gate is formed in the insulating layer 130. Viewed from above, the opening 130G is located between openings 130S and 130D. The gate electrode 50 is disposed on the insulating layer 130 and makes a Schottky contact with the nitride semiconductor stack 120 through the opening 130G. The current flowing between the drain electrode 44D and the source electrode 44S via the channel region 155 changes according to the voltage applied to the gate electrode 50.

[0049] In the first embodiment, the barrier layer 126 has a second nitride semiconductor layer 126A comprising Sc. Sc has the following effect: reducing the crystallinity of the second nitride semiconductor layer 126A. a The lattice constant along the axial direction causes the second nitride semiconductor layer 126A to exhibit strong spontaneous polarization. Therefore, compared to the case where the second nitride semiconductor layer 126A does not contain Sc, the electron density in the channel region 155 is higher.

[0050] Figure 3 This is a graph showing the relationship between composition and spontaneous polarization in Al-containing nitride semiconductors. When x is set to a real number greater than 0 and less than 1, such as... Figure 3 As shown, in aluminum gallium nitride (Ga) x Al 1－x N) and indium aluminum nitride (In) x Al 1－x In N), the larger the value of x, the closer the spontaneous polarization is to 0C / m. 2 On the other hand, in scandium aluminum nitride (Sc... x Al 1－x In N), the larger the value of x, the further the spontaneous polarization is from 0C / m. 2 That is, in Sc x Al 1－xIn N, the larger the value of x, the stronger the spontaneous polarization. It should be noted that, for the sake of comparison between Sc, In, and Ga, Ga is listed at the beginning of the chemical formula for aluminum gallium nitride.

[0051] Furthermore, in the first embodiment, the channel layer 124 includes tensile strain parallel to the first main surface 161, and the barrier layer 126 includes compressive strain parallel to the first main surface 161. Therefore, as... Figure 1 As shown, the effective mass of electrons in the channel layer 124 can be reduced, and the electron mobility in the channel region 155 can be increased. On the other hand, as described above, since the second nitride semiconductor layer 126A contains Sc, the electron density in the channel region 155 is high due to the increase in spontaneous polarization. Therefore, according to the first embodiment, the conductivity increases through the synergistic effect of the increased electron mobility and the increased electron density, thereby reducing the sheet resistance.

[0052] Figure 4 This graph shows the relationship between the composition of Al-containing nitride semiconductors and the lattice constant along the a-axis. The lattice constant of GaN in bulk is 3.19 Å along the a-axis. Figure 4 As shown, in scandium aluminum nitride (Sc x Al 1－x In the case of N), when the value of x is greater than 0.18, the lattice constant along the a-axis in the bulk phase is greater than 3.19 Å. When the value of x is 0.18, the value of 1-x is 0.82, and the number of Sc atoms N(Sc) is 0.22 times the number of Al atoms N(Al). Therefore, in the case where the first nitride semiconductor layer 124A is a GaN layer and the second nitride semiconductor layer 126A is a Sc layer... x Al 1－x In the case of an N-layer, if the number N(Sc) of Sc atoms contained in the second nitride semiconductor layer 126A is greater than 0.22 times the number N(Al) of Al atoms contained in the second nitride semiconductor layer 126A, then through lattice matching with the second nitride semiconductor layer 126A, the channel layer 124 includes tensile strain parallel to the first main surface 161, and the barrier layer 126 includes compressive strain parallel to the first main surface 161. In the second nitride semiconductor layer 126A, the number N(Sc) of atoms can also be greater than or equal to 0.25 times the number N(Al) of atoms, and can also be greater than or equal to 0.30 times the number N(Al) of atoms.

[0053] Figure 5 This is a graph showing the relationship between the composition of the second nitride semiconductor layer 126A and the sheet resistance. Figure 5The diagram shows the calculated sheet resistance of the channel region 155 when the first nitride semiconductor layer 124A is a GaN layer and the second nitride semiconductor layer 126A is a ScAlN layer. It should be noted that in this calculation, it is assumed that the spontaneous polarization of ScAlN is constant regardless of its composition, and the electron density in the channel region 155 is set to 1 × 10⁻⁶. 13 cm -3 The electron mobility in GaN without strain is set to 1800 cm⁻¹. 2 / Vs. For example... Figure 5 As shown, the larger the ratio R of the number of atoms N(Sc) to the number of atoms N(Al), the smaller the thin-film resistance. This is because a larger ratio R reduces the compression of the lattice of the first nitride semiconductor layer 124A parallel to the first principal plane 161, thereby reducing the effective mass and increasing the electron mobility. It should be noted that GaN does not produce strain when the ratio R is 0.22. Therefore, when the ratio R is greater than 0.22, tensile strain occurs in the first nitride semiconductor layer 124A parallel to the first principal plane 161.

[0054] It should be noted that, as Figure 4 As shown, in indium aluminum nitride (In x Al 1－x In N), when the value of x is greater than 0.17, the lattice constant in the a-axis direction is greater than 3.19 Å. Alternatively, the second nitride semiconductor layer 126A may contain In in addition to Sc and Al. When the second nitride semiconductor layer 126A contains In, even when the value of "atoms N(Sc) / atoms N(Al)" is less than 0.22, the channel layer 124 may include tensile strain parallel to the first principal surface 161, and the barrier layer 126 may also include compressive strain parallel to the first principal surface 161. However, as... Figure 3 As shown, the spontaneous polarization of the second nitride semiconductor layer 126A decreases when the amount of In increases. When the second nitride semiconductor layer 126A contains In, if the number of In atoms N(In) is less than the number of Sc atoms N(Sc), the decrease in spontaneous polarization is suppressed, making it easier to obtain stronger spontaneous polarization. If the number of In atoms N(In) is less than or equal to 0.5 times the number of Sc atoms, it is even easier to obtain stronger spontaneous polarization.

[0055] Alternatively, the second nitride semiconductor layer 126A may contain Ga in addition to Sc and Al. By including Ga in the second nitride semiconductor layer 126A, excellent flatness can be easily obtained on its upper surface. In the first nitride semiconductor layer 124A, GaN is a layer, and in the second nitride semiconductor layer 126A, scandium aluminum gallium nitride (Sc) is a layer of GaN.x（1－y） Al （1－x）（1－y） Ga y In the case of the N) layer, if the number of atoms N(Sc) is greater than 0.22 times the number of atoms N(Al), then the channel layer 124 will also include tensile strain parallel to the first main surface 161, and the barrier layer 126 will also include compressive strain parallel to the first main surface 161. The value of y is, for example, less than or equal to 0.9. The second nitride semiconductor layer 126A may also contain Sc, Al, Ga, and In.

[0056] The composition of the second nitride semiconductor layer 126A is not limited, but when the proportion of Sc in the elements other than N is greater than or equal to 0.02 atomic%, it is easier to reduce the sheet resistance. When the proportion of Sc is greater than or equal to 0.1 atomic%, it is even easier to reduce the sheet resistance, and when the proportion of Sc is greater than or equal to 0.2 atomic%, it is even easier to reduce the sheet resistance.

[0057] When the second nitride semiconductor layer 126A contains Ga, when the number of Ga atoms N (Ga) contained in the second nitride semiconductor layer 126A is greater than or equal to 0.3 times the sum of the number of Ga atoms N (Sc), the number of Ga atoms N (Al), and the number of Ga atoms N (Ga), it is easy to obtain particularly excellent flatness on the upper surface of the barrier layer 126. Figure 6 It means Sc 0.18（1－y） Al 0.82（1－y） Ga y A graph showing the relationship between the composition of layer N and the arithmetic mean roughness Ra. (See figure) Figure 6 As shown, a particularly small arithmetic mean roughness Ra is obtained when the value of y is greater than or equal to 0.3. In the second nitride semiconductor layer 126A, the number of atoms N (Ga) can also be greater than or equal to 0.5 times the sum of the number of atoms N (Sc), N (Al), and N (Ga), and can also be greater than or equal to 0.7 times the sum of the number of atoms N (Sc), N (Al), and N (Ga). To improve the crystallinity of the cap layer 128, it is preferable that the arithmetic mean roughness Ra of the upper surface of the barrier layer 126 is small.

[0058] (Second Implementation) The second embodiment will be described. The second embodiment differs from the first embodiment in the composition of the barrier layer. Figure 7 This is a diagram showing the composition of the barrier layer in the high electron mobility transistor of the second embodiment.

[0059] In the HEMT of the second embodiment, as Figure 7As shown, the ratio R of the number of atoms N (Sc) to the number of atoms N (Al) in the second nitride semiconductor layer 126A is greater than 0.22 at the second principal surface 162 (lower surface) and decreases as it moves away from the second principal surface 162 toward the cap layer 128. The ratio R at the upper surface of the second nitride semiconductor layer 126A can be greater than 0.22, less than 0.22, or equal to 0.22.

[0060] The other components of the second embodiment are the same as those of the first embodiment.

[0061] In the second embodiment, similar to the first embodiment, the sheet resistance can be reduced by increasing electron mobility and electron density. As described above, when the ratio R is greater than 0.22, the channel layer 124 includes tensile strain parallel to the first main surface 161, but the same effect can be obtained even if the lattice constant of the portion far from the second main surface 162 is small. Furthermore, compared to the first embodiment, the charge near the upper surface of the barrier layer 126 (the main surface opposite to the second main surface 162) can be reduced. Therefore, gate leakage can be reduced. Moreover, compared to the case where the ratio R is high throughout the second nitride semiconductor layer 126A, it is easier to stably grow the second nitride semiconductor layer 126A.

[0062] (Third Implementation) The third embodiment will be described. The third embodiment differs from the first embodiment in the composition of the barrier layer. Figure 8 This is a diagram showing the composition of the barrier layer in the high electron mobility transistor of the third embodiment.

[0063] In the HEMT of the third embodiment, such as Figure 8 As shown, the ratio R of the number of atoms N(Sc) to the number of atoms N(Al) in the second nitride semiconductor layer 126A is greater than 0.22 at the second principal surface 162 (lower surface) and increases with distance from the second principal surface 162.

[0064] The other components of the third embodiment are the same as those of the first embodiment.

[0065] In the third embodiment, similar to the first embodiment, the sheet resistance can also be reduced by increasing the electron mobility and electron density. To achieve lattice matching between the second nitride semiconductor layer 126A and the cap layer 128, it is sometimes preferable to increase the ratio R in this way.

[0066] (Fourth Implementation) The fourth embodiment will be described. The fourth embodiment differs from the first embodiment in the composition of the barrier layer. Figure 9 This is a cross-sectional view showing the high electron mobility transistor of the fourth embodiment. Figure 10 This is a diagram showing the composition of the barrier layer in the high electron mobility transistor of the fourth embodiment.

[0067] In the fourth embodiment of HEMT4, such as Figure 9 and Figure 10 As shown, the barrier layer 126 has a second nitride semiconductor layer 126A and a third nitride semiconductor layer 126B. The third nitride semiconductor layer 126B contains aluminum (Al) but not scandium (Sc). The third nitride semiconductor layer 126B is, for example, an n-type AlN layer or an AlGaN layer. The composition of the third nitride semiconductor layer 126B is fixed from the lower surface to the upper surface. For example, the thickness of the second nitride semiconductor layer 126A is greater than or equal to 2 nm and less than or equal to 10 nm, and the thickness of the third nitride semiconductor layer 126B is greater than or equal to 2 nm and less than or equal to 10 nm. The third nitride semiconductor layer 126B is located above the second nitride semiconductor layer 126A, which is located between the channel layer 124 and the third nitride semiconductor layer 126B. In the second nitride semiconductor layer 126A, the ratio R is greater than 0.22, therefore the second nitride semiconductor layer 126A has compressive strain parallel to the first main surface 161. Furthermore, tensile stress acts from the second nitride semiconductor layer 126A to the first nitride semiconductor layer 124A. In the third nitride semiconductor layer 126B, the ratio R is less than 0.22, therefore the third nitride semiconductor layer 126B has tensile strain relative to the first nitride semiconductor layer 124A, parallel to the first main surface 161. Furthermore, compressive stress acts from the third nitride semiconductor layer 126B to the first nitride semiconductor layer 124A. However, the third nitride semiconductor layer 126B may also include tensile strain, and the effect of the tensile strain in the third nitride semiconductor layer 126B on the compressive strain of the second nitride semiconductor layer 126A parallel to the first main surface 161 is small.

[0068] The other components of the fourth embodiment are the same as those of the first embodiment.

[0069] In the fourth embodiment, similar to the first embodiment, the sheet resistance can also be reduced by increasing electron mobility and electron density. Furthermore, by having a third nitride semiconductor layer 126B in the barrier layer 126, gate leakage can be reduced, and good flatness can be easily obtained on the upper surface of the barrier layer 126.

[0070] (Fifth Implementation) The fifth embodiment will be described. The fifth embodiment differs from the fourth embodiment in the composition of the third nitride semiconductor layer. Figure 11 This is a diagram showing the composition of the barrier layer in the high electron mobility transistor of the fifth embodiment.

[0071] In the HEMT of the fifth embodiment, such as Figure 11 As shown, the third nitride semiconductor layer 126B comprises scandium (Sc) and aluminum (Al). The third nitride semiconductor layer 126B is, for example, an n-type ScAlN layer or a ScAlGaN layer. The composition of the third nitride semiconductor layer 126B is fixed from the lower surface to the upper surface. The first ratio R1 of the number of atoms N (Sc) to the number of atoms N (Al) in the second nitride semiconductor layer 126A is greater than the second ratio R2 of the number of atoms N (Sc) to the number of atoms N (Al) in the third nitride semiconductor layer 126B.

[0072] The other components of the fifth embodiment are the same as those of the fourth embodiment.

[0073] In the fifth embodiment, similar to the first embodiment, the sheet resistance can be reduced by increasing electron mobility and electron density. Furthermore, by having a third nitride semiconductor layer 126B in the barrier layer 126, gate leakage can be reduced, and good flatness can be easily obtained on the upper surface of the barrier layer 126. Moreover, compared to the case where the first ratio R1 and the second ratio R2 are equal, it is easier to stably perform crystal growth on the third nitride semiconductor layer 126B.

[0074] The strain of the channel layer and the strain of the barrier layer can be determined, for example, by observation using a high-resolution transmission electron microscope (TEM) at the atomic image level. More specifically, it can be estimated with high precision based on the deviation from the lattice constant of the original crystal (bulk phase) by averaging and calculating the interatomic spacing of the observed region.

[0075] The composition of the elements contained in the barrier layer can be inferred, for example, by X-ray photoelectron spectroscopy (XPS) or energy dispersive X-ray spectroscopy (EDX).

[0076] The embodiments have been described in detail above, but this disclosure is not limited to specific embodiments, and various modifications and alterations can be made within the scope of the claims.

Claims

1. A high electron mobility transistor, having: The channel layer has a first principal surface; and The barrier layer has a second main surface opposite to the first main surface. The channel layer has a first nitride semiconductor layer comprising gallium. The barrier layer has a second nitride semiconductor layer comprising scandium and aluminum. The channel layer includes tensile strain parallel to the first master surface.

2. The high electron mobility transistor according to claim 1, wherein, The number of scandium atoms contained in the second nitride semiconductor layer is greater than 0.22 times the number of aluminum atoms contained in the second nitride semiconductor layer.

3. The high electron mobility transistor according to claim 1 or 2, wherein, The second nitride semiconductor layer contains gallium.

4. The high electron mobility transistor according to claim 3, wherein, The number of gallium atoms contained in the second nitride semiconductor layer is greater than or equal to 0.3 times the sum of the number of scandium atoms contained in the second nitride semiconductor layer, the number of aluminum atoms contained in the second nitride semiconductor layer, and the number of gallium atoms contained in the second nitride semiconductor layer.

5. The high electron mobility transistor according to claim 1 or 2, wherein, The further away from the second main surface, the smaller the ratio of the number of scandium atoms in the second nitride semiconductor layer to the number of aluminum atoms in the second nitride semiconductor layer.

6. The high electron mobility transistor according to claim 1 or 2, wherein, The barrier layer has a third nitride semiconductor layer comprising aluminum. The second nitride semiconductor layer is located between the channel layer and the third nitride semiconductor layer. The third nitride semiconductor layer includes tensile strain parallel to the first main surface.

7. The high electron mobility transistor according to claim 6, wherein, The third nitride semiconductor layer contains scandium. The first ratio of the number of scandium atoms contained in the second nitride semiconductor layer to the number of aluminum atoms contained in the second nitride semiconductor layer is greater than the second ratio of the number of scandium atoms contained in the third nitride semiconductor layer to the number of aluminum atoms contained in the third nitride semiconductor layer.

8. The high electron mobility transistor according to claim 1 or 2, wherein, The second nitride semiconductor layer contains indium.

Images