Group iii nitride lamination structure and high electron mobility transistor

The Group III nitride stacked structure with a diffusion suppressing layer containing In effectively prevents impurity diffusion, enhancing the conductivity and reducing resistance fluctuations in high electron mobility transistors by controlling impurity concentration.

JP2024156332APending Publication Date: 2024-11-06SUMITOMO CHEM CO LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
JP2023070705
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-04-24
Publication Date
2024-11-06

AI Technical Summary

Technical Problem

The diffusion of impurities that increase electrical resistance from one layer to another in Group III nitride layered structures is not effectively suppressed, affecting the electrical properties of the structure.

Method used

A Group III nitride stacked structure with a diffusion suppressing layer composed of Group III nitride containing In is introduced, where the impurity concentration is lower on the third layer side than on the first layer side, and the impurity concentration in the third layer is maintained at a low level over a certain thickness, preventing impurity diffusion into the channel layer.

Benefits of technology

This structure effectively suppresses the diffusion of high-resistance impurities, maintaining the conductivity of the channel layer and reducing fluctuations in sheet resistance and threshold voltage, thereby improving the electrical insulation and performance of high electron mobility transistors.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024156332000001_ABST
    Figure 2024156332000001_ABST
Patent Text Reader

Abstract

To provide a technique capable of suppressing the diffusion of impurities that increase electrical resistance from a layer containing the impurities to other layers in a lamination structure of group III nitride.SOLUTION: A group III nitride lamination structure has a first layer formed of group III nitride, a second layer arranged on the first layer and formed of group III nitride containing In, and a third layer arranged on the second layer and formed of group III nitride, wherein the first layer contains impurities which electrically make the group III nitride higher in resistance in a boundary surface nearby a region of 50 nm in thickness at least from the boundary surface of the first layer with the second layer, the impurities decrease in concentration more on the third layer side than the first layer side across the second layer, and the density of the impurities reaches less than 2×1015 cm-3 continuously over a thickness of at least 3 nm in the third layer, so the third layer contains substantially none of the impurity.SELECTED DRAWING: Figure 3
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to III-nitride stack structures and high electron mobility transistors. [Background technology]

[0002] Impurities (e.g., iron) are added to Group III nitrides (e.g., gallium nitride) to increase electrical resistance (see, for example, Patent Document 1). In a Group III nitride stack structure, a technique capable of suppressing the diffusion of an impurity that increases electrical resistance from a layer containing the impurity to other layers is preferable from the viewpoint of improving the controllability of the electrical properties of the stack structure. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2019 / 181391 Summary of the Invention [Problem to be solved by the invention]

[0004] An object of the present invention is to provide a technique capable of suppressing diffusion of an impurity, which increases electrical resistance, from a layer containing the impurity to other layers in a Group III nitride stack structure. [Means for solving the problem]

[0005] According to one aspect of the present invention, a first layer made of a group III nitride; a second layer disposed on the first layer and made of an In-containing Group III nitride; a third layer disposed on the second layer and composed of a Group III nitride; having the first layer contains an impurity that makes the III nitride electrically resistive at least in a region near the interface between the first layer and the second layer, the region being a thickness of 50 nm from the interface; a concentration of the impurity is lower on the third layer side than on the first layer side across the second layer, In the third layer, the concentration of the impurity is 2×10 15 cm -3 The third layer is substantially free of the impurities. III-nitride stack structure is provided. Effect of the Invention

[0006] A technique is provided that can suppress the diffusion of an impurity that increases electrical resistance from a layer containing the impurity to other layers in a layered structure of group III nitrides. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a schematic cross-sectional view showing an example of a laminate structure according to an embodiment of the present invention. [Diagram 2] FIG. 2 is an overall SIMS profile of a sample of the first example of the embodiment. [Diagram 3] FIG. 3 is a partial SIMS profile of a sample of the first example of the embodiment. [Figure 4] FIG. 4 is an overall SIMS profile of a sample of the second example of the embodiment. [Diagram 5] FIG. 5 is a partial SIMS profile of a sample according to the second embodiment. [Figure 6] FIG. 6 is a global SIMS profile of a sample of the comparative morphology. [Figure 7] FIG. 7 is a partial SIMS profile of a sample of the comparative configuration. [Figure 8]Fig. 8(a) is a graph plotting the relationship between the In composition and the Fe concentration in the diffusion suppression layer in the samples of the embodiment and the comparative embodiment, showing a sample with a design value of the additive peak concentration C1 of 1 x 1017 cm-3. Fig. 8(b) is a graph plotting the relationship between the In composition and the Fe concentration in the diffusion suppression layer in the samples of the embodiment and the comparative embodiment, showing a sample with a design value of the additive peak concentration C1 of 1 x 1018 cm-3. [Figure 9] FIG. 9 is a SIMS profile showing the In amount distribution in a diffusion suppression layer having a designed In composition of 15.6% and a designed thickness of 2 nm. [Figure 10] FIG. 10(a) is a graph plotting the relationship between the In composition of the diffusion suppression layer and the sheet resistance in a HEMT, and FIG. 10(b) is a graph plotting the relationship between the In composition of the diffusion suppression layer and the threshold voltage in a HEMT. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] A III-nitride stacked structure 100 (hereinafter referred to as stacked structure 100) according to an embodiment of the present invention will be described. Fig. 1 is a schematic cross-sectional view showing an example of the stacked structure 100. The stacked structure 100 is characterized by including a diffusion suppression part 1 configured by stacking three III-nitride layers. The diffusion suppression part 1 will be described in detail later.

[0009] The stacked structure 100 may be in the form of a semiconductor wafer including the diffusion suppression portion 1, or in the form of a semiconductor element in which other members (electrodes, etc.) are further provided on the semiconductor wafer. Hereinafter, a high electron mobility transistor (HEMT) is exemplified as one embodiment of the stacked structure 100. Hereinafter, the stacked structure 100 will also be referred to as HEMT 100.

[0010] The HEMT 100 includes a wafer 110 and an electrode 120. The wafer 110 includes a substrate 10, a nucleation layer 20, a buffer layer 30, a diffusion suppression layer 40, a channel layer 50, a barrier layer 60, and a cap layer 70. The electrode 120 is disposed on the cap layer 70 and includes a source electrode 121, a gate electrode 122, and a drain electrode 123.

[0011] The substrate 10 is, for example, a silicon carbide (SiC) substrate. A stack of III-nitride layers from a nucleation layer 20 to a cap layer 70 is formed on the substrate 10. The nucleation layer 20 is disposed on the substrate 10 and is made of, for example, aluminum nitride (AlN). The buffer layer 30 is disposed on the nucleation layer 20 and is made of, for example, gallium nitride (GaN). The diffusion suppression layer 40 is disposed on the buffer layer 30 and is made of, for example, indium gallium nitride (InGaN). The channel layer 50 is disposed on the diffusion suppression layer 40 and is made of, for example, GaN. The barrier layer 60 is disposed on the channel layer 50 and is made of, for example, aluminum gallium nitride (AlGaN). The cap layer 70 is disposed on the barrier layer 60 and is made of, for example, GaN.

[0012] The HEMT 100 according to this embodiment is characterized by including a diffusion suppression section 1 configured by stacking a buffer layer 30, a diffusion suppression layer 40, and a channel layer 50. The buffer layer 30, the diffusion suppression layer 40, and the channel layer 50 are examples of a first layer, a second layer, and a third layer, respectively, which are the three Group III nitride layers that constitute the diffusion suppression section 1.

[0013] The buffer layer 30 contains impurities that make the III nitride electrically high-resistance (hereinafter, referred to as high-resistance impurities) at least in an interface vicinity region 33 (see the description of Figs. 2 to 7 described later), which is a region 50 nm thick from the interface with the diffusion suppression layer 40. As the high-resistance impurities, a transition metal is preferably used, and examples of the transition metal include iron (Fe) and manganese (Mn). The buffer layer 30 is divided into a buffer lower layer 31 and a buffer upper layer 32, as will be described in detail later.

[0014] The diffusion suppression layer 40 is disposed on the buffer layer 30 (directly on the buffer layer 30) and is a layer that suppresses the diffusion of high resistance impurities from the buffer layer 30 to the channel layer 50. In other words, the concentration of high resistance impurities is lower on the channel layer 50 side than on the buffer layer 30 side across the diffusion suppression layer 40. The diffusion suppression layer 40 is made of a group III nitride containing indium (In). The inventor of the present application speculates that, as one idea, the diffusion suppression layer 40 is made of a group III nitride containing In, has a larger lattice constant than the group III nitride constituting the buffer layer 30, and receives compressive strain from the buffer layer 30, thereby suppressing the diffusion of high resistance impurities present in the buffer layer 30 to the channel layer 50.

[0015] The channel layer 50 is disposed on the diffusion suppression layer 40 (directly on the diffusion suppression layer 40), that is, on the buffer layer 30 via the diffusion suppression layer 40, and thus is a layer that does not substantially contain high-resistance impurities resulting from the buffer layer 30. The definition of "substantially not containing high-resistance impurities" will be described later.

[0016] In the HEMT 100 according to this embodiment, a two-dimensional electron gas (2DEG) is generated in the channel layer 50 near the interface with the barrier layer 60, and the 2DEG functions as a channel of the HEMT. Since the channel layer 50 does not substantially contain high-resistance impurities, it is possible to suppress fluctuations in, for example, sheet resistance and, for example, threshold voltage caused by the high-resistance impurities.

[0017] A method for manufacturing the HEMT 100 will be described by way of example. For example, a SiC substrate is prepared as the substrate 10. Above the substrate 10, layers 20 to 70 made of group III nitrides are epitaxially grown by, for example, metal-organic vapor phase epitaxy (MOVPE). For example, Fe is added to the buffer layer 30 as a high-resistance impurity.

[0018] Among the group III source gases, the aluminum (Al) source gas may be, for example, trimethylaluminum (Al(CH3)3, TMA) gas. Among the group III source gases, the gallium (Ga) source gas may be, for example, trimethylgallium (Ga(CH3)3, TMG) ​​gas. Among the group III source gases, the indium (In) source gas may be, for example, trimethylindium (In(CH3)3, TMI) gas. Among the Fe doping source gases, for example, Fe(C2H5)2(Cp2Fe) gas is used. Among the group V source gases, the nitrogen (N) source gas may be, for example, ammonia (NH3). As the carrier gas, for example, at least one of nitrogen gas (N2 gas) and hydrogen gas (H2 gas) is used.

[0019] The growth temperature can be selected, for example, in the range of 700°C to 1400°C, and the V / III ratio, which is the flow rate ratio of the Group V source gas to the Group III source gas, can be selected, for example, in the range of 10 to 5000. The ratio of the supply amount of each source gas is adjusted according to the composition of each layer to be formed. The thickness of each layer to be formed can be controlled by the growth time, for example, by calculating the growth time corresponding to the designed thickness from the growth rate obtained in a preliminary experiment.

[0020] For example, an AlN layer is grown on the substrate 10 to a design thickness of 6 nm to 40 nm to form the nucleation layer 20. For example, a GaN layer is grown on the nucleation layer 20 to a design thickness of 400 nm to 3000 nm to form the buffer layer 30.

[0021] GaN is grown from near the bottom surface of the buffer layer 30 to an intermediate thickness while supplying Fe doping source gas, thereby forming a buffer lower layer 31. GaN is grown from the intermediate thickness to the top surface of the buffer layer 30 after the supply of Fe doping source gas is stopped, thereby forming a buffer upper layer 32. The designed thickness of the buffer lower layer 31 is, for example, 200 nm or more and 2000 nm or less, and the designed thickness of the buffer upper layer 32 is, for example, 200 nm or more and 1000 nm or less.

[0022] Fe is added to the lower buffer layer 31, but diffuses to the upper surface of the upper buffer layer 32. As a result, the buffer layer 30 contains Fe from the lower buffer layer 31 to the upper buffer layer 32. As described later, the wafer 110 is measured by secondary ion mass spectrometry (SIMS) to obtain the Fe concentration distribution in the thickness direction in the diffusion suppression section 1 (buffer layer 30, diffusion suppression layer 40, and channel layer 50). The Fe concentration distribution in the buffer layer 30 has a peak Fe concentration near the growth thickness at which the supply of Fe doping source gas is stopped. This peak Fe concentration is called an added peak concentration C1. The added peak concentration C1 is preferably 1×10 17 cm -3 More than 1×10 19 cm -3 Fe is added to the buffer layer 30 as follows:

[0023] The diffusion suppression layer 40 is formed on the buffer layer 30 by growing, for example, an InGaN layer with a design thickness of 1 nm to 10 nm. The channel layer 50 is formed on the diffusion suppression layer 40 by growing, for example, a GaN layer with a design thickness of 5 nm to 100 nm.

[0024] The In composition of the InGaN constituting the diffusion suppression layer 40 is set to a level such that the channel layer 50 does not substantially contain Fe. As described later, the In composition for the channel layer 50 to be substantially free of Fe can vary depending on the Fe concentration of the buffer layer 30, but as a guideline, it is more than 15%, and preferably more than 16%.

[0025] For example, an AlGaN layer is grown to a design thickness of 5 nm to 80 nm on the channel layer 50 to form a barrier layer 60. For example, a GaN layer is grown to a design thickness of 10 nm or less on the barrier layer 60 as necessary to form a cap layer 70. In this manner, the wafer 110 is manufactured.

[0026] Thereafter, electrodes 120 (source electrode 121, gate electrode 122, and drain electrode 123) and, if necessary, other components such as insulating films are formed on wafer 110 using known techniques related to HEMT manufacturing, thereby manufacturing HEMT 100.

[0027] The diffusion suppression portion 1 will be further described below with reference to SIMS measurement results (SIMS profiles) for the sample of the embodiment and the sample of the comparative embodiment. SIMS measurement was performed on the samples of the embodiment and the comparative embodiment to obtain the thickness direction distributions of Fe concentration, Ga amount, Al amount, and In amount. The Fe concentration distribution, Ga amount distribution, Al amount distribution, and In amount distribution were measured using a magnetic field type SIMS device. However, the measurement of the In amount distribution for quantifying the In composition was performed using a quadrupole type SIMS device (see FIG. 9).

[0028] 2 and 3 are SIMS profiles of a sample of a first embodiment, FIGs. 4 and 5 are SIMS profiles of a sample of a second embodiment, and FIGs. 6 and 7 are SIMS profiles of a sample of a comparative embodiment. FIGs. 2, 4, and 6 are SIMS profiles of the entire diffusion suppression portion 1 in the HEMT 100, including the entire thickness of the buffer layer 30. FIGs. 3, 5, and 7 are SIMS profiles of parts of the diffusion suppression portion 1 in the HEMT 100, showing the vicinity of the diffusion suppression layer 40.

[0029] Figures 2, 4, and 6 show the Al amount distribution together with the Fe concentration distribution. The Fe concentration distribution is shown by a solid line, and the Al amount distribution is shown by a dotted line. In Figures 2, 4, and 6, the left axis shows the Fe concentration, and the right axis shows the secondary ion intensity related to Al.

[0030] 3, 5, and 7 show the In amount distribution together with the Fe concentration distribution. The Fe concentration distribution is shown by a solid line, and the In amount distribution is shown by a dotted line. In FIG. 3, 5, and 7, the left axis shows the Fe concentration. FIG. 3, 5, and 7 show the shape of the In amount distribution, and the In amount is shown in an arbitrary unit. Here, the In amount is shown on a linear scale, not a logarithmic scale. A calculation method of the In composition based on SIMS measurement will be described later with reference to FIG. 9.

[0031] The nucleation layer 20 is made of AlN, and the buffer layer 30 is made of GaN. The thickness position where the amount of Al on the buffer layer 30 side is reduced to half of the peak amount of Al in the nucleation layer 20 is defined as the interface between the nucleation layer 20 and the buffer layer 30.

[0032] The buffer layer 30 is made of GaN, the diffusion suppression layer 40 is made of InGaN, and the channel layer 50 is made of GaN. The thickness position where the amount of In on the buffer layer 30 side is reduced to half of the peak amount of In in the diffusion suppression layer 40 is defined as the interface between the buffer layer 30 and the diffusion suppression layer 40. The thickness position where the amount of In on the channel layer 50 side is reduced to half of the peak amount of In in the diffusion suppression layer 40 is defined as the interface between the diffusion suppression layer 40 and the channel layer 50.

[0033] The channel layer 50 is made of GaN, and the barrier layer 60 is made of AlGaN. The thickness position where the amount of Al on the channel layer 50 side is reduced to half of the peak amount of Al in the barrier layer 60 is defined as the interface between the channel layer 50 and the barrier layer 60.

[0034] The Fe concentration distribution in the buffer layer 30 has a peak Fe concentration (peak additive concentration C1) near the growth thickness at which the supply of the Fe doping source gas is stopped. In the buffer layer 30, the thickness position at which the peak Fe additive concentration C1 exists is used as a boundary, and the lower side is defined as the lower buffer layer 31 and the upper side is defined as the upper buffer layer 32. A 50 nm thick region on the buffer layer 30 side from the interface between the buffer layer 30 and the diffusion suppression layer 40 is defined as a region 33 of the buffer layer 30 near the interface with the diffusion suppression layer 40.

[0035] The thickness of each layer can be obtained based on the interface position and boundary position obtained from the SIMS measurement. Such thickness of each layer obtained from the SIMS measurement may not match the design thickness. For example, the thickness of the diffusion suppression layer 40 obtained from the SIMS measurement tends to be several times thicker than the design thickness of the diffusion suppression layer 40. Since the design thickness, which is a numerical value assumed at the time of design, cannot be known from the fabricated wafer 110 itself, the interface position and boundary position obtained from the SIMS measurement and the thickness of each layer are used here.

[0036] In both the first embodiment shown in FIG. 2 and FIG. 3 and the comparative embodiment shown in FIG. 6 and FIG. 7, the design value of the added peak concentration C1 is set to 1×10 17 cm -3 As an example, this is a sample in which Fe is added to the buffer layer 30. However, the first embodiment and the comparative example differ in the In composition of the InGaN that constitutes the diffusion suppression layer 40. The designed In composition of the diffusion suppression layer 40 in the comparative example is low at 10.8%, while the designed In composition of the diffusion suppression layer 40 in the first embodiment is high at 15.6%.

[0037] In the second embodiment shown in FIG. 4 and FIG. 5, the design value of the added peak concentration C1 is set to 1×10 18 cm -3 As a result, the buffer layer 30 was doped with Fe, and the In design composition of the diffusion suppression layer 40 was high at 27.1%.

[0038] In the first and second embodiments and the comparative embodiment, the distributions of Fe concentration, Al amount, and In amount obtained by SIMS measurement have similar characteristics as follows. The Al amount decreases from the nucleation layer 20 toward the buffer layer 30 and is almost constant in the buffer layer 30. The Fe concentration tends to be observed at a high concentration in the nucleation layer 20, decreases as the Al amount decreases toward the buffer layer 30, and reaches a minimum value near the bottom surface of the buffer layer 30.

[0039] The high Fe concentration observed near the interface between the buffer layer 30 and the nucleation layer 20 is not due to the addition of Fe to the buffer layer 30, but is considered to be due to the nucleation layer 20 containing Al. For this reason, here, the portion with high Fe concentration near the interface due to the nucleation layer 20 (the portion on the nucleation layer 20 side of the minimum value of the Fe concentration distribution) is excluded from the Fe concentration distribution in the buffer layer 30. In other words, the Fe concentration distribution above the minimum value is treated as the Fe concentration distribution in the buffer layer 30.

[0040] The Fe concentration reaches a minimum near the bottom surface of buffer layer 30, i.e., near the bottom surface of buffer lower layer 31, and then increases upward as Fe is added to buffer lower layer 31, reaching a peak addition concentration C1 at the top surface of buffer lower layer 31.

[0041] The Fe concentration decreases upward from the lower surface of buffer upper layer 32, starts to increase in near-interface region 33, increases in near-interface region 33 toward diffusion suppression layer 40, and reaches a peak near the interface between buffer upper layer 32 and diffusion suppression layer 40. The concentration at which Fe starts to increase in near-interface region 33, i.e., the minimum value of the Fe concentration in near-interface region 33, is referred to as near-interface minimum concentration C2. The peak Fe concentration near the interface between buffer upper layer 32 and diffusion suppression layer 40 is referred to as interface peak concentration C3.

[0042] After reaching the interface peak concentration C3, the Fe concentration decreases toward the channel layer 50. The In amount peaks at approximately the center of the diffusion suppression layer 40, and the In amount distribution is approximately symmetrical in a bell shape on the buffer layer 30 side and the channel layer 50 side.

[0043] The Fe concentration measured by SIMS in each sample is as follows: In the first embodiment, the added peak concentration C1 is 1.54×10 17 cm -3 The minimum concentration C2 near the interface is 5.47×10 16 cm -3 and the interface peak concentration C3 is 7.96 × 10 17 cm -3 In the second embodiment, the peak concentration C1 is 1.1×10 18 cm -3 The minimum concentration C2 near the interface is 4.85×10 17 cm -3 and the interface peak concentration C3 is 1.04 × 10 19 cm -3 In the comparative example, the peak concentration C1 is 1.09×10 17 cm -3 The minimum concentration C2 near the interface is 9.9×10 15 cm -3 and the interface peak concentration C3 is 7.12 × 10 17 cm -3 It is.

[0044] Here, the first embodiment and the comparative embodiment have significantly different characteristics of the Fe concentration distribution in the channel layer 50. In the first embodiment (see FIG. 3 in particular), the Fe concentration is below the lower detection limit (4.44×10 14 cm -3 ), the low level that reaches the detection limit is maintained until the vicinity of the upper surface of the channel layer 50. That is, in the first example of the embodiment, the diffusion of Fe originating from the buffer layer 30 into the channel layer 50 is prevented.

[0045] In contrast, in the comparative example (see FIG. 7 in particular), the Fe concentration is somewhat reduced in the channel layer 50, but is not so reduced that it maintains a stable low level within the thickness range of the channel layer 50. In other words, in the comparative example, the diffusion of Fe into the channel layer 50 due to the buffer layer 30 is not prevented.

[0046] In this way, in both the first embodiment and the comparative example, the design value of the added peak concentration C1 is set to 1×10 17 cm -3 In the first embodiment, Fe diffusion from the buffer layer 30 to the channel layer 50 is prevented, whereas in the comparative example, Fe diffusion from the buffer layer 30 to the channel layer 50 is not prevented. This difference is believed to be due to the fact that the In composition of the diffusion suppression layer 40 is higher in the first embodiment (In design composition 15.6%) than in the comparative example (In design composition 10.8%).

[0047] In the second embodiment (see FIG. 5 in particular), the Fe concentration was below the detection limit (1.22×10 15 cm -3 ), the low level that reaches the detection limit is generally maintained up to the upper surface of the channel layer 50. Some parts of the channel layer 50 where the Fe concentration is slightly high are observed in the middle of the thickness, but these are considered to be noise. In the second embodiment, as in the first embodiment, the diffusion of Fe into the channel layer 50 due to the buffer layer 30 is prevented. This is considered to be due to the fact that the In composition of the diffusion suppression layer 40 is high in the second embodiment (In design composition 27.1%).

[0048] In the channel layer 50 of the first and second embodiments, a characteristic is observed in which the Fe concentration continuously reaches the detection limit over a certain thickness range. From this, it is concluded that the channel layer 50 does not substantially contain Fe.

[0049] The detection limit of Fe concentration may vary slightly for each SIMS measurement. Here, we define a low Fe concentration level close to the detection limit of Fe concentration as 2×10 15 cm -3 (A lower guideline is 1.5×10 15 cm -3 The above "certain thickness range" is set as a guideline of 3 nm (5 nm as a thicker guideline).

[0050] Using these guidelines, the Fe concentration in the channel layer 50 is set to 2×10 15 cm -3 Less than (preferably 1.5 × 10 15 cm -3 When the Fe content reaches less than 100%, it is defined that the channel layer 50 is substantially free of Fe.

[0051] In addition to the samples whose SIMS profiles are shown in Figures 2 to 7, various samples were produced as embodiments and comparative examples. Figures 8(a) and 8(b) are graphs plotting the relationship between the In composition and the Fe concentration of the diffusion suppression layer 40 in these samples. Figure 8(a) shows the results when the design value of the additive peak concentration C1 was 1 x 10 17 cm -3 Fig. 8(b) shows a sample with a design value of the added peak concentration C1 of 1×10 18 cm -3 The sample is shown.

[0052] 8(a) and 8(b), the horizontal axis represents the In design composition, and the vertical axis represents the Fe concentration. The Fe interface peak concentration C3 ([Fe] at peak) is shown by a square plot, and the average Fe concentration in the channel layer 50 ([Fe] at Channel) is shown by a circle plot.

[0053] The average Fe concentration in the channel layer 50 is defined as the average Fe concentration in a range of 10 nm above the interface with the diffusion suppression layer 40. 15 cm -3When the Fe content is less than 100%, it is judged as Fe not detected (ND). The cases where Fe was not detected are shown with open circles.

[0054] The measurement shown in Fig. 8(a), that is, the design value of the added peak concentration C1 is 1 × 10 17 cm -3 For the measurement, samples were prepared with In design compositions of 0% (i.e., no diffusion suppression layer 40 was provided), 5.9%, 10.8%, and 15.6%. The sample with an In design composition of 10.8% corresponds to the comparative embodiment shown in Figs. 6 and 7, and the sample with an In design composition of 15.6% corresponds to the first example of the embodiment shown in Figs. 2 and 3.

[0055] In the sample with 0% In design composition, the characteristic of the formation of an interface peak concentration C3 of Fe near the interface between the buffer layer 30 and the diffusion suppression layer 40 and the decrease in the Fe concentration on the channel layer 50 side, which was observed in the sample with the diffusion suppression layer 40, was not observed. For this reason, the interface peak concentration C3 was not plotted for the sample with 0% In design composition, and the average Fe concentration in the channel layer 50 was 1×10, which is approximately the same as the design value of the added peak concentration C1. 17 cm -3 It is.

[0056] The Fe interface peak concentration C3 in the samples with In design compositions of 5.9%, 10.8%, and 15.6% was 1×10 18 cm -3 , 7×10 17 cm -3 , and 8×10 17 cm -3 and the average Fe concentration in the channel layer 50 is 7×10 15 cm -3 , 7×10 15 cm -3 , and 4×10 14 cm -3 It is.

[0057] In the samples with In design compositions of 5.9%, 10.8%, and 15.6%, the Fe interface peak concentration C3 is approximately the same, but by increasing the In design composition to 15.6%, it is found that Fe is not detected in the channel layer 50. The determination that Fe is not detected based on the average Fe concentration in the channel layer 50 of the sample with an In design composition of 15.6% (first example of the embodiment) is consistent with the determination that Fe is not substantially contained based on the distribution shape of the SIMS profile as described above.

[0058] The measurement shown in Fig. 8(b), that is, the design value of the added peak concentration C1 is 1 × 10 18 cm -3 For the measurement, samples were prepared with In design compositions of 0% (i.e., no diffusion suppression layer 40 was provided), 15.6%, 19.9%, and 27.1%. The sample with an In design composition of 27.1% corresponds to the second example of the embodiment shown in FIGS. 4 and 5.

[0059] For the sample with 0% In design composition, the interface peak concentration C3 is not plotted, and the average Fe concentration in the channel layer 50 is 4×10 17 cm -3 It is.

[0060] The Fe interface peak concentration C3 in the samples with In design compositions of 15.6%, 19.9%, and 27.1% was 1×10 19 cm -3 , 1×10 19 cm -3 , and 1×10 19 cm -3 and the average Fe concentration in the channel layer 50 is 7×10 15 cm -3 , 1×10 15 cm -3 , and 2×10 15 cm -3 It is.

[0061] In the samples with In design compositions of 15.6%, 19.9%, and 27.1%, the interface peak concentration C3 of Fe is approximately the same, but by increasing the In design composition to 19.9% ​​and 27.1%, it is found that Fe is not detected in the channel layer 50. The determination that Fe is not detected based on the average Fe concentration in the channel layer 50 of the sample with an In design composition of 27.1% (second example of the embodiment) is consistent with the determination that Fe is not substantially contained based on the distribution shape of the SIMS profile as described above.

[0062] The sample with an In design composition of 19.9% ​​is referred to as the third embodiment. As with the first and second embodiment samples, the third embodiment sample was also determined to contain substantially no Fe based on the distribution shape of the SIMS profile, and the determination based on the average Fe concentration and the determination based on the distribution shape of the SIMS profile are consistent.

[0063] As described above, by increasing the In composition in the diffusion suppression layer 40, it is possible to prevent the diffusion of Fe into the channel layer 50 due to the buffer layer 30. The In composition required to prevent the diffusion of Fe into the channel layer 50 varies depending on the Fe concentration added to the buffer layer 30, more specifically, the Fe concentration near the interface between the buffer layer 30 and the diffusion suppression layer 40 (interface peak concentration C3 or interface minimum concentration C2). The higher the Fe concentration in the buffer layer 30, the higher the In composition of the diffusion suppression layer 40 is preferable.

[0064] In other words, it is preferable that the In composition of the diffusion suppression layer 40 is appropriately increased according to the Fe concentration of the buffer layer 30 so as to prevent the diffusion of Fe into the channel layer 50. A specific guideline for the In composition of the diffusion suppression layer 40 is more than 15%, and preferably more than 16%.

[0065] From the viewpoint of preventing the diffusion of Fe into the channel layer 50, the upper limit of the In composition of the diffusion suppression layer 40 is not particularly limited, but a numerical value of (at least) 31.4% can be cited as one guideline. This is because good results were obtained in measurements of the HEMT characteristics described below with a sample having an In design composition of 31.4%.

[0066] The designed In composition is a value assumed at the time of design, and cannot be known from the fabricated wafer 110 itself. However, it is believed that an In composition close to the designed composition is realized in a properly fabricated diffusion suppression layer 40. Here, a method for calculating the In composition based on SIMS measurement will be described.

[0067] 9 is a SIMS profile showing the In amount distribution of the diffusion suppression layer 40 having a designed In composition of 15.6% and a designed thickness of 2 nm. The left axis shows the In composition. As described above, the In amount distribution for quantifying the In composition is measured by a quadrupole SIMS device.

[0068] The total amount of In contained in the diffusion suppression layer 40 corresponds to the total area of ​​the In amount distribution. In this method, first, the In amount distribution is fitted with a Gaussian distribution. The actually measured In amount distribution is shown by a solid line, and the fitted Gaussian distribution is shown by a circle plot.

[0069] The parameters of the fitted Gaussian distribution are then used to calculate the area of ​​the Gaussian distribution. In this example, the fitted Gaussian distribution has a height of 0.0773, a center value of 54.6, a Gaussian width of 1.56, and an area of ​​0.303.

[0070] A rectangle having an area equal to the area of ​​the fitted Gaussian distribution and a width of the design thickness of 2 nm is shown by a dotted line. The height of the rectangle is considered to indicate the In composition in a mode in which the diffusion suppression layer 40 is formed with a uniform composition within the range of the design thickness of 2 nm. In this example, the height of the rectangle is 15.1% (area 0.303 / thickness 2 nm), which is approximately equal to the In design composition of 15.6%.

[0071] When the design thickness is unknown, the thickness of the diffusion suppression layer 40 measured by observation with a transmission electron microscope is treated as the design thickness, and the In composition based on SIMS measurement can be calculated by dividing the area of ​​the fitted Gaussian distribution by this thickness.

[0072] Hereinafter, detailed characteristics of the diffusion suppression portion 1 will be described by way of example. As described above, by using the diffusion suppression layer 40 of this embodiment, it is possible to prevent the diffusion of high resistance impurities resulting from the buffer layer 30 into the channel layer 50. This makes it possible to increase the concentration of high resistance impurities in the interface vicinity region 33 of the buffer layer 30 (i.e., the region near the top surface of the buffer layer 30) while suppressing a decrease in the conductivity of the channel layer 50, thereby improving electrical insulation. In the example of the HEMT 100, it is possible to suppress the leakage current of the HEMT flowing through the buffer layer 30.

[0073] The minimum concentration C2 near the interface (the minimum value of the high resistance impurity concentration in the interface near region 33) is a concentration slightly lower than the peak concentration C1. The peak concentration C1 is preferably, for example, 1×10 17 cm -3 The minimum concentration C2 near the interface is preferably 2×10 16 cm -3 That's all.

[0074] In order to make the interface vicinity minimum concentration C2 high, it is preferable that the interface vicinity minimum concentration C2 is not excessively decreased from the additive peak concentration C1. The ratio (C2 / C1) of the interface vicinity minimum concentration C2 to the additive peak concentration C1 is, for example, preferably 1 / 30 or more (approximately 0.033 or more), and more preferably 1 / 10 or more (0.1 or more). The additive peak concentration C1 can be said to be a peak concentration that exists in a region on the lower surface side of the buffer layer 30 relative to the interface vicinity region 33 and is higher than the interface vicinity minimum concentration C2.

[0075] In the first embodiment, the peak concentration C1 is 1.54×10 17 cm-3 The minimum concentration C2 near the interface is 5.47×10 16 cm -3 In the second embodiment, the peak concentration C1 is 1.1×10 18 cm -3 The minimum concentration C2 near the interface is 4.85×10 17 cm -3 The ratio (C2 / C1) of the minimum concentration C2 near the interface to the added peak concentration C1 is 0.355 in the first embodiment, and 0.441 in the second embodiment.

[0076] The thicker the buffer upper layer 32, the lower the minimum concentration C2 near the interface becomes relative to the peak concentration C1. In order not to excessively lower the minimum concentration C2 near the interface from the peak concentration C1, it is preferable to make the thickness of the buffer upper layer 32 thinner. For example, the thickness of the buffer upper layer 32 is thinner than the thickness of the buffer lower layer 31. For example, the thickness of the buffer upper layer 32 is thinner than the thickness of the buffer lower layer 31. For example, the thickness of the buffer upper layer 32 is 250 nm or less. Here, the thickness of the buffer lower layer 31 can be said to be the thickness of the buffer layer 30 from the bottom surface of the buffer layer 30 to the thickness position where the peak concentration C1 exists. Also, the thickness of the buffer upper layer 32 can be said to be the thickness of the buffer layer 30 from the thickness position where the peak concentration C1 exists to the interface with the diffusion suppression layer 40.

[0077] The fact that the diffusion of the high-resistance impurity into the channel layer 50 is suppressed can also be evaluated by comparing the minimum concentration C2 near the interface with the high-resistance impurity concentration in the channel layer 50. For example, a feature can be given that the high-resistance impurity concentration on the upper surface of the channel layer 50 is lower than the minimum concentration C2 near the interface.

[0078] As described below, the interface peak concentration C3 tends to be about 15 to 20 times higher than the interface vicinity minimum concentration C2 in the first and second examples of the embodiment. As the interface vicinity minimum concentration C2 is higher, the interface peak concentration C3 also becomes higher. The interface vicinity minimum concentration C2 is preferably set to, for example, 2×10 16 cm -3 By setting the interface peak concentration C3 at 3×10 or more, the interface peak concentration C3 is preferably set at 3×10 17 cm -3 That's all.

[0079] In the first embodiment, the interface peak concentration C3 is 7.96×10 17 cm -3 and the minimum concentration near the interface C2 (5.47×10 16 cm -3 ) is about 15 times. In the second embodiment, the interface peak concentration C3 is 1.04×10 19 cm -3 and the minimum concentration near the interface C2 (4.85×10 17 cm -3 ) is about 20 times as large.

[0080] The diffusion suppression layer 40 according to the present embodiment has a high barrier property against high resistance impurities. As a result, the interface peak concentration C3 is 3×10 17 cm -3 Even if the impurity concentration for increasing resistance in the channel layer 50 is increased to 2×10 15 cm -3 It can be as follows:

[0081] In other words, the high resistance impurity concentration is the interface peak concentration C3 (3×10 17 cm -3 or more) to 2 x 10 15 cm -3 The reduction ratio of the interface peak concentration C3 to 15 cm -3 Ratio (2×10 15 / C3) can be set to 1 / 150 or less (approximately 0.0067 or less).

[0082] In the first embodiment, the interface peak concentration C3 (7.96 × 10 17 cm -3 ) for 2×10 15 cm -3 In the second embodiment, the interface peak concentration C3 (1.04×10 19 cm -3 ) for 2×10 15 cm -3 The ratio is 0.000192.

[0083] Furthermore, the diffusion suppression layer 40 according to this embodiment has the function of rapidly decreasing the high resistance impurity concentration upward in the thickness direction. As described below, in the first and second examples of the embodiment, the rapid decrease in the high resistance impurity concentration typically occurs within a thickness of about 12 nm. The interface peak concentration C3 (3×10 17 cm -3 or more) to 2 x 10 15 cm -3 If the decrease in thickness occurs over a 12-nm thickness, the average decrease in slope is 2.5×10 16 cm -3 / nm or more.

[0084] In the first embodiment, the interface peak concentration C3 (7.96 × 10 17 cm -3 ) to 2 × 10 15 cm -3 The average slope of the decrease is 7.10×10 16 cm -3 / nm. In the second embodiment, the interface peak concentration C3 (1.04 × 10 19 cm -3 ) to 2 × 10 15 cm -3 The average slope of the decrease is 7.88×10 17 cm -3 / nm.

[0085] When the ratio of the minimum concentration C2 near the interface to the peak concentration C1 is large, for example 1 / 10 or more, that is, when the minimum concentration C2 near the interface is not significantly lower than the peak concentration C1, the peak concentration C3 at the interface is higher than the peak concentration C1 at the doping level. Since the interface of the buffer layer 30 with the diffusion suppression layer 40 is located near the peak concentration C3 at the interface, the high resistance impurity concentration in the buffer layer 30 is highest at the interface with the diffusion suppression layer 40. The SIMS profiles of the first and second examples of the embodiment show such an aspect.

[0086] In the SIMS profiles of the first and second embodiments, the interface peak concentration C3 is located slightly toward the diffusion suppression layer 40 from the interface between the buffer layer 30 and the diffusion suppression layer 40, that is, within the diffusion suppression layer 40. Accordingly, the interface of the buffer layer 30 with the diffusion suppression layer 40 is located midway along the slope where the high resistance impurity concentration increases toward the diffusion suppression layer 40 in the interface vicinity region 33, and the high resistance impurity concentration in the buffer layer 30 is at its highest value at that interface.

[0087] Due to an error in the SIMS measurement, it is also possible that the interface peak concentration C3 is located slightly toward the buffer layer 30 from the interface between the buffer layer 30 and the diffusion suppression layer 40, i.e., within the buffer layer 30. In such a case, the interface between the buffer layer 30 and the diffusion suppression layer 40 is located at a thickness position slightly above the thickness position of the interface peak concentration C3, where the high resistance impurity concentration is slightly lower than the interface peak concentration C3. However, even in such a case, the interface of the buffer layer 30 with the diffusion suppression layer 40 is still located near the interface peak concentration C3, so it can be said that the high resistance impurity concentration in the buffer layer 30 is highest at the interface with the diffusion suppression layer 40.

[0088] The following is an exemplary description of the characteristics of the HEMT 100 including the diffusion suppression section 1. Specifically, the characteristics of the HEMT 100 relating to the sheet resistance and threshold voltage are described. Fig. 10(a) is a graph plotting the relationship between the In composition of the diffusion suppression layer 40 in the HEMT 100 and the sheet resistance, with the horizontal axis indicating the In design composition and the vertical axis indicating the sheet resistance. Fig. 10(b) is a graph plotting the relationship between the In composition of the diffusion suppression layer 40 in the HEMT 100 and the threshold voltage, with the horizontal axis indicating the In design composition and the vertical axis indicating the threshold voltage.

[0089] In Fig. 10(a) and Fig. 10(b), the embodiment in which no Fe was added (w / o Fe doping) is shown by a circle plot, and the design value of the added peak concentration C1 of Fe is 1 × 10 17 cm -3 The aspect ([Fe] = 1e17 cm-3) is shown by a square plot, and the design value of the added peak concentration C1 of Fe is 1 × 10 18 cm -3 The aspect ([Fe]=1e18 cm−3) is shown by the diamond plot.

[0090] No Fe was added. The design value of the Fe peak concentration C1 was 1×10 17 cm -3 (Hereinafter, "Fe concentration is 1×10 17 cm -3 "), and the design value of the added peak concentration C1 of Fe is 1×10 18 cm -3 (Hereinafter, "Fe concentration is 1×10 18 cm -3 In each of the embodiments (hereinafter referred to as "without diffusion barrier layer 40"), samples were fabricated with various In design compositions of the diffusion barrier layer 40. The In design composition values ​​shown in FIG. 10(a) are 0% (i.e., no diffusion barrier layer 40 was provided), 5.9%, 10.8%, 15.6%, 18.6%, 19.9%, 27.1%, and 31.4%.

[0091] Here, the Fe concentration is 1×10 17 cm -3 In this embodiment, the interface peak concentration C3 is, for example, 3×1017 cm -3 (Alternatively, the minimum concentration C2 near the interface is, for example, 2×10 16 cm -3 (The Fe concentration is 1×10 18 cm -3 In this embodiment, the interface peak concentration C3 is, for example, 3×10 18 cm -3 (Alternatively, the minimum concentration C2 near the interface is, for example, 2×10 17 cm -3 corresponds to the above).

[0092] The sheet resistance (hereinafter referred to as the reference sheet resistance) and threshold voltage (hereinafter referred to as the reference threshold voltage) of a sample with no added Fe and with a 0% In design composition (i.e., a HEMT in which the diffusion suppression layer 40 is omitted and the buffer layer 30 does not contain high-resistance impurities) are used as the evaluation criteria.

[0093] The characteristics related to the sheet resistance will be described with reference to Fig. 10(a). In the embodiment in which Fe is not added, even if the In composition of the diffusion suppression layer 40 is increased, the sheet resistance does not change particularly from the reference sheet resistance. In other words, providing the diffusion suppression layer 40 made of a group III nitride containing In does not particularly change the sheet resistance.

[0094] Fe addition state (Fe concentration is 1×10 17 cm -3 and the Fe concentration is 1×10 18 cm -3 In the embodiment, the sheet resistance of the sample with an In design composition of 0% is significantly higher than the reference sheet resistance. 18 cm -3 In the embodiment, the Fe concentration is 1×10 17 cm -3 The sheet resistance is significantly higher than that of the embodiment in 2. This is believed to be because Fe originating from the buffer layer 30 diffuses directly into the channel layer 50 without passing through the diffusion suppression layer 40, thereby decreasing the conductivity of the channel layer 50.

[0095] Fe addition state (Fe concentration is 1×10 17 cm -3 and the Fe concentration is 1×10 18 cm -3 In each of the above embodiments, the sheet resistance decreases and approaches the reference sheet resistance as the In composition of the diffusion suppression layer 40 increases. This is believed to be because the effect of suppressing the diffusion of Fe into the buffer layer 30 increases as the In composition of the diffusion suppression layer 40 increases, thereby suppressing the decrease in conductivity of the channel layer 50.

[0096] The standard sheet resistance is 430 ohm / sq. Fe concentration is 1×10 17 cm -3 and the Fe concentration is 1×10 18 cm -3 In the case of the sample with 0% In design composition, the sheet resistance is 478 ohm / sq and 597 ohm / sq, respectively. From the viewpoint of suppressing the fluctuation of the sheet resistance, for example, a sheet resistance less than 1.1 times the reference sheet resistance (less than 473 ohm / sq in this example) is determined to be good.

[0097] From Fig. 10(a), the Fe concentration is 1×10 17 cm -3 (more specifically, the interface peak concentration C3 is 3×10 17 cm -3 or more, or the minimum concentration C2 near the interface is 2×10 16 cm -3 Even in the above embodiment, the Fe concentration is 1×10 18 cm -3 (more specifically, the interface peak concentration C3 is 3×10 18 cm -3 or more, or the minimum concentration C2 near the interface is 2×10 17 cm -3 It can be seen that even in the above embodiment, by increasing the In composition of the diffusion suppression layer 40, the sheet resistance can be made less than 1.1 times the reference sheet resistance.

[0098] The characteristics related to the threshold voltage will be described with reference to Fig. 10(b). In the embodiment in which Fe is not added, even if the In composition of the diffusion suppression layer 40 is increased, the threshold voltage does not change from the reference threshold voltage. In other words, providing the diffusion suppression layer 40 made of a group III nitride containing In does not change the threshold voltage.

[0099] Fe addition state (Fe concentration is 1×10 17 cm -3 and the Fe concentration is 1×10 18 cm -3 In the embodiment, in the sample with an In design composition of 0%, the magnitude of the threshold voltage is significantly smaller than the magnitude of the reference threshold voltage. 18 cm -3 In the embodiment, the Fe concentration is 1×10 17 cm -3 In comparison with the embodiment of FIG. 1, the magnitude of the threshold voltage is smaller.

[0100] The threshold voltage is a negative voltage for turning the HEMT 100 off, and a small magnitude (absolute value) of the threshold voltage means that the threshold voltage is close to 0 V. In other words, a smaller magnitude of the threshold voltage compared to the reference threshold voltage means that the HEMT 100 is easier to turn off. This is believed to be because Fe originating from the buffer layer 30 diffuses directly into the channel layer 50 without passing through the diffusion suppression layer 40, thereby reducing the conductivity of the channel layer 50.

[0101] Fe addition state (Fe concentration is 1×10 17 cm -3 and the Fe concentration is 1×10 18 cm -3 In the above embodiment, the threshold voltage increases closer to the reference threshold voltage, that is, increases in the negative voltage direction, as the In composition of the diffusion suppression layer 40 increases. This is because the effect of suppressing the diffusion of Fe into the buffer layer 30 increases as the In composition of the diffusion suppression layer 40 increases, and the decrease in conductivity of the channel layer 50 is suppressed.

[0102] The magnitude of the reference threshold voltage is 3 V. The Fe concentration is 1×10 17 cm -3 and the Fe concentration is 1×10 18 cm -3 In the sample with 0% In design composition of the embodiment, the threshold voltage magnitude is 2.6 V and 2.3 V. From the viewpoint of suppressing the fluctuation of the threshold voltage magnitude, for example, a threshold voltage greater than 0.9 times the reference threshold voltage magnitude (greater than 2.7 V in this example) is determined to be a good threshold voltage magnitude.

[0103] From Fig. 10(b), the Fe concentration is 1×10 17 cm -3 (more specifically, the interface peak concentration C3 is 3×10 17 cm -3 or more, or the minimum concentration C2 near the interface is 2×10 16 cm -3 Even in the above embodiment, the Fe concentration is 1×10 18 cm -3 (more specifically, the interface peak concentration C3 is 3×10 18 cm -3 or more, or the minimum concentration C2 near the interface is 2×10 17 cm -3 It can be seen that even in the above embodiment, by increasing the In composition of the diffusion suppression layer 40, the magnitude of the threshold voltage can be made more than 0.9 times the reference threshold voltage.

[0104] Although the stacked structure 100 including the diffusion suppression portion 1 according to the embodiment has been described above using a HEMT as an example, the application of the stacked structure 100 is not limited to a HEMT. The structure of the diffusion suppression portion 1 is not limited to the above example. For example, the first layer of the diffusion suppression portion 1 is not limited to an epitaxial layer (a buffer layer in the example of a HEMT) grown on a growth substrate, and may be a free-standing substrate made of a group III nitride (when FIG. 1 shows such an embodiment, attention should be paid to the stacked portion of the first layer 30, the second layer 40, and the third layer 50 constituting the diffusion suppression portion 1, and the other portions may be omitted, and the first layer 30 represents a free-standing substrate).

[0105] <Preferred embodiment of the present invention> Preferred aspects of the present invention will be described below.

[0106] (Appendix 1) a first layer made of a group III nitride; a second layer disposed on the first layer and made of an In-containing Group III nitride; a third layer disposed on the second layer and composed of a Group III nitride; having the first layer contains an impurity that makes the III nitride electrically resistive at least in a region near the interface between the first layer and the second layer, the region being a thickness of 50 nm from the interface; a concentration of the impurity is lower on the third layer side than on the first layer side across the second layer, In the third layer, the concentration of the impurity is 2×10 15 cm -3 Less than (preferably 1.5 × 10 15 cm -3 The third layer is substantially free of the impurities. III-nitride stack structure.

[0107] (Appendix 2) The second layer is composed of a group III nitride having an In composition of at least 15% or more. 2. The III-nitride stack structure according to claim 1.

[0108] (Appendix 3) The minimum concentration of the impurity in the region near the interface is 2×10 16 cm -3 That's all. 3. The III-nitride stack structure according to claim 1 or 2.

[0109] (Appendix 4) a concentration of the impurity increases toward the second layer in the region near the interface; After the increase, the concentration of the impurity reaches a first peak concentration; The concentration of the impurity is 2×10 15 cm -3 2×10 to the first peak concentration, which is the reduction ratio when the concentration is reduced to 15 cm -3 The ratio is 1 / 150 or less. 3. The III-nitride stack structure according to claim 1 or 2.

[0110] (Appendix 5) a concentration of the impurity increases toward the second layer in the region near the interface; After the increase, the concentration of the impurity reaches a first peak concentration; The concentration of the impurity is 2×10 15 cm -3 The average slope of the decrease is 2.5×10 16 cm -3 / nm or more, 3. The III-nitride stack structure according to claim 1 or 2.

[0111] (Appendix 6) In the third layer, the average concentration of the impurity in a range of 10 nm above the interface with the second layer is 3×10 15 cm -3 is less than 3. The III-nitride stack structure according to claim 1 or 2.

[0112] (Appendix 7) a concentration of the impurity in the upper surface of the third layer is lower than a minimum value of a concentration of the impurity in the interface vicinity region of the first layer; 3. The III-nitride stack structure according to claim 1 or 2.

[0113] (Appendix 8) the concentration of the impurity in the first layer is highest at the interface with the second layer; 3. The III-nitride stack structure according to claim 1 or 2.

[0114] (Appendix 9) a second peak concentration that is higher than a minimum value of the impurity concentration in the interface vicinity region is present in a region on the lower surface side of the first layer than the interface vicinity region; The ratio of the minimum value to the second peak concentration is 1 / 30 or more. 3. The III-nitride stack structure according to claim 1 or 2.

[0115] (Appendix 10) a second peak concentration that is higher than a minimum value of the impurity concentration in the interface vicinity region is present in a region on the lower surface side of the first layer than the interface vicinity region; the thickness of the first layer from the thickness position where the second peak concentration exists to the interface with the second layer is thinner than the thickness of the first layer from the lower surface of the first layer to the thickness position where the second peak concentration exists; 3. The III-nitride stack structure according to claim 1 or 2.

[0116] (Appendix 11) a second peak concentration that is higher than a minimum value of the impurity concentration in the interface vicinity region is present in a region on the lower surface side of the first layer than the interface vicinity region; a thickness of the first layer from a thickness position where the second peak concentration exists to an interface with the second layer is 250 nm or less; 3. The III-nitride stack structure according to claim 1 or 2.

[0117] (Appendix 12) a concentration of the impurity increases toward the second layer in the region near the interface; After the increase, the concentration of the impurity reaches a first peak concentration; a second peak concentration that is higher than a minimum value of the impurity concentration in the interface vicinity region is present in a region on the lower surface side of the first layer than the interface vicinity region; the first peak concentration is greater than the second peak concentration; 3. The III-nitride stack structure according to claim 1 or 2.

[0118] (Appendix 13) the III-nitride stack structure being included as part of a high electron mobility transistor; the third layer is a channel layer of the high electron mobility transistor; 3. The III-nitride stack structure according to claim 1 or 2.

[0119] (Appendix 14) The first layer is a free-standing substrate made of a Group III nitride. 3. The Group III nitride stack structure according to claim 1 or 2.

[0120] (Appendix 15) The impurities are preferably transition metals (more preferably, for example, iron, and also for example, manganese); 3. The III-nitride stack structure according to claim 1 or 2.

[0121] (Appendix 16) 1. A high electron mobility transistor, comprising: a first layer made of a group III nitride; a second layer disposed on the first layer and made of an In-containing Group III nitride; a third layer disposed on the second layer, the third layer being made of a group III nitride and being a channel layer of the high electron mobility transistor; having the first layer contains an impurity that makes the III nitride electrically resistive at least in a region near the interface between the first layer and the second layer, the region being a thickness of 50 nm from the interface; a concentration of the impurity increases toward the second layer in the region near the interface; Through this increase, the concentration of the impurity becomes 3×10 17 cm -3 a first peak concentration is reached that is equal to or greater than a concentration of the impurity is lower on the third layer side than on the first layer side across the second layer, The sheet resistance of the high electron mobility transistor is less than 1.1 times the sheet resistance of a high electron mobility transistor in which the second layer is omitted and the first layer does not contain the impurity. High electron mobility transistor.

[0122] (Appendix 17) 1. A high electron mobility transistor, comprising: a first layer made of a group III nitride; a second layer disposed on the first layer and made of an In-containing Group III nitride; a third layer disposed on the second layer, the third layer being made of a group III nitride and being a channel layer of the high electron mobility transistor; having the first layer contains an impurity that makes the III nitride electrically resistive at least in a region near the interface between the first layer and the second layer, the region being a thickness of 50 nm from the interface; a concentration of the impurity increases toward the second layer in the region near the interface; Through this increase, the concentration of the impurity becomes 3×10 17 cm -3 a first peak concentration is reached that is equal to or greater than a concentration of the impurity is lower on the third layer side than on the first layer side across the second layer, the magnitude of the threshold voltage of the high electron mobility transistor is more than 0.9 times the magnitude of the threshold voltage of a high electron mobility transistor in an embodiment in which the second layer is omitted and the first layer does not contain the impurity; High electron mobility transistor.

[0123] (Appendix 18) The first peak concentration is 3×10 18 cm -3 That's all. 18. The high electron mobility transistor of claim 16 or 17. [Explanation of symbols]

[0124] 1...diffusion suppression portion, 10...substrate, 20...nucleation layer, 30...buffer layer, 31...lower buffer layer, 32...upper buffer layer, 40...diffusion suppression layer, 50...channel layer, 60...barrier layer, 70...cap layer, 100...laminated structure (HEMT), 110...wafer, 120...electrode, 121...source electrode, 122...gate electrode, 123...drain electrode

Claims

1. a first layer comprised of a Group III nitride; a second layer disposed on the first layer and composed of an In-containing Group III nitride; a third layer disposed on the second layer and comprised of a Group III nitride; having the first layer contains an impurity that makes a Group III nitride electrically resistive at least in a region near the interface between the first layer and the second layer, the region being a thickness of 50 nm from the interface; a concentration of the impurity is lower on the third layer side than on the first layer side across the second layer, In the third layer, the concentration of the impurity is 2×10 15 cm -3 The third layer is substantially free of the impurities. Group III nitride stack structure.

2. The second layer is composed of a Group III nitride having an In composition of at least 15%. The III-nitride stack structure of claim 1 .

3. The minimum concentration of the impurity in the region near the interface is 2×10 16 cm -3 That's all. The Group III nitride stack structure according to claim 1 or 2.

4. a concentration of the impurity increases toward the second layer in the region near the interface; Through this increase, the concentration of the impurity reaches a first peak concentration, The impurity concentration is 2×10 15 cm -3 2×10 to the first peak concentration, which is the reduction ratio when the concentration is reduced to 15 cm -3 is 1 / 150 or less; The Group III nitride stack structure according to claim 1 or 2.

5. a concentration of the impurity increases toward the second layer in the region near the interface; Through this increase, the concentration of the impurity reaches a first peak concentration, The impurity concentration is 2×10 15 cm -3 The average slope of the decrease is 2.5 x 10 16 cm -3 / nm or more, The Group III nitride stack structure according to claim 1 or 2.

6. In the third layer, the average concentration of the impurity in a range of 10 nm above the interface with the second layer is 3×10 15 cm -3 is less than The Group III nitride stack structure according to claim 1 or 2.

7. a concentration of the impurity in the upper surface of the third layer is lower than a minimum value of a concentration of the impurity in the interface vicinity region of the first layer; The Group III nitride stack structure according to claim 1 or 2.

8. the concentration of the impurity in the first layer is highest at the interface with the second layer; The Group III nitride stack structure according to claim 1 or 2.

9. a second peak concentration that is higher than a minimum value of the impurity concentration in the interface vicinity region is present in a region on the lower surface side of the first layer than the interface vicinity region; the ratio of the minimum value to the second peak concentration is 1 / 30 or more; The Group III nitride stack structure according to claim 1 or 2.

10. a second peak concentration that is higher than a minimum value of the impurity concentration in the interface vicinity region is present in a region on the lower surface side of the first layer than the interface vicinity region; a thickness of the first layer from the thickness position where the second peak concentration exists to the interface with the second layer is smaller than a thickness of the first layer from the lower surface of the first layer to the thickness position where the second peak concentration exists; The Group III nitride stack structure according to claim 1 or 2.

11. a second peak concentration that is higher than a minimum value of the impurity concentration in the interface vicinity region is present in a region on the lower surface side of the first layer than the interface vicinity region; a thickness of the first layer from a thickness position at which the second peak concentration exists to an interface with the second layer is 250 nm or less; The Group III nitride stack structure according to claim 1 or 2.

12. a concentration of the impurity increases toward the second layer in the region near the interface; Through this increase, the concentration of the impurity reaches a first peak concentration, a second peak concentration that is higher than a minimum value of the impurity concentration in the interface vicinity region is present in a region on the lower surface side of the first layer than the interface vicinity region; the first peak concentration is greater than the second peak concentration; The Group III nitride stack structure according to claim 1 or 2.

13. the III-nitride stack is included as part of a high electron mobility transistor; the third layer is a channel layer of the high electron mobility transistor; The Group III nitride stack structure according to claim 1 or 2.

14. The first layer is a free-standing substrate composed of a Group III nitride. The Group III nitride stack structure according to claim 1 or 2.

15. The impurity is a transition metal. The Group III nitride stack structure according to claim 1 or 2.

16. 1. A high electron mobility transistor, comprising: a first layer comprised of a Group III nitride; a second layer disposed on the first layer and composed of an In-containing Group III nitride; a third layer disposed on the second layer, the third layer being made of a Group III nitride and being a channel layer of the high electron mobility transistor; having the first layer contains an impurity that makes a Group III nitride electrically resistive at least in a region near the interface between the first layer and the second layer, the region being a region of 50 nm in thickness from the interface between the first layer and the second layer; a concentration of the impurity increases toward the second layer in the region near the interface; Through this increase, the concentration of the impurity becomes 3×10 17 cm -3 a first peak concentration is reached that is equal to or greater than a concentration of the impurity is lower on the third layer side than on the first layer side across the second layer, the sheet resistance of the high electron mobility transistor is less than 1.1 times the sheet resistance of a high electron mobility transistor in which the second layer is omitted and the first layer does not contain the impurity; High electron mobility transistor.

17. 1. A high electron mobility transistor, comprising: a first layer comprised of a Group III nitride; a second layer disposed on the first layer and composed of an In-containing Group III nitride; a third layer disposed on the second layer, the third layer being made of a Group III nitride and being a channel layer of the high electron mobility transistor; having the first layer contains an impurity that makes a Group III nitride electrically resistive at least in a region near the interface between the first layer and the second layer, the region being a region of 50 nm in thickness from the interface between the first layer and the second layer; a concentration of the impurity increases toward the second layer in the region near the interface; Through this increase, the concentration of the impurity becomes 3×10 17 cm -3 a first peak concentration is reached that is equal to or greater than a concentration of the impurity is lower on the third layer side than on the first layer side across the second layer, the magnitude of the threshold voltage of the high electron mobility transistor is more than 0.9 times the magnitude of the threshold voltage of a high electron mobility transistor in an embodiment in which the second layer is omitted and the first layer does not contain the impurity; High electron mobility transistor.

18. The first peak concentration is 3×10 18 cm -3 That's all.

18. A high electron mobility transistor according to claim 16 or 17.

Citation Information

Patent Citations

  • Compound semiconductor device and manufacturing method of the same

    JP2019021704A

  • HEMT Structure with Iron-Doping-Stop Component and Methods of Forming

    US20150001582A1

  • Nitride semiconductor device

    WO2019181391A1