JP7339096B2 - Manufacturing method of nitride semiconductor substrate and nitride semiconductor substrate - Google Patents

Description

The present invention relates to a method for manufacturing a nitride semiconductor substrate, a nitride semiconductor substrate and a laminated structure.

Various methods have been disclosed for obtaining a free-standing substrate (hereinafter also referred to as a "nitride semiconductor substrate") made of a single crystal of a group III nitride semiconductor (for example, Patent Document 1).

JP 2003-178984 A

An object of the present invention is to easily and stably obtain a nitride semiconductor substrate having good crystal quality.

According to one aspect of the invention,
A method for manufacturing a nitride semiconductor substrate using a vapor deposition method, comprising:
preparing an underlying substrate;
forming a first underlying layer made of a group III nitride semiconductor on the underlying substrate;
forming a metal layer on the first underlayer;
heat-treating to form voids in the first underlayer;
A second underlayer made of a single crystal of a Group III nitride semiconductor and having a main surface whose closest low-index crystal plane is the (0001) plane is epitaxially grown above the first underlayer, and the second underlayer is mirroring the main surface of the stratum;
A single crystal of a group III nitride semiconductor having a top surface with an exposed (0001) plane is directly epitaxially grown on the main surface of the second underlayer, and is composed of an inclined interface other than the (0001) plane. A plurality of concave portions are formed on the top surface, the inclined interface is gradually expanded upward from the main surface of the second underlayer, and the (0001) plane disappears from the top surface at least once. , growing a three-dimensional growth layer;
slicing the three-dimensional growth layer to form the nitride semiconductor substrate;
A method for manufacturing a nitride semiconductor substrate having

According to another aspect of the invention,
A nitride semiconductor substrate having a diameter of 2 inches or more, made of a single crystal of a Group III nitride semiconductor, and having a main surface whose closest low-index crystal plane is the (0001) plane,
Having a high oxygen concentration region having an oxygen concentration of 9×10 ¹⁷ cm ⁻³ or more,
A nitride semiconductor substrate is provided in which the area ratio of the high oxygen concentration region in the main surface is 80% or more.

According to yet another aspect of the invention,
an underlying layer made of a single crystal of a Group III nitride semiconductor, having a mirror-finished main surface, and having a (0001) plane as the closest low-index crystal plane to the main surface;
a three-dimensional growth layer provided on the underlayer, made of a single crystal of a group III nitride semiconductor, and having a plurality of recesses on the surface thereof, the recesses being formed by tilted interfaces other than the (0001) plane;
has
The three-dimensional growth layer has a high oxygen concentration region made of a group III nitride semiconductor single crystal having an oxygen concentration of 9×10 ¹⁷ cm ⁻³ or more,
The layered structure is provided, wherein the three-dimensionally grown layer has a creeping cross section cut along the main surface, and the high oxygen concentration region occupies an area ratio of 80% or more.

According to the present invention, a nitride semiconductor substrate having good crystal quality can be obtained easily and stably.

1 is a flow chart showing a method for manufacturing a nitride semiconductor substrate according to one embodiment of the present invention; 4 is a flow chart showing a base structure fabricating process. (a) to (e) are schematic cross-sectional views showing part of a method for manufacturing a nitride semiconductor substrate according to an embodiment of the present invention. (a) to (c) are schematic cross-sectional views showing part of the method for manufacturing a nitride semiconductor substrate according to one embodiment of the present invention. 1 is a schematic perspective view showing part of a method for manufacturing a nitride semiconductor substrate according to one embodiment of the present invention; FIG. (a) to (b) are schematic cross-sectional views showing part of a method for manufacturing a nitride semiconductor substrate according to an embodiment of the present invention. It is a schematic sectional drawing which shows a part of manufacturing method of the nitride semiconductor substrate which concerns on one Embodiment of this invention. (a) is a schematic cross-sectional view showing the growth process under standard growth conditions in which neither the tilted interface nor the c-plane expands or shrinks; 1 is a schematic cross-sectional view showing a growth process under No. 1 growth conditions; FIG. 1A is a schematic top view showing a nitride semiconductor substrate according to an embodiment of the present invention, and FIG. 1B is a schematic cross section along the m-axis of the nitride semiconductor substrate according to an embodiment of the present invention; FIG. FIG. 2C is a schematic cross-sectional view along the a-axis of the nitride semiconductor substrate according to one embodiment of the present invention; 1 is a schematic diagram showing a cathodoluminescence image of a main surface of a nitride semiconductor substrate according to an embodiment of the present invention observed with a scanning electron microscope; FIG. It is a figure which shows the observation image which observed the surface of the laminated structure of the Example with the optical microscope. It is a figure which shows the observation image which observed the cross section of the laminated structure of an Example with the fluorescence microscope. (a) and (b) are diagrams showing observation images obtained by observing a cross section of another portion of the laminated structure of the example with a fluorescence microscope. It is a figure which shows the observation image which observed the main surface of the nitride semiconductor substrate of an Example with the multiphoton excitation microscope. 14. It is a figure which shows the observation image which expanded the dotted-line square part of FIG. FIG. 10 is a diagram showing an observation image observed with a multiphoton excitation microscope when the focus is changed in the thickness direction in the nitride semiconductor substrate of the example. (a) is a diagram showing the result of rocking curve measurement of X-ray diffraction in the direction along the m-axis of the nitride semiconductor substrate of the example, and (b) is the nitride semiconductor of the example. It is a figure which shows the result of having performed the rocking-curve measurement of X-ray diffraction with respect to the direction along a-axis orthogonal to the m-axis of a board|substrate. (a) is a diagram showing absorption spectra of samples 1 and 2, and (b) is a diagram obtained by adding a theoretical formula to (a). (a) is a diagram showing the transmission spectrum of Sample 1, and (b) is a diagram showing the reflection spectrum of Sample 1. FIG.

<Knowledge acquired by the inventors, etc.>
First, the findings obtained by the inventors will be described.

The method in Patent Document 1 mentioned above is called a VAS method (Void-assisted Separation Method). The VAS method is performed, for example, as follows.

First, a first crystal layer made of a Group III nitride semiconductor is formed on a predetermined underlying substrate. After forming the first crystal layer, a metal layer is formed on the first crystal layer. After forming the metal layer, heat treatment is performed in an atmosphere containing a predetermined gas to form fine holes in the metal layer and to create voids in the first crystal layer through the mesh (nanonet) of the metal layer. to form After the voids are formed in the first crystal layer, a second crystal layer made of a Group III nitride semiconductor is formed above the first crystal layer. At this time, some of the voids described above remain. After the second crystal layer is formed, the second crystal layer is separated from the base substrate due to the remaining voids described above when the temperature is lowered from the growth temperature of the second crystal layer. After separating the second crystal layer, the second crystal layer is sliced and polished to obtain a nitride semiconductor substrate having high crystal quality.

Here, when a crystal layer made of a single crystal of a Group III nitride semiconductor is grown thick with the c-plane as the growth plane, the dislocation density on the surface of the crystal layer tends to be inversely proportional to the thickness of the crystal layer. be.

Therefore, in order to reduce the dislocation density of the second crystal layer in the above-described VAS method, it is conceivable to grow the second crystal layer thicker above the first crystal layer.

However, in the above-described VAS method, the tensile stress generated by the attraction of the initial nuclei of the crystal accumulates on the underlying substrate side of the second crystal layer. On the other hand, the c-plane of the second crystal layer is curved into a concave spherical shape due to the tensile stress generated in the second crystal layer. When the second crystal layer is grown thick on the concavely curved c-plane, the stress applied to the second crystal layer gradually changes to compressive stress as the second crystal layer becomes thicker. Therefore, the stress difference between the base substrate side and the surface side of the second crystal layer gradually increases. If the stress difference becomes excessive, cracks may occur in the second crystal layer.

Thus, in the VAS method, it is difficult to grow the second crystal layer thick using the c-plane as the growth plane.

The present invention is based on the above knowledge discovered by the inventor.

<One embodiment of the present invention>
An embodiment of the present invention will be described below with reference to the drawings.

(1) Method for Manufacturing Nitride Semiconductor Substrate A method for manufacturing a nitride semiconductor substrate according to the present embodiment will be described with reference to FIGS. 1 to 8. FIG. FIG. 1 is a flow chart showing a method for manufacturing a nitride semiconductor substrate according to this embodiment. FIG. 2 is a flow chart showing the base structure fabricating process. 3(a) to (e), FIGS. 4(a) to (c), and FIGS. 6(a) to 7 are schematic cross-sectional views showing part of the method for manufacturing a nitride semiconductor substrate according to this embodiment. is. FIG. 5 is a schematic perspective view showing part of the method for manufacturing a nitride semiconductor substrate according to this embodiment. 3A to 3E, the lower side of the underlying substrate 1 is omitted. 5 corresponds to a perspective view at the time of FIG. 4(b) and shows part of the three-dimensional growth layer 30 growing on the base structure 10. FIG. 4(c) and 6(a) to 7, dotted lines indicate dislocations.

In the following, the <0001> axis (for example, the [0001] axis) is referred to as the "c-axis" and the (0001) plane is referred to as the "c-plane" in the group III nitride semiconductor crystal having the wurtzite structure. The (0001) plane is sometimes referred to as the "+c plane (group III element polar plane)", and the (000-1) plane is sometimes called the "-c plane (nitrogen (N) polar plane)". Also, the <1-100> axis (for example, the [1-100] axis) is called the "m-axis", and the {1-100} plane is called the "m-plane". Note that the m-axis may also be expressed as the <10-10> axis. The <11-20> axis (for example, the [11-20] axis) is called the "a-axis", and the {11-20} plane is called the "a-plane".

As shown in FIG. 1, the method for manufacturing a nitride semiconductor substrate according to this embodiment includes, for example, a base structure forming step S100, a three-dimensional growth step S200, a peeling step S300, a slicing step S400, and a polishing step. S500 and.

(S100: Underlying Structure Preparing Step)
First, for example, the base structure 10 is manufactured by the above-described VAS method.

Specifically, the underlying structure fabricating step S100 includes, for example, an underlying substrate preparing step S110, a first underlying layer forming step S120, a metal layer forming step S130, a void forming step S140, and a second underlying layer forming step. and S150.

(S110: base substrate preparation step)
First, as shown in FIG. 3A, a base substrate 1 is prepared. The underlying substrate 1 is made of, for example, a material different from the Group III nitride semiconductor. Specifically, the underlying substrate 1 is, for example, a sapphire substrate. Note that the base substrate 1 may be, for example, a Si substrate or a gallium arsenide (GaAs) substrate.

The diameter of the underlying substrate 1 is, for example, 2 inches (50.8 mm) or more. Further, the thickness of the underlying substrate 1 is, for example, 300 μm or more and 1 mm or less.

The underlying substrate 1 has, for example, a main surface 1s that serves as an epitaxial growth surface. A low-index crystal plane closest to the principal plane 1s is, for example, the c-plane.

In this embodiment, the c-plane of the underlying substrate 1 is inclined with respect to the main surface 1s. The c-axis of the underlying substrate 1 is inclined at a predetermined off angle with respect to the normal to the main surface 1s. The off angle within the main surface 1s of the underlying substrate 1 is uniform over the entire main surface 1s. The magnitude of the off-angle at the center of the main surface 1s of the underlying substrate 1 is, for example, greater than 0° and less than or equal to 1°. The off angle within the main surface 1s of the base substrate 1 affects the off angle at the center of the main surface 6s of the second base layer 6, which will be described later.

(S120: first base layer forming step)
Next, as shown in FIG. 3B, on the main surface 1s of the underlying substrate 1, a first underlying layer (first crystal layer) 2 made of group III nitride is formed.

Specifically, for example, a Group III source gas and a nitrogen source gas are supplied to the base substrate 1 heated to a predetermined growth temperature by metalorganic vapor phase epitaxy (MOVPE). For example, by supplying trimethylgallium (TMG) gas as a III-group source gas and ammonia gas (NH ₃ ) as a nitrogen source gas, the first underlayer 2 is a low-temperature-grown GaN buffer layer, a GaN layers are grown on the main surface 1s of the base substrate 1 in this order. During the growth of the GaN layer, the GaN layer may be doped with an n-type impurity by supplying monosilane (SiH ₄ ) gas as an n-type dopant gas, for example.

At this time, the growth conditions for the low-temperature-grown GaN buffer layer as the first underlayer 2 are adjusted so as to obtain the desired crystal quality of the GaN layer. Specifically, the growth temperature of the low-temperature-grown GaN buffer layer is, for example, 400° C. or higher and 600° C. or lower.

Also, at this time, the growth conditions of the GaN layer as the first underlayer 2 are adjusted so that desired voids are formed in the void forming step S140 described later. Specifically, the growth temperature of the GaN layer is, for example, 1,000° C. or higher and 1,200° C. or lower.

Moreover, at this time, for example, the surface of the first underlayer 2 is mirror-finished. The term “mirror surface” as used herein refers to a surface on which the maximum value of the height difference between adjacent unevenness on the surface is equal to or less than the wavelength of visible light. facets exposed hillocks may occur. Specifically, the root-mean-square roughness RMS of the surface of the first underlayer 2 is, for example, less than 10 nm, preferably less than 1 nm. By mirror-finishing the surface of the first underlayer 2 in this manner, the appearance of voids can be made substantially uniform over the entire main surface 1s of the undersubstrate 1 in the void formation step S140 described later.

(S130: Metal layer forming step)
Next, as shown in FIG. 3(c), a metal layer 3 is formed on the first underlayer 2. Next, as shown in FIG. For example, the metal layer 3 is formed by vacuum deposition or sputtering.

The metal layer 3 is heat-treated in the void forming step S140 described later to form fine holes in itself, promote decomposition of the first underlayer 2, and form voids in the first underlayer 2. It is preferably made of a material that Specifically, examples of the metal layer 3 satisfying this condition include titanium (Ti), scandium (Sc), yttrium (Y), zirconium (Zr), hafnium (Hf), vanadium (V), and niobium (Nb). ), tantalum (Ta), chromium (Cr), molybdenum (Mo), tungsten (W), rhenium (Re), iron (Fe), ruthenium (Ru), osmium (Os), cobalt (Co), rhodium (Rh ), iridium (Ir), nickel (Ni), palladium (Pd), manganese (Mn), copper (Cu), platinum (Pt) or gold (Au). In this embodiment, the metal layer 3 is, for example, a Ti layer.

At this time, the thickness of the metal layer 3 is, for example, 1 μm or less, preferably 300 nm or less, more preferably 100 nm or less. By setting the thickness of the metal layer 3 to 1 μm or less, preferably 300 nm or less, and more preferably 100 nm or less, it is possible to suppress deterioration in flatness of the metal layer 3 (metal nitride layer 5) in the void formation step S140 described later. can be done. Thereby, deterioration of the crystal quality of the second underlayer 6 formed on the metal nitride layer 5 can be suppressed. Note that the lower limit of the thickness of the metal layer 3 is not particularly limited. However, from the viewpoint of stably forming the metal nitride layer 5 in the void forming step S140 described later, the thickness of the metal layer 3 is preferably 0.5 nm or more, for example.

(S140: void formation step)
After forming the metal layer 3 , a predetermined heat treatment is performed to form voids in the first underlayer 2 .

Specifically, the base substrate 1 described above is put into an electric furnace, and the base substrate 1 is placed on a susceptor having a predetermined heater. After the base substrate 1 is placed on the susceptor, the base substrate 1 is heated by a heater and heat-treated in an atmosphere containing a predetermined gas.

At this time, for example, the heat treatment is performed in an atmosphere containing at least one of nitrogen (N ₂ ) gas and nitrogen-containing gas. Examples of the nitrogen-containing gas include at least one of NH ₃ gas and hydrazine (N ₂ H ₄ ) gas. As a result, the metal layer 3 is aggregated and nitrided to form a network-like metal nitride layer 5 (also referred to as a nanonet of the metal nitride layer 5) having high-density fine holes (through holes) on the surface. can.

Furthermore, at this time, for example, the heat treatment is performed in an atmosphere containing at least one of hydrogen (H ₂ ) gas and hydrogen-containing gas. The hydrogen-containing gas includes, for example, at least one of the NH ₃ gas and N ₂ H ₄ gas described above. As a result, a portion of the first underlayer 2 can be etched through the mesh of the metal nitride layer 5 to form high-density voids in the first underlayer 2 .

At this time, the step of forming the nanonet of the metal nitride layer 5 and the step of forming the voids in the first underlayer 2 may be performed simultaneously, or may be performed separately in different atmospheres. good.

Also, at this time, the heat treatment conditions are adjusted so that the first underlayer 2 has a predetermined void ratio. Specifically, the partial pressure ratio of the hydride gas in the atmosphere is, for example, 10% or more and 95% or less, preferably 50% or less. Also, the heat treatment temperature is, for example, 1,000° C. or higher and 1,100° C. or lower. Also, the heat treatment time is set to, for example, 1 minute or more and 100 minutes or less.

Through the void forming step S140 described above, the void-containing first underlayer 4 is formed as shown in FIG. 3(d).

(S150: second base layer forming step)
After forming the void-containing first underlayer 4, as shown in FIG. layer) 6 is epitaxially grown.

Specifically, for example, the underlying substrate 1 is carried into a predetermined vapor phase growth apparatus. Next, a Group III raw material gas and a nitrogen raw material gas are supplied to the underlying substrate 1 heated to a predetermined growth temperature by hydride vapor phase epitaxy (HVPE). For example, gallium chloride (GaCl) gas and NH ₃ gas are supplied to epitaxially grow a GaN layer as the second underlayer 6 on the void-containing first underlayer 4 and the metal nitride layer 5 . When growing the GaN layer as the second underlayer 6, for example, by supplying at least one of dichlorosilane (SiH ₂ Cl ₂ ) gas and tetrachlorogermane (GeCl ₄ ) gas as an n-type dopant gas, , the GaN layer may be doped with an n-type impurity.

In the initial stage of growth of the second underlayer 6, initial nuclei of island crystals are generated. At this time, the frequency of the island-like crystals in the second underlayer 6 depends on the degree of supersaturation when growing on the nanonets of the metal nitride layer 5 and the variation in the opening width of the nanonets. As the second underlayer 6 grows, initial nuclei of island crystals grow laterally. After that, when the distance between the adjacent initial nuclei becomes closer, it is more energetically stable when the surfaces of the initial nuclei are bonded to one than when there are two surfaces of the initial nuclei. Therefore, adjacent initial nuclei are forced to attract each other (associate). The primary surface 6 s of the second underlayer 6 is formed by the association of the initial nuclei over the entire second underlayer 6 .

At this time, for example, the main surface 6s of the second underlayer 6 is mirror-finished. The term “mirror surface” as used herein refers to a surface on which the maximum value of the height difference between adjacent irregularities on the surface is equal to or less than the wavelength of visible light, as described above. may have hillocks where facets other than the c-plane are exposed. Specifically, the root-mean-square roughness RMS of the main surface 6s of the second underlayer 6 is, for example, less than 10 nm, preferably less than 1 nm. Details will be described later in (4), but by mirror-finishing the main surface 6s of the second underlayer 6, the growth mode of the three-dimensional growth layer 30 described later can be changed to the islands generated in the initial growth of the second underlayer 6. It can be changed from the growth morphology of the crystals. As a result, the average distance between the nearest vertexes of the three-dimensional growth layer 30, which will be described later, can be controlled based on the first growth conditions in the three-dimensional growth step S200, and the distance between the nearest vertices can be increased.

At this time, since the low-index crystal plane closest to the main surface 1 s of the underlying substrate 1 is the c-plane, the low-index crystal plane closest to the mirror-finished main surface 6 s of the second underlayer 6 is also the c-plane. face.

Also, at this time, the growth conditions of the second underlayer 6 are adjusted so that initial nuclei of island-like crystals are generated at a predetermined frequency. Specifically, the growth temperature of the second underlayer 6 is, for example, 1,000° C. or higher and 1,100° C. or lower. Also, the ratio of the partial pressure of the flow rate of _NH3 gas as the nitrogen source gas to the partial pressure of GaCl gas as the group III source gas during the growth of the second underlayer 6 (hereinafter also referred to as "V/III ratio") is, for example, 1 or more and 50 or less.

At this time, the thickness of the second underlayer 6 is, for example, 100 μm or more and 1.5 mm or less, preferably 200 μm or more and 500 μm or less.

If the thickness of the second underlayer 6 is less than 100 μm, it becomes difficult to make the main surface 6s of the second underlayer 6 a mirror surface. Further, if the thickness of the second underlayer 6 is less than 100 μm, the second underlayer 6 is not robust. Cracks may occur in the second underlayer 6 due to the short-range stress difference between the . On the other hand, by setting the thickness of the second underlayer 6 to 100 μm or more, the main surface 6s of the second underlayer 6 can be easily mirror-finished. As a result, the growth mode of the three-dimensional growth layer 30 is reliably changed from the growth mode of the island-shaped crystals generated in the initial stage of growth of the second underlayer 6, and the three-dimensional growth is performed in the three-dimensional growth step S200 described later. The distance between the nearest tops of the growth layer 30 can be easily increased. Further, by setting the thickness of the second underlayer 6 to 100 μm or more, the second underlayer 6 can be provided with predetermined robustness. As a result, in the three-dimensional growth step S200, which will be described later, it is possible to suppress the occurrence of cracks in the second underlayer 6 due to the short-distance stress difference between the second underlayer 6 and the three-dimensional growth layer 30. . Furthermore, by setting the thickness of the second underlayer 6 to 200 μm or more, the main surface 6s of the second underlayer 6 can be stably mirror-finished. Further, by setting the thickness of the second underlayer 6 to 200 μm or more, the occurrence of cracks in the second underlayer 6 can be stably suppressed.

On the other hand, when the thickness of the second underlayer 6 exceeds 1.5 mm, cracks may occur in the second underlayer 6 due to thickening of the second underlayer 6 itself. . On the other hand, by setting the thickness of the second underlayer 6 to 1.5 mm or less, it is possible to suppress the occurrence of cracks in the second underlayer 6 due to the thickening of the second underlayer 6 itself. . Furthermore, by setting the thickness of the second underlayer 6 to 500 μm or less, the occurrence of cracks in the second underlayer 6 can be stably suppressed.

Also, at this time, the second underlayer 6 grows from the void-containing first underlayer 4 through the holes in the metal nitride layer 5 onto the void-containing first underlayer 4 and the metal nitride layer 5 . Some of the voids in the void-containing first underlayer 4 are filled by the second underlayer 6, but other parts of the voids in the void-containing first underlayer 4 remain. A flat gap is formed between the second underlayer 6 and the metal nitride layer 5 due to the voids remaining in the void-containing first underlayer 4 . This gap causes the peeling of the second underlayer 6, which will be described later.

At this time, a tensile stress is introduced into the second underlayer 6 due to attraction between initial nuclei generated during the growth process. Therefore, due to the tensile stress generated in the second underlayer 6, an internal stress acts such that the c-plane of the second underlayer 6 is depressed. Further, the dislocation density on the main surface 6s side of the second underlayer 6 is low, while the dislocation density on the undersubstrate 1 side of the second underlayer 6 is high. Therefore, due to the difference in dislocation density in the thickness direction of the second underlayer 6 as well, internal stress works so that the c-plane of the second underlayer 6 is depressed. Therefore, the c-plane of the second underlayer 6 is curved in a concave spherical shape with respect to the main surface 6s. The term "spherical" as used herein means a curved surface that is approximated to a spherical surface. Also, the term "spherical surface approximation" as used herein means approximating a perfect circular spherical surface or an elliptical spherical surface within a predetermined error range. Since the c-plane is curved as described above, the off-angle formed by the c-axis with respect to the normal to the center of the main surface 6s of the second underlayer 6 has a predetermined distribution.

Also, at this time, although not shown in detail, the entire second underlayer 6 is slightly warped following the curvature of the c-plane described above. Specifically, the principal surface 6s of the second underlayer 6 (as viewed from the principal surface 6s side) curves slightly concave following the curvature of the c-plane. On the other hand, the back surface of the second base layer 6 (the surface facing the base substrate 1) is peeled off from the outer peripheral side due to the tensile stress described above. For these reasons, the back surface of the second underlayer 6 is slightly convex (as viewed from the back surface side).

The underlying structure 10 is obtained by the underlying structure manufacturing step S100 described above.

(S200: Three-dimensional growth step)
Thereafter, as shown in FIGS. 4B, 4C, and 5, a group III nitride semiconductor single crystal having a top surface 30u with an exposed c-plane 30c is applied to the second underlayer 6 as the main layer. It is epitaxially grown directly on the surface 6s. Thereby, the three-dimensional growth layer 30 is grown.

At this time, a plurality of recesses 30p surrounded by inclined interfaces 30i other than the c-plane are formed on the top surface 30u of the single crystal. , and the c-plane 30c is gradually reduced. As a result, the c-plane 30c disappears from the top surface 30u at least once. As a result, a three-dimensional growth layer 30 whose surface consists only of the inclined interface 30i is grown.

At this time, in the three-dimensional growth layer 30, an inclined interface growth region 70 (gray portion in the drawing) is formed by growing the inclined interface 30i other than the c-plane as a growth plane. Further, as will be described later, the area occupied by the inclined interface growth region 70 in the three-dimensional growth layer 30 along the main surface 6s of the second underlayer 6 is, for example, 80% or more.

As described above, in the three-dimensional growth step S200, the three-dimensional growth layer 30 is grown three-dimensionally so as to intentionally roughen the main surface 6s of the mirror-finished second underlayer 6 . Note that the three-dimensional growth layer 30 is grown as a single crystal as described above even if such a growth form is formed. In this respect, the three-dimensional growth layer 30 is different from a so-called low-temperature-grown buffer layer formed as amorphous or polycrystalline on a foreign substrate such as sapphire before epitaxially growing a Group III nitride semiconductor on the foreign substrate. is.

In this embodiment, as the three-dimensional growth layer 30, for example, a layer made of the same Group III nitride semiconductor as the Group III nitride semiconductor forming the second underlayer 6 is epitaxially grown. Specifically, for example, the underlying structure 10 is heated by the HVPE method, and GaCl gas and NH ₃ gas are supplied to the heated underlying structure 10 to form a GaN layer as the three-dimensional growth layer 30. is epitaxially grown.

Here, in the three-dimensional growth step S200, the three-dimensional growth layer 30 is grown under predetermined first growth conditions, for example, in order to develop the growth process described above.

First, referring to FIG. 8(a), reference growth conditions under which neither the inclined interface 30i nor the c-plane 30c expand or contract will be described. FIG. 8(a) is a schematic cross-sectional view showing the growth process under standard growth conditions in which the tilted interface and the c-plane, respectively, do not expand or contract.

In FIG. 8A, the thick solid line indicates the surface of the three-dimensional growth layer 30 for each unit time. The inclined interface 30i shown in FIG. 8A is assumed to be the most inclined interface with respect to the c-plane 30c. In FIG. 8A, the growth rate of the c-plane 30c in the three-dimensional growth layer 30 is _Gc0 , the growth rate of the inclined interface 30i in the three-dimensional growth layer 30 is _Gi , and the three-dimensional growth Let γ be the angle formed by the c-plane 30c and the inclined interface 30i in the layer 30 . Also, in FIG. 8A, it is assumed that the three-dimensional growth layer 30 is grown while maintaining the angle γ between the c-plane 30c and the inclined interface 30i. It is assumed that the off-angle of the c-plane 30c of the three-dimensional growth layer 30 is negligible compared to the angle γ formed between the c-plane 30c and the inclined interface 30i.

As shown in FIG. 8A, when the inclined interface 30i and the c-plane 30c neither expand nor contract, the trajectory of the intersection of the inclined interface 30i and the c-plane 30c is perpendicular to the c-plane 30c. . Therefore, the reference growth condition under which the tilted interface 30i and the c-plane 30c do not expand or contract satisfies the following formula (a).
G _c0 = G _i /cosγ (a)

Next, the first growth condition under which the inclined interface 30i expands and the c-plane 30c contracts will be described with reference to FIG. 8(b). FIG. 8(b) is a schematic cross-sectional view showing the growth process under the first growth condition in which the tilted interface expands and the c-plane shrinks.

In FIG. 8(b), similarly to FIG. 8(a), the thick solid line indicates the surface of the three-dimensional growth layer 30 for each unit time. The inclined interface 30i shown in FIG. 8B is also assumed to be the most inclined interface with respect to the c-plane 30c. In FIG. 8B, the growth rate of the c-plane 30c of the three-dimensional growth layer 30 is assumed to be _Gc1 , and the progress of the trajectory of the intersection of the inclined interface 30i and the c-plane 30c of the three-dimensional growth layer 30 is Let the rate be _R1 . Also, let γ _R1 be the narrower angle between the locus of the intersection of the inclined interface 30i and the c-plane 30c and the c-plane 30c. When the angle between the _R1 direction and the _Gi direction is γ', γ'=γ+90− _γR1 . It is assumed that the off-angle of the c-plane 30c of the three-dimensional growth layer 30 is negligible compared to the angle γ formed between the c-plane 30c and the inclined interface 30i.

As shown in FIG. 8B, the traveling rate _R1 of the trajectory of the intersection between the inclined interface 30i and the c-plane 30c is represented by the following equation (b).
R ₁ =G _i /cos γ' (b)

Also, the growth rate _Gc1 of the c-plane 30c of the three-dimensional growth layer 30 is represented by the following equation (c).
G _c1 =R ₁ sin γ _R1 (c)

By substituting the formula (b) into the formula (c), G _c1 is represented by the following formula (d) using G _i .
G _c1 = G _i sinγ _R1 /cos(γ+90−γ _R1 ) (d)

In order for the inclined interface 30i to expand and the c-plane 30c to contract, it is preferable that γ _R1 <90°. Therefore, the first growth condition under which the inclined interface 30i expands and the c-plane 30c shrinks preferably satisfies the following formula (1) based on the formula (d) and γ _R1 <90°.
G _c1 >G _i /cosγ (1)
However, as described above, G _i is the growth rate of the tilted interface 30i that is most tilted with respect to the c-plane 30c, and γ is the growth rate of the tilted interface 30i that is most tilted with respect to the c-plane 30c. is the angle formed by

Alternatively, it can be considered that G _c1 under the first growth condition is preferably greater than G _c0 under the reference growth condition. From this also, the formula (1) can be derived by substituting the formula (a) for G _c1 >G _c0 .

Since the growth condition for enlarging the inclined interface 30i that is most inclined with respect to the c-plane 30c is the strictest condition, if the first growth condition satisfies the formula (1), the other inclined interfaces 30i should also be enlarged. becomes possible.

Specifically, for example, when the most inclined interface 30i with respect to the c-plane 30c is the {10-11} plane, γ=61.95°. Therefore, the first growth condition preferably satisfies, for example, the following formula (1').
G _c1 >2.13 G _i (1′)

Alternatively, as will be described later, for example, when the inclined interface 30i is the {11-2m} plane with m≧3, the inclined interface 30i most inclined with respect to the c-plane 30c is the {11-23} plane. Therefore, γ=47.3°. Therefore, the first growth condition preferably satisfies, for example, the following formula (1'').
G _c1 >1.47 G _i (1″)

As the first growth condition of the present embodiment, for example, the growth temperature in the three-dimensional growth step S200 is typically set to be lower than the growth temperature when the group III nitride layer is grown with the c-plane as the growth plane. . Specifically, the growth temperature in the three-dimensional growth step S200 is, for example, 980° C. or higher and 1,020° C. or lower, preferably 1,000° C. or higher and 1,020° C. or lower.

Further, as the first growth condition of the present embodiment, for example, the V/III ratio in the three-dimensional growth step S200 is set to V/III when the group III nitride layer is grown typically using the c-plane as the growth plane. It may be larger than the ratio. Specifically, the V/III ratio in the three-dimensional growth step S200 is, for example, 2 or more and 20 or less, preferably 2 or more and 15 or less.

In practice, as the first growth condition, at least one of the growth temperature and the V/III ratio is adjusted within the above range so as to satisfy the formula (1).

Other conditions of the first growth conditions of the present embodiment are, for example, as follows.
Growth pressure: 90-105 kPa, preferably 90-95 kPa
GaCl gas partial pressure: 1.5 to 15 kPa
N ₂ gas flow rate/H ₂ gas flow rate: 0 to 1

Here, the three-dimensional growth step S200 of this embodiment is classified into two steps, for example, based on the shape of the three-dimensional growth layer 30 during growth. Specifically, the three-dimensional growth step S200 of this embodiment includes, for example, an inclined interface enlarging step S220 and an inclined interface maintaining step S240. Through these steps, the three-dimensional growth layer 30 will have, for example, the inclined interface enlarging layer 32 and the inclined interface maintaining layer 34 .

(S220: Inclined interface enlarging step)
First, as shown in FIGS. 4(b), 4(c) and 5, the inclined interface enlarging layer 32 of the three-dimensionally grown layer 30 made of single crystal of group III nitride semiconductor is grown under the first growth conditions described above. Below, it is epitaxially grown on the second underlayer 6 .

In the initial stage of growth of the inclined interface enlarging layer 32, the inclined interface is enlarged by a predetermined thickness in the normal direction (direction along the c-axis) of the main surface 6s of the second underlayer 6 with the c-plane 30c as the growth plane. Layer 32 is step-flow grown (two-dimensional growth). That is, the inclined interface enlarging layer 32 having a mirror-finished surface is continuously formed in the direction along the main surface 6s of the second underlayer 6 (over the entire main surface 6s) with a predetermined thickness. . At this time, the thickness of a part of the inclined interface enlarging layer 32 grown with the c-plane 30c as the growth plane (also referred to as the “initial layer”) is, for example, 1 μm or more and 100 μm or less, preferably 1 μm or more and 20 μm or less.

After that, by gradually growing the inclined interface enlarging layer 32 under the first growth conditions, as shown in FIGS. A plurality of concave portions 30p composed of inclined interfaces 30i other than the c-plane are formed in 30u. A plurality of recesses 30p composed of inclined interfaces 30i other than the c-plane are randomly formed on the top surface 30u. As a result, an inclined interface enlarging layer 32 is formed in which the c-plane 30c and the inclined interface 30i other than the c-plane coexist on the surface.

The term "inclined interface 30i" as used herein means a growth interface that is inclined with respect to the c-plane 30c. contains slopes that cannot be represented by Note that facets other than the c-plane are, for example, {11-2m}, {1-10n}, and the like. However, m and n are integers other than 0.

In the present embodiment, the tilted interface enlarging layer 32 is grown on the mirror-finished second underlayer 6, and the first growth conditions are adjusted so as to satisfy the formula (1). , for example {11-2m} planes with m≧3. This makes it possible to moderate the angle of inclination of the {11-2m} plane with respect to the c-plane 30c. Specifically, the inclination angle can be 47.3° or less.

By further growing the inclined interface enlarging layer 32 under the first growth conditions, as shown in FIG. , the inclined interface 30i other than the c-plane is gradually enlarged, and the c-plane 30c is gradually reduced. At this time, the inclination angle formed by the inclined interface 30i with respect to the main surface 6s of the second underlayer 6 gradually decreases as the second underlayer 6 is moved upward. As a result, most of the inclined interfaces 30i eventually become {11-2m} planes with m≧3 as described above.

As the inclined interface enlarging layer 32 is further grown, the c-plane 30c of the inclined interface enlarging layer 32 disappears from the top surface 30u at least once, and the outermost surface (uppermost surface) of the inclined interface enlarging layer 32 is only the inclined interface 30i. Configured.

In this way, a plurality of recesses 30p constituted by inclined interfaces 30i other than the c-plane are formed on the top surface 30u of the inclined interface enlarging layer 32, and the c-plane 30c disappears. Then, a plurality of valleys 30v and a plurality of tops 30t are formed on the surface of the inclined interface enlarging layer 32. As shown in FIG. Each of the plurality of valleys 30v is a downwardly convex inflection point on the surface of the inclined interface enlarging layer 32, and is formed above the position where each of the inclined interfaces 30i other than the c-plane is generated. On the other hand, each of the plurality of apexes 30t is an upwardly convex inflection point in the surface of the inclined interface enlarging layer 320, and is a c-plane across a pair of inclined interfaces 30i enlarged in opposite directions. 30c is formed at or above the (last) vanished position. The valleys 30v and the tops 30t are alternately formed in the direction along the principal surface 6s of the second underlayer 6 .

In the present embodiment, as described above, the second underlayer 6 having the mirror-finished main surface 6s is grown on the main surface 1s of the base substrate 1 in the second underlayer forming step S150. Furthermore, as described above, in the initial stage of growth of the inclined interface enlarging layer 32, the inclined interface enlarging layer is grown on the main surface 6s of the second underlayer 6 with the c-plane 30c as the growth surface without forming the inclined interface 30i. 32 was grown with a given thickness. Thereafter, in the inclined interface enlarging step S220, an inclined interface 30i other than the c-plane is generated on the surface of the inclined interface enlarging layer 32. As shown in FIG. As a result, the plurality of valleys 30v are formed at positions spaced upward from the main surface 6s of the second underlayer 6 .

Through the growth process of the tilted interface enlarging layer 32 as described above, dislocations bend and propagate as follows. Specifically, as shown in FIG. 4C, the plurality of dislocations extending in the direction along the c-axis in the second underlayer 6 spread from the second underlayer 6 to the inclined interface enlarging layer 32 . propagates along the c-axis of . In the region of the inclined interface enlarging layer 32 grown with the c-plane 30c as the growth plane, dislocations propagate in the direction along the c-axis of the inclined interface enlarging layer 32 . However, in the tilted interface enlarging layer 32, when the growth interface where the dislocation is exposed changes from the c-plane 30c to the tilted interface 30i, the dislocation bends and propagates in a direction substantially perpendicular to the tilted interface 30i. That is, the dislocation propagates while bending in a direction tilted with respect to the c-axis. As a result, dislocations are locally collected above substantially the center between the pair of top portions 30t in the steps after the inclined interface enlarging step S220. As a result, the dislocation density on the surface of the later-described graded interface maintaining layer 34 can be reduced.

At this time, in the present embodiment, when an arbitrary cross section perpendicular to the main surface 6s of the second underlayer 6 is viewed, one of the plurality of valleys 30v is interposed between the plurality of tops 30t, which are the most The average distance (also referred to as “average distance between closest apexes”) L by which a pair of apexes 30t that are close to each other are separated in the direction along the main surface 6s of the second underlayer 6 is, for example, more than 100 μm. When the average distance L between the nearest apexes is 100 μm or less, as in the case where fine hexagonal pyramidal crystal nuclei are generated on the second underlayer 6 from the initial stage of the inclined interface enlarging step S220, the enlarging of the inclined interface does not occur. In the steps after step S220, the dislocation bends and propagates a shorter distance. For this reason, dislocations are not sufficiently collected above the approximate center between the pair of top portions 30t in the inclined interface enlarging layer 32 . As a result, the dislocation density on the surface of the later-described graded interface maintaining layer 34 may not be sufficiently reduced. On the other hand, in the present embodiment, by setting the average distance L between the nearest vertexes to more than 100 μm, in the steps after the tilted interface enlarging step S220, the distance in which the dislocation bends and propagates is ensured to be at least more than 50 μm. be able to. As a result, dislocations can be sufficiently collected above the approximate center between the pair of top portions 30 t in the inclined interface enlarging layer 32 . As a result, it is possible to sufficiently reduce the dislocation density on the surface of the later-described graded interface maintaining layer 34 .

In the present embodiment, for example, in the three-dimensional growth layer 30, it is preferable that there is no portion where the average distance between the closest tops is 100 μm or less. In other words, when an arbitrary cross section perpendicular to the main surface 6s of the second underlayer 6 is viewed, the average distance between the closest tops is more than 100 μm over the entire surface of the three-dimensional growth layer 30. is preferred. As a result, the dislocation density can be substantially uniformly reduced over the entire surface of the later-described graded interface preserving layer 34 .

On the other hand, in the present embodiment, the average distance L between the closest tops is less than 800 μm. If the average distance L between the closest tops is 800 μm or more, it takes a long time to eliminate the c-plane 30c over the entire in-plane because dislocations are collected over the entire in-plane. Therefore, the productivity of the substrate 50 is lowered. Moreover, when the average distance L between the closest top portions is 800 μm or more, the height from the valley portion 30v to the top portion 30t of the inclined interface enlarging layer 32 may become excessively high. The valleys 30v on the crystal surface cause penetrating pits when the substrate 50 is sliced unless measures such as burying growth are taken. Therefore, the yield of obtaining the substrate 50 may decrease. On the other hand, in the present embodiment, by setting the average distance L between the closest tops to less than 800 μm, it is possible to shorten the time for the c-plane 30c to disappear over the entire plane. Thereby, the productivity of the substrate 50 can be improved. Further, by setting the average distance L between the closest tops to less than 800 μm, the height from the valleys 30v to the tops 30t of the inclined interface enlarging layer 32 can be reduced. As a result, when the substrate 50 is sliced, it is possible to minimize the area where a non-defective substrate cannot be obtained due to the occurrence of penetrating pits caused by the valley portion 30v. As a result, the yield of obtaining the substrate 50 can be improved.

At this time, the inclined interface enlarging layer 320 includes a c-plane growth region (first c-plane growth region) 60 grown with the c-plane 30c as the growth plane and a c-plane An inclined interface growth region 70 (gray portion in the drawing) is formed by growing using the inclined interface 30i other than the above as a growth surface.

At this time, in the c-plane growth region 60, valleys 60a are formed at positions where the inclined interfaces 30i are generated, and peaks 60b are formed at positions where the c-plane 30c disappears. In the c-plane growth region 60, a pair of inclined portions 60i are formed on both sides of the mountain portion 60b as loci of intersections between the c-plane 30c and the inclined interface 30i.

Further, at this time, when the first growth condition satisfies the formula (1), the angle β formed by the pair of inclined portions 60i when looking at the cross section passing through the respective centers of the two adjacent valley portions 60a is, for example, , 70° or less.

Details of these areas will be described later.

(S240: inclined interface maintenance step)
After the c-plane 30c has disappeared from the surface of the inclined interface enlarging layer 32, the growth conditions in the inclined interface maintaining step S240 are maintained at the first growth conditions described above, similarly to the inclined interface enlarging step S220.

As a result, as shown in FIG. 6(a), the three-dimensional growth layer 30 is grown over a predetermined thickness while maintaining the state where the inclined interface growth region 70 occupies 80% or more of the creeping cross section. continue. As a result, the graded interface maintaining layer 34 is formed on the graded interface enlarging layer 32 .

Here, in the three-dimensional growth step S200, in order to reliably bend the dislocation propagation direction and reduce the dislocation density as described above, when the history of the growth interface is observed at an arbitrary position of the three-dimensional growth layer 30, Furthermore, it is important that the c-plane 30c disappears at least once. Therefore, it is desirable that the c-plane 30c disappears at least once in an early stage of the three-dimensional growth step S200 (for example, the above-described inclined interface enlarging step S220).

However, in the inclined interface maintaining step S240, the c-plane 30c may reappear on a part of the surface of the inclined interface maintaining layer 34 after the c-plane 30c has disappeared at least once. However, the sloped interface maintaining layer is added so that the sloped interface growth region 70 accounts for 80% or more of the creepage cross section along the main surface 6s of the second underlayer 6 (hereinafter also simply referred to as the "creepage cross section"). At the surface of 34, it is preferable to expose mainly the inclined interface 30i. If the area ratio occupied by the inclined interface growth region 70 is less than 80% in the creeping cross section, cracks may occur during growth. Moreover, it may be difficult to perform processing such as slicing and polishing. On the other hand, in the present embodiment, by setting the area ratio of the inclined interface growth region 70 to 80% or more in the creeping cross section, it is possible to suppress the occurrence of cracks during the growth, and the slicing, polishing, and the like can be suppressed. Processing can be easily applied.

It should be noted that the higher the ratio of the area occupied by the inclined interface growth region 70 in the creeping cross section, the better, and is preferably 100%.

However, as described above, in the three-dimensional growth process 200, for example, the c-plane 30c reappears in part of the surface of the inclined interface sustaining layer 34, and the area ratio of the inclined interface growth region 70 in the creeping cross section is 100%. may be less than In this case, part of the three-dimensional growth layer 30 includes both the inclined interface growth region 70 and the c-plane growth region (second c-plane growth region) 80 . In the inclined interface growth region 70, oxygen as an n-type impurity is relatively easily incorporated, whereas in the mixed c-plane growth region 80, the incorporation of oxygen is relatively suppressed. Therefore, the oxygen concentration in the c-plane growth region 80 becomes lower than the oxygen concentration in the inclined interface growth region 70, and the carrier concentration in the c-plane growth region 80 becomes lower than the carrier concentration in the inclined interface growth region 70. . As a result, in the substrate 50 sliced from the region where the inclined interface growth region 70 and the c-plane growth region 80 are mixed, in-plane variations in carrier concentration may occur.

Therefore, in the three-dimensional growth step S200 of the present embodiment, for example, it is preferable to add the conductivity type impurity at a concentration higher than the concentration of oxygen taken into the inclined interface growth region 70 . Conductive impurities include, for example, at least one of Si and Ge as n-type impurities. For example, at least in the inclined interface maintaining step S240 of the three-dimensional growth step S200, when the inclined interface maintaining layer 34 is grown at the planned slicing position of the substrate 50, the conductivity type impurity may be added at the above concentration. Incidentally, the conductive impurity may be added to the entire three-dimensional growth layer 30 at the concentration described above. Addition of such a conductivity type impurity can suppress a relative decrease in the carrier concentration in the c-plane growth region. As a result, in-plane variations in carrier concentration can be suppressed in the substrate 50 .

At this time, in the inclined interface maintaining step S240, the inclined interface maintaining layer 34 is grown mainly using the inclined interface 30i as the growth surface under the first growth conditions, so that the inclined interface enlarging layer 32 At the position where the tilted interface 30i is exposed, the dislocation that has propagated while bending in the direction tilted with respect to the c-axis continues to propagate in the tilted interface maintaining layer 34 in the same direction. As a result, dislocations are locally collected at the meeting portions of the adjacent inclined interfaces 30 i in the inclined interface maintaining layer 34 . Of the plurality of dislocations gathered at the meeting portion of the adjacent inclined interfaces 30i in the inclined interface maintaining layer 34, the dislocations having mutually opposite Burgers vectors disappear at the time of meeting. Moreover, some of the plurality of dislocations collected at the meeting portion of the adjacent inclined interfaces 30i can form loops and propagate in the direction along the c-axis (that is, the surface side of the inclined interface maintaining layer 34). Suppressed. The other portion of the plurality of dislocations collected at the meeting portion of the adjacent inclined interfaces 30i in the inclined interface maintaining layer 34 propagates in a direction along the c axis from a direction inclined with respect to the c axis. It is changed again and propagates to the surface side of the inclined interface maintaining layer 34 . By eliminating some of the plurality of dislocations or suppressing propagation of some of the plurality of dislocations to the surface side of the c-plane enlarging layer 42 in this way, the dislocations on the surface of the inclined interface maintaining layer 34 Density can be reduced. In addition, by locally collecting dislocations, a low dislocation density region can be formed above a portion of the tilted interface maintaining layer 34 where dislocations are propagated in a direction tilted with respect to the c-axis.

At this time, in the three-dimensional growth step S200, the three-dimensional growth layer 30 having the inclined interface enlarging layer 32 and the inclined interface maintaining layer 34 is three-dimensionally grown on the underlying structure 10, thereby forming the underlying structure 10. A stress cancellation effect (stress relaxation effect) can be obtained in which the tensile stress accumulated in the second underlayer 6 is canceled out by the three-dimensional growth layer 30 .

One of the reasons why the three-dimensional growth layer 30 can obtain the stress canceling effect is, for example, as follows.

Compared to the c-plane growth region 60, the inclined interface growth region 70 formed in the three-dimensional growth step S200 incorporates oxygen more easily. Therefore, the oxygen concentration in the inclined interface growth region 70 is higher than the oxygen concentration in the c-plane growth region 60 described above. That is, the inclined interface growth region 70 can be considered as a high oxygen concentration region.

By incorporating oxygen into the high oxygen concentration region in this way, the lattice constant of the high oxygen concentration region can be made larger than the lattice constant of the regions other than the high oxygen concentration region (reference: Chris G. Van de Walle, Physical Review B vol.68, 165209 (2003)). In the c-plane growth region 60 grown from the second underlayer 6 or the three-dimensional growth layer 30 with the c-plane 30c as the growth plane, the curvature of the c-plane of the second underlayer 6 causes the c-plane to grow toward the center of curvature of the c-plane. There is a stress that concentrates on On the other hand, by relatively increasing the lattice constant of the high oxygen concentration region, it is possible to generate stress in the high oxygen concentration region that expands the c-plane 30c outward in the surface direction. As a result, the stress that concentrates toward the center of curvature of the c-plane 30c below the high oxygen concentration region and the stress that spreads the c-plane 30c in the high oxygen concentration region outward in the surface direction can be offset.

By obtaining the stress canceling effect of the three-dimensional growth layer 30 in this way, even if the three-dimensional growth layer 30 is grown thick while the tensile stress is accumulated in the second underlayer 6, the three-dimensional growth layer 30 will not be damaged. It is possible to suppress the occurrence of a stress difference between the underlying structure 10 side and the surface side. Thereby, it is possible to suppress the occurrence of cracks or the like in the three-dimensional growth layer 30 .

Moreover, at this time, the thickness of the inclined interface preserving layer 34 can be increased without causing cracks by obtaining the above-described stress canceling effect. Specifically, the thickness of the inclined interface maintaining layer 34 is, for example, a thickness obtained by adding 300 μm or more and 10 mm or less to the unevenness height from the valley portion 30 v to the top portion 30 t of the inclined interface enlarging layer 32 . By setting the thickness of the inclined interface preserving layer 34 added to the height of the unevenness to 300 μm or more, at least one or more substrates 50 can be sliced from the inclined interface preserving layer 34 in the slicing step S400 described later. On the other hand, by setting the thickness of the sloped interface preserving layer 34 to be added to the height of the unevenness to 10 mm, the final thickness is 650 μm. , at least 10 substrates 50 can be obtained even if the kerf loss of about 200 μm is considered.

At this time, by obtaining the above-described stress canceling effect, as the thickness of the inclined interface maintenance layer 34 is increased, the c-plane of the inclined interface maintenance layer 34 curved in a concave spherical shape is gradually flattened. can go That is, the radius of curvature of the c-plane of the inclined interface maintaining layer 34 can be gradually increased.

At this time, the lattice constant of the high oxygen concentration region of the three-dimensional growth layer 30 is relatively increased, and the three-dimensional growth layer 30 is grown while gradually flattening the c-plane 30c. As the c-plane 30c of the growth layer 30 is flattened, a stress for flattening the c-plane can be applied to the second underlayer 6 under the three-dimensional growth layer 30 . As a result, the entire second underlayer 6, which is concavely curved toward the main surface 6s in the as-grown state, can be gradually flattened.

Further, at this time, if the growth of the inclined interface preserving layer 34 is maintained mainly with the inclined interface 30 i as the growth surface at the end of the growth of the inclined interface preserving layer 34 , the second underlayer 6 is mainly formed on the surface of the inclined interface preserving layer 34 . The inclination angle formed by the inclined interface 30i with respect to the surface 6s may not necessarily be maintained. For example, at the end of the growth of the graded interface preserving layer 34, at least part of the recesses 30p of the graded interface preserving layer 34 may be embedded. In this case, the inclination angle of the inclined interface 30i may be gradually moderated, and the index m of the {11-2m} plane may be gradually increased.

Through the three-dimensional growth step S200 described above, the three-dimensional growth layer 30 having the inclined interface enlarging layer 32 and the inclined interface maintaining layer 34 is formed.

(S300: peeling step)
After the growth of the three-dimensional growth layer 30 is completed, the laminated structure 90 having the second base layer 6 and the three-dimensional growth layer 30 is separated from the base substrate 1 as shown in FIG. 6(b).

Specifically, in the process of cooling the inside of the chamber of the vapor phase growth apparatus, stress is generated between the laminated structure 90 and the underlying substrate 1 due to the difference in linear expansion coefficients. For example, the base substrate 1 made of sapphire has a larger coefficient of linear expansion than the laminated structure 90 mainly made of GaN, so that stress is generated that causes the base substrate 1 to contract with respect to the laminated structure 90 .

As a result, due to the flat gap formed between the metal nitride layer 5 and the second underlayer 6 , the laminate structure 90 described above is naturally peeled off from the undersubstrate 1 . As a result, the laminated structure 90 can be separated from the underlying substrate 1 without causing cracks in the laminated structure 90 .

At this time, as described above, by following the flattening of the c-plane 30c of the three-dimensional growth layer 30 during the growth process of the three-dimensional growth layer 30, the whole of the second underlayer 6 is gradually flattened. In the laminated structure 90 separated from the substrate 1, the back surface of the second base layer 6 can be flattened.

Through the steps described above, the laminated structure 90 of the present embodiment is obtained.

The steps from the second base layer forming step S150 to the peeling step S300 described above are continuously performed in the same vapor phase growth apparatus without exposing the base substrate 1 to the atmosphere. This prevents the formation of an unintended high oxygen concentration region (region having an excessively higher oxygen concentration than the inclined interface growth region 70) at the interface between the second underlayer 6 and the three-dimensional growth layer 30. can be suppressed.

(S400: Slicing step)
Next, as shown in FIG. 7, for example, the three-dimensional growth layer 30 is sliced with a wire saw along a cut plane substantially parallel to the main surface 6s of the second underlayer 6. Next, as shown in FIG. Thus, at least one nitride semiconductor substrate 50 (also referred to as substrate 50) is formed as an as-sliced substrate. At this time, the thickness of the substrate 50 is, for example, 300 μm or more and 700 μm or less.

At this time, for example, the substrate 50 is formed by slicing the inclined interface maintaining layer 34 . Further, for example, the inclined interface maintaining layer 34 is sliced at a position above the position where the c-plane growth region 60 continuing from the second underlayer 6 disappears last (that is, the crest 60b of the c-plane growth region 60). do. Thereby, the substrate 50 with reduced dislocations can be stably obtained.

(S500: polishing step)
Next, both surfaces of the substrate 50 are polished by a polishing device. At this time, the final thickness of the substrate 50 is, for example, 250 μm or more and 650 μm or less.

Through the steps S100 to S500 described above, the substrate 50 according to the present embodiment is manufactured.

(Semiconductor Laminate Manufacturing Process and Semiconductor Device Manufacturing Process)
After the substrate 50 is manufactured, for example, a semiconductor functional layer made of a Group III nitride semiconductor is epitaxially grown on the substrate 50 to produce a semiconductor laminate. After manufacturing the semiconductor laminate, electrodes and the like are formed using the semiconductor laminate, the semiconductor laminate is diced, and chips of a predetermined size are cut out. Thus, a semiconductor device is manufactured.

(2) Laminated Structure Next, a laminated structure 90 according to the present embodiment will be described with reference to FIG. 6(b).

The laminated structure 90 of this embodiment has, for example, the second base layer 6 and the three-dimensional growth layer 30 . In addition, the second underlayer 6 can be simply referred to as "underlayer".

The second underlayer 6 is made of, for example, a group III nitride single crystal. The second underlayer 6 has, for example, (a trace of) a mirror-finished main surface 6s. The closest low-index crystal plane to the main surface 6s of the second underlayer 6 is, for example, the c-plane.

The second underlayer 6 is formed, for example, on the back surface side opposite to the main surface 6s based on the difference in the growth surface in the second underlayer forming step S150. (not shown). The underlayer-side inclined interface growth region is a region grown using the inclined interface as a growth surface when the initial nucleus of the second underlayer 6 grows laterally. The underlayer-side inclined interface growth region has, for example, an oxygen concentration equivalent to that of an inclined interface growth region 70, which will be described later.

The three-dimensional growth layer 30 is grown on the second underlayer 6, for example.

The three-dimensional growth layer 30 is formed, for example, by forming a plurality of recesses 30p constituted by inclined interfaces 30i other than the c-plane on the top surface 30u of the single crystal of the group III nitride semiconductor, thereby eliminating the c-plane 30c. It has a plurality of valleys 30v and a plurality of crests 30t formed therein. When viewing an arbitrary cross section perpendicular to the main surface 6s of the second underlayer 6, the average distance between the closest tops is, for example, more than 100 μm.

Also, the three-dimensional growth layer 30 is divided into, for example, a c-plane growth region (first low oxygen concentration region) 60 and an inclined interface growth region (high oxygen concentration region) 70 based on the difference in the growth plane during the growth process. ,have.

The c-plane grown region 60 is a region grown using the c-plane 30c as a growth plane. As described above, in the c-plane growth region 60, oxygen uptake is suppressed compared to the inclined interface growth region 70. FIG. Therefore, the oxygen concentration in the c-plane growth region 60 is lower than that in the inclined interface growth region 70 . Specifically, the oxygen concentration in the c-plane growth region 60 is, for example, 5×10 ¹⁶ cm ⁻³ or less, preferably 3×10 ¹⁶ cm ⁻³ or less.

The c-plane growth region 60 is provided on the major surface 6s of the second underlayer 6 .

The c-plane growth region 60 has, for example, a plurality of valleys 60a and a plurality of peaks 60b in a cross-sectional view. Here, each of the valley portion 60a and the peak portion 60b means a part of the shape observed based on the difference in emission intensity when the cross section of the laminated structure 90 is observed with a fluorescence microscope or the like, and is three-dimensionally grown. It does not mean a part of the shape of the outermost surface that occurs during the growth of the layer 30 . Each of the plurality of troughs 60a is formed at a position where the inclined interface 30i is generated, which is a downwardly convex inflection point in the c-plane growth region 60 in a cross-sectional view. At least one of the plurality of troughs 60a is provided at a position spaced upward from the main surface 6s of the second underlayer 6 . On the other hand, each of the plurality of peaks 60b is an upwardly convex inflection point in the c-plane growth region 60 in a cross-sectional view, and sandwiches a pair of inclined interfaces 30i expanded in opposite directions. is formed at the position where the c-plane 30c (finally) disappears. The trough portions 60a and the peak portions 60b are alternately formed in the direction along the main surface 6s of the second underlayer 6 .

When viewing an arbitrary cross section perpendicular to the main surface 6s of the second underlayer 6, the average distance between the nearest peaks during the growth process of the three-dimensional growth layer 30 is Corresponds to the average distance. The average distance between peaks 60b of the c-plane growth region 60 is, for example, over 100 μm.

The c-plane growth region 60 has a pair of inclined portions 60i provided as trajectories of intersections of the c-plane 30c and the inclined interface 30i on both sides of one of the plurality of mountain portions 60b in a cross-sectional view. ing. Note that the inclined portion 60i here means a portion of the shape observed based on the difference in emission intensity when the cross section of the laminated structure 90 is observed with a fluorescence microscope or the like. It does not mean the resulting uppermost inclined interface 30i.

An angle β formed by a pair of inclined portions 60i when viewing a cross section passing through the respective centers of two adjacent valley portions 60a is, for example, 70° or less, preferably 20° or more and 65° or less. The fact that the angle β formed by the pair of inclined portions 60i is 70° or less means that the growth rate G _i of the inclined interface 30i most inclined with respect to the c-plane 30c in the three-dimensional growth layer 30 under the first growth condition is , means that the ratio G _c1 /G _i of the growth rate G _c1 of the c-plane 30 c of the three-dimensional growth layer 30 was high. This makes it possible to easily generate the inclined interface 30i other than the c-plane. As a result, the dislocation can be easily bent at the position where the inclined interface 30i is exposed. Further, by setting the angle β formed by the pair of inclined portions 60i to be 70° or less, it is possible to easily generate the plurality of valley portions 30v and the plurality of peak portions 30t above the second underlayer 6 . Furthermore, by setting the angle β formed by the pair of inclined portions 60i to 65° or less, the inclined interface 30i other than the c-plane can be more easily generated, and above the second underlayer 6, a plurality of valleys can be formed. 30v and multiple tops 30t can be produced more easily. By setting the angle β formed by the pair of inclined portions 60i to 20° or more, it is possible to suppress the height from the valley portion 30v to the top portion 30t of the three-dimensional growth layer 30 from increasing. As a result, when the substrate 50 is sliced, it is possible to minimize the area where a non-defective substrate cannot be obtained due to the occurrence of penetrating pits caused by the valley portion 30v.

On the other hand, the inclined interface growth region 70 is a region grown using the inclined interface 30i other than the c-plane as a growth plane. The inclined interface growth region 70 incorporates oxygen more easily than the c-plane growth region 60 . Therefore, the oxygen concentration in the inclined interface growth region 70 is higher than that in the c-plane growth region 60 . The oxygen taken into the inclined interface growth region 70 is, for example, oxygen that is unintentionally mixed into the vapor phase growth apparatus, or oxygen that is released from members (such as quartz members) constituting the vapor phase growth apparatus. is. Specifically, the oxygen concentration in the graded interface growth region 70 is, for example, 9×10 ¹⁷ cm ⁻³ or more and 5×10 ¹⁹ cm ^{−3 or} less.

The sloped interface growth region 70 is provided on the c-plane growth region 60 . The lower surface of the inclined interface growth region 70 is formed following the shape of the c-plane growth region 60, for example.

At least part of the inclined interface growth region 70 is provided continuously along the main surface 6 s of the second underlayer 6 . That is, when viewing a plurality of creepage cross sections of the three-dimensional growth layer 30 cut along the main surface 6s of the second underlayer 6, the cross section that does not include the c-plane growth region grown with the c-plane 30c as the growth plane is It is desirable to exist in at least part of the three-dimensional growth layer 30 in the thickness direction.

By continuously growing the inclined interface sustaining layer 34 mainly using the inclined interface 30i as the growth surface, the three-dimensional growth layer 30 is grown at the inclined interface in the creeping cross section cut along the main surface 6s of the second underlayer 6. The area ratio occupied by the region 70 is, for example, 80% or more.

As described above, the area ratio of the inclined interface growth region 70 may be less than 100% in a predetermined creeping cross section. In other words, a creeping cross section in which the inclined interface growth region 70 and the c-plane growth region (second low oxygen concentration region) 80 are mixed may occur. The c-plane growth region 80 has, for example, the same oxygen concentration as that of the c-plane growth region 60 described above.

During the growth process of the three-dimensional growth layer 30, the c-plane growth regions 80 appear and disappear. It varies randomly towards the surface of layer 30 .

In addition, since the c-plane 30c disappears at least once during the growth process of the three-dimensional growth layer 30, the c-plane growth region 80 is continuous from the second underlayer 6 to the surface (uppermost surface) of the three-dimensional growth layer 30. not.

Further, in the present embodiment, during the growth process of the three-dimensional growth layer 30, the dislocation bends in the direction substantially perpendicular to the inclined interface 30i at the position where the inclined interface 30i other than the c-plane is exposed. By propagating, in the graded interface preserving layer 34, some of the plurality of dislocations are eliminated, or propagation of some of the plurality of dislocations to the surface side of the graded interface preservation layer 34 is suppressed. Thereby, the dislocation density on the surface of the graded interface maintaining layer 34 is lower than the dislocation density on the main surface 6 s of the second underlayer 6 .

In addition, in the present embodiment, the entire surface of the three-dimensional growth layer 30 is oriented along the +c plane, and the three-dimensional growth layer 30 does not include polarity inversion domains (inversion domains). In this respect, the laminated structure 90 of the present embodiment is different from the laminated structure formed by the so-called DEEP (Dislocation Elimination by the Epitaxial-growth with inverse-pyramidal pits) method, that is, it is positioned at the center of the pit. It is different from a laminated structure that contains a polarity reversal area in the core.

(3) Nitride semiconductor substrate (nitride semiconductor free-standing substrate, nitride crystal substrate)
Next, the nitride semiconductor substrate 50 according to this embodiment will be described with reference to FIGS. 9 and 10. FIG. 9A is a schematic top view showing the nitride semiconductor substrate according to this embodiment, and FIG. 9B is a schematic cross-sectional view along the m-axis of the nitride semiconductor substrate according to this embodiment, (c) is a schematic cross-sectional view along the a-axis of the nitride semiconductor substrate according to the present embodiment. FIG. 10 is a schematic diagram showing a cathodoluminescence image of the main surface of the nitride semiconductor substrate according to this embodiment observed with a scanning electron microscope (SEM).

In this embodiment, the substrate 50 obtained by the manufacturing method described above is, for example, a self-supporting substrate made of a single crystal of a Group III nitride semiconductor. In this embodiment, the substrate 50 is, for example, a GaN free-standing substrate.

The diameter of substrate 50 is, for example, 2 inches or more. Also, the thickness of the substrate 50 is, for example, 300 μm or more and 1 mm or less.

The substrate 50 has, for example, a main surface 50s that serves as an epitaxial growth surface. In the present embodiment, the closest low-index crystal plane to the principal plane 50s is, for example, the c-plane 50c.

The main surface 50s of the substrate 50 is, for example, mirror-finished, and the root-mean-square roughness RMS of the main surface 50s of the substrate 50 is, for example, less than 1 nm.

Also, in this embodiment, the substrate 50 does not include, for example, inversion domains, as described above.

(impurity concentration)
Next, with reference to FIG. 10, features related to the impurity concentration in the main surface 50s of the substrate 50 will be described.

As shown in FIG. 10, the substrate 50 has a high oxygen concentration region 70, for example. The high oxygen concentration region 70 is, for example, an inclined interface growth region 70 grown using the inclined interface 30i as a growth surface as shown in FIG. 6(a). In addition, the same code|symbol is used about these. The oxygen concentration in the high oxygen concentration region 70 is, for example, 9×10 ¹⁷ cm ⁻³ or more and 5×10 ¹⁹ cm ^{−3 or} less as described above.

Since the high-oxygen-concentration region 70 contains high-concentration oxygen as described above, cathodoluminescence when imaged in a wavelength range including at least part of the wavelength of light emitted near the bandgap energy of the group III nitride semiconductor In the image (CL image) (or multiphoton excitation microscope image (2PPL image)), the high oxygen concentration region 70 is observed relatively bright.

The shape of the high oxygen concentration region 70 in plan view reflects, for example, the shape in plan view of the recess 30p generated in the growth process of the three-dimensional growth layer 30, and has at least a portion of a substantially hexagonal shape. . Among the shapes of the high oxygen concentration region 70 in a plan view, one substantially hexagon may intersect another substantially hexagon. In the 2PPL image, traces of the ridgelines of the concave portions 30p generated during the growth process of the three-dimensional growth layer 30 may be seen in the high oxygen concentration region 70. FIG.

In the present embodiment, the area ratio of the high oxygen concentration region 70 in the main surface 50s is, for example, 80% or more. In other words, the low oxygen concentration region 80 may exist at an area ratio of 20% or less in the main surface 50s.

The low oxygen concentration region 80 is, for example, a c-plane growth region 80 in which the c-plane 30c reappears in the graded interface preserving layer 34 shown in FIG. 6(a). In addition, the same code|symbol is used about these. The oxygen concentration in the low oxygen concentration region 80 is, for example, 5×10 ¹⁶ cm ⁻³ or less, preferably 3×10 ¹⁶ cm ^{−3 or} less, as described above.

The size of the low-oxygen-concentration region 80 in plan view varies, for example, randomly from the back surface opposite to the main surface 50s of the substrate 50 toward the main surface 50s.

For example, the low oxygen concentration region 80 is not continuously connected from the back surface opposite to the main surface 50s of the substrate 50 toward the main surface 50s.

On the other hand, the area ratio of the high oxygen concentration region 70 in the main surface 50s may be 100%, that is, the substrate 50 may not have the low oxygen concentration region 80 .

Based on the area ratio of the high oxygen concentration region 70 in the main surface 50s, the average oxygen concentration of the entire main surface 50s of the substrate 50 is, for example, 7×10 ¹⁷ cm ⁻³ or more and 5×10 ¹⁹ cm ^{−3 .} It is below.

In this embodiment, the substrate 50 obtained by the manufacturing method described above is, for example, n-type. The substrate 50 of the present embodiment contains at least one of Si and Ge as well as the oxygen (O) described above, for example, as n-type impurities. The total n-type impurity concentration in the substrate 50 is, for example, 1.0×10 ¹⁸ cm ⁻³ or more and 1.0×10 ²⁰ cm ⁻³ or less.

In the substrate 50 of this embodiment, not only at least one of Si and Ge, but also O is activated. Therefore, the free electron concentration in the substrate 50 is equivalent to the total concentration of O, Si and Ge in the substrate 50, for example.

In this embodiment, the concentration of impurities other than the n-type impurities (conductivity-type impurities) in the substrate 50 obtained by the manufacturing method described above is low.

For example, the hydrogen concentration in the substrate 50 obtained by the above manufacturing method is lower than that of the substrate obtained by the flux method, the ammonothermal method, or the like.

Specifically, the hydrogen concentration in the substrate 50 is, for example, less than 1.times.10.sup.17 ^{cm.sup. -3} , preferably ^{5.times.10.sup.16} cm.sup. ^-3 or less.

(optical properties)
Next, optical properties of the substrate 50 will be described.

The substrate 50 of this embodiment has the high oxygen concentration region 70 as described above. However, the absorption coefficient of the substrate 50 in the visible light region is equivalent to that of a high-quality nitride semiconductor substrate obtained by the VAS method, for example.

Specifically, in the substrate 50 of the present embodiment, the absorption coefficient in the wavelength range from 500 nm to 700 nm is, for example, 0.15 cm ⁻¹ or less. Note that the absorption coefficient of the substrate 50 in the wavelength range may be lower than, for example, the absorption coefficient of the nitride semiconductor substrate obtained by the conventional VAS method.

In addition, the substrate 50 manufactured in this embodiment has a small crystal strain and hardly contains impurities other than conductive impurities (for example, impurities compensating for n-type impurities). As a result, in the substrate 50 of the present embodiment, the absorption coefficient in the infrared region based on free carrier absorption can be approximated as a function of free carrier concentration and wavelength as follows.

Specifically, the wavelength is λ (μm), the absorption coefficient of the substrate 50 at 27° C. is α (cm ⁻¹ ), the free electron concentration in the substrate 50 is n (cm ⁻³ ), and K and a are constants. Then, in the substrate 50 of the present embodiment, the absorption coefficient α in the wavelength range of at least 1 μm to 2.5 μm is approximated by the following equation (3) by the least-squares method.
α= ^nKλa (3)
(However, 1.5×10 ⁻¹⁹ ≦K≦6.0×10 ⁻¹⁹ , a=3)

At a wavelength of 2 μm, the error of the actually measured absorption coefficient with respect to the absorption coefficient α obtained from Equation (3) is, for example, within ±0.1α, preferably within ±0.01α.

In the substrate 50 of this embodiment, the high oxygen concentration region 70 and the low oxygen concentration region 80 may coexist as described above. However, even in this case, the high-oxygen concentration regions 70 and the low-oxygen concentration regions 80 are evenly and randomly dispersed within the main surface 50s of the substrate 50 . The size of the low-oxygen concentration region 80 in plan view varies randomly in the thickness direction. Also, the low oxygen concentration region 80 is not continuous in the thickness direction. Therefore, when the optical characteristics of the substrate 50 are measured with a spectrometer using a predetermined light irradiation size sufficiently larger than the size of the low oxygen concentration region 80 in plan view, the in-plane oxygen concentration in the substrate 50 is Information is obtained that averages the effect of the distribution and the effect of the oxygen concentration distribution in the thickness direction. Therefore, in the substrate 50 of the present embodiment, in-plane variations in transmittance and reflectance are reduced.

Specifically, when the transmittance of light in the wavelength range of 500 nm or more and 700 nm or less is measured at a plurality of different points in the main surface 50s of the substrate 50, the transmission at each wavelength within the wavelength range of 500 nm or more and 700 nm or less The variation of the ratio within the main surface 50s is, for example, within ±0.5%, preferably within ±0.3%.

However, as a condition for transmittance measurement, the light irradiation size is, for example, 3 mm×3 mm. Also, the incident angle of light with respect to the normal to the main surface 50s is assumed to be 0°. Also, there is no reference in the transmittance measurement, that is, the transmittance is assumed to be 100% when all the light is transmitted. The number of measurement points is, for example, 3 or more, preferably 5 or more.

Further, when the reflectance of light in the wavelength range of 500 nm or more and 700 nm or less is measured at a plurality of different points in the main surface 50 s of the substrate 50, the reflectance at each wavelength within the wavelength range of 500 nm or more and 700 nm or less, The variation within the main surface 50s is, for example, within ±0.5%, preferably within ±0.3%.

However, as a condition for reflectance measurement, the light irradiation size is set to, for example, 10 mm×5 mm. Also, the incident angle of light with respect to the normal to the main surface 50s is 8°. An Al mirror is used as a reference in reflectance measurement. The number of measurement points is, for example, 3 or more, preferably 5 or more.

Details of the above absorption coefficient, transmittance and reflectance will be described later in Examples.

(scotoma)
Next, dark spots on the main surface 50s of the substrate 50 of this embodiment will be described with reference to FIG. The term "dark point" as used herein means a point having a low emission intensity observed in an observation image of the main surface 50s with a multiphoton excitation microscope, a cathodoluminescence image of the main surface 50s, or the like. However, it also contains non-radiative centers caused by foreign matter or point defects. Note that the “multiphoton excitation microscope” is also called a two-photon excitation fluorescence microscope.

In the present embodiment, since the substrate 50 is manufactured using the base structure 10 manufactured by the VAS method that can obtain high-purity GaN single crystals, non-uniformity caused by foreign matter or point defects in the substrate 50 is present. Fewer luminescent centers.

Therefore, in the CL image (or 2PPL image) of the main surface 50s of the substrate 50 shown in FIG. 10, 95% or more, preferably 99% or more of the dark spots are not non-luminous centers caused by foreign matter or point defects. , dislocation (threading dislocation) d.

In addition, in the present embodiment, dislocations are locally collected in the graded interface maintaining layer 34 by the manufacturing method described above, and the dislocation density on the surface of the graded interface maintaining layer 34 is reduced to the dislocation density on the main surface 6s of the second underlayer 6. reduced than density. As a result, the dislocation density is also reduced in the main surface 50s of the substrate 50 formed by slicing the inclined interface maintaining layer 34 .

Further, in this embodiment, as shown in FIG. 10, dislocations d are locally concentrated at the center of the substantially hexagonal high oxygen concentration region 70 by the above-described manufacturing method. Hereinafter, the region in which the dislocations d are relatively concentrated in the high oxygen concentration region 70 is also referred to as "dislocation concentrated region dca". A low dislocation density region is widely formed outside the dislocation concentrated region dca.

In addition, in the present embodiment, the three-dimensional growth step S200 is performed using the underlying structure 10 that has not been patterned by the above-described manufacturing method, so that the three-dimensional growth layer 30 is sliced and formed. In the main surface 50 s of the substrate 50 , the high dislocation density regions that are regularly generated due to the patterning of the underlying structure 10 are not formed. In other words, in the substrate 50 of the present embodiment, even if dislocation concentrated regions dca exist, the dislocation concentrated regions dca are randomly arranged. Further, the dislocation density in the dislocation concentrated region dca of the substrate 50 of this embodiment is lower than that in the case where the underlying structure 10 is patterned.

Specifically, in the present embodiment, when the main surface 50s of the substrate 50 is observed with a multiphoton excitation microscope in a field of view of 250 μm square and the dislocation density is obtained from the dark spot density, the dislocation density is 3×10 ⁶ cm ⁻ There are no regions above ² . Further, the region having a dislocation density of less than 1×10 ⁶ cm ⁻² is present in 80% or more, preferably 90% or more, more preferably 95% or more of the main surface 50s. Observation of the cathodoluminescence image also yields the same result as obtained with the multiphoton excitation microscope.

Note that when the manufacturing method of the present embodiment is used, the upper limit of the proportion of the region with a dislocation density of less than 1×10 ⁶ cm ⁻² may be 99% of the main surface 50s, for example.

In other words, in the present embodiment, the average dislocation density of the entire main surface 50s of the substrate 50 is, for example, less than 1×10 ⁶ cm ⁻² , preferably less than 5.5×10 ⁵ cm ⁻² . , more preferably 3×10 ⁵ cm ⁻² or less.

Further, the main surface 50s of the substrate 50 of the present embodiment includes a dislocation-free region of at least 50 μm square based on the average distance L between the closest tops in the three-dimensional growth step S200 described above, for example. Dislocation-free regions of 50 μm square are scattered over the entire main surface 50s of the substrate 50 except for the dislocation concentrated regions dca described above, for example. In addition, the main surface 50s of the substrate 50 of the present embodiment has, for example, 100/cm ² or more, preferably 800/cm ² or more, more preferably 1600/cm ^{2 or} more dislocation-free regions of 50 μm square that do not overlap. has a density of When the density of non-overlapping 50 μm-square dislocation-free regions is 1600/cm ² or more, for example, the main surface 50s has at least one 50 μm-square dislocation-free region within an arbitrary 250 μm-square field of view. do.

The upper limit of the density of non-overlapping 50 μm square dislocation-free regions is, for example, about 30000/cm ² .

For reference, in a substrate obtained by a conventional manufacturing method that does not perform a special step of collecting dislocations, the size of the dislocation-free regions is smaller than 50 μm square, or the density of the dislocation-free regions of 50 μm square is 100 pieces. / ^cm2 . Also, in the substrate obtained by the conventional ELO method, the size of the dislocation-free regions is smaller than 50 μm square, or the density of the dislocation-free regions of 50 μm square is lower than 100/cm ² .

Further, in the present embodiment, even when the substrate 50 has the low oxygen concentration region 80 as described above, the c-plane 30c disappears at least once during the growth process of the three-dimensional growth layer 30. Below the low oxygen concentration region 80 on the 50s side, dislocations are bent. As a result, for example, the low-oxygen-concentration region 80 in the main surface 50s may include a dislocation-free region of at least 50 μm square as described above.

Further, in the present embodiment, the dislocation density in the dislocation concentrated regions dca is low as described above, so that the main surface 50s can be observed with a multiphoton excitation microscope in a 50 μm square field of view including the dislocation concentrated regions dca. The dislocation density obtained from the dark spot density is, for example, less than 3×10 ⁶ cm ⁻² . Since it is considered that at least one dark spot exists in the 50 μm-square field of view including the dislocation concentrated region dca, the lower limit of the dislocation density in the 50 μm-square field of view including the dislocation concentrated region dca is, for example, 4× 10 ⁴ cm ⁻² .

(basal plane dislocation)
Next, basal plane dislocations bpd on the main surface 50s of the substrate 50 of this embodiment will be described with reference to FIG.

In the present embodiment, crystal strain is applied to the three-dimensional growth layer 30 due to the stress cancellation effect of the tilted interface growth region 70 described later, and basal plane dislocations may occur. Therefore, as shown in FIG. 10 , basal plane dislocations bpd may be observed in the CL image of the main surface 50s of the substrate 50 formed by slicing the three-dimensional growth layer 30 .

However, in this embodiment, the number of basal plane dislocations bpd in the main surface 50s of the substrate 50 is less than that of a substrate obtained by the ELO method using a mask layer.

Specifically, in the CL image of the main surface 50s, when an arbitrary virtual line segment ls having a length of 200 μm is drawn, the number of intersections between the line segment ls and the basal plane dislocations bpd is, for example, 10. points or less, preferably 5 points or less. The minimum number of intersections between the line segment ls and the basal plane dislocations bpd is, for example, 0 points.

(Bergers vector)
Next, the Burgers vector of dislocations in the substrate 50 of this embodiment will be described.

In the present embodiment, since the dislocation density in the main surface 6s of the second underlayer 6 used in the manufacturing method described above is low, a plurality of dislocations are generated when the three-dimensional growth layer 30 is grown on the second underlayer 6. Less likely to combine (mix). Thereby, generation of dislocations having a large Burgers vector can be suppressed in the substrate 50 obtained from the three-dimensional growth layer 30 .

Specifically, in the substrate 50 of the present embodiment, for example, there are many dislocations whose Burgers vector is <11-20>/3, <0001>, or <11-23>/3. The “Bergers vector” here can be measured by, for example, a large angle convergence electron diffraction method (LACBED method) using a transmission electron microscope (TEM). Further, a dislocation with a Burgers vector of <11-20>/3 is an edge dislocation, a dislocation with a Burgers vector of <0001> is a screw dislocation, and a Burgers vector of <11-23>/3 is a screw dislocation. Some dislocations are mixed dislocations in which edge dislocations and screw dislocations are mixed.

In the present embodiment, when 100 dislocations on the main surface 50s of the substrate 50 are randomly extracted, the Burgers vector is <11-20>/3, <0001> or <11-23>/3. is, for example, 50% or more, preferably 70% or more, and more preferably 90% or more. Dislocations having a Burgers vector of 2<11-20>/3 or <11-20> may be present in at least a portion of the main surface 50s of the substrate 50 .

(curvature radius of c-plane)
Next, the radius of curvature of the c-plane 50c of the substrate 50 of this embodiment will be described with reference to FIG.

As shown in FIGS. 9B and 9C, in the present embodiment, the c-plane 50c as the low-index crystal plane closest to the major surface 50s of the substrate 50 is, for example, the substrate 50 manufactured as described above. Due to the method, it is either flat or spherically curved. In addition, the c-plane 50c of the substrate 50 may be curved in a concave spherical shape with respect to the main surface 50s, or may be curved in a convex spherical shape, for example. Alternatively, if the c-surface 50c of the substrate 50 is substantially flat, for example, a portion curved in a spherical shape concave with respect to the main surface 50s and a portion curved in a spherical shape convex with respect to the main surface 50s and may have

In this embodiment, the c-plane 50c of the substrate 50 is, for example, flat in each of the cross section along the m-axis and the cross-section along the a-axis, or has a curved surface that approximates a spherical surface. .

Let “x” be the coordinate in the direction along the m-axis among the positions within the main surface 50 s of the substrate 50 . On the other hand, among the positions within the main surface 50s of the substrate 50, let the coordinate in the direction along the a-axis be "y". The coordinates (x, y) of the center of the main surface 50s of the substrate 50 are set to (0, 0). Of the off angle θ of the c-axis 50ca with respect to the normal to the main surface 50s, the directional component along the m-axis is defined as "θ _m ", and the directional component along the a-axis is defined as "θ _a ". Note that θ ² =θ _m ² +θ _a ² .

In the present embodiment, since the c-plane 50c of the substrate 50 is flat or spherically curved as described above, the off-angle m-axis component _θm and the off-angle a-axis component _θa are , can be approximately represented by a linear function of x and a linear function of y, respectively.

Specifically, for example, the X-ray rocking curve measurement of the (0002) plane is performed at each position on a straight line passing through the center in the main surface 50s, and the peak formed by the X-ray incident on the main surface 50s and the main surface 50s When the angle ω is plotted against position on a line, the peak angle ω can be approximated by a linear function of position. Here, the “peak angle ω” is the angle formed by the X-ray incident on the main surface 50s and the main surface 50s, and is the angle at which the diffraction intensity is maximized. The radius of curvature of the c-plane 50c can be obtained from the reciprocal of the slope of the linear function approximated as described above.

In this embodiment, the radius of curvature of the c-plane 50c of the substrate 50 is, for example, larger than the radius of curvature of the c-plane of the nitride semiconductor substrate obtained by the conventional VAS method.

Specifically, when the peak angle ω is approximated by a linear function of position in the X-ray rocking curve measurement of the c-plane 50c, the radius of curvature of the c-plane 50c obtained by the reciprocal of the slope of the linear function is, for example, , 15 m or more, preferably 20 m or more, more preferably 30 m or more, and even more preferably 40 m or more.

In the present embodiment, the upper limit of the radius of curvature of the c-plane 50c of the substrate 50 is not particularly limited because the larger the better. When the c-plane 50c of the substrate 50 is substantially flat, it can be considered that the radius of curvature of the c-plane 50c is infinite.

In the present embodiment, since the radius of curvature of the c-plane 50c of the substrate 50 is large, the variation in the off-angle θ of the c-axis 50ca with respect to the normal to the main surface 50s of the substrate 50 is measured by the conventional VAS method. It can be made smaller than the variation of the off-angle of the c-axis of the substrate.

In addition, in the present embodiment, when the peak angle ω is approximated by a linear function of the position in the X-ray rocking curve measurement of the c-plane 50c, it is 1 The error of ω with respect to the following function is small.

Specifically, the error of the measured peak angle ω with respect to the linear function approximated as described above is, for example, 0.05° or less, preferably 0.02° or less, more preferably 0.01° or less. is. Note that the minimum value of the error is 0° because at least part of the peak angle ω may match the linear function.

(Photoluminescence characteristics in semiconductor laminate)
Next, the photoluminescence characteristics of the semiconductor layer when the semiconductor laminate is produced using the substrate 50 will be described.

In this embodiment, the substrate 50 has the high oxygen concentration region 70 as described above. , crystal distortion caused by the substrate 50 hardly occurs. Therefore, in the photoluminescence of the semiconductor layer grown on the substrate 50, the peak shift caused by the crystal strain of the semiconductor layer is small or absent.

Specifically, the following characteristics are obtained. First, a laminate having a substrate 50 and a semiconductor layer formed by epitaxially growing a non-doped single crystal of a predetermined Group III nitride semiconductor on the main surface 50s of the substrate 50 is produced. On the other hand, a reference substrate made of a single crystal which is the same group III nitride semiconductor as the substrate 50 and has an in-plane uniform oxygen concentration lower than that of the high oxygen concentration region 70, and a reference substrate which is the same single crystal as the semiconductor layer described above. and a reference semiconductor layer epitaxially grown thereon. The term "non-doped" as used herein means that the semiconductor does not contain intentionally added impurities, and includes the case where the semiconductor contains unavoidable impurities. Next, the photoluminescence in each of the semiconductor layer of the laminate of this embodiment and the reference semiconductor layer of the reference laminate is measured with a temperature difference of less than 1° C. (for example, at 27° C.). In this case, the difference between the maximum peak wavelength in the semiconductor layer of the laminate of this embodiment and the maximum peak wavelength in the semiconductor layer of the reference laminate is, for example, 1 nm or less.

(4) Effects obtained by this embodiment According to this embodiment, one or more of the following effects can be obtained.

(a) In the three-dimensional growth step S200, by forming an inclined interface 30i other than the c-plane on the surface of the single crystal constituting the three-dimensional growth layer 30, at the position where the inclined interface 30i is exposed, Dislocations can be bent and propagated in a direction substantially perpendicular to them. This allows dislocations to be locally collected. By locally gathering dislocations, dislocations having mutually opposite Burgers vectors can be eliminated. Alternatively, locally collected dislocations form loops, thereby suppressing propagation of dislocations to the surface side of the three-dimensional growth layer 30 . In this way, the dislocation density on the surface of the three-dimensional growth layer 30 can be reduced. As a result, the substrate 50 having a lower dislocation density than the main surface 6s of the second underlayer 6 can be obtained.

(b) In the three-dimensional growth step S200, the c-plane 30c disappears from the top surface 30u of the three-dimensional growth layer 30 at least once. Thereby, the dislocations propagating from the second underlayer 6 can be reliably bent at the position where the inclined interface 30i in the three-dimensional growth layer 30 is exposed.

Here, consider the case where the c-plane remains in the three-dimensional growth process. In this case, in the portion where the c-plane remains, the dislocation propagated from the second underlayer propagates substantially vertically upward without being bent, and reaches the surface of the three-dimensional growth layer. Therefore, dislocations are not reduced above the portion where the c-plane remains, and a high dislocation density region is formed.

On the other hand, according to the present embodiment, in the three-dimensional growth step S200, the c-plane 30c disappears from the top surface 30u of the three-dimensional growth layer 30 at least once, thereby changing the surface of the three-dimensional growth layer 30 at least once. It can be composed only of the inclined interface 30i other than the c-plane. Thereby, dislocations propagating from the second underlayer 6 can be reliably bent over the entire surface of the three-dimensional growth layer 30 . By reliably bending the dislocations, some of the plurality of dislocations can be made easier to disappear, or some of the plurality of dislocations can be less likely to propagate to the surface side of the three-dimensional growth layer 30 . As a result, it becomes possible to reduce the dislocation density over the entire main surface 50s of the substrate 50 obtained from the three-dimensional growth layer 30 .

(c) By three-dimensionally growing the three-dimensional growth layer 30 on the underlying structure 10, the tensile stress accumulated in the second underlying layer 6 of the underlying structure 10 is offset by the three-dimensional growth layer 30 as described above. A stress canceling effect (stress relaxation effect) can be obtained.

By obtaining the stress canceling effect of the three-dimensional growth layer 30 in this way, even if the three-dimensional growth layer 30 is grown thick while the tensile stress is accumulated in the second underlayer 6, the three-dimensional growth layer 30 will not be damaged. It is possible to suppress the occurrence of a stress difference between the underlying structure 10 side and the surface side. Thereby, it is possible to suppress the occurrence of cracks or the like in the three-dimensional growth layer 30 .

(d) In the present embodiment, by obtaining the above-described stress canceling effect of the three-dimensional growth layer 30, the radius of curvature of the c-plane 50c of the substrate 50 obtained from the three-dimensional growth layer 30 can be easily controlled. For example, the radius of curvature of the c-plane 50c of the substrate 50 is the same as that of the nitride semiconductor obtained by growing the second underlayer to the same thickness as the three-dimensional growth layer 30 and slicing the second underlayer without performing the three-dimensional growth step. The radius of curvature of the substrate, that is, the nitride semiconductor substrate obtained by the conventional VAS method can be made larger. As a result, the variation in the off-angle θ of the c-axis 50ca with respect to the normal to the main surface 50s of the substrate 50 can be made smaller than the variation in the off-angle of the c-axis of the nitride semiconductor substrate obtained by the conventional VAS method. .

As described in (a) to (d) above, in this embodiment, the substrate 50 with good crystal quality can be obtained easily and stably.

(e) In the present embodiment, the second underlayer formation step S150 is performed between the void formation step S140 and the three-dimensional growth step S200 to mirror-finish the main surface 6s of the second underlayer 6 . As a result, the growth mode of the three-dimensional growth layer 30 can be changed from the growth mode of the island-shaped crystals generated in the initial stage of growth of the second underlayer 6 .

Here, if the three-dimensional growth step S200 is performed without performing the second underlayer formation step S150 immediately after the void formation step S140, the metal nitride layer is formed above the void-containing first underlayer from the initial stage of growth of the three-dimensional growth layer. A three-dimensional growth layer grows as island crystals via the . In this case, the frequency of island crystals in the three-dimensional growth layer depends on the degree of supersaturation when growing on the nanonets of the metal nitride layer and the variation in the opening width of the nanonets, as described above. Therefore, the tops of the three-dimensional growth layer are densely formed, and the average distance between the nearest tops of the three-dimensional growth layer is shortened. If the average distance between the nearest apexes of the three-dimensional grown layer is short, dislocations cannot be collected sufficiently. As a result, it may not be possible to sufficiently reduce the dislocation density on the surface of the three-dimensional growth layer.

On the other hand, in the present embodiment, after the void forming step S140, the main surface 6s of the second underlayer 6 is mirror-finished, so that the growth mode of the three-dimensional growth layer 30 is changed to that of the second underlayer 6 as described above. 6 can be changed from the island-like crystal growth morphology generated in the initial stage of growth.

That is, the three-dimensional growth layer 30 grown on the mirror-finished second underlayer 6 is not grown as island crystals from the initial stage of growth, but depends on the first growth conditions in the three-dimensional growth step S200. can be grown three-dimensionally. At this time, the average distance between the nearest peaks of the three-dimensional growth layer 30 is the growth rate G _c0 of the c-plane 30c in the c-axis direction and the growth rate G _i depends on the difference between This makes it possible to control the average distance between the nearest apexes of the three-dimensional growth layer 30 based on the first growth condition and lengthen the distance between the nearest apexes.

In addition, in the present embodiment, after the void formation step S140, the main surface 6s of the second underlayer 6 is mirror-finished so that the second underlayer 6 can be formed regardless of the states of the void-containing first underlayer 4 and the metal nitride layer 5 The morphology of the main surface 6s of the stratum 6 can be made substantially uniform throughout. By making the morphology of the main surface 6s of the second underlayer 6 substantially uniform, the state of occurrence of the inclined interface 30i in the three-dimensional growth step S200 can be made substantially uniform over the entire surface of the three-dimensional growth layer 30. can. As a result, it is possible to suppress the formation of a region having a short distance between the nearest vertices on a part of the surface of the three-dimensional growth layer 30, and the distance between the nearest vertexes is reduced over the entire surface of the three-dimensional growth layer 30. It can be lengthened substantially uniformly.

As a result, in the present embodiment, formation of a region with an excessively high dislocation density on a part of the surface of the three-dimensional growth layer 30 is suppressed, and dislocations are formed over the entire surface of the three-dimensional growth layer 30. Density can be lowered.

(f) In the present embodiment, among the planarization of the main surface 6s of the second underlayer 6, the formation of the mask layer on the main surface 6s of the second underlayer 6, and the formation of the uneven pattern on the main surface 6s, , the three-dimensional growth step S200 is performed on the underlying structure 10 in the as-grown state that is not patterned.

The term "mask layer" as used herein means a mask layer which is used in, for example, the so-called ELO (Epitaxial Lateral Overgrowth) method and which is made of silicon oxide or the like and has a predetermined opening. Further, the "concavo-convex pattern" referred to here means at least one of trenches and ridges directly patterned on the main surface, which is used, for example, in the so-called pendeo-epitaxy method. The height difference of the uneven pattern here is, for example, 100 nm or more.

In the above-described ELO method and pendeo-epitaxy method, a step of patterning a photoresist on the main surface of the substrate is performed in order to form a mask layer or an uneven pattern. At this time, if the substrate is in an as-grown state and the main surface is uneven, the exposure gap may vary, resulting in pattern variations within the surface. Therefore, the main surface of the substrate must be planarized before performing the above-described photoresist patterning step. As a result, the ELO method and pendeoepitaxy method require many steps.

In contrast, in the present embodiment, the three-dimensional growth layer 30 can be three-dimensionally grown directly on the underlying structure 10 in the as-grown state without patterning by the above-described manufacturing method. As a result, predetermined processing steps (slicing, planarization (polishing), photolithography, mask layer formation, uneven pattern processing, etc.) can be eliminated, and the number of steps in the present embodiment can be reduced. be able to. As a result, the manufacturing cost can be reduced while improving the yield of the present embodiment.

Further, in the present embodiment, the three-dimensional growth step S200 is performed using the as-grown base structure 10 that is not patterned by the above-described manufacturing method, thereby slicing the three-dimensional growth layer 30. In the main surface 50s of the substrate 50 formed in this way, a high dislocation density region that is regularly generated due to the patterning of the underlying structure 10 is not formed. In other words, in the substrate 50 of the present embodiment, even if the dislocation concentrated regions dca exist, the dislocation concentrated regions dca can be randomly arranged to reduce the dislocation density in the dislocation concentrated regions dca.

(g) In the three-dimensional growth step S200, in the initial stage of growth of the inclined interface enlarging layer 32, the inclined interface enlarging layer 32 is two-dimensionally grown to a predetermined thickness using the c-plane 30c as the growth surface, and then the inclined interface enlarging layer 32 is A plurality of recesses 30p are formed in the top surface 30u of the interface enlarging layer 32. As shown in FIG. In other words, the tilted interface enlarging layer 32 (initial layer) having a mirror-finished surface is formed with a predetermined thickness before the tilted interface enlarging layer 32 begins to grow three-dimensionally. As a result, in the three-dimensional growth step S200, the appearance of the inclined interface 30i in the three-dimensional growth layer 30 can be stably made uniform within the plane.

(h) In the three-dimensional growth step S200, after the c-plane 30c has disappeared from the surface of the three-dimensional growth layer 30, while maintaining a state in which the inclined interface growth region 70 occupies an area of 80% or more of the creeping cross section, a predetermined continue to grow the three-dimensional growth layer 30 over a thickness of . As a result, it is possible to secure a sufficient time for bending the dislocations at the position where the inclined interface 30i is exposed. Here, if the c-plane is grown immediately after the c-plane disappears, there is a possibility that the dislocations will propagate substantially vertically toward the surface of the three-dimensional growth layer without being bent sufficiently. On the other hand, in the present embodiment, by securing a sufficient time for bending the dislocations at the position where the tilted interface 30i other than the c-plane is exposed, It is possible to suppress the dislocation from propagating in the substantially vertical direction. Thereby, concentration of dislocations in the three-dimensional growth layer 30 can be suppressed.

(i) In the three-dimensional growth step S200, the stress canceling effect of the three-dimensional growth layer 30 is stably exhibited as described above by setting the area ratio of the inclined interface growth region 70 to 80% or more in the creeping cross section. be able to. Thereby, the occurrence of cracks during the growth of the three-dimensional growth layer 30 can be suppressed. As a result, the three-dimensional growth layer 30 can be easily grown thick.

(j) In the three-dimensional growth step S200, by setting the area ratio of the inclined interface growth region 70 to 80% or more in the creeping cross section, the three-dimensional growth layer 30 has a larger area than when the c-plane growth region occupies a large area. On the other hand, processing such as slicing and polishing can be easily applied.

(k) In the present embodiment, the three-dimensional growth layer 30 is grown on the mirror-finished second underlayer 6, and the first growth conditions are set so as to satisfy the formula (1) in the three-dimensional growth step S200. By adjusting, in the three-dimensional growth step S200, a {11-2m} plane with m≧3 can be generated as the inclined interface 30i. This makes it possible to moderate the angle of inclination of the {11-2m} plane with respect to the c-plane 30c. Specifically, the inclination angle can be 47.3° or less. By making the inclination angle of the {11-2m} plane gentle with respect to the c-plane 30c, the period of the plurality of top portions 30t can be lengthened. Specifically, when viewing an arbitrary cross section perpendicular to the main surface 1 s of the base substrate 1 , the average distance L between the closest tops can be more than 100 μm.

For reference, when etch pits are formed in a nitride semiconductor substrate using a predetermined etchant, etch pits composed of {1-10n} planes are formed on the surface of the substrate. On the other hand, on the surface of the three-dimensional growth layer 30 grown under predetermined conditions in this embodiment, {11-2m} planes where m≧3 can be generated. Therefore, it is considered that the sloped interface 30i peculiar to the manufacturing method is formed in this embodiment as compared with a normal etch pit.

(l) In the present embodiment, when an arbitrary cross section perpendicular to the main surface 1s of the base substrate 1 is viewed, the average distance L between the closest tops is set to be more than 100 μm, so that the dislocation bends and propagates. can be ensured to be at least greater than 50 μm. As a result, dislocations can be sufficiently collected above the approximate center between the pair of top portions 30 t in the three-dimensional growth layer 30 . As a result, the dislocation density on the surface of the three-dimensional growth layer 30 can be sufficiently reduced.

(m) In the present embodiment, warpage of the laminated structure 90 separated from the base substrate 1 can be suppressed in the separation step S300.

Here, consider the case where a substrate obtained by the conventional VAS method is used and a three-dimensional growth process is performed. In the conventional VAS method, the major surface of the substrate is polished flat, but the c-plane of the substrate is curved like a concave spherical surface as described above. When the three-dimensional growth layer is grown on the flat main surface of the substrate while gradually flattening the c-plane, the three-dimensional growth layer below the three-dimensional growth layer grows along with the flattening of the c-plane of the three-dimensional growth layer. A stress acts on the substrate to flatten the c-plane. Therefore, the substrate, which was flat before the growth of the three-dimensional growth layer, may warp convexly toward the main surface during the growth process of the three-dimensional growth layer. If the substrate warps convexly toward the main surface side, the back surface of the laminated structure warps concavely. If the back surface warps concavely, the laminated structure cannot be stably fixed to a jig for slicing in the slicing process. As a result, the yield of processing such as slicing may decrease.

In contrast, in the present embodiment, the lattice constant of the high oxygen concentration region of the three-dimensional growth layer 30 is relatively increased, and the three-dimensional growth layer 30 is grown while gradually flattening the c-plane 30c. As the c-plane 30c of the three-dimensional growth layer 30 is flattened, stress can be applied to the second underlayer 6 under the three-dimensional growth layer 30 to flatten the c-plane. As a result, the entire second underlayer 6, which is concavely curved toward the main surface 6s in the as-grown state, can be gradually flattened. By gradually flattening the entire second base layer 6, the back surface of the second base layer 6 can be flattened in the laminated structure 90 separated from the base substrate 1 in the peeling step S300. By flattening the back surface of the second base layer 6, the laminated structure can be stably fixed to the jig for slicing in the slicing process. As a result, it is possible to improve the yield of processing such as slicing.

(n) In the slicing step S400, the substrate 50 is formed by slicing the inclined interface preserving layer . In addition, the inclined interface preserving layer 34 is sliced at a position above the position where the c-plane growth region 60 continuing from the second underlayer 6 has finally disappeared. This makes it possible to avoid the portion where dislocations are collected in the three-dimensional growth layer 30 . As a result, the substrate 50 with reduced dislocations can be stably obtained.

<Other embodiments>
The embodiments of the present invention have been specifically described above. However, the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the present invention.

In the above-described embodiment, the case where the substrate 50 is a GaN free-standing substrate has been described, but the substrate 50 is not limited to a GaN free-standing substrate, and may be aluminum nitride (AlN), aluminum gallium nitride (AlGaN), indium nitride (InN), or the like. ), indium gallium nitride (InGaN), _{aluminum indium gallium} nitride (AlInGaN), etc. It may be a self-supporting substrate made of a group III nitride semiconductor represented by a composition formula of 0≦x+y≦1).

In the above embodiments, the substrate 50 is of n-type, but the substrate 50 may be of p-type or semi-insulating. For example, when manufacturing a semiconductor device as a high electron mobility transistor (HEMT) using the substrate 50, the substrate 50 preferably has semi-insulating properties.

In the above-described embodiment, in the three-dimensional growth step S200, the case where the n-type impurity is added at a concentration higher than the concentration of oxygen taken into the inclined interface growth region 70 has been described. 50, in the three-dimensional growth step S200, a p-type impurity may be added at a concentration equal to or higher than the concentration of oxygen taken into the inclined interface growth region .

In the above-described embodiment, the case where the growth temperature is mainly adjusted as the first growth condition in the three-dimensional growth step S200 has been described. Alternatively, the growth conditions other than the growth temperature may be adjusted, or the growth temperature and the growth conditions other than the growth temperature may be combined for adjustment.

In the above-described embodiment, the case where the growth conditions in the inclined interface maintaining step S240 are maintained at the above-described first growth conditions as in the inclined interface expanding step S220 has been described, but the growth conditions in the inclined interface maintaining step S240 satisfies the first growth condition, the growth condition in the inclined interface maintaining step S240 may be different from the growth condition in the inclined interface expanding step S220.

In the above-described embodiment, in the slicing step S400, a wire saw is used to slice the main growth layer 44. However, for example, an outer peripheral blade slicer, an inner peripheral blade slicer, an electric discharge machine, or the like may be used.

In the above-described embodiment, the case where the substrate 50 is obtained by slicing the inclined interface preserving layer 34 of the laminated structure 90 has been described, but the present invention is not limited to this case. For example, the stacked structure 90 may be used as it is to manufacture a semiconductor stack for manufacturing a semiconductor device. Specifically, after the laminated structure 90 is produced, a semiconductor functional layer is epitaxially grown on the laminated structure 90 in a semiconductor laminated structure producing step to produce a semiconductor laminated structure. After the semiconductor laminate is produced, the back side of the laminated structure 90 is polished to remove the second base layer 6 and the inclined interface enlarging layer 32 from the laminated structure 90 . As a result, a semiconductor laminate having the graded interface maintaining layer 34 and the semiconductor functional layer is obtained as in the above-described embodiments. In this case, the slicing step S400 and polishing step S500 for obtaining the substrate 50 can be omitted.

Various experimental results supporting the effects of the present invention will be described below.

(1) Experiment 1
(1-1) Fabrication of Nitride Semiconductor Substrate Nitride semiconductor substrates of Examples and Comparative Examples were fabricated as follows. For the examples, a laminated structure was also fabricated before slicing the nitride semiconductor substrate. Hereinafter, "nitride semiconductor substrate" may be abbreviated as "substrate".

[Conditions for manufacturing nitride semiconductor substrates of Examples]
(underlying substrate)
Material: Sapphire Diameter: 2 inches Thickness: 400 μm
Low-index crystal plane closest to the principal plane: c-plane No patterning such as a mask layer for the principal plane.
(First underlayer)
・Low temperature growth buffer layer:
Material: GaN
Growth temperature: 550°C
Thickness: 50nm
・GaN layer:
Material: Undoped (unintentionally doped) GaN
Growth temperature: 1,050°C
Thickness: 350nm
Note that the surface of the GaN layer was mirror-finished.
(metal layer)
Material: Ti
Thickness: 20nm
(Heat treatment conditions)
A two-stage heat treatment was performed.
First heat treatment:
Atmosphere: _N2 gas 80%, _NH3 gas 20%
Temperature: 1,050°C
Time: 10 minutes Second heat treatment:
Atmosphere: _H2 gas 80%, _NH3 gas 20%
Temperature: 1,050°C
Time: 20 minutes (second underlayer)
Material: Undoped (unintentionally doped) GaN
Growth temperature: 1,050°C
Thickness: about 1000 μm
Note that the main surface of the second underlayer was mirror-finished.
No patterning such as a mask layer on the main surface.
(three-dimensional growth layer)
Material: GaN
Growth method: HVPE method First growth condition:
The growth temperature is 980° C. or higher and 1,020° C. or lower, and the V/III ratio is 2 or higher and 20 or lower. At this time, at least one of the growth temperature and the V/III ratio was adjusted within the above range so that the first growth condition satisfies the formula (1).
Three-dimensional growth layer thickness: about 1.45 mm
A predetermined amount of Si was added to the graded interface maintaining layer.
(slicing conditions)
Thickness of nitride semiconductor substrate: about 400 μm
Calf loss: 200 μm
In the example, two substrates were sliced from the inclined interface maintaining layer.
(polishing conditions)
Polishing thickness: 200 μm

[Conditions for manufacturing nitride semiconductor substrate of Comparative Example 1]
In Comparative Example 1, a substrate was produced under the same conditions as the conventional VAS method.
(underlying substrate)
Same as Example.
(First crystal layer, metal layer, and heat treatment conditions)
Same as the first underlayer, metal layer, and heat treatment conditions in the example.
(Second crystal layer)
Same as the second underlayer of the example (same thickness).
(three-dimensional growth layer)
none.
(Peeling conditions)
Spontaneous exfoliation when the temperature is lowered from the growth temperature of the second crystal layer.
(slicing and polishing conditions)
The conditions are the same as in the example except that one substrate is sliced from the second crystal layer.

[Conditions for manufacturing nitride semiconductor substrate of Comparative Example 2]
(underlying substrate)
Same as Example.
(First crystal layer, metal layer, and heat treatment conditions)
Same as the first underlayer, metal layer, and heat treatment conditions in the example.
(Second crystal layer)
The conditions are the same as those of the second underlayer of the example except that the thickness is 1700 μm.
(three-dimensional growth layer)
none.

(1-2) Evaluation (observation of laminated structure)
Using a fluorescence microscope, the cross section of the laminated structure before slicing the substrate of the example was observed. The surface of the laminated structure was also observed using an optical microscope.

(Observation of nitride semiconductor substrate by multiphoton excitation microscope)
Using a multiphoton excitation microscope, the principal surfaces of the substrates of Example and Comparative Example 1 were observed. At this time, the dislocation density was measured by measuring the dark spot density over the entire main surface at every 250 μm visual field. It is confirmed that all the dark spots in these substrates are threading dislocations by performing measurements while shifting the focal point in the thickness direction. At this time, the ratio of the number of regions having a dislocation density of less than 1×10 ⁶ cm ⁻² (low dislocation density regions) to the total number of measured regions in a field of view of 250 μm was obtained.

In addition, the substrate of the example was also observed with the focal point shifted from the main surface.

(X-ray rocking curve measurement)
The X-ray rocking curves of the (0002) plane of each of the substrates of Example and Comparative Example 1 were measured. At this time, within the main surface of each substrate, on a straight line passing through the center along the m-axis direction and on a straight line along the a-axis direction passing through the center and perpendicular to the m-axis, a plurality of substrates set at intervals of 5 mm The measurement was performed at the measurement point of At this time, the X-ray was made incident from the side defined positively as the position within the main surface of the substrate. As a result of the measurement, the peak angle ω between the X-ray incident on the main surface and the main surface was plotted against the position on the straight line, and the peak angle ω was approximated by a linear function of the position. The radius of curvature of the c-plane was obtained from the reciprocal of the slope of the linear function. In the arrangement of the device used in this experiment, when the slope of the linear function is negative, it means that the c-plane is convex.

(1-3) Results <Observation Results of Laminated Structures of Examples>
FIG. 11 is a diagram showing an observation image obtained by observing the surface of the laminated structure of the example with an optical microscope. FIG. 12 is a diagram showing an observation image obtained by observing the cross section of the laminated structure of the example with a fluorescence microscope. Note that FIG. 12 is a cross section along the <11-20> axis. FIGS. 13A and 13B are diagrams showing observation images obtained by observing a cross section of another portion of the laminated structure of the example with a fluorescence microscope.

As shown in FIGS. 12, 13(a) and (b), in the laminated structure of the example, the three-dimensionally grown layer is grown on the basis of the difference in the growth surface (that is, the difference in oxygen concentration) during the growth process. , a c-plane growth region grown with the c-plane as the growth plane, and an inclined interface growth region grown with the inclined interface as the growth plane.

At least part of the inclined interface growth region was provided continuously along the main surface of the second underlayer. That is, when viewing a plurality of creepage cross sections of the three-dimensional growth layer cut along the main surface of the second underlayer, the cross section that does not include the c-plane growth region grown with the c-plane as the growth plane is the three-dimensional growth layer. It was confirmed that it existed in at least a part of the thickness direction.

The c-plane growth region on the second underlayer had multiple valleys and multiple peaks. The average angle formed by the pair of inclined portions in the c-plane grown region was about 43°. Also, the average distance between the closest tops was about 145.6 μm.

As shown in FIGS. 12, 13(a) and 13(b), the c-plane growth region was also present at a position overlapping the inclined interface growth region in a cross-sectional view. Since the c-plane growth region appeared and disappeared during the growth process of the three-dimensional growth layer, the width of the c-plane growth region in cross-section (in other words, the size of the c-plane growth region in plan view) It varied randomly from the underlayer toward the surface of the three-dimensional growth layer.

In addition, since the c-plane disappeared at least once, the c-plane growth region was not continuous from the second underlayer to the surface (uppermost surface) of the three-dimensional growth layer.

As shown in FIG. 11, the surface of the three-dimensionally grown layer had a plurality of depressions formed by inclined interfaces other than the c-plane. Six shiny surfaces were formed in the recesses formed on the surface of the three-dimensional growth layer, that is, the recesses had six inclined interfaces.

Considering the direction of the orientation flat of the peeled base substrate, the ridgeline in the recess is along the <1-100> axis direction, and the inclined interface forming the recess is inclined from the <11-20> axis. It was a plane (that is, the {11-2m} plane) whose normal direction was the direction of the plane.

In the cross section along the <11-20> axis shown in FIG. 12, the angle of the inclined interface in the three-dimensional growth layer with respect to the main surface of the underlying substrate was about 43° or more and 47° or less.

Here, the angle of {11-2m} with respect to the {0001} plane of GaN is as follows.
{11-21} plane: 72.9°
{11-22} plane: 58.4°
{11-23} plane: 47.3°
{11-24} plane: 39.1°

From the above, it was confirmed that the inclined interface generated on the surface of the three-dimensional growth layer was the {11-2m} plane with m≧3. It was also confirmed that most of the inclined interfaces were {11-23} planes.

<Comparison of Example and Comparative Examples 1 and 2>
[Comparison between Example and Comparative Example 2]
First, Example and Comparative Example 2 are compared.

In Comparative Example 2 in which the second crystal layer was grown thick, fine cracks were found in the second crystal layer when the temperature was lowered to room temperature after the growth of the crystal layer. In addition, crystals grew abnormally from the cracked cross section of the second crystal layer. From this, it is considered that the second crystal layer was cracked during the growth of the second crystal layer. In Comparative Example 2, no nitride semiconductor substrate was fabricated because a second crystal layer having a sufficient thickness and diameter could not be obtained.

On the other hand, in Example, when the temperature was lowered to room temperature after the growth of the three-dimensional growth layer was completed, the laminated structure separated from the underlying substrate at the boundary between the metal nitride layer and the second underlying layer. Ta. No cracks were observed in the laminated structure after peeling.

[Comparison between Example and Comparative Example 1]
Next, Example and Comparative Example 1 are compared. Table 1 shows the results of Example and Comparative Example 1.

(Observation result of nitride semiconductor substrate by multiphoton excitation microscope)
FIG. 14 is a schematic diagram showing an observation image of the main surface of the nitride semiconductor substrate of the example, observed with a multiphoton excitation microscope. In addition, the thick-line square indicates a dislocation-free region of 50 μ square. FIG. 15 is a diagram showing an observation image in which the dotted square portion of FIG. 14 is enlarged. FIG. 16 is a view showing images observed with a multiphoton excitation microscope when the focus is changed in the thickness direction in the nitride semiconductor substrate of the example.

As shown in FIG. 14, the substrate of Example had a relatively bright high oxygen concentration region. The area ratio occupied by the high oxygen concentration region in the main surface of the substrate of the example was 80% or more.

As shown in Table 1, in the substrate of Example, the average dislocation density on the main surface was significantly reduced compared to the substrate of Comparative Example 1, and was less than 5.5×10 ⁵ cm ⁻² .

Moreover, in the substrate of the example, there was no region where the dislocation density exceeded 3×10 ⁶ cm ⁻² . Further, in the substrates of the examples, a region having a dislocation density of less than 1×10 ⁶ cm ⁻² (low dislocation density region) was present in 90% or more of the main surface.

In addition, as shown in FIG. 14, the main surface of the substrate of Example included a dislocation-free region of at least 50 μm square. In addition, the main surface of the substrate of the example had non-overlapping dislocation-free regions of 50 μm square at a density of 100/cm ² or more. Specifically, in the main surface of the substrate of the example, the density of non-overlapping 50 μm square dislocation-free regions was about 5000/cm ² .

Further, as shown in FIG. 15, when the low-oxygen-concentration region observed dark was enlarged and confirmed, the low-oxygen-concentration region had a dislocation-free region of at least 50 μm square.

Further, in FIG. 14, when the main surface of the substrate of the example was observed with a multiphoton excitation microscope in a field of view of 50 μm square including the dislocation concentrated region and the dislocation density was obtained from the dark spot density, the dislocation density was 3×. It was less than 10 ⁶ cm ⁻² .

As shown in FIG. 14, the example substrate had basal plane dislocations (bpd). Observation of the CL image of the substrate of the example showed that when an arbitrary virtual line segment with a length of 200 μm was drawn in the CL image of the main surface, the number of intersections between the line segment and the basal plane dislocations was , was less than 10 points.

As shown in FIG. 16, when observing the substrate of the example while changing the focus in the thickness direction, the region A appears on the surface (main surface) side and disappears at a position deeper than the surface. Was. Also, the region B appeared or disappeared depending on the position in the thickness direction. Also, region C disappeared near the surface. From the above, the size of the low-oxygen-concentration region in plan view changed randomly from the back surface opposite to the main surface of the substrate toward the main surface. In addition, the low-oxygen-concentration region was not continuously connected from the back surface on the opposite side of the main surface of the substrate to the main surface because there was a missing portion.

(curvature radius of c-plane)
Next, the radius of curvature of the c-plane will be described with reference to Table 1 and FIGS. 17(a) and (b).

FIG. 17(a) is a diagram showing the results of rocking curve measurement of X-ray diffraction in the direction along the m-axis of the nitride semiconductor substrate of the example, and (b) is a nitride semiconductor substrate of the example. FIG. 10 is a diagram showing the result of rocking curve measurement of X-ray diffraction in the direction along the a-axis perpendicular to the m-axis of the semiconductor substrate. In addition, R in the figure indicates the radius of curvature of the c-plane, and a negative R means that the c-plane was curved convexly with respect to the main surface, as described above.

As shown in Table 1, FIGS. 17A and 17B, the radius of curvature (absolute value) of the c-plane in the substrate of Example is larger than the radius of curvature of the c-plane in the substrate of Comparative Example 1, It was over 22m.

The substrates shown in FIGS. 17(a) and 17(b) are obtained by slicing from the side near the surface of the laminated structure of the example. In the substrate, the c-plane was curved in a convex spherical shape with respect to the main surface.

On the other hand, although not shown, in the substrate sliced from the side close to the second underlayer of the laminated structure of the example, the c-plane was slightly concave with respect to the main surface, but the curvature of the c-plane was small. The radius was large and the c-plane was almost flat. The absolute value of the radius of curvature of the c-plane in the substrate sliced from the side closer to the second underlayer was larger than the absolute value of the radius of curvature of the c-plane in the substrate sliced from the side closer to the surface. The radius of curvature of the c-plane in the substrate was 1053 m.

For these substrates of the example, when the peak angle ω was approximated by a linear function of the position in the X-ray rocking curve measurement of the c-plane, the error with respect to the linear function was small. Specifically, the error of the measured peak angle ω with respect to the linear function approximated as described above was less than or equal to 0.01°.

(Summary of Experiment 1)
According to the above examples, in the three-dimensional growth process, the first growth conditions were adjusted so as to satisfy the formula (1). As a result, the c-plane could be reliably eliminated at least once during the growth process of the three-dimensional growth layer. By eliminating the c-plane at least once, the dislocation could be reliably bent at the position where the tilted interface in the three-dimensional growth layer was exposed. As a result, it was confirmed that the dislocation density on the main surface of the substrate could be reduced.

Further, according to the embodiment, by growing the three-dimensional growth layer on the mirror-finished second underlayer and by adjusting the first growth conditions so as to satisfy the formula (1), m {11-2m} planes with ≧3 could be produced. As a result, in the three-dimensional growth layer, the average distance between the closest tops was able to exceed 100 μm. As a result, it was confirmed that the dislocation density on the main surface of the substrate could be sufficiently reduced. Moreover, it was confirmed that a dislocation-free region of at least 50 μm square could be formed by setting the average distance between the closest tops to more than 100 μm.

Moreover, according to the example, it was confirmed that the occurrence of cracks during the growth of the three-dimensional growth layer could be suppressed by the stress cancellation effect of the inclined interface growth region.

Further, according to the example, the radius of curvature of the c-plane of the substrate of the example could be made larger than the radius of curvature of the c-plane of the substrate of the comparative example 1 due to the stress cancellation effect of the inclined interface growth region. As a result, it was confirmed that the variation in the off-angle of the c-axis in the substrate of Example and the variation in off-angle of the c-axis in Comparative Example 1 could be reduced.

In the examples, it was confirmed that the c-plane changed from concave to convex toward the surface side from the second underlayer side of the laminated structure due to the stress cancellation effect of the inclined interface growth region. Therefore, it is considered that a substrate having an extremely flat c-plane can be obtained by optimizing the position in the thickness direction of slicing from the laminated structure.

(2) Experiment 2
(2-1) Fabrication of Nitride Semiconductor Substrate Substrates of Samples 1 and 2 were fabricated as follows.

[Preparation conditions for nitride semiconductor substrate of sample 1]
A substrate was fabricated in the same manner as in Experiment 1, except that Si was not added in the three-dimensional growth process. The thickness of the substrate was set to 323 μm.

[Preparation conditions for nitride semiconductor substrate of sample 2]
A substrate was fabricated in the same manner as in Comparative Example 1 of Experiment 1, except that 1×10 ¹⁸ cm ⁻³ of Si was added in the second crystal layer growth step.

(2-2) Evaluation (secondary ion mass spectrometry (SIMS))
The oxygen concentration in the substrate of sample 1 was measured by SIMS.

(Hall measurement)
By Hall measurement, the mobility and resistivity of the substrate of Sample 1, and the free electron concentration in the substrate of Sample 1 were measured.

(optical measurement)
The transmittance and reflectance of each of the substrates of Samples 1 and 2 were measured.

A Shimadzu SolidSpec-3700 DUV UV-Vis-NIR spectrophotometer was used to measure transmittance and reflectance. The measurement conditions are as follows.
Slit width: 8 nm (wavelength 720 nm or less), 32 nm (wavelength 720 nm or more)
Measurement speed: Slow Light source: Halogen lamp (wavelength 310 nm or more)
Detector:
Photomultiplier tube (PMT) (wavelength 870 nm or less)
InGaAs (wavelength 870 nm to 1650 nm)
PbS (wavelength 1650 nm or more)
Attached equipment Large sample chamber Integrating sphere (60 mmφ) Spectralon Incident angle: (incident from Ga surface)
Transmittance measurement: 0°
Reflectance measurement: 8°
Light irradiation size:
Transmittance measurement: 3mm x 3mm
Reflectance measurement: 10mm x 5mm
reference:
Transmittance measurement: None Reflectance measurement: Al mirror

For the substrate of sample 1, the transmittance and reflectance were measured at the center of the main surface and four points at positions 15 mm apart from the center in the m-axis direction and the a-axis direction. After that, the absorption coefficient was obtained based on the transmittance and reflectance at the center of the main surface of the substrate.

For the substrate of sample 2, the transmittance and reflectance were measured at the center of the main surface, and the absorption coefficient was determined based on the results.

(2-3) Results <SIMS measurement results>
As a result of SIMS measurement, the oxygen concentration in the substrate of sample 1 was 1×10 ¹⁸ cm ⁻³ .

<Hall measurement result>
As a result of Hall measurement, the mobility of the substrate of sample 1 was 254 cm ² /V·s. The specific resistance of the substrate of sample 1 was 0.0232 Ω·cm.

Also, the free electron concentration in the substrate of sample 1 was 1.06×10 ¹⁸ cm ⁻³ . The free electron concentration in the substrate of sample 1 was almost the same as the oxygen concentration in the substrate.

From this, it is considered that the free electron concentration in the substrate of Sample 1 was almost the same as the free electron concentration in the substrate of Sample 2.

<Optical measurement result>
The absorption coefficients of the substrates of Samples 1 and 2 will be described with reference to FIGS. 18(a) and 18(b). FIG. 18(a) is a diagram showing absorption spectra of samples 1 and 2, and (b) is a diagram obtained by adding a theoretical formula to (a). The theoretical formula in FIG. 18(b) is the above formula (3) when K=2.2×10 ⁻¹⁹ and a=3.

As shown in FIG. 18A, the absorption coefficient of the substrate of sample 1 was equal to or lower than that of the substrate of sample 2 in the visible light region. Specifically, the substrate of Sample 1 had an absorption coefficient of 0.15 cm ⁻¹ or less in the wavelength range of 500 nm to 700 nm.

In the infrared region, the absorption coefficient of the substrate of sample 1 due to free carrier absorption almost coincided with the absorption coefficient of the substrate of sample 2. This confirms that the free electron concentration in the substrate of Sample 1 was substantially the same as the free electron concentration in the substrate of Sample 2.

Furthermore, as shown in FIG. 18(b), in the substrate of sample 1 (and 2), when K=2.2×10 ⁻¹⁹ and a=3, the wavelength range is at least 1 μm or more and 2.5 μm or less. can be approximated by the above equation (3) by the method of least squares. At this time, at a wavelength of 2 μm, the error of the actually measured absorption coefficient with respect to the absorption coefficient α obtained from Equation (3) was, for example, within ±0.1α.

Next, the in-plane distribution of transmittance and reflectance on the substrate of Sample 1 will be described with reference to FIGS. 19A is a diagram showing the transmission spectrum of Sample 1, and FIG. 19B is a diagram showing the reflection spectrum of Sample 1. FIG. Note that FIG. 19A shows the transmission spectra of two points with the largest transmittance difference among the five measurement points of Sample 1. FIG. FIG. 19B shows reflection spectra (solid line and dashed line) of two points with the largest difference in reflectance among the five measurement points of sample 1. FIG.

As shown in FIG. 19A, the in-plane variation in transmittance of sample 1 was small. Specifically, in the substrate of Sample 1, the variation in the transmittance within the main surface at each wavelength within the wavelength range of 500 nm to 700 nm was within ±0.3%.

In addition, as shown in FIG. 19B, the in-plane variation in reflectance of sample 1 was small. Specifically, in the substrate of Sample 1, the variation in the reflectance within the main surface at each wavelength within the wavelength range of 500 nm to 700 nm was within ±0.3%.

(Summary of Experiment 2)
According to Experiment 2 described above, the absorption coefficient of the substrate of Sample 1 was equal to or lower than that of the substrate of Sample 2 in the visible light region. Although the substrate of sample 1 has a high oxygen concentration region, no peculiar absorption occurs, and the substrate of sample 1 has a high crystal quality equivalent to the substrate obtained by the VAS method like sample 2. I confirmed that I did. In other words, it was confirmed that the substrate of Sample 1 is sufficiently applicable as an LED substrate.

Further, according to Experiment 2, the absorption coefficient of the substrate of Sample 1 due to free carrier absorption substantially coincided with the absorption coefficient of the substrate of Sample 2. Also, in the substrate of sample 1, the absorption coefficient in the infrared region based on free carrier absorption could be approximated as a function of free carrier concentration and wavelength. Also from these facts, it was confirmed that the substrate of Sample 1 had a high crystal quality equivalent to that of the substrate obtained by the VAS method, such as Sample 2. Further, in the substrate of Sample 1, the absorption coefficient in the infrared region can be approximated by the formula (3) without variation, so that when the substrate of Sample 1 is heated by infrared irradiation, the substrate can be accurately and It was confirmed that heating can be performed with good reproducibility.

According to the results of Experiment 1 described above, in the substrate of Example, the high oxygen concentration region and the low oxygen concentration region were evenly and randomly distributed within the main surface of the substrate. The size of the low-oxygen-concentration region in plan view varied randomly in the thickness direction. Also, the low oxygen concentration region was not continuous in the thickness direction. The results of Experiment 2 supported the results of Experiment 1, and it was confirmed that the substrate of Sample 1 had small in-plane variations in transmittance and reflectance.

<Preferred embodiment of the present invention>
Preferred embodiments of the present invention are described below.

(Appendix 1)
A method for manufacturing a nitride semiconductor substrate using a vapor deposition method, comprising:
preparing an underlying substrate;
forming a first underlying layer made of a group III nitride semiconductor on the underlying substrate;
forming a metal layer on the first underlayer;
heat-treating to form voids in the first underlayer;
A second underlayer made of a single crystal of a Group III nitride semiconductor and having a main surface whose closest low-index crystal plane is the (0001) plane is epitaxially grown above the first underlayer, and the second underlayer is mirroring the main surface of the stratum;
A single crystal of a group III nitride semiconductor having a top surface with an exposed (0001) plane is directly epitaxially grown on the main surface of the second underlayer, and is formed of an inclined interface other than the (0001) plane. A plurality of concave portions are formed on the top surface, the inclined interface is gradually expanded upward from the main surface of the second underlayer, and the (0001) plane disappears from the top surface at least once. , growing a three-dimensional growth layer;
slicing the three-dimensional growth layer to form the nitride semiconductor substrate;
A method for manufacturing a nitride semiconductor substrate having

(Appendix 2)
In the step of growing the three-dimensional growth layer,
forming an inclined interface growth region grown in the three-dimensional growth layer using the inclined interface as a growth surface;
The method for manufacturing a nitride semiconductor substrate according to Supplementary Note 1, wherein the sloped interface growth region accounts for 80% or more of the surface area of the three-dimensional growth layer cut along the main surface.

(Appendix 3)
The step of forming the three-dimensional growth layer includes:
After the (0001) plane has disappeared from the top surface, the single crystal is grown over a predetermined thickness while maintaining a state in which the inclined interface growth region occupies an area of 80% or more of the creeping cross section. 3. The method for manufacturing a nitride semiconductor substrate according to appendix 2, further comprising the step of forming the graded interface preserving layer.

(Appendix 4)
In the step of slicing the three-dimensional growth layer,
3. The method of manufacturing a nitride semiconductor substrate according to appendix 3, wherein the inclined interface maintaining layer is sliced.

(Appendix 5)
In the step of forming the three-dimensional growth layer,
5. The method for manufacturing a nitride semiconductor substrate according to any one of Appendices 2 to 4, wherein the conductive impurity is added at a concentration equal to or higher than the concentration of oxygen introduced into the inclined interface growth region.

(Appendix 6)
In the step of forming the three-dimensional growth layer,
6. The method according to any one of appendices 1 to 5, wherein after the single crystal is grown to a predetermined thickness using the (0001) plane as a growth surface, the plurality of recesses are formed in the top surface of the single crystal. A method for manufacturing a nitride semiconductor substrate.

(Appendix 7)
In the step of forming the three-dimensional growth layer,
7. The method for manufacturing a nitride semiconductor substrate according to any one of appendices 1 to 6, wherein the inclined interface is a {11-2m} plane where m≧3.

(Appendix 8)
In the step of forming the three-dimensional growth layer,
forming a plurality of valleys and a plurality of tops on the surface of the three-dimensional growth layer by forming the plurality of recesses on the top surface of the single crystal and eliminating the (0001) plane;
When viewing an arbitrary cross section perpendicular to the main surface, a pair of apexes closest to each other among the plurality of apexes sandwiching one of the plurality of troughs is in a direction along the main surface. 8. The method for manufacturing a nitride semiconductor substrate according to any one of Appendices 1 to 7, wherein the average distance between the two is more than 100 μm.

(Appendix 9)
In the step of forming the three-dimensional growth layer,
The method for manufacturing a nitride semiconductor substrate according to appendix 8, wherein the average distance between the pair of apexes that are closest to each other is less than 800 μm.

(Appendix 10)
In the step of slicing the nitride semiconductor substrate,
The radius of curvature of the (0001) plane of the nitride semiconductor substrate is reduced by growing the second underlayer to the same thickness as the three-dimensional growth layer and forming the three-dimensional growth layer without performing the step of forming the three-dimensional growth layer. 2. The method of manufacturing a nitride semiconductor substrate according to any one of Appendices 1 to 9, wherein the radius of curvature of the (0001) plane of the nitride semiconductor substrate when the underlying layer is sliced is larger than that of the substrate.

(Appendix 11)
11. The method of manufacturing a nitride semiconductor substrate according to appendix 10, wherein the radius of curvature of the (0001) plane of the nitride semiconductor substrate is 15 m or more.

(Appendix 12)
A nitride semiconductor substrate having a diameter of 2 inches or more, made of a single crystal of a Group III nitride semiconductor, and having a main surface whose closest low-index crystal plane is the (0001) plane,
Having a high oxygen concentration region having an oxygen concentration of 9×10 ¹⁷ cm ⁻³ or more,
The nitride semiconductor substrate, wherein the area ratio of the high oxygen concentration region in the main surface is 80% or more.

(Appendix 13)
13. The nitride semiconductor substrate according to appendix 12, which has a low oxygen concentration region having an oxygen concentration of 5×10 ¹⁶ cm ⁻³ or less.

(Appendix 14)
14. The nitride semiconductor substrate according to appendix 13, wherein the low oxygen concentration region in the main surface includes a dislocation-free region of at least 50 μm square.

(Appendix 15)
15. The nitride semiconductor substrate according to Supplementary Note 13 or 14, wherein the size of the low oxygen concentration region in plan view changes from the back surface opposite to the main surface toward the main surface.

(Appendix 16)
16. The nitride semiconductor substrate according to any one of Appendices 13 to 15, wherein the low oxygen concentration region is not continuously connected from the back surface opposite to the main surface toward the main surface.

(Appendix 17)
17. The nitride semiconductor substrate according to any one of Appendices 13 to 16, which does not have a low oxygen concentration region having an oxygen concentration of 5×10 ¹⁶ cm ⁻³ or less.

(Appendix 18)
18. The nitride semiconductor substrate according to any one of appendices 12 to 17, wherein the absorption coefficient in the wavelength range of 500 nm or more and 700 nm or less is 0.15 cm ^{−1 or} less.

(Appendix 19)
λ (μm) is the wavelength, α (cm ⁻¹ ) is the absorption coefficient of the nitride semiconductor substrate at 27° C., n (cm ⁻³ ) is the free electron concentration in the nitride semiconductor substrate, and K and a are constants. when
The absorption coefficient α in the wavelength range of at least 1 μm or more and 2.5 μm or less is approximated by the following formula (3) by the least squares method,
19. The nitride semiconductor according to any one of appendices 12 to 18, wherein at a wavelength of 2 μm, the error of the actually measured absorption coefficient with respect to the absorption coefficient α obtained from the formula (3) is within ±0.1α. substrate.
α= ^nKλa (3)
(However, 1.5×10 ⁻¹⁹ ≦K≦6.0×10 ⁻¹⁹ , a=3)

(Appendix 20)
When measuring the transmittance of light in a wavelength range of 500 nm or more and 700 nm or less at a plurality of different points in the main surface,
20. The nitride semiconductor substrate according to any one of appendices 12 to 19, wherein the transmittance at each wavelength within a wavelength range of 500 nm or more and 700 nm or less has a variation within the main surface of ±0.5%. .

(Appendix 21)
When measuring the reflectance of light in a wavelength range of 500 nm or more and 700 nm or less at a plurality of different points in the main surface,
21. The nitride semiconductor substrate according to any one of appendices 12 to 20, wherein the reflectance at each wavelength within the wavelength range of 500 nm or more and 700 nm or less has a variation within the main surface of ±0.5%. .

(Appendix 22)
An X-ray rocking curve measurement of the (0002) plane is performed at each position on a straight line passing through the center in the main surface, and a peak angle ω formed between the X-ray incident on the main surface and the main surface is measured on the straight line. When plotting against the position of and approximating the peak angle ω with a linear function of the position,
The curvature radius of the (0001) plane obtained by the reciprocal of the slope of the linear function is 15 m or more,
22. The nitride semiconductor substrate according to any one of appendices 12 to 21, wherein an error of the measured peak angle ω with respect to the linear function is 0.05° or less.

(Appendix 23)
When the principal surface is observed with a multiphoton excitation microscope in a field of view of 250 μm square and the dislocation density is obtained from the dark spot density, there is no region on the principal surface where the dislocation density exceeds 3×10 ⁶ cm ⁻² , 23. The nitride semiconductor substrate according to any one of Appendices 12 to 22, wherein 80% or more of the main surface has a region with a dislocation density of less than 1×10 ⁶ cm ⁻² .

(Appendix 24)
24. The nitride semiconductor substrate according to any one of Appendices 12 to 23, wherein the main surface has non-overlapping dislocation-free regions of 50 μm square at a density of 100/cm ² or more.

(Appendix 25)
The main surface has a dislocation concentrated region in which dislocations are relatively concentrated,
Note that the dislocation density is less than 3×10 ⁶ cm ⁻² when the main surface is observed with a multiphoton excitation microscope in a 50 μm square field of view including the dislocation concentrated region and the dislocation density is obtained from the dark spot density. 25. The nitride semiconductor substrate according to any one of 12 to 24.

(Appendix 26)
Any one of Appendices 12 to 25, wherein when an arbitrary virtual line segment with a length of 200 μm is drawn in the cathodoluminescence image of the main surface, the number of intersections between the line segment and the basal plane dislocations is 10 or less. 1. The nitride semiconductor substrate according to claim 1.

(Appendix 27)
fabricating a laminate having the nitride semiconductor substrate and a semiconductor layer obtained by epitaxially growing a non-doped single crystal of a predetermined Group III nitride semiconductor on the main surface;
A reference substrate made of a single crystal of the same III-nitride semiconductor as the nitride semiconductor substrate and having an oxygen concentration lower than that of the high oxygen concentration region, and a single crystal of the same semiconductor layer as the semiconductor layer are epitaxially grown on the reference substrate. creating a reference stack having a reference semiconductor layer;
When the photoluminescence in each of the semiconductor layer of the laminate and the reference semiconductor layer of the reference laminate is measured with a temperature difference of less than 1° C.,
27. The nitride semiconductor according to any one of appendices 12 to 26, wherein a difference between the maximum peak wavelength in the semiconductor layer of the laminate and the maximum peak wavelength in the reference semiconductor layer of the reference laminate is 1 nm or less. substrate.

(Appendix 28)
an underlying layer made of a single crystal of a Group III nitride semiconductor, having a mirror-finished main surface, and having a (0001) plane as the closest low-index crystal plane to the main surface;
a three-dimensional growth layer provided on the underlayer, made of a single crystal of a group III nitride semiconductor, and having a plurality of recesses on the surface thereof, the recesses being formed by tilted interfaces other than the (0001) plane;
has
The three-dimensional growth layer has a high oxygen concentration region made of a group III nitride semiconductor single crystal having an oxygen concentration of 9×10 ¹⁷ cm ⁻³ or more,
The laminated structure, wherein the three-dimensionally grown layer has a creeping cross section cut along the main surface, the cross section having an area ratio of 80% or more of the high oxygen concentration region.

(Appendix 29)
The three-dimensional growth layer is provided on the main surface of the underlayer and has a low oxygen concentration region made of a group III nitride semiconductor single crystal having an oxygen concentration of 5×10 ¹⁶ cm ⁻³ or less,
29. The laminated structure according to appendix 28, wherein the low oxygen concentration region is discontinuous from the underlayer to the surface of the three-dimensionally grown layer.

(Appendix 30)
29. The laminated structure according to Supplementary Note 29, wherein, when viewing a plurality of the creeping cross sections, a cross section that does not include the low oxygen concentration region exists in at least a portion of the thickness direction of the three-dimensionally grown layer.

(Appendix 31)
When looking at any cross section perpendicular to the main surface,
The upper surface of the low oxygen concentration region has a plurality of valleys and a plurality of peaks,
Supplementary note 29 that the average distance between the pair of peaks that are closest to each other among the plurality of peaks across one of the plurality of valleys in the direction along the main surface is more than 100 μm. 31. or the laminated structure according to 30.

(Appendix 32)
31. Any one of Appendices 28 to 30, wherein the underlying layer has, on the side opposite to the main surface, a high oxygen concentration region made of a group III nitride semiconductor single crystal having an oxygen concentration of 9×10 ¹⁷ cm ⁻³ or more. The laminated structure according to 1.

1 base substrate 2 first base layer 3 metal layer 4 void-containing first base layer 5 metal nitride layer 6 second base layer 10 base structure 30 three-dimensional growth layer 50 nitride semiconductor substrate (substrate)

Claims (18)

A method for manufacturing a nitride semiconductor substrate using a vapor deposition method, comprising:
preparing an underlying substrate;
forming a first underlying layer made of a group III nitride semiconductor on the underlying substrate;
forming a metal layer on the first underlayer;
heat-treating to form voids in the first underlayer;
A second underlayer made of a single crystal of a Group III nitride semiconductor and having a main surface whose closest low-index crystal plane is the (0001) plane is epitaxially grown above the first underlayer, and the second underlayer is mirroring the main surface of the stratum;
A single crystal of a group III nitride semiconductor having a top surface with an exposed (0001) plane is directly epitaxially grown on the main surface of the second underlayer, and is composed of an inclined interface other than the (0001) plane. A plurality of concave portions are formed on the top surface, the inclined interface is gradually expanded upward from the main surface of the second underlayer, and the (0001) plane disappears from the top surface at least once. , growing a three-dimensional growth layer;
slicing the three-dimensional growth layer to form the nitride semiconductor substrate;
A method for manufacturing a nitride semiconductor substrate having In the step of growing the three-dimensional growth layer,
forming an inclined interface growth region grown in the three-dimensional growth layer using the inclined interface as a growth surface;
2. The method of manufacturing a nitride semiconductor substrate according to claim 1, wherein the sloped interface growth region occupies 80% or more of the surface area of the three-dimensional growth layer cut along the main surface. The step of forming the three-dimensional growth layer includes:
After the (0001) plane has disappeared from the top surface, the single crystal is grown over a predetermined thickness while maintaining a state in which the inclined interface growth region occupies an area of 80% or more of the creeping cross section. 3. The method of manufacturing a nitride semiconductor substrate according to claim 2, further comprising the step of forming the graded interface sustaining layer. In the step of forming the three-dimensional growth layer,
4. The method of manufacturing a nitride semiconductor substrate according to claim 2, wherein a conductivity type impurity is added at a concentration equal to or higher than the concentration of oxygen introduced into said inclined interface growth region. In the step of forming the three-dimensional growth layer,
5. The method according to any one of claims 1 to 4, wherein after the single crystal is grown to a predetermined thickness using the (0001) plane as a growth plane, the plurality of recesses are formed on the top surface of the single crystal. A method for manufacturing a nitride semiconductor substrate of In the step of forming the three-dimensional growth layer,
6. The method for manufacturing a nitride semiconductor substrate according to claim 1, wherein a {11-2m} plane in which m≧3 is generated as the inclined interface. In the step of forming the three-dimensional growth layer,
forming a plurality of valleys and a plurality of tops on the surface of the three-dimensional growth layer by forming the plurality of recesses on the top surface of the single crystal and eliminating the (0001) plane;
When viewing an arbitrary cross section perpendicular to the main surface, a pair of apexes closest to each other among the plurality of apexes sandwiching one of the plurality of troughs is in a direction along the main surface. The method for manufacturing a nitride semiconductor substrate according to any one of claims 1 to 6, wherein the average distance between the two is more than 100 µm. In the step of slicing the nitride semiconductor substrate,
The radius of curvature of the (0001) plane of the nitride semiconductor substrate is reduced by growing the second underlayer to the same thickness as the three-dimensional growth layer and forming the three-dimensional growth layer without performing the step of forming the three-dimensional growth layer. 8. The method of manufacturing a nitride semiconductor substrate according to claim 1, wherein the radius of curvature of the (0001) plane of the nitride semiconductor substrate obtained by slicing the second underlying layer is larger than that of the sliced underlying layer. A nitride semiconductor substrate having a diameter of 2 inches or more, made of a single crystal of a Group III nitride semiconductor, and having a main surface whose closest low-index crystal plane is the (0001) plane,
Having a high oxygen concentration region having an oxygen concentration of 9×10 ¹⁷ cm ⁻³ or more,
The area ratio occupied by the high oxygen concentration region in the main surface is 80% or more,
the main surface has non-overlapping 50 μm square dislocation-free regions at a density of 100/cm ² or more;
When measuring the transmittance of light in a wavelength range of 500 nm or more and 700 nm or less at a plurality of different points in the main surface,
Variation in the main surface of the transmittance at each wavelength within a wavelength range of 500 nm or more and 700 nm or less is within ±0.5%.
Nitride semiconductor substrate. When measuring the reflectance of light in a wavelength range of 500 nm or more and 700 nm or less at a plurality of different points in the main surface,
10. The nitride semiconductor substrate according to claim 9 , wherein the reflectance at each wavelength within a wavelength range of 500 nm or more and 700 nm or less has a variation within the main surface of ±0.5%. A nitride semiconductor substrate having a diameter of 2 inches or more, made of a single crystal of a Group III nitride semiconductor, and having a main surface whose closest low-index crystal plane is the (0001) plane,
9x10 ¹⁷ cm ^-3 Having a high oxygen concentration region having an oxygen concentration of more than
The area ratio occupied by the high oxygen concentration region in the main surface is 80% or more,
The main surface has 100 dislocation-free regions of 50 μm square that do not overlap each other. ² having a density of greater than or equal to
When measuring the reflectance of light in a wavelength range of 500 nm or more and 700 nm or less at a plurality of different points in the main surface,
Variation of the reflectance within the main surface at each wavelength within a wavelength range of 500 nm or more and 700 nm or less is within ±0.5%.
Nitride semiconductor substrate. 12. The nitride semiconductor substrate according to claim 9, having a low oxygen concentration region having an oxygen concentration of 5×10 ¹⁶ cm ⁻³ or less . 13. The nitride semiconductor substrate according to claim 12 , wherein said low oxygen concentration region in said main surface includes a dislocation-free region of at least 50 [mu]m square. 14. The nitride semiconductor substrate according to claim 12 , wherein the size of said low oxygen concentration region in plan view changes from the back surface opposite to said main surface toward said main surface. 15. The nitride semiconductor substrate according to any one of claims 12 to 14 , wherein said low oxygen concentration region is not continuously connected from the back surface opposite to said main surface toward said main surface. 10. The nitride semiconductor substrate according to claim 9, which does not have a low oxygen concentration region having an oxygen concentration of 5×10 ¹⁶ cm ⁻³ or less. 17. The nitride semiconductor substrate according to claim 9, wherein the absorption coefficient in the wavelength range of 500 nm or more and 700 nm or less is 0.15 cm ⁻¹ or less. λ (μm) is the wavelength, α (cm ⁻¹ ) is the absorption coefficient of the nitride semiconductor substrate at 27° C., n (cm ⁻³ ) is the free electron concentration in the nitride semiconductor substrate, and K and a are constants. when
The absorption coefficient α in the wavelength range of at least 1 μm or more and 2.5 μm or less is approximated by the following formula (3) by the least squares method,
The nitride according to any one of claims 9 to 17 , wherein the error of the measured absorption coefficient with respect to the absorption coefficient α obtained from the formula (3) at a wavelength of 2 μm is within ±0.1α. semiconductor substrate.
α= ^nKλa (3)
(However, 1.5×10 ⁻¹⁹ ≦K≦6.0×10 ⁻¹⁹ , a=3)

Images