JP2025071301A - Crystal substrate and method for producing the same - Google Patents

Abstract

To provide a technique for further reducing a dislocation density in a crystal substrate composed of a group III nitride.SOLUTION: A crystal substrate is a substrate composed of a group III nitride single crystal, a low-index crystal plane most proximate to a principal surface of the substrate is a c-plane, a threading dislocation included in the single crystal and penetrating the c-plane extends in a thickness direction without bending in a direction orthogonal to the thickness direction of the substrate, and a ratio of edge dislocations in the threading dislocations is less than 60%.SELECTED DRAWING: Figure 12

Description

The present invention relates to a crystal substrate and a method for manufacturing a crystal substrate.

Crystal substrates made of Group III nitrides such as gallium nitride (GaN) are used as substrates for manufacturing semiconductor devices such as light-emitting elements and transistors. Various techniques have been proposed to reduce the dislocation density in such crystal substrates (see, for example, Non-Patent Document 1).

Yuichi Oshima and 5 others, "GaN Substrates by Void Formation and Peeling Method", Hitachi Cable, No. 26 (2007-1), pp. 31-36

One objective of the present invention is to provide a technique for further reducing dislocation density in a crystal substrate made of Group III nitrides.

According to one aspect of the present invention, a first crystal body made of a single crystal of a Group III nitride;
a second crystal body grown on the first crystal body and composed of a single crystal of a Group III nitride;
having
the second crystal body has an intermediate layer including a plurality of closed spaces containing an alkali metal;
the intermediate layer has a plurality of closed space rows formed by arranging a plurality of the closed spaces on each of a plurality of linear crystal defects included in the second crystal body,
A crystal substrate is provided in which a certain row of closed spaces is selected from the plurality of rows of closed spaces, a first closed space and a second closed space included in the row of closed spaces are selected, and the heights from the first closed space and the second closed space to the first crystal body are designated as h1 and h2, respectively, and there are a plurality of other rows of closed spaces having closed spaces at the same heights as h1 and h2.

According to another aspect of the present invention,
A substrate made of a single crystal of a Group III nitride,
A crystal substrate is provided in which the single crystal contains threading dislocations in which the ratio of edge dislocations is less than 60%.

According to yet another aspect of the present invention,
preparing a first crystal body composed of a single crystal of a Group III nitride;
growing a second crystal body composed of a single crystal of a Group III nitride on the first crystal body in a mixed melt containing an alkali metal and a Group III element;
having
In the step of growing the second crystal body,
the second crystal body has an intermediate layer including a plurality of closed spaces containing the alkali metal;
a plurality of closed space rows are formed by arranging a plurality of the closed spaces on each of a plurality of linear crystal defects included in the second crystal body in the intermediate layer,
A method for manufacturing a crystal substrate is provided, which comprises selecting a certain row of closed spaces from the plurality of rows of closed spaces, selecting a first closed space and a second closed space included in the row of closed spaces, and growing the second crystal body so that when the heights from the first closed space and the second closed space to the first crystal body are h1 and h2, respectively, there are a plurality of other rows of closed spaces having closed spaces at the same heights as h1 and h2.

A technology is provided to further reduce dislocation density in crystal substrates composed of group III nitrides.

FIG. 1 is a flow chart showing an overall method for manufacturing a crystal substrate according to one embodiment of the present invention. FIG. 2 is a flowchart showing details of step S100 in the method for manufacturing a crystal substrate according to one embodiment. 3(a) to 3(g) are schematic cross-sectional views showing the process of preparing seed substrate 21 in step S100 of one embodiment. 4A to 4C are schematic cross-sectional views showing the process of producing the substrate 31 in step S200 of one embodiment. 5(a) and 5(b) are fluorescence microscope images of a crystal 30 produced as an experimental example of one embodiment. 6(a) and 6(b) are MPPL images of a crystal 30 produced as an experimental example of one embodiment. FIG. 7 is an MPPL image of a crystal 30 produced as an experimental example of one embodiment. FIG. 8 is a flowchart showing an overall method for manufacturing a crystal substrate according to a modified embodiment. 9A to 9C are schematic cross-sectional views showing the process of producing a substrate 41 in step S300 according to a modified example of the embodiment. FIG. 10 is a schematic diagram illustrating an example of an HVPE apparatus. FIG. 11 is a schematic diagram illustrating a flux liquid phase growth apparatus. FIG. 12 is a schematic perspective view of a crystal 30 according to one embodiment.

<One embodiment of the present invention>

An embodiment of the present invention will be described below. In this embodiment, a technology for growing a crystal substrate 31 (hereinafter also referred to as substrate 31) using a seed crystal substrate 21 (hereinafter also referred to as seed substrate 21) as a seed crystal will be described. The seed substrate 21 and the substrate 31 are each composed of a single crystal of a group III nitride. An example of the group III nitride that constitutes the seed substrate 21 and the substrate 31 is gallium nitride (GaN).

In this embodiment, a GaN substrate grown by the void-assisted separation (VAS) method is exemplified as the seed substrate 21, but a GaN substrate grown by another method may be used as the seed substrate 21. However, from the viewpoint of forming the closed space array 25C in a dispersed state in the plane as described later, the GaN substrate used as the seed substrate 21 preferably has a low dislocation density region (region with dislocation density less than 1×10 ⁷ /cm ² , for example) with a non-excessive dislocation density on the main surface 21s serving as the base for the growth of the substrate 31, and more preferably does not have a dislocation concentration region (region with dislocation density equal to or greater than 1×10 7 /cm 2, for example) with an excessively high dislocation density. In other words, it is more preferable that the maximum dislocation density on the main surface 21s is less than 1× ^{10 7} /cm ^{2 ,} for example. The VAS method is preferably used as one method for obtaining a seed substrate 21 having a maximum dislocation density in the main surface 21s of less than 1×10 ⁷ /cm ² , for example. In this specification, dislocation means threading dislocation, and to avoid complication of expression, threading dislocation may be simply referred to as dislocation.

(1) Manufacturing Method of Crystal Substrate A manufacturing method of substrate 31 according to the present embodiment will be described. Fig. 1 is a flow chart showing the entire manufacturing method of substrate 31 according to the present embodiment. This manufacturing method includes step S100 of preparing seed substrate 21, and step S200 of producing substrate 31 by a liquid phase method, specifically, a flux method.

(S100: Preparation of seed crystal substrate)
In step S100, a seed substrate 21 is prepared. Fig. 2 is a flow chart showing details of step S100 in an example of the VAS method. Figs. 3(a) to 3(g) are schematic cross-sectional views showing the process of preparing the seed substrate 21 in step S100 in an example of the VAS method.

(S110: Preparation of void-forming substrate)
Step S100 includes steps S110 to S140. In step S110, a void-formed substrate 15 is prepared. More specifically, step S110 includes steps S111 to S114. In step S111, as shown in Fig. 3A, a base substrate 10 is prepared. An example of the base substrate 10 is a sapphire substrate.

In step S112, as shown in FIG. 3B, an underlayer 11 is formed on the undersubstrate 10. The underlayer 11 is, for example, a stack of a buffer layer made of low-temperature grown GaN and a single crystal layer of GaN. The buffer layer and the single crystal layer are formed, for example, by metal-organic vapor phase epitaxy (MOVPE). For example, trimethylgallium (TMG) is used as the group III raw material, and for example, ammonia (NH ₃ ) is used as the group V raw material. The thickness of the buffer layer and the single crystal layer are, for example, 20 nm and 0.5 μm, respectively.

In step S113, as shown in FIG. 3(c), a metal layer 12 is formed on the base layer 11. The metal layer 12 is formed, for example, by vapor-depositing titanium (Ti) to a thickness of 20 nm.

In step S114, as shown in FIG. 3(d), the metal layer 12 is nitrided by heat treatment to form a nanomask 14, and voids are formed in the underlayer 11 to form a void-containing layer 13. The heat treatment is performed, for example, as follows. The undersubstrate 10 on which the underlayer 11 and the metal layer 12 are formed is placed in an electric furnace and placed on a susceptor having a heater. Then, the undersubstrate 10 is heated in an atmosphere containing hydrogen gas ( _H2 gas) or hydride gas. Specifically, for example, in a _H2 gas flow containing 20% _NH3 gas as a nitriding gas, heat treatment is performed for 20 minutes at a predetermined temperature, for example, a temperature of 850° C. or more and 1100° C. or less.

The heat treatment nitrides the metal layer 12, forming a nanomask 14 with a high density of fine holes on its surface. Also, a part of the base layer 11 is etched through the fine holes of the nanomask 14, causing voids in the base layer 11 and forming a void-containing layer 13. In this way, in step S110, a void-forming substrate 15 is prepared, which has the void-containing layer 13 and nanomask 14 formed on the base substrate 10.

It is preferable that the heat treatment is performed so that the distribution of the micropores in the nanomask 14 and the distribution of the voids in the void-containing layer 13 are uniform within the surface. For this reason, for example, it is preferable that the heating condition of the heater is adjusted so that the temperature distribution becomes closer to uniform within the surface, and it is also preferable that the heat treatment is performed while rotating the susceptor.

(S120: Crystal growth by HVPE)
In step S120, as shown in Fig. 3(e), a crystal 20 is grown on the nanomask 14 of the void-forming substrate 15. The crystal 20 is grown by a vapor phase method, specifically, a hydride vapor phase epitaxy (HVPE) method. Here, an HVPE apparatus 200 will be described. Fig. 10 is a schematic diagram illustrating the HVPE apparatus 200.

The HVPE apparatus 200 is made of a heat-resistant material such as quartz, and includes an airtight container 203 with a film-forming chamber 201 formed therein. A susceptor 208 is provided in the film-forming chamber 201 to hold a substrate 250 to be processed. The susceptor 208 is connected to a rotating shaft 215 of a rotating mechanism 216, and is configured to be freely rotatable. Gas supply pipes 232a to 232c are connected to one end of the airtight container 203, which supply hydrochloric acid (HCl) gas, NH ₃ gas, and nitrogen gas (N ₂ gas) into the film-forming chamber 201. A gas supply pipe 232d is connected to the gas supply pipe 232c, which supplies hydrogen (H ₂ ) gas. The gas supply pipes 232a to 232d are provided with flow rate controllers 241a to 241d and valves 243a to 243d, respectively, in this order from the upstream side. A gas generator 233a that contains Ga melt as a raw material is provided downstream of the gas supply pipe 232a. A nozzle 249a that supplies gallium chloride (GaCl) gas generated by the reaction between HCl gas and Ga melt toward the substrate 250 held on the susceptor 208 is connected to the gas generator 233a. Nozzles 249b and 249c that supply various gases supplied from these gas supply pipes toward the substrate 250 held on the susceptor 208 are connected to the downstream sides of the gas supply pipes 232b and 232c, respectively. An exhaust pipe 230 that exhausts the inside of the film formation chamber 201 is provided at the other end of the airtight container 203. A pump 231 is provided to the exhaust pipe 230. A zone heater 207 for heating the inside of the gas generator 233a and the substrate 250 held on the susceptor 208 to a desired temperature is provided on the outer periphery of the airtight container 203, and a temperature sensor 209 for measuring the temperature inside the film formation chamber 201 is provided inside the airtight container 203. Each member of the HVPE apparatus 200 is connected to a controller 280 configured as a computer, and the processing procedures and processing conditions described below are controlled by a program executed on the controller 280.

Step S120 can be performed, for example, by the following procedure using the HVPE apparatus 200. First, Ga is accommodated as a raw material in the gas generator 233a. Also, the void-forming substrate 15 is held on the susceptor 208 as the substrate 250 to be processed. Then, a mixed gas of _H2 gas and _N2 gas is supplied into the film-forming chamber 201 while the film-forming chamber 201 is heated and evacuated. Then, when the film-forming chamber 201 reaches a desired film-forming temperature and pressure, and the atmosphere in the film-forming chamber 201 becomes a desired atmosphere, gas is supplied from the gas supply pipes 232a and 232b to supply GaCl gas and _NH3 gas as film-forming gas to the void-forming substrate 15.

The processing conditions for carrying out step S120 are exemplified as follows.
Growth temperature Tg: 980 to 1100° C., preferably 1050 to 1100° C.
Pressure in the film formation chamber 201: 90 to 105 kPa, preferably 90 to 95 kPa
Partial pressure of GaCl gas: 0.2 to 15 kPa
Partial pressure of _NH3 gas/partial pressure of GaCl gas: 4 to 20
Flow rate of _N2 gas/flow rate of _H2 gas: 1 to 20

In this growth process, GaN crystals that start growing from the void-containing layer 13 pass through the fine holes in the nanomask 14 and appear on the surface, forming initial nuclei on the nanomask 14. The initial nuclei grow in the thickness direction (vertical direction) and in-plane direction (horizontal direction) and bond to each other within the plane to form a crystal 20 of a continuous film made of GaN single crystals. Since a mutual attractive force acts between the bonded initial nuclei, a tensile stress is introduced into the crystal 20. In addition, in areas where no initial nuclei are formed, gaps 16 are formed between the nanomask 14 and the crystal 20 due to the voids in the void-containing layer 13.

It is preferable that the growth process is performed so that the distribution of the initial nuclei and the distribution of the voids 16 are uniform in the plane. For this reason, for example, it is preferable to adjust the heating condition of the zone heater 207 so that the temperature distribution approaches uniformity in the plane, and it is also preferable that the growth process is performed while rotating the susceptor 208. It is preferable that the distribution of the fine holes in the nanomask 14 and the distribution of the voids in the void-containing layer 13 are uniform in the plane so that the distribution of the initial nuclei and the distribution of the voids 16 are uniform in the plane. By uniformly distributing the initial nuclei, a dislocation concentration region with a locally very high dislocation density (a region with a dislocation density of, for example, 1×10 ⁷ /cm ² or more) does not occur in the plane, and the distribution of the dislocation density becomes uniform. Furthermore, by uniformly distributing the initial nuclei, the tensile stress becomes uniform in the plane.

The thickness of the crystal 20 to be grown is preferably a thickness that allows at least one free-standing seed substrate 21 to be obtained from the crystal 20, for example, a thickness of 0.2 mm or more. There is no particular upper limit to the thickness of the crystal 20 to be grown.

(S130: Peeling)
3(f), the crystal 20 is peeled off from the void-forming substrate 15. This peeling is performed by the crystal 20 naturally peeling off from the void-forming substrate 15 at the gaps 16 formed between the crystal 20 and the nanomask 14 during the growth of the crystal 20 or in the process of cooling the inside of the film formation chamber 201 after the growth of the crystal 20. The gaps 16 are uniformly distributed, so that the peeling is performed in a state where stress concentration in a specific area is unlikely to occur.

After the crystal 20 has grown to a predetermined thickness, the temperature inside the deposition chamber 201 is lowered to a level at which removal is possible, and the void-forming substrate 15 and the crystal 20 are removed from the deposition chamber 201.

Due to the tensile stress introduced during the growth of the crystal 20, the exfoliated crystal 20 warps so that the growth surface is concave. In the crystal 20, the warping occurs uniformly because the tensile stress is uniform within the surface. As a result, the c-plane 120 of the GaN single crystal constituting the exfoliated crystal 20 curves in a concave spherical shape toward the inside of the crystal 20 when the main surface 20s is viewed from the +c side. Here, "spherical" means a curved surface that is approximated by a spherical surface. Also, "spherical approximation" here means approximation within a specified range of error to a perfect sphere (true sphere) or an elliptical sphere (prolate spheroid).

In addition, in a plan view from the normal direction of the center of the main surface 20s, the uniformity of the growing crystals is likely to decrease at the outermost periphery of the crystal body 20. For this reason, the above-mentioned "c-face 120 is spherically curved" may not be true at the outermost periphery of the crystal body 20. The above-mentioned "c-face 120 is spherically curved" only needs to be true in a region (hereinafter also referred to as the main region) that accounts for 80% or more of the area of the center side of the main surface 20s in a plan view. This also applies to the c-face 121 of the seed substrate 21, the c-face 130 of the crystal body 30, the c-face 131 of the substrate 31, the c-face 140 of the crystal body 40, and the c-face 141 of the substrate 41, which will be described later.

To avoid complication, the c-face 120 of the crystal 20 is simply expressed as "curved in a concave spherical shape toward the inside of the crystal 20 when the main surface 20s is viewed from the +c side in the main region." The same expression is also used for the c-face 121 of the seed substrate 21, the c-face 130 of the crystal 30, the c-face 131 of the substrate 31, the c-face 140 of the crystal 40, and the c-face 141 of the substrate 41, which will be described later.

The off-angle of the crystal 20 is defined as the angle between the c-axis direction of the GaN single crystal constituting the crystal 20 and the normal direction to the center of the main surface 20s. The off-angles of the seed substrate 21, crystal 30, substrate 31, crystal 40, and substrate 41 described below are defined in the same manner.

The off-angle of the crystal 20 varies depending on the position on the main surface 20s because the c-plane 120 is curved. In other words, the peeled crystal 20 has an off-angle distribution. The off-angle at the center of the main surface 20s is the central off-angle. The central off-angle can be controlled by adjusting the off-angle of the base substrate 10, and may be set to tilt in a specific direction (e.g., the a-axis direction or the m-axis direction).

When the central off-angle is tilted in a specific direction, the radius of curvature of c-face 120 may differ between the direction parallel to the tilt direction and the direction perpendicular to it. In other words, the curved shape of c-face 120 may be approximated not only by a perfect circular sphere, but also by an elliptical sphere. When approximated by a perfect circular sphere, the shape of c-face 120 is represented by one radius of curvature. When approximated by an elliptical sphere, the shape of c-face 120 is represented by two radii of curvature.

In addition, in crystals grown using a method that creates dislocation concentration regions by unevenly distributing the density of initial nuclei, such as ELO (Epitaxially Lateral Overgrowth) using a stripe mask, the shape of the c-plane is distorted. That is, in such crystals, the radius of curvature of the c-plane varies within the plane. It is also possible that the direction of the unevenness of the c-plane is reversed depending on the location. For this reason, the shape of the c-plane in such crystals cannot be appropriately approximated as a sphere with a constant radius of curvature.

(S140: Machining and polishing)
In step S140, as shown in FIG. 3(g), the crystal body 20 peeled off in step S130 is mechanically processed and/or polished as necessary to obtain a seed substrate 21 (crystal body 20 as a self-supporting substrate). The mechanical processing is, for example, cutting with a wire saw. For example, one seed substrate 21 may be obtained from the entire crystal body 20. Also, for example, the crystal body 20 may be sliced into multiple pieces to obtain multiple seed substrates 21 from the entire crystal body 20. The obtained seed substrate 21 may or may not be polished as necessary. One of the two main surfaces may be polished. The crystal body 20 peeled off in step S130 may be used as the seed substrate 21 as it is.

Due to the machining or polishing in step S140, the radius of curvature of the c-face 121 of the seed substrate 21 may change from the radius of curvature of the c-face 120 of the peeled crystal body 20.

Figure 3 (g) illustrates a seed substrate 21 configured in a flat plate shape. The seed substrate 21 is configured such that the lowest-index crystal plane closest to the main surface 21s is the c-plane 121, and the c-plane 121 is curved in a concave spherical shape relative to the main surface 21s (relative to one of the main surfaces 21s of the seed substrate 21). The c-plane 121 has a constant radius of curvature (in the main region) because distortion from the spherical shape is suppressed. In this manner, in step S100, the seed substrate 21 (crystal mass 20) is prepared.

The fact that the c-plane 121 of the seed substrate 21 has a shape that is approximated to a spherical shape means that the in-plane distribution of dislocation density is uniform (i.e., there is no region where the dislocation density is locally very high). In the main surface 21s of the seed substrate 21, the maximum dislocation density is less than 1× ¹⁰ / ^cm2 , preferably, for example, 5× ¹⁰ / ^cm2 or less, and the average dislocation density is preferably, for example, 3× ¹⁰ / ^cm2 or less.

The dislocation density is evaluated, for example, by emission intensity mapping near the band edge of GaN using multiphoton excited photoluminescence (hereinafter abbreviated as MPPL). Since dislocation defects are non-radiative recombination centers, the emission intensity is weak around the dislocation defects, and the dislocation defects are observed as dark spots in the emission intensity mapping image. This can be used to evaluate the dislocation density. Specifically, for example, an image is taken of an observation area of 250 μm square in one field of view at any location on the sample, and the dark spot density obtained as the evaluation result can be treated as the dislocation density. In this specification, the dark spot density measured by MPPL is treated as the dislocation density.

(S200: Preparation of crystal substrate by liquid phase method)
After the seed substrate 21 is prepared in step S100, the substrate 31 is fabricated in step S200 by a liquid phase method, specifically, a flux method. Figures 4(a) to 4(c) are schematic cross-sectional views showing the process of fabricating the substrate 31 in step S200. Figure 4(a) shows the seed substrate 21 prepared in step S100.

(S210: Crystal growth by flux method)
Step S200 includes steps S210 and S220. In step S210, as shown in FIG. 4B, a crystal 30 is grown on a seed substrate 21 by a flux method. In the flux method, a group III nitride (GaN in this example) is grown in a mixed melt containing an alkali metal used as a flux (solvent) and a group III element (Ga in this example). As the alkali metal to be the flux, sodium (Na) is preferably used, but other alkali metal elements such as lithium (Li) and potassium (K) may be used. These elements may also be mixed and used. The metal used as the flux may be an alkali metal plus an alkaline earth metal. As the alkaline earth metal, magnesium (Mg), calcium (Ca), etc. may be used alone or in a mixture. Here, a flux liquid phase growth apparatus 300 will be described. FIG. 11 is a schematic diagram illustrating the flux liquid phase growth apparatus 300.

The flux liquid phase epitaxy apparatus 300 includes a pressure vessel 303. The pressure vessel 303 is made of stainless steel (SUS) or the like, and a pressurizing chamber 301 capable of increasing the pressure to a high pressure state of, for example, about 10 MPa is configured inside the pressure vessel 303. The pressurizing chamber 301 includes a crucible 308, a lid for the crucible 308, a reaction vessel 310, a heater 307 for heating the inside of the crucible 308, and a temperature sensor 309 for measuring the temperature inside the pressurizing chamber 301. The crucible 308 is configured to accommodate a mixed melt composed of Na and Ga, for example, with Na as a flux, and to allow a substrate 350 to be processed to be immersed in the mixed melt with its main surface (base surface for crystal growth) facing upward. The reaction vessel 310 is composed of a reaction vessel body and a reaction vessel lid, and a crucible 308 is accommodated inside the reaction vessel 310. A heater 307 is provided on the outside of the reaction vessel 310. The flux liquid phase epitaxy apparatus 300 also includes a rotation mechanism 320, which can rotate the reaction vessel 310, i.e., the crucible 308 accommodated therein. A gas supply pipe 332 that supplies _N2 gas to the pressurizing chamber 301 is connected to the pressure-resistant vessel 303. The gas supply pipe 332 is provided with a pressure control device 333, a flow rate controller 341, and a valve 343 in this order from the upstream side. Each member of the flux liquid phase epitaxy apparatus 300 is connected to a controller 380 configured as a computer, and is configured to control the processing procedure and processing conditions described later by a program executed on the controller 380.

Step S210 is performed, for example, by the following procedure using the flux liquid phase epitaxy apparatus 300. First, the seed substrate 21, which is the substrate 350 to be processed, and the raw materials (Na and Ga) of the mixed melt are accommodated in the crucible 308, and additives are also accommodated as necessary, and the pressure-resistant container 303 is sealed. Then, _N2 gas is supplied into the pressurizing chamber 301, and heating is started after a predetermined gas pressure is reached. By starting heating with the heater 307, Na and Ga in the crucible 308 are melted, and a mixed melt (Ga melt with Na as a medium, mixed melt containing Na and Ga) is formed. After raising the temperature to the crystal growth temperature, the gas pressure is adjusted to the crystal growth pressure. Nitrogen (N) is dissolved in the mixed melt, and this state is maintained for a predetermined time. In this specification, the "mixed melt" is not limited to one in which nitrogen is not dissolved, but also includes one in which nitrogen is dissolved.

The processing conditions for performing step S210 are exemplified as follows.
Growth temperature (temperature of mixed melt): 700 to 1000°C, preferably 800 to 900°C, more preferably 870 to 890°C
Growth pressure (pressure in the pressure chamber): 0.1 to 10 MPa, preferably 1 to 6 MPa, more preferably 2.5 to 4.0 MPa
Na concentration in the mixed melt [[Na]/([Na]+[Ga])]: 10-90%, preferably 40-85%, more preferably 70-85% (the Na concentration is a molar concentration)
Here, for example, the number of moles of Na is represented as [Na]. This notation of the number of moles is also used for other elements in the same manner.
In this embodiment, the mixed molten liquid used is one in which Na (an example of an alkali metal used as a flux) and Ga (an example of a Group III element, i.e., a Group 13 typical metal element) are doped with another metal M, which will be described below. The concentration (molar concentration) of the other metal M is defined as the ratio of the content of the other metal M to the combined content of Ga and the other metal M in the mixed molten liquid.
Concentration of other metal M [[M]/([M]+[Ga])]: 3 to 15%, preferably 4 to 12%
In this case, the other metal M may be a single element or multiple elements. In the case of multiple elements, the combination of the multiple elements is called M. [M] is the total number of moles of the metal elements added other than sodium and gallium.
As another additive to the mixed molten liquid, for example, carbon (C) may be added. C concentration [[C]/([C]+[Ga]+[Na])]: for example, 0.1 to 1.0%
Distance between the gas-liquid interface of the mixed melt and the nitrogen gas and the main surface of the seed substrate 21: 3 to 70 mm, preferably 5 to 40 mm, and more preferably 20 to 35 mm
Rotation speed: 1 to 30 rpm, preferably 5 to 20 rpm, more preferably 7 to 15 rpm

The other metal M added to the mixed melt includes at least one metal selected from group 4 transition metals, group 5 transition metals, group 6 transition metals, group 7 transition metals, group 8 transition metals, group 9 transition metals, group 10 transition metals, group 11 transition metals, group 12 typical metals, group 14 typical metals, and group 15 typical metals. As the other metal M, specifically, for example, one of Ti, Nb, Cr, Fe, Ni, Zn, Ge, Sn, Sb, and Bi, or a combination thereof, may be used. The other metal M added to the mixed melt may include a group 13 typical metal element (e.g., Al or In in the case of growing GaN) different from the group III element contained in the group III nitride to be grown. Even when other group III elements are added in this way, if the concentration is low as described above, a mixed crystal will not be formed.

In step S210, as shown in FIG. 4(b), a GaN single crystal grows on the main surface 21s of the seed substrate 21, forming a crystal 30. In addition, in the crystal 30, an alkali metal (Na in this example) contained in the mixed molten liquid is incorporated as an inclusion 22 near the interface 27 between the crystal 30 and the seed substrate 21, forming an intermediate layer 24 that includes (at least) a plurality of closed spaces 23 that contain the alkali metal. Above the intermediate layer 24, an upper layer 26 is formed that is composed of crystals 30 that do not include closed spaces 23 (inclusions 22).

Figure 12 is a schematic perspective view of the crystal 30 according to this embodiment. This schematic perspective view is a hypothetical model inferred from knowledge obtained from the experiments shown in Figures 5(a) to 7 described below. Starting from threading dislocations in the seed substrate 21, linear crystal defects 51 extending in the growth direction (thickness direction) are formed in the intermediate layer 24 of the crystal 30, and multiple closed spaces 23 (inclusions 22) are formed discretely on each linear crystal defect 51, thereby forming a closed space array 25C. (For example, in a seed substrate 21 produced by the VAS method) Since threading dislocations are randomly distributed on the main surface, the closed space array 25C is randomly distributed within the surface.

As will be understood from the growth method of the crystal 30 described below, the closed spaces 23 are formed multiple times at predetermined intervals in the growth direction (thickness direction). Therefore, the closed spaces 23 formed simultaneously are arranged at the same height in each closed space row 25C. FIG. 12 shows a virtual plane 52 that is approximately parallel to the interface 27 between the seed substrate 21 and the crystal 30, that is, has an approximately constant height from the interface 27. Here, "approximately parallel to one surface" means that the angle between the normal direction of the one surface and the normal direction of the other surface is 2° C. or less. The closed spaces 23 formed simultaneously are arranged on the same virtual plane 52.

When the closed spaces 23 included in each closed space column 25C are projected onto a cross section 53 perpendicular to the interface 27, the projected closed spaces 23 are aligned on the line where the cross section 53 intersects with the virtual plane 52, that is, aligned in a direction approximately parallel to the interface 27, forming a closed space row 25R. Therefore, the closed spaces 23 projected onto the cross section 53 form a matrix structure 25 consisting of the closed space row 25R and the closed space column 25C. Figure 4(b) shows a schematic diagram of the matrix structure 25 formed by the closed spaces 23 projected onto a cross section (paper surface) perpendicular to the interface 27.

The inventors of the present application have found that the density of threading dislocations 54 in the upper layer 26 is reduced compared to the density of threading dislocations in the seed substrate 21, as described below. From this, it is considered that the closed space array 25C formed on the linear crystal defect 51 in the intermediate layer 24 has the function of terminating threading dislocations (the function of preventing the threading dislocations in the seed substrate 21 from extending to the upper layer 26). In order to make the illustration easier to understand, the threading dislocations 55 (i.e., the amount reduced by the intermediate layer 24) that would be terminated by the intermediate layer 24 if they were assumed to extend to the upper layer 26 are shown by dashed lines.

More specifically, in step S210, the crystal 30 is grown, for example, as follows: Using the above-mentioned mixed melt containing the other metal M, the temperature of the mixed melt (growth temperature) is set to, for example, 880°C, and the nitrogen gas pressure (growth pressure) is set to, for example, 3.4 MPa, and maintained for, for example, 10 hours, to grow an initial layer of the crystal 30 on the seed substrate 21.

Then, an intermediate layer 24 having a series of closed spaces 25C is formed. Methods for forming the intermediate layer 24 include, for example, a method for controlling the growth temperature, and a method for controlling the growth pressure. These methods may be used in combination as appropriate.

In the method of controlling the growth temperature, after the growth of the initial layer, the temperature of the mixed melt is lowered to, for example, 875°C while keeping the nitrogen gas pressure at, for example, 3.4 MPa, and maintained for, for example, 1 hour. This allows the crystal 30 to grow so that the alkali metal (Na in this example) contained in the mixed melt is incorporated as an inclusion 22, that is, so that a closed space 23 containing the alkali metal is formed. After that, the temperature of the mixed melt is again raised to, for example, 880°C while keeping the nitrogen gas pressure at, for example, 3.4 MPa, and maintained for, for example, 1 hour. This allows the crystal 30 to grow so that the closed space 23 (inclusion 22) is not formed. By alternately lowering and raising the growth temperature in this way, an intermediate layer 24 having a closed space row 25C is formed.

In the method of controlling the growth pressure, after the growth of the initial layer, the nitrogen gas pressure is lowered to, for example, 3 MPa while keeping the temperature of the mixed melt at, for example, 880°C, and maintained for, for example, one hour. This allows the crystal 30 to grow so that the closed spaces 23 (inclusions 22) are formed. After that, the nitrogen gas pressure is increased to, for example, 3.4 MPa while keeping the temperature of the mixed melt at, for example, 880°C, and maintained for, for example, one hour. In this way, the crystal 30 is grown so that the closed spaces 23 (inclusions 22) are not formed. By alternately lowering and raising the growth pressure in this way, an intermediate layer 24 having a series of closed spaces 25C is formed.

After the intermediate layer 24 is formed, the temperature of the mixed melt is set to, for example, 880°C, the nitrogen gas pressure is set to, for example, 3.4 MPa, and held for, for example, 100 hours, in the same manner as in the growth of the initial layer, thereby growing a crystal 30 in which no closed space 23 (no inclusions 22) is formed, thereby forming the upper layer 26.

In this way, in step S210, a crystal 30 having an intermediate layer 24 is grown. Specifically, for example, the growth of a crystal 30 in which a closed space 23 is formed and the growth of a crystal 30 in which a closed space 23 is not formed are alternately performed, so that the intermediate layer 24 is formed, by alternately lowering and raising the growth temperature or alternately lowering and raising the growth pressure.

Figures 5(a) and 5(b) are fluorescent microscope images of a crystal 30 produced as an experimental example of this embodiment, showing a cross section perpendicular to the interface 27 between the seed substrate 21 and the crystal 30. Figure 5(b) is an enlarged image of a portion of Figure 5(a). The bright area at the bottom represents the seed substrate 21, and the dark area at the top represents the crystal 30. A linear interface 27 is observed as the boundary between the seed substrate 21 and the crystal 30. Figures 5(a) and 5(b) show an area of about 700 μm and an area of about 100 μm, respectively, in the direction along the interface.

A matrix structure 25, in which multiple brightly indicated closed spaces 23 are arranged in a matrix, i.e., an intermediate layer 24, is observed in the crystal 30 slightly above the interface 27. The multiple closed spaces 23 are arranged in rows in a direction along the interface (parallel direction), and in columns in a direction intersecting the interface (perpendicular direction). Note that in the matrix structure 25, closed spaces 23 do not necessarily exist at all positions in a row, and similarly, closed spaces 23 do not necessarily exist at all positions in a column. In other words, the matrix structure 25 (each closed space column 25C) may have a missing closed space 23.

In the example shown in FIG. 5(a), the matrix structure 25 is arranged in rows, for example, over 500 μm or more. As can be seen from FIG. 5(a), the matrix structure 25 is formed over almost the entire surface of the crystal body 30, for example, over 80% or more of the area, in a plan view seen from a direction perpendicular to the interface. In the example shown in FIG. 5(a) or FIG. 5(b), the multiple closed spaces 23 are arranged in rows, for example, three or more rows, for example, four or more rows, or further, for example, five or more rows, and the multiple closed spaces 23 are arranged in columns, for example, three or more columns, for example, four or more columns, or further, for example, five or more columns.

As can be seen from FIG. 5(a), the matrix structure 25, i.e., the intermediate layer 24, is formed in a region within a predetermined height from the interface 27, and above the intermediate layer 24, an upper layer 26 is formed, which is a region that does not include a closed space 23 (inclusions 22).

Figures 6(a) and 6(b) are MPPL images (multiphoton microscope images) of a crystal 30 produced as an experimental example of this embodiment, showing a cross section perpendicular to the interface 27 between the seed substrate 21 and the crystal 30. The dark area at the bottom represents the seed substrate 21, and the bright area at the top represents the crystal 30. Figures 6(a) and 6(b) are MPPL images of the same area displayed with different contrasts. Figure 6(a) is an MPPL image displayed with a contrast that makes it easier to observe the structure in the crystal 30 than in the seed substrate 21, and Figure 6(b) is an MPPL image displayed with a contrast that makes it easier to observe the structure in the seed substrate 21 than in the crystal 30.

As can be seen from FIG. 6(a), a plurality of closed spaces 23 (inclusions 22) are arranged on (along) a linear crystal defect 51 contained in the crystal body 30, forming a closed space row 25C. The linear crystal defect 51 extends in the growth direction (in a direction roughly perpendicular to the interface 27). Furthermore, a matrix structure 25 of closed spaces 23 is formed by arranging multiple rows of such closed space rows 25C in a direction along the interface 27.

As can be seen from FIG. 6(b), there are threading dislocations in the seed substrate 21 that extend in the thickness direction (vertical direction), and in the crystal body 30, a closed space array 25C is formed on the extension of the threading dislocation. In other words, the linear crystal defect 51 contained in the crystal body 30 is formed with the threading dislocation contained in the seed substrate 21 as its starting point, and it can be seen that the closed space array 25C is formed on such a linear crystal defect 51.

Figure 7 is a figure in which a straight line (i.e., a line intersecting the observed cross section and the virtual plane 52) is drawn at a certain height from the interface 27 between the seed substrate 21 and the crystal mass 30 to the same MPPL image as in Figure 6 (b). Figure 7 shows three rows of closed spaces 25C1, 25C2, and 25C3 as an example. For example, the heights to the three closed spaces 23 included in the closed space row 25C1 are h1, h2, and h3, respectively. Of the closed spaces 23 included in the closed space rows 25C1 to 25C3, the closed spaces 23 arranged at height h1 form the closed space row 25R1, the closed spaces 23 arranged at height h2 form the closed space row 25R2, and the closed spaces 23 arranged at height h3 form the closed space row 25R3.

Based on the contents explained with reference to Figures 4(b), 5(a) to 7, and 12, the closed space array 25C of the intermediate layer 24 according to this embodiment can be described as follows.

For example, when a certain closed space row 25C is selected from among the multiple closed space rows 25C, a first closed space 23 and a second closed space 23 included in the closed space row 25C are selected, and the heights from the first closed space 23 and the second closed space 23 to the seed substrate 21 are set to h1 and h2, respectively, there are multiple other closed space rows 25C that have closed spaces 23 at the same heights as h1 and h2.

For example, when a certain closed space row 25C is selected from among the multiple closed space rows 25C, and a first closed space 23 and a second closed space 23 included in the closed space row 25C are selected, and a first virtual plane 52 and a second virtual plane 52 are considered that pass through the first closed space 23 and the second closed space 23, respectively, and are approximately parallel to the interface 27 between the seed substrate 21 and the crystal body 30, there are multiple closed space rows 25C including the closed space 23 arranged on the first virtual plane 52, and there are multiple closed space rows 25C including the closed space 23 arranged on the second virtual plane 52.

For example, when the closed spaces 23 included in the multiple closed space rows 25C are projected onto a cross section 53 perpendicular to the interface 27 between the seed substrate 21 and the crystal body 30, multiple closed space rows 25R are present within the cross section 53, each of which is formed by lining up the multiple projected closed spaces 23 in a direction approximately parallel to the interface 27.

As shown in FIG. 4(b), in a cross section perpendicular to the interface 27 (cross section parallel to the thickness direction) of the laminated crystal substrate 100 composed of the seed substrate 21 and the crystal 30, the intermediate layer 24 contains a plurality of spatially scattered closed spaces 23 (inclusions 22), and the intermediate layer 24 is a transition region for the GaN single crystal constituting the crystal 30 to grow uniformly by the flux method from the GaN single crystal constituting the seed substrate 21. For example, by performing the growth process under the above-mentioned processing conditions, the closed spaces 23 (inclusions 22) can be distributed uniformly or in a dispersed state without localization within the surface of the intermediate layer 24.

The crystal 30 grows reflecting the crystallinity of the seed substrate 21, and a GaN single crystal is obtained as the crystal 30. The inventors of the present application have found that by growing the crystal 30 so that the intermediate layer 24 having the closed space array 25C is formed, the dislocation density in the upper layer 26 of the crystal 30 can be significantly reduced compared to the dislocation density in the seed substrate 21. For example, it has been confirmed that the average dislocation density of about 3×10 ⁶ /cm ² in the main surface 21s of the seed substrate 21 can be significantly reduced to an average dislocation density of about 3×10 ⁵ /cm ² in the main surface 30s of the upper layer 26 of the crystal 30. In the main surface 30s of the upper layer 26 of the crystal 30, the maximum dislocation density is preferably, for example, 5×10 ⁵ /cm ² or less, and the average dislocation density is preferably, for example, 3×10 ⁵ /cm ² or less. The minimum dislocation density in the main surface 30s is not particularly limited. The ratio of the maximum dislocation density to the minimum dislocation density can be larger as the minimum dislocation density is lower, but as a guideline, it is, for example, 100 times or less, or, for example, 10 times or less.

The inventors of the present application have also found that the ratio of types of dislocations (threading dislocations) (ratio of edge dislocations, mixed dislocations, and screw dislocations) changes between the seed substrate 21 grown by the VAS method (hereinafter, in the explanation of the types of dislocations, the seed substrate 21 means the seed substrate 21 grown by the VAS method) and the crystal body 30 (of the upper layer 26). In the main surface 21s of the seed substrate 21, the ratio of edge dislocations, mixed dislocations, and screw dislocations is, for example, 64:36:0. In contrast, in the main surface 30s of the upper layer 26 of the crystal body 30 according to this embodiment, the ratio of edge dislocations, mixed dislocations, and screw dislocations is, for example, 56:44:0. No screw dislocations have been detected in the seed substrate 21 and the crystal body 30 (of the upper layer 26).

It is speculated that the growth of the crystal 30, which involves the formation of an intermediate layer 24 having a series of closed spaces 25C, reduces dislocations (threading dislocations) overall, with edge dislocations being preferentially reduced.

The crystal 30 is characterized in that the ratio of edge dislocations is lower than that of the seed substrate 21. The crystal 30 is characterized in that the ratio of mixed dislocations is higher than that of the seed substrate 21. The characteristics of the dislocations contained in the seed substrate 21 (grown by the VAS method) include, for example, a ratio of edge dislocations of 60% or more, and, for example, a ratio of mixed dislocations of less than 40%. In contrast, in the crystal 30 grown in this embodiment, the characteristics of the dislocations contained in the crystal 30 constituting the upper layer 26 include, for example, a ratio of edge dislocations of less than 60%, and, for example, a ratio of mixed dislocations of 40% or more.

The type of dislocations (threading dislocations) contained in a crystal can be evaluated, for example, as follows. The main surface (surface) of the crystal is treated with a molten alkaline solution of, for example, a mixture of potassium hydroxide and sodium hydroxide, to form etch pits at positions corresponding to each dislocation. The main surface on which the etch pits are formed is observed, for example, with a differential interference microscope, and the etch pits are classified into a number of groups according to their size. In addition, the cross sections of the etch pits in each group are observed, for example, with a transmission electron microscope (TEM), to confirm what type of dislocation each group of etch pits corresponds to. Using the confirmed correspondence, the type of dislocation contained in the crystal can be evaluated from the size of each etch pit observed with the differential interference microscope.

More specifically, the type of dislocations contained in the crystal 30 was evaluated as follows. On the substrate 31 obtained from the upper layer 26 of the crystal 30, a GaN single crystal was epitaxially grown by the HVPE method in the same manner as in the modified example described later (see FIG. 8), to form the crystal 40. The type of dislocations was then evaluated for the crystal 40. This is to accurately compare and evaluate the ratio of the types of dislocations in the seed substrate 21 and the crystal 40, which were formed by the same HVPE method, by evaluating the types of dislocations. Since the crystal 40 is epitaxially grown inheriting the dislocations of the crystal 30, the dislocation densities of the crystal 40 and the crystal 30 are approximately the same, and the ratio of the types of dislocations does not change. Therefore, it can be said that the evaluation result of the ratio of the types of dislocations for the crystal 40 is the same as the evaluation result of the ratio of the types of dislocations for the crystal 30.

It should be noted that by growing the crystal using the flux method, it is possible to suppress the hydrogen concentration and the oxygen concentration in the crystal 30. The hydrogen concentration in the crystal 30 is, for example, less than 1×10 ¹⁷ /cm ³ , and preferably 5×10 ¹⁶ /cm ³ or less. The oxygen concentration in the crystal 30 is, for example, less than 1×10 ¹⁷ /cm ³ , and preferably 8×10 ¹⁶ /cm ³ or less.

The c-plane 130 of the GaN single crystal constituting the crystal 30 is curved in a concave spherical shape with respect to the main surface 30s of the crystal 30. The inventors of the present application have also discovered that by using a seed substrate 21 grown by the VAS method as a seed crystal and growing the crystal 30 by the flux method, the radius of curvature of the c-plane 130 of the crystal 30 can be made larger than the radius of curvature of the c-plane 121 of the seed substrate 21. The reason for this is presumably that the intermediate layer 24 present near the interface between the seed substrate 21 and the crystal 30 weakens the force that binds the crystal 30 to the seed substrate 21, and the compressive stress generated in the crystal 30 as the crystal 30 grows is alleviated.

The thickness of the crystal 30 to be grown is not particularly limited, but in this embodiment, the thickness is set to a thickness that allows at least one free-standing substrate 31 to be obtained from the crystal 30, for example, a thickness of 0.2 mm or more. After growing the crystal 30 to a predetermined thickness, the pressure vessel 303 is returned to room temperature and atmospheric pressure, and the seed substrate 21 on which the crystal 30 has been formed is removed from the crucible 308.

(S220: Machining and polishing)
In step S220, as shown in FIG. 4C, the crystal 30 grown in step S210 is cut by mechanical processing, for example, a wire saw, to separate it from the seed substrate 21. By cutting at a position appropriately away from the interface 27 between the seed substrate 21 and the crystal 30, it is possible to prevent the root portion of the crystal 30 including the intermediate layer 24 from being included in the separated crystal 30. In other words, it is possible to selectively separate the crystal 30 of the upper layer 26 that does not include the closed space 23 (inclusions 22). The layered crystal substrate 110 consisting of the seed substrate 21 and the root portion of the crystal 30 including the intermediate layer 24 that remains after the separation of the crystal 30 may be reused as a seed crystal for crystal growth by the flux method or the like.

The separated crystal 30 (upper layer 26) is machined and/or polished as necessary to obtain a substrate 31 (crystal 30 as a free-standing substrate). For example, one substrate 31 may be obtained from the entire crystal 30. Alternatively, for example, multiple substrates 31 may be obtained from the entire crystal 30 by slicing the crystal 30 into multiple pieces. The obtained substrate 31 may or may not be polished as necessary. One of the two main surfaces may be polished. The separated crystal 30 may be used as the substrate 31 as is.

Due to the separation of the crystal 30 from the seed substrate 21, the radius of curvature of the c-face 130 of the crystal 30 may change from the radius of curvature of the c-face 130 of the crystal 30 before separation. Also, due to machining or polishing of the separated crystal 30, the radius of curvature of the c-face 131 of the substrate 31 may change from the radius of curvature of the c-face 130 of the separated crystal 30. However, even if such a change in the radius of curvature occurs, the tendency that the radius of curvature of the c-face 130 of the separated crystal 30 and the radius of curvature of the c-face 131 of the substrate 31 are larger than the radius of curvature of the c-face 121 of the seed substrate 21 is maintained.

Figure 4 (c) illustrates a substrate 31 configured in a flat plate shape. The substrate 31 is configured such that the lowest index crystal plane closest to the main surface 31s is the c-plane 131, and the c-plane 131 is curved in a concave spherical shape relative to the main surface 31s (relative to one of the main surfaces 31s of the substrate 31). The c-plane 131 has a constant radius of curvature (in the main region) because distortion from the spherical shape is suppressed. As described above, in step S200, the substrate 31 (crystal 30) is produced by the flux method using the seed substrate 21 (crystal 20) as a seed crystal.

(2) Effects Obtained by the Present Embodiment According to the present embodiment, one or more of the following effects can be obtained.

(a) By growing the crystal 30 on the seed substrate 21 so that an intermediate layer 24 having a closed space array 25C is formed (or an intermediate layer 24 having a matrix structure 25 is formed), the dislocation density of the GaN single crystal constituting the crystal 30 of the upper layer 26 formed above the intermediate layer 24 can be significantly reduced compared to the dislocation density of the GaN single crystal constituting the seed substrate 21.

The substrate 31 made of the crystal 30 has a very low dislocation density, for example, about 3×10 ⁵ /cm ² , uniformly distributed over the entire surface of the substrate, and is of high quality. For this reason, the substrate 31 is expected to contribute to improving the performance of semiconductor devices manufactured using the substrate, and is expected to be used, for example, as a substrate for manufacturing power devices that are more energy-efficient than existing devices.

By growing the crystal 30 with the formation of the intermediate layer 24 having a closed space array 25C (a matrix structure 25) (i.e., reflecting the formation of the crystal 30 by such growth), a crystal 30 is obtained in the upper layer 26 in which the ratio of dislocation types is changed relative to the seed substrate 21 (grown by the VAS method).

(b) The intermediate layer 24 having the closed space array 25C (matrix structure 25) can be formed over almost the entire surface of the crystal 30, for example, over 80% or more of the area, in a plan view seen from a direction perpendicular to the interface between the seed substrate 21 and the crystal 30. This makes it possible to obtain the above-mentioned effect of reducing dislocation density uniformly over almost the entire surface of the crystal 30.

The closed spaces 23 constituting the closed space array 25C of the intermediate layer 24 are aligned on the linear crystal defects 51 of the crystal body 30 originating from threading dislocations of the seed substrate 21. In order for the closed space array 25C to be formed in a state in which it is dispersed within the plane to such an extent that the matrix structure 25 is formed, it is considered preferable that the interval between the dislocations on the main surface 21s of the seed substrate 21 is relatively wide. This is because if the interval between the dislocations is excessively narrow, the closed spaces 23 (inclusions 22) formed on adjacent dislocations collide with each other and are not properly aligned. Therefore, it is preferable that the seed substrate 21 has a low dislocation density region (where the dislocation density is, for example, less than 1×10 ⁷ /cm ² ) on the main surface 21s as a region where the closed space array 25C is likely to be formed in a dispersed state.

Moreover, it is more preferable that the seed substrate 21 does not have a dislocation concentrated region with an excessively high dislocation density (a region with a dislocation density of, for example, 1×10 ⁷ /cm ² or more) in the main surface 21s, that is, the maximum dislocation density in the main surface 21s is, for example, less than 1×10 ⁷ /cm ^2. This makes it easy to form the closed space array 25C (the matrix structure 25) uniformly in the plane so as to cover a wide area above the main surface 21s of the seed substrate 21.

(c) By growing the substrate 31 by the flux method using the seed substrate 21 grown by the VAS method as a seed crystal, the radius of curvature of the c-plane 131 in the substrate 31 can be made larger than the radius of curvature of the c-plane 121 in the seed substrate 21. The larger the radius of curvature of the c-plane 131, the less the curvature of the c-plane 131 is, so that the width of the off-angle distribution in the substrate 31 can be made smaller. In other words, the substrate 31 with improved uniformity of the off-angle in the plane can be obtained. The spherical c-plane 131 of the substrate 31 is obtained by expanding the radius of curvature from the spherical c-plane 121 of the seed substrate 21. Therefore, the substrate 31 has a gentler change in the off-angle overall in the plane than the seed substrate 21. The substrate 31 has a gentler change in the off-angle overall in the plane than the seed substrate 21, so that the width of the off-angle distribution can be made smaller.

The larger the diameter of the substrate 31, the larger the width of the off-angle distribution tends to be, so this technology, which can reduce the width of the off-angle distribution, is particularly preferably applied to a substrate 31 with a large diameter. The radius of curvature of the c-face 121 of the seed substrate 21 is, for example, 3 m or more, and, for example, 15 m or less. The radius of curvature of the c-face 131 of the substrate 31 is larger than the radius of curvature of the c-face 121 of the seed substrate 21, and is preferably 15 m or more, more preferably 20 m or more, and even more preferably 30 m or more. The size of the substrate 31 is preferably, for example, 4 inches (10.16 cm) or more in diameter, in response to market demand for large diameter substrates. Correspondingly, the size of the seed substrate 21 is also preferably, for example, 4 inches or more in diameter. Since the c-face 131 is spherical, the off-angle does not change irregularly within a certain region of the main surface 31s (for example, within a region corresponding to one semiconductor device), and the change in the off-angle is gradual. In addition, because the c-plane 131 is spherical, the overall off-angle distribution within the main surface 31s can be easily grasped. In other words, because the radius of curvature of the c-plane 131 is large, the off-angle distribution within the plane can be made small even for a large-diameter substrate, and because the c-plane 131 is spherical, the change in the off-angle changes regularly.

By utilizing the phenomenon in which inclusions 22 are incorporated near the interface between the seed substrate 21 and the crystal 30 during crystal growth by the flux method, an intermediate layer 24 can be formed and the radius of curvature of the c-face 131 of the substrate 31 can be increased. By distributing the inclusions 22, i.e., the closed spaces 23, uniformly or in a non-localized, dispersed state within the interface between the seed substrate 21 and the crystal 30, the effect of increasing the radius of curvature can be obtained uniformly within the surface.

The concave spherical c-face 131 of the substrate 31 is obtained by expanding the radius of curvature of the concave spherical c-face 121 of the seed substrate 21. Therefore, the c-face 131 does not have a locally convex curvature, and local distortion is suppressed within the substrate 31. As a result, the substrate 31 is less susceptible to cracks and chips and is easy to process. This improves the yield in the manufacturing process of the substrate 31 and in the device manufacturing process using the substrate 31.

<Modification>
A modified example of the above embodiment will be described. In the embodiment, the manufacturing method for producing the substrate 31 using the seed substrate 21 as a seed crystal has been described. In this modified example, a manufacturing method for producing the crystal substrate 41 (hereinafter also referred to as substrate 41) using the substrate 31 as a seed crystal will be described.

Figure 8 is a flow chart showing a method for manufacturing a substrate 41 according to this modified example. This manufacturing method includes step S100 of preparing a seed substrate 21, step S200 of fabricating a substrate 31 by a liquid phase method, specifically a flux method, and step S300 of fabricating a substrate 41 by a gas phase method, for example, an HVPE method. Steps S100 and S200 are similar to those described with reference to Figure 1 in the above embodiment.

(S300: Preparation of crystal substrate by vapor phase method)
After the substrate 31 is prepared in step S200, the substrate 41 is fabricated by a gas phase method, for example, an HVPE method, in step S300. Figures 9(a) to 9(c) are schematic cross-sectional views showing the process of fabricating the substrate 41 in step S300. Figure 9(a) shows the substrate 31 prepared in step S200.

(S310: Crystal growth by HVPE method)
Step S300 includes steps S310 and S320. In step S310, as shown in FIG. 9B, a GaN single crystal is epitaxially grown on a substrate 31 by the HVPE method to form a crystal 40. Step S310 can be performed by the HVPE apparatus 200 using the substrate 31 as the substrate 250 to be processed, in a similar procedure to step S120 for producing the seed substrate 21. The processing conditions for the crystal growth process in step S310 are exemplified by the same conditions as those in step S120. Note that the substrate 31 used for the seed crystal of the crystal 40 may be the crystal 30 (the laminated crystal substrate 100 shown in FIG. 4B) laminated on the seed substrate 21.

The thickness of the crystal 40 to be grown is preferably a thickness that allows at least one free-standing substrate 41 to be obtained from the crystal 40, for example, a thickness of 0.2 mm or more. There is no particular upper limit to the thickness of the crystal 20 to be grown.

By growing the crystal 40 using the HVPE method, it is possible to suppress the hydrogen concentration and the oxygen concentration in the crystal 40. The hydrogen concentration in the crystal 40 is, for example, less than 1×10 ¹⁷ /cm ³ , and preferably 5×10 ¹⁶ /cm ³ or less. The oxygen concentration in the crystal 40 is, for example, 5×10 ¹⁶ /cm ³ or less, preferably 3×10 ¹⁶ /cm ³ or less, and more preferably 1×10 ¹⁶ /cm ³ or less.

(S320: Machining and polishing)
9C, in step S320, the crystal 40 grown in step S310 is cut by mechanical processing, for example, a wire saw, and separated from the substrate 31. The laminated crystal substrate 130 consisting of the substrate 31 and the root portion of the crystal 40 that remains after the separation of the crystal 40 may be reused as a seed crystal for crystal growth by the HVPE method or the like.

The separated crystal 40 is machined and/or polished as necessary to obtain a substrate 41 (crystal 40 as a free-standing substrate). For example, one substrate 41 may be obtained from the entire crystal 40. Alternatively, for example, multiple substrates 41 may be obtained from the entire crystal 40 by slicing the crystal 40 into multiple pieces. The obtained substrate 41 may or may not be polished as necessary. One of the two main surfaces may be polished. The separated crystal 40 may be used as the substrate 41 as is.

Due to the separation of the crystal 40 from the substrate 31, the radius of curvature of the c-face 140 of the crystal 40 may change from the radius of curvature of the c-face 140 of the crystal 40 before separation. Also, due to machining or polishing of the separated crystal 40, the radius of curvature of the c-face 141 of the substrate 41 may change from the radius of curvature of the c-face 140 of the separated crystal 40.

Figure 9 (c) illustrates a substrate 41 configured in a flat plate shape. The substrate 41 is configured such that the closest low-index crystal plane to the main surface 41s is the c-plane 141, and the c-plane 141 is curved in a concave spherical shape relative to the main surface 41s (relative to one of the main surfaces 41s of the substrate 41). The c-plane 141 has a constant radius of curvature (in the main region) because distortion from the spherical shape is suppressed.

As described above, in step S300, substrate 41 is produced by a gas phase method, for example, HVPE, using substrate 31 as a seed crystal. HVPE can grow crystals at a higher growth rate than flux methods. For this reason, this modified example is preferably used as a technique for obtaining thick-film crystal 40 using substrate 31 as a seed crystal after obtaining substrate 31 with enhanced crystallinity by flux methods.

The crystallinity of the substrate 41 (crystal 40) is as high as that of the substrate 31 (crystal 30). The radius of curvature of the c-plane 141 of the substrate 41 is at least as large as the radius of curvature of the c-plane 131 of the substrate 31, and is larger than the radius of curvature of the c-plane 121 of the seed substrate 21. As an experimental example of this modification, the free-standing substrate 41 was fabricated by the HVPE method. It has been confirmed that in the substrate 41 of the experimental example, a very low dislocation density of, for example, about 3×10 ⁵ /cm ² is obtained over the entire substrate surface. It has also been confirmed that in the substrate 41 of the experimental example, the radii of curvature of the c-plane in the m-axis direction and the a-axis direction are 20 m or more (preferably 30 m or more), for example, very large values of 76.4 m and 65.5 m, are obtained.

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

For example, in a modified example, one or more substrates may be obtained from a laminated crystal substrate 120 of the substrate 31 and the crystal 40, without separating the crystal 40 from the substrate 31. Also, for example, when growing the crystal 20, 30, or 40, impurities such as conductivity-determining impurities may be added as necessary.

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

(Appendix 1)
a first crystal body composed of a single crystal of a Group III nitride;
a second crystal body grown on the first crystal body and composed of a single crystal of a Group III nitride;
having
the second crystal body has an intermediate layer including a plurality of closed spaces (inclusions) containing an alkali metal;
the intermediate layer has a plurality of closed space rows formed by arranging a plurality of the closed spaces on (along) each of a plurality of linear crystal defects included in the second crystal body,
A crystal substrate, in which a certain row of closed spaces is selected from the plurality of rows of closed spaces, a first closed space and a second closed space included in the row of closed spaces are selected, and when the heights from the first closed space and the second closed space to the first crystal body are h1 and h2, respectively, there are a plurality of other rows of closed spaces having closed spaces at the same heights as h1 and h2.

(Appendix 2)
the second crystal body has an upper layer that is disposed above the intermediate layer and does not include the closed space,
2. The crystal substrate of claim 1, wherein the threading dislocation density in the main surface of the second crystal body of the upper layer is lower than the threading dislocation density in the main surface of the first crystal body that serves as the base for growth of the second crystal body.

(Appendix 3)
3. The crystal substrate according to claim 1, wherein the linear crystal defect contained in the second crystal body of the intermediate layer is formed with a threading dislocation contained in the first crystal body as an origin.

(Appendix 4)
4. The crystal substrate according to claim 1, wherein a main surface of the first crystal body, which serves as a base for growth of the second crystal body, has a region having a threading dislocation density of less than 1×10 ⁷ /cm ² .

(Appendix 5)
5. The crystal substrate according to claim 1, wherein the main surface of the first crystal body, which serves as a base for growth of the second crystal body, does not have an area having a threading dislocation density of 1×10 ⁷ /cm ² or more (the maximum dislocation density is less than 1×10 ⁷ /cm ² ).

(Appendix 6)
The crystal substrate according to any one of claims 1 to ^{5, wherein the second crystal body has an upper layer that is disposed above the intermediate layer and does not include the closed space, and a maximum threading dislocation density in a main surface of the second crystal body of the upper layer is 5 x 105} / ^cm2 or less.

(Appendix 7)
The crystal substrate according to any one of claims 1 to ^{6, wherein the second crystal body has an upper layer that is disposed above the intermediate layer and does not include the closed space, and an average threading dislocation density in a main surface of the second crystal body of the upper layer is 3 x 105} / ^cm2 or less.

(Appendix 8)
the c-plane of the single crystal constituting the first crystal body is curved into a concave spherical shape with a predetermined curvature,
the second crystal body has an upper layer that is disposed above the intermediate layer and does not include the closed space,
9. The crystal substrate according to claim 1, wherein a radius of curvature of a c-plane of the single crystal constituting the second crystal body of the upper layer is larger than a radius of curvature of a c-plane of the single crystal constituting the first crystal body.

(Appendix 9)
the second crystal body has an upper layer that is disposed above the intermediate layer and does not include the closed space,
the second crystal body of the upper layer is free of screw dislocations;
9. The crystal substrate according to claim 1, wherein a ratio of edge dislocations among threading dislocations contained in the second crystal body of the upper layer is lower than a ratio of edge dislocations among threading dislocations contained in the first crystal body.

(Appendix 10)
10. The crystal substrate of claim 9, wherein a ratio of edge dislocations among the threading dislocations contained in the second crystal body of the upper layer is less than 60%, and a ratio of edge dislocations among the threading dislocations contained in the first crystal body is 60% or more.

(Appendix 11)
a first crystal body composed of a single crystal of a Group III nitride;
a second crystal body grown on the first crystal body and composed of a single crystal of a Group III nitride;
having
the second crystal body has an intermediate layer including a plurality of closed spaces (inclusions) containing an alkali metal;
the intermediate layer has a plurality of closed space rows formed by arranging a plurality of the closed spaces on (along) each of a plurality of linear crystal defects included in the second crystal body,
A crystal substrate in which, when a certain row of closed spaces is selected from the plurality of rows of closed spaces, a first closed space and a second closed space included in the row of closed spaces are selected, and a first imaginary plane and a second imaginary plane are considered which pass through the first closed space and the second closed space, respectively, and are approximately parallel to the interface between the first crystal body and the second crystal body, there are a plurality of rows of closed spaces including closed spaces arranged on the first imaginary plane, and there are a plurality of rows of closed spaces including closed spaces arranged on the second imaginary plane.

(Appendix 12)
a first crystal body composed of a single crystal of a Group III nitride;
a second crystal body grown on the first crystal body and composed of a single crystal of a Group III nitride;
having
the second crystal body has an intermediate layer including a plurality of closed spaces (inclusions) containing an alkali metal;
the intermediate layer has a plurality of closed space rows formed by arranging a plurality of the closed spaces on (along) each of a plurality of linear crystal defects included in the second crystal body,
A crystal substrate in which, when the closed spaces included in the multiple closed space rows are projected onto a cross section perpendicular to the interface between the first crystal body and the second crystal body, there are multiple closed space rows within the cross section, each of which is formed by the multiple projected closed spaces being arranged in a direction approximately parallel to the interface.

(Appendix 13)
a first crystal body composed of a single crystal of a Group III nitride;
a second crystal body grown on the first crystal body and composed of a single crystal of a Group III nitride;
having
the second crystal body has an intermediate layer including a plurality of closed spaces (inclusions) containing an alkali metal;
The intermediate layer is a crystal substrate having a matrix structure of closed spaces formed by arranging a plurality of the closed spaces in rows in a direction along the interface between the first crystal body and the second crystal body, and arranging a plurality of the closed spaces in a direction intersecting the interface, in a cross section intersecting the interface.

(Appendix 14)
14. The crystal substrate of claim 13, wherein in the matrix structure, the closed spaces are arranged in three or more rows (four or more, five or more rows).

(Appendix 15)
15. The crystal substrate according to claim 13, wherein in the matrix structure, the closed spaces are arranged in three or more rows (four or more, five or more rows).

(Appendix 16)
16. The crystal substrate according to any one of claims 13 to 15, wherein the matrix structure is formed over 500 μm or more in a direction along the interface.

(Appendix 17)
17. The crystal substrate according to any one of claims 13 to 16, wherein the matrix structure is formed over 80% or more of the area of the second crystal mass in a plan view.

(Appendix 18)
A substrate made of a single crystal of a Group III nitride,
A crystal substrate, wherein the ratio of edge dislocations among the threading dislocations contained in the single crystal is less than 60%.

(Appendix 19)
19. The crystal substrate of claim 18, wherein the threading dislocations contained in the single crystal are free of screw dislocations.

(Appendix 20)
Has a diameter of 4 inches or more;
20. The crystal substrate according to claim 18 or 19, wherein the c-plane of the single crystal is curved inwardly of the substrate with a radius of curvature of 15 m or more in a concave spherical shape when the main surface of the substrate is viewed from the +c side, and has a constant radius of curvature in a region of 80% or more of the main surface in a planar view.

(Appendix 21)
21. The crystal substrate according to any one of claims 18 to 20, wherein both of a radius of curvature of the c-plane of the single crystal in the a-axis direction and a radius of curvature of the c-plane of the single crystal in the m-axis direction are 20 m or more (preferably 30 m or more).

(Appendix 22)
22. The crystal substrate according to any one of claims 18 to 21, wherein the maximum threading dislocation density in the main surface of the single crystal is 5×10 ⁵ /cm ² or less.

(Appendix 23)
23. The crystal substrate according to any one of claims 18 to 22, wherein the average threading dislocation density in the main surface of the single crystal is 3×10 ⁵ /cm ² or less.

(Appendix 24)
24. The crystal substrate according to any one of claims 18 to 23, wherein the ratio of the maximum dislocation density to the minimum threading dislocation density in the main surface of the single crystal is 100 times or less.

(Appendix 25)
25. The crystal substrate according to any one of claims 18 to 24, wherein the hydrogen concentration in the single crystal is less than 1×10 ¹⁷ /cm ³ (preferably 5×10 ¹⁶ /cm ³ or less).

(Appendix 26)
The crystal substrate ^{according to any one of appendices 18 to 25, wherein the oxygen concentration in the single crystal is less than 1×10 17 /cm 3 (preferably 8×10 16 / cm 3 or less} , more preferably 5×10 ¹⁶ /cm ³ or less, even more preferably 3×10 ¹⁶ /cm ³ or less, and even more preferably 1×10 16 /cm 3 or less).

(Appendix 27)
preparing a first crystal body composed of a single crystal of a Group III nitride;
growing a second crystal body composed of a single crystal of a Group III nitride on the first crystal body in a mixed melt containing an alkali metal and a Group III element;
having
In the step of growing the second crystal body,
the second crystal body has an intermediate layer including a plurality of closed spaces (inclusions) containing the alkali metal;
a plurality of closed spaces are arranged on each of the plurality of linear crystal defects included in the second crystal body (along each of the plurality of linear crystal defects) in the intermediate layer, thereby forming a plurality of closed space rows;
A method for manufacturing a crystal substrate, comprising the steps of: selecting a certain row of closed spaces from the plurality of rows of closed spaces; selecting a first closed space and a second closed space included in the row of closed spaces; and growing the second crystal body so that, when the heights from the first closed space and the second closed space to the first crystal body are h1 and h2, respectively, there are a plurality of other rows of closed spaces having closed spaces at the same heights as h1 and h2.

(Appendix 28)
the mixed molten liquid contains, in addition to the alkali metal and the Group III element which is a Group 13 typical metal element, other metals;
the other metal includes at least one metal selected from among a Group 4 transition metal, a Group 5 transition metal, a Group 6 transition metal, a Group 7 transition metal, a Group 8 transition metal, a Group 9 transition metal, a Group 10 transition metal, a Group 11 transition metal, a Group 12 typical metal, a Group 13 typical metal element different from the Group III element, a Group 14 typical metal, and a Group 15 typical metal;
In the step of growing the second crystal body,
A method for manufacturing a crystal substrate as described in Appendix 27, in which the growth temperature is alternately decreased and increased, or the growth pressure is alternately decreased and increased, so that the intermediate layer is formed by alternating between growth of the second crystal body in which the closed space is formed and growth of the second crystal body in which the closed space is not formed.

(Appendix 29)
29. The method for producing a crystal substrate according to claim 28, wherein the other metal includes at least one of Ti, Nb, Cr, Fe, Ni, Zn, Ge, Sn, Sb, and Bi.

(Appendix 30)
30. The method for producing a crystal substrate according to claim 28 or 29, wherein a concentration of the other metal, which is a ratio of a content of the other metal to a combined content of the Group III element and the other metal in the mixed molten liquid, is set to 3% or more and 15% or less (preferably 4% or more and 12% or less).

(Appendix 31)
In the step of growing the second crystal body,
31. The method for manufacturing a crystal substrate according to any one of claims 28 to 30, wherein after the intermediate layer is formed, the second crystal body is grown in such a way that the closed space is not formed, thereby forming an upper layer that is disposed above the intermediate layer and does not include the closed space.

(Appendix 32)
32. The method for producing a crystal substrate according to any one of claims 28 to 31, further comprising the step of growing a third crystal body composed of a single crystal of a Group III nitride on the second crystal body by a vapor phase method.

10...underlying substrate, 11...underlying layer, 12...metal layer, 13...void-containing layer, 14...nanomask, 15...void-formed substrate, 16...void, 20, 30, 40...crystal, 21, 31, 41...substrate, 20s, 21s, 30s, 31s, 40s, 41s...main surface, 120, 121, 130, 131, 140, 141...c-plane, 100, 110, 120, 130...laminated crystal substrate, 22...inclusion, 23...closed space, 24...intermediate layer, 25...matrix-like structure, 25R...closed space row, 25C...closed space column, 26...upper layer, 27...interface (between seed substrate 21 and crystal 30), 51...linear crystal defect, 52...virtual plane, 53...cross section, 54...threading dislocation (in upper layer)

Claims (13)

A substrate made of a single crystal of a Group III nitride, the low-index crystal plane closest to a primary surface of the substrate being a c-plane;
A crystal substrate, wherein threading dislocations penetrating the c-plane contained in the single crystal extend in the thickness direction of the substrate without bending in a direction perpendicular to the thickness direction of the substrate, and a ratio of edge dislocations among the threading dislocations is less than 60%. The crystal substrate according to claim 1 , wherein the oxygen concentration in the single crystal is less than 1×10 ¹⁷ /cm ³ . The crystal substrate according to claim 1 or 2, wherein the threading dislocations contained in the single crystal do not include screw dislocations. The crystal substrate according to any one of claims 1 to 3, wherein both the radius of curvature of the c-plane of the single crystal in the a-axis direction and the radius of curvature of the c-plane of the single crystal in the m-axis direction are 20 m or more. 5. The crystal substrate according to claim 1, wherein the maximum threading dislocation density in the main surface of the single crystal is 5×10 ⁵ /cm ² or less. 6. The crystal substrate according to claim 1, wherein the average threading dislocation density in the main surface of the single crystal is 3×10 ⁵ /cm ² or less. The crystal substrate according to any one of claims 1 to 6, wherein the ratio of the maximum dislocation density to the minimum threading dislocation density on the main surface of the single crystal is 100 times or less. 8. The crystal substrate according to claim 1, wherein the hydrogen concentration in the single crystal is less than 1×10 ¹⁷ /cm ³ . a first crystal body composed of a single crystal of a Group III nitride;
a second crystal body grown on the first crystal body and composed of a single crystal of a Group III nitride;
having
the second crystal body has an intermediate layer including a plurality of closed spaces containing an alkali metal;
the intermediate layer has a plurality of closed space rows formed by arranging a plurality of the closed spaces on each of a plurality of linear crystal defects included in the second crystal body,
A crystal substrate, in which a certain row of closed spaces is selected from the plurality of rows of closed spaces, a first closed space and a second closed space included in the row of closed spaces are selected, and when the heights from the first closed space and the second closed space to the first crystal body are h1 and h2, respectively, there are a plurality of other rows of closed spaces having closed spaces at the same heights as h1 and h2. the second crystal body has an upper layer that is disposed above the intermediate layer and does not include the closed space,
10. The crystal substrate of claim 9, wherein the threading dislocation density in the main surface of the second crystal body of the upper layer is lower than the threading dislocation density in the main surface of the first crystal body that serves as the base for growth of the second crystal body. The crystal substrate according to claim 9 or 10, wherein the linear crystal defects contained in the second crystal body of the intermediate layer are formed starting from threading dislocations contained in the first crystal body. the second crystal body has an upper layer that is disposed above the intermediate layer and does not include the closed space,
the second crystal body of the upper layer is free of screw dislocations;
12. The crystal substrate according to claim 9, wherein a ratio of edge dislocations among threading dislocations contained in the second crystal body of the upper layer is lower than a ratio of edge dislocations among threading dislocations contained in the first crystal body. preparing a first crystal body composed of a single crystal of a Group III nitride;
growing a second crystal body composed of a single crystal of a Group III nitride on the first crystal body in a mixed melt containing an alkali metal and a Group III element;
having
In the step of growing the second crystal body,
the second crystal body has an intermediate layer including a plurality of closed spaces containing the alkali metal;
a plurality of closed space rows are formed by arranging a plurality of the closed spaces on each of a plurality of linear crystal defects included in the second crystal body in the intermediate layer,
A method for manufacturing a crystal substrate, comprising the steps of: selecting a certain row of closed spaces from the plurality of rows of closed spaces; selecting a first closed space and a second closed space included in the row of closed spaces; and growing the second crystal body so that, when the heights from the first closed space and the second closed space to the first crystal body are h1 and h2, respectively, there are a plurality of other rows of closed spaces having closed spaces at the same heights as h1 and h2.

Images