Patent Application
20130040442
The present invention relates to a method of manufacturing a GaN-based film capable of obtaining a GaN-based film having a large main surface area and less warpage.
A GaN-based film is suitably used as a substrate and a semiconductor layer in a semiconductor device such as a light emitting device and an electronic device. A GaN substrate is best as a substrate for manufacturing such a GaN-based film, from a point of view of match or substantial match in lattice constant and coefficient of thermal expansion between the substrate and the GaN-based film. A GaN substrate, however, is very expensive, and it is difficult to obtain such a GaN substrate having a large diameter that a diameter of a main surface exceeds 2 inches.
Therefore, a sapphire substrate is generally used as a substrate for forming a GaN-based film. A sapphire substrate and a GaN crystal are significantly different from each other in lattice constant and coefficient of thermal expansion. Therefore, in order to mitigate unmatch in lattice constant between a sapphire substrate and a GaN crystal and to grow a GaN crystal having good crystallinity, for example, Japanese Patent Laying-Open No. 04-297023 discloses forming a GaN buffer layer on a sapphire substrate and growing a GaN crystal layer on the GaN buffer layer, in growing GaN crystal on the sapphire substrate.
In addition, in order to obtain a GaN film less in warpage by employing a substrate having a coefficient of thermal expansion close to that of GaN crystal, for example, Japanese National Patent Publication No. 2007-523472 discloses a composite support substrate having one or more pairs of layers having substantially the same coefficient of thermal expansion with a central layer lying therebetween and having an overall coefficient of thermal expansion substantially the same as a coefficient of thermal expansion of GaN crystal.
According to Japanese Patent Laying-Open No. 04-297023 above, GaN crystal grows as warping in a shape recessed in a direction of growth of crystal, probably because crystal defects such as dislocation disappear as a result of association during growth of the GaN crystal.
As described above, however, the sapphire substrate is much higher in coefficient of thermal expansion than GaN crystal, and hence grown GaN crystal greatly warps in a shape projecting in a direction of growth of crystal during cooling after crystal growth and a GaN film great in warpage in a shape projecting in the direction of growth of crystal is obtained. Here, as the main surface of the sapphire substrate has a greater diameter, warpage of the GaN crystal during cooling above becomes greater (specifically, warpage of the obtained GaN film is substantially in proportion to a square of a diameter of the main surface of the sapphire substrate). Therefore, it becomes difficult to obtain a GaN film less in warpage as the main surface has a greater diameter.
The composite support substrate disclosed in Japanese National Patent Publication No. 2007-523472 above has a coefficient of thermal expansion substantially the same as that of the GaN crystal and hence warpage of the GaN layer grown thereon can be less. Such a composite support substrate, however, has a complicated structure, and design and formation of the structure is difficult. Therefore, cost for design and manufacturing becomes very high and cost for manufacturing a GaN film becomes very high.
An object of the present invention is to solve the problems above and to provide a method of manufacturing a GaN-based film capable of manufacturing a GaN-based film having a large main surface area and less warpage.
According to one aspect, the present invention is directed to a method of manufacturing a GaN-based film including the steps of preparing a composite substrate, the composite substrate including a support substrate in which a coefficient of thermal expansion in a main surface is more than 0.8 time and less than 1.2 times as high as a coefficient of thermal expansion of GaN crystal in a direction of a axis and a single crystal film arranged on a side of a main surface of the support substrate, the single crystal film having threefold symmetry with respect to an axis perpendicular to a main surface of the single crystal film, and forming a GaN-based film on a main surface of the single crystal film in the composite substrate.
In the method of manufacturing a GaN-based film according to the present invention, the main surface of the single crystal film in the composite substrate can have an area equal to or greater than 45 cm2. In addition, the support substrate of the composite substrate can be made of a sintered body containing an oxide. Moreover, the support substrate of the composite substrate can contain yttria stabilized zirconia and an Al2O3—SiO2 composite oxide formed of alumina and silica. Here, a content of yttria stabilized zirconia to the total of the Al2O3—SiO2 composite oxide and yttria stabilized zirconia can be not lower than 20 mass % and not higher than 40 mass %. Further, a content of yttria to yttria stabilized zirconia can be not lower than 5 mol %. Furthermore, the step of forming a GaN-based film can include sub steps of forming a GaN-based buffer layer on the main surface of the single crystal film and forming a GaN-based single crystal layer on a main surface of the GaN-based buffer layer.
According to the present invention, a method of manufacturing a GaN-based film capable of manufacturing a GaN-based film having a large main surface area and less warpage can be provided.
Referring to
According to the method of manufacturing a GaN-based film in the present embodiment, by employing a composite substrate including a support substrate in which a coefficient of thermal expansion in a main surface is more than 0.8 time and less than 1.2 times as high as a coefficient of thermal expansion of GaN crystal in a direction of a axis and a single crystal film arranged on a main surface side of the support substrate, the single crystal film having threefold symmetry with respect to an axis perpendicular to a main surface of the crystal film, a GaN-based film having a large main surface area (that is, a large diameter) and less warpage can be obtained.
(Step of Preparing Composite Substrate)
Referring to
Composite substrate 10 above includes support substrate 11 in which a coefficient of thermal expansion in main surface 11m matches with or is close to (specifically, more than 0.8 time and less than 1.2 times as high as) a coefficient of thermal expansion of GaN crystal in the direction of a axis and single crystal film 13 arranged on the main surface 11m side of support substrate 11, and single crystal film 13 has threefold symmetry with respect to the axis perpendicular to main surface 13m of single crystal film 13. Therefore, a GaN-based film less in warpage, low in dislocation density, and having a large diameter can be grown on main surface 13m of single crystal film 13 of composite substrate 10.
From a point of view of growing a GaN-based film less in warpage, low in dislocation density, and having a large diameter on single crystal film 13 of composite substrate 10, support substrate 11 included in composite substrate 10 above should have a coefficient of thermal expansion in main surface 11m more than 0.8 time and less than 1.2 times, preferably more than 0.9 time and less than 1.15 times and further preferably more than 0.95 time and less than 1.1 times, as high as a coefficient of thermal expansion of GaN crystal in the direction of a axis.
Here, support substrate 11 is not particularly restricted, so long as a substrate has a coefficient of thermal expansion in main surface 11m more than 0.8 time and less than 1.2 times as high as a coefficient of thermal expansion of GaN crystal in the direction of a axis, and a substrate may be monocrystalline, polycrystalline, or non-crystalline. Support substrate 11 is preferably made of a sintered body containing an oxide, from a point of view of ease in adjustment of a coefficient of thermal expansion based on variation in type and ratio of source materials and ease in obtaining a coefficient of thermal expansion in the range above, resistance to an atmosphere in which a GaN-based film is grown, and allowing selective etching by such a solvent as hydrofluoric acid (hydrogen fluoride) after growth of a GaN-based film. For example, preferred examples of the sintered bodies include an Al2O3—SiO2-based sintered body, an MgO—SiO2-based sintered body, a ZrO2—SiO2-based sintered body, a ZrO2—Al2O3—SiO2-based sintered body, a Y2O3—Al2O3—SiO2-based sintered body, a Y2O3—ZrO2—Al2O3—SiO2-based sintered body, and the like.
Alternatively, support substrate 11 particularly preferably contains yttria stabilized zirconia (Y2O3 stabilized ZrO2, hereinafter also referred to as YSZ) and an Al2O3—SiO2 composite oxide formed of alumina (Al2O3) and silica (SiO2). Support substrate 11 containing the Al2O3—SiO2 composite oxide and YSZ (yttria stabilized zirconia) is dissoluble in hydrofluoric acid (hydrogen fluoride) and it allows growth of a GaN-based film excellent in crystallinity on single crystal film 13 arranged on the main surface side of support substrate 11 of composite substrate 10. Here, the Al2O3—SiO2 composite oxide is not particularly restricted, and mullite (3Al2O3.2SiO2 to 2Al2O3.SiO2 or Al6O13Si2) or the like is suitable.
From a point of view of suppression of cracks generated in a GaN-based film during growth of a GaN-based film excellent in crystallinity on single crystal film 13 on support substrate 11 containing an Al2O3—SiO2 composite oxide and YSZ (yttria stabilized zirconia), a content of YSZ to the total of the Al2O3—SiO2 composite oxide and YSZ is preferably not lower than 20 mass % and not higher than 40 mass % and more preferably not lower than 25 mass % and not higher than 35 mass %. Further, from a point of view the same as above, a content of yttria (Y2O3) to YSZ is preferably not lower than 5 mol % and more preferably not lower than 6 mol % and not higher than 50 mol %.
Here, since a coefficient of thermal expansion of each of support substrate 11 and GaN crystal generally greatly fluctuates depending on a temperature thereof, it is important at which temperature or in which temperature region determination should be made based on a coefficient of thermal expansion. The present invention aims to manufacture a GaN-based film less in warpage on a composite substrate. A GaN-based film is formed on the composite substrate at a film formation temperature for a GaN-based film with a temperature being increased from room temperature, thereafter the temperature is lowered to room temperature, and then the GaN-based film formed on the composite substrate is taken out. Therefore, it is considered as appropriate to handle an average coefficient of thermal expansion of each of the support substrate and the GaN crystal from room temperature to the film formation temperature for the GaN-based film as the coefficient of thermal expansion of each of the support substrate and the GaN crystal. The GaN crystal, however, decomposes even in an inert gas atmosphere if a temperature exceeds 800° C. Therefore, in the present invention, the coefficient of thermal expansion of each of the support substrate and the GaN crystal is determined by an average coefficient of thermal expansion from room temperature (specifically, 25° C.) to 800° C.
In addition, from a point of view of growing a GaN-based film less in warpage, low in dislocation density, and having a large diameter on single crystal film 13 of composite substrate 10, single crystal film 13 arranged on the main surface 11m side of support substrate 11 included in composite substrate 10 above should have threefold symmetry with respect to the axis perpendicular to main surface 13m of single crystal film 13, and preferred examples of the single crystal film include a sapphire film having a (0001) plane as main surface 13m, an SiC film having a (0001) plane as main surface 13m, an Si film having a (111) plane as main surface 13m, a GaAs film having a (111) plane as main surface 13m, and the like. Here, the single crystal film having threefold symmetry with respect to the axis perpendicular to the main surface of the single crystal film does not mean having threefold symmetry strict in terms of crystal geometry but having substantial threefold symmetry in an actual single crystal film, and specifically means that an absolute value of an angle between a threefold symmetry axis strict in terms of crystal geometry of the single crystal film and an axis perpendicular to the main surface of the single crystal film being not greater than 10° suffices.
From a point of view of lessening warpage and lowering dislocation density in composite substrate 10, main surface 11m of support substrate 11 and main surface 13m of single crystal film 13 are preferably substantially parallel to each other. Here, two surfaces being substantially parallel to each other means that an absolute value of an angle formed by these two surfaces is not greater than 10°.
In addition, a method of arranging single crystal film 13 on the main surface 11m side of support substrate 11 of composite substrate 10 is not particularly restricted, and exemplary methods include a method of directly growing single crystal film 13 on main surface 11m of support substrate 11 (a first method), a method of bonding single crystal film 13 formed on a main surface of an underlying substrate to main surface 11m of support substrate 11 and thereafter removing the underlying substrate (a second method), a method of bonding single crystal (not shown) to main surface 11m of support substrate 11 and thereafter separating the single crystal at a plane at a prescribed depth from a bonding surface to thereby form single crystal film 13 on main surface 11m of support substrate 11 (a third method), and the like. In a case where a support substrate is made of a polycrystalline sintered body, the first method above is difficult and hence any of the second and third methods above is preferably employed. A method of bonding single crystal film 13 to support substrate 11 in the second method above is not particularly restricted, and exemplary methods include a method of directly bonding single crystal film 13 to main surface 11m of support substrate 11, a method of bonding single crystal film 13 to main surface 11m of support substrate 11 with an adhesive layer 12 being interposed, and the like. A method of bonding single crystal to support substrate 11 in the third method above is not particularly restricted, and exemplary methods include a method of directly bonding single crystal to main surface 11m of support substrate 11, a method of bonding single crystal to main surface 11m of support substrate 11 with adhesive layer 12 being interposed, and the like.
The step of preparing composite substrate 10 above is not particularly restricted. From a point of view of efficient preparation of composite substrate 10 of high quality, however, for example, referring to
In
A specific technique for bonding support substrate 11 and single crystal film 13 to each other is not particularly restricted. From a point of view of ability to hold joint strength even at a high temperature after bonding, however, a direct joint method of washing a bonding surface, performing bonding, and thereafter increasing a temperature to about 600° C. to 1200° C. for joint, a surface activation method of washing a bonding surface, activating the bonding surface with plasma, ions or the like, and thereafter performing joint at a low temperature from around room temperature (for example, 25° C.) to 400° C., and the like are preferably employed.
(Step of Forming GaN-Based Film)
Referring to
Composite substrate 10 prepared in the step of preparing a composite substrate above includes support substrate 11 in which a coefficient of thermal expansion in main surface 11m matches with or is close to (specifically, more than 0.8 time and less than 1.2 times as high as) a coefficient of thermal expansion of GaN crystal in the direction of a axis and single crystal film 13 arranged on the main surface 11m side of support substrate 11, and single crystal film 13 has threefold symmetry with respect to the axis perpendicular to main surface 13m of single crystal film 13. Therefore, GaN-based film 20 less in warpage, low in dislocation density, and having a large diameter can be formed on main surface 13m of single crystal film 13 of composite substrate 10.
Though a method of forming a GaN-based film is not particularly restricted, from a point of view of forming a GaN-based film low in dislocation density, a vapor phase epitaxy method such as an MOCVD (Metal Organic Chemical Vapor Deposition) method, an HVPE (Hydride Vapor Phase Epitaxy) method, an MBE (Molecular Beam Epitaxy) method, and a sublimation method, a liquid phase epitaxy method such as a flux method and a high nitrogen pressure solution method, and the like are preferably exemplified.
The step of forming a GaN-based film is not particularly restricted. From a point of view of forming a GaN-based film low in dislocation density, however, the step preferably includes a sub step of forming a GaN-based buffer layer 21 on main surface 13m of single crystal film 13 of composite substrate 10 and a sub step of forming a GaN-based single crystal layer 23 on a main surface 21m of GaN-based buffer layer 21. Here, GaN-based buffer layer 21 refers to a layer low in crystallinity or non-crystalline, that is a part of GaN-based film 20 and grown at a temperature lower than a growth temperature of GaN-based single crystal layer 23 which is another part of GaN-based film 20.
By forming GaN-based buffer layer 21, unmatch in lattice constant between GaN-based single crystal layer 23 formed on GaN-based buffer layer 21 and single crystal film 13 is mitigated, and hence crystallinity of GaN-based single crystal layer 23 improves and dislocation density thereof is lowered. Consequently, crystallinity of GaN-based film 20 improves and dislocation density thereof is lowered.
GaN-based single crystal layer 23 can also be grown as GaN-based film 20 on single crystal film 13, without growing GaN-based buffer layer 21. Such a method is suitable for a case where unmatch in lattice constant between single crystal film 13 and GaN-based film 20 formed thereon is less.
1. Measurement of Coefficient of Thermal Expansion of GaN Crystal
A sample for evaluation having a size of 2×2×20 mm (having the a axis in a longitudinal direction and having any of a c plane and an m plane as a plane in parallel to the longitudinal direction, with accuracy in plane orientation being within ±0.1° was cut from GaN single crystal grown with the HVPE method and having dislocation density of 1×106 cm−2, Si concentration of 1×1018 cm2, oxygen concentration of 1×1017 cm−2, and carbon concentration of 1×1016 cm−2.
An average coefficient of thermal expansion of the sample for evaluation above when a temperature was increased from a room temperature (25° C.) to 800° C. was measured with TMA (thermomechanical analysis). Specifically, using TMA8310 manufactured by Rigaku Corporation, the coefficient of thermal expansion of the sample for evaluation was measured with differential dilatometry in an atmosphere in which a nitrogen gas flows. An average coefficient of thermal expansion αGaN-a from 25° C. to 800° C. of GaN crystal in the direction of the a axis obtained by such measurement was 5.84×106/° C.
2. Step of Preparing Composite Substrate
(1) Sub Step of Preparing Support Substrate
Referring to
Al2O3—SiO2-based sintered body A had a molar ratio between Al2O3 and SiO2 of 53:47 and attained average coefficient of thermal expansion αS from 25° C. to 800° C. (hereinafter simply referred to as average coefficient of thermal expansion αS) of 4.2×10−6/° C., and a ratio of average coefficient of thermal expansion αS of the sintered body to an average coefficient of thermal expansion αGaN-a from 25° C. to 800° C. of the GaN crystal in the direction of a axis (hereinafter simply referred to as average coefficient of thermal expansion αGaN-a) (hereinafter denoted as an αS/αGaN-a ratio) was 0.719. Al2O3—SiO2-based sintered body B had a molar ratio between Al2O3 and SiO2 of 45:55, and attained average coefficient of thermal expansion αS of 4.7×10−6/° C. and the αS/αGaN-a ratio of 0.805. Al2O3—SiO2-based sintered body C had a molar ratio between Al2O3 and SiO2 of 40:60, and attained average coefficient of thermal expansion αS of 5.2×10−6/° C. and the αS/αGaN-a ratio of 0.890. Al2O3—SiO2-based sintered body D had a molar ratio between Al2O3 and SiO2 of 36:64, and attained average coefficient of thermal expansion αS of 5.4×10−6/° C. and the αS/αGaN-a ratio of 0.925. Al2O3—SiO2-based sintered body E had a molar ratio between Al2O3 and SiO2 of 35:65, and attained average coefficient of thermal expansion as of 5.5×10−6/° C. and the αS/αGaN-a ratio of 0.942. Al2O3—SiO2-based sintered body F had a molar ratio between Al2O3 and SiO2 of 34:66, and attained average coefficient of thermal expansion αS of 5.6×10−6/° C. and the αS/αGaN, ratio of 0.959. Al2O3—SiO2-based sintered body G had a molar ratio between Al2O3 and SiO2 of 32:68, and attained average coefficient of thermal expansion αS of 5.7×10−6/° C. and the αS/αGaN-a ratio of 0.976. Al2O3—SiO2-based sintered body H had a molar ratio between Al2O3 and SiO2 of 26:74, and attained average coefficient of thermal expansion αS of 6.1×10−6/° C. and the αS/αGaN-a ratio of 1.045. Al2O3—SiO2-based sintered body I had a molar ratio between Al2O3 and SiO2 of 24:76, and attained average coefficient of thermal expansion αS of 6.4×10−6/° C. and the αS/αGaN, ratio of 1.096. Al2O3—SiO2-based sintered body J had a molar ratio between Al2O3 and SiO2 of 20:80, and attained average coefficient of thermal expansion αS of 6.6×10−6/° C. and the αS/αGaN-a ratio of 1.130. Al2O3—SiO2-based sintered body K had a molar ratio between Al2O3 and SiO2 of 16:84, and attained average coefficient of thermal expansion αS of 7.0×10−6/° C. and the αS/αGaN-a ratio of 1.199. Al2O3—SiO2-based sintered body L had a molar ratio between Al2O3 and SiO2 of 14:86, and attained average coefficient of thermal expansion αS of 7.2×10−6/° C. and the αS/αGaN-a ratio of 1.233. Al2O3—SiO2-based sintered body M had a molar ratio between Al2O3 and SiO2 of 9:91, and attained average coefficient of thermal expansion αS of 7.5×10−6/° C. and the αS/αGaN-a ratio of 1.284.
A support substrate having a diameter of 4 inches (101.6 mm) and a thickness of 1 mm was cut from each of the 13 types of Al2O3—SiO2-based sintered bodies A to M above, and opposing main surfaces of each support substrate were mirror-polished to thereby obtain support substrates A to M. Namely, an average coefficient of thermal expansion of each of support substrates A to M from 25° C. to 800° C. was equal to an average coefficient of thermal expansion of each corresponding one of Al2O3—SiO2-based sintered bodies A to M from 25° C. to 800° C. Table 1 summarizes the results.
(2) Sub Step of Forming Single Crystal Film on Underlying Substrate
Referring to
An SiC film having a thickness of 0.4 μM was formed as single crystal film 13 on main surface 30n of the Si substrate (underlying substrate 30) above with a CVD (chemical vapor deposition) method. Regarding film formation conditions, an SiH4 gas and a C3H8 gas were used as source gases, an H2 gas was used as a carrier gas, a film formation temperature was set to 1300° C., and a film formation pressure was set to an atmospheric pressure. Main surface 13m of the SiC film (single crystal film 13) thus obtained had a plane orientation having an off angle within ±1° with respect to the (0001) plane, and included an Si atomic plane (a (0001) plane) and a C atomic plane (a (000-1) plane) as mixed like mosaic.
(3) Sub Step of Bonding Support Substrate and Single Crystal Film to Each Other
Referring to (C1) in
Referring also to (C2) in
Referring next to (C3) in
(4) Sub Step of Removing Underlying Substrate
Referring to
3. Step of Forming GaN-Based Film
Referring to
Regarding wafers A to M and R taken out of a film formation apparatus after cooling to room temperature, warpage of the wafer and crack count and dislocation density of the GaN film were determined. Here, a shape of warpage and an amount of warpage of the wafer were determined based on interference fringes observed at the main surface of the GaN film with FM200EWafer of Corning Tropel. Regarding crack count in the GaN film, the number of cracks per unit length was counted with a Nomarski microscope, and evaluation as “extremely few”, “few”, “many”, and “extremely many” was made when the count was smaller than 1 count/mm, when the count was not smaller than 1 count/mm and smaller than 5 counts/mm, when the count was not smaller than 5 counts/mm and smaller than 10 counts/mm, and when the count was not smaller than 10 counts/mm, respectively. Dislocation density of the GaN film was measured with CL (cathode luminescence) based on the number of dark points per unit area. It is noted that cracks generated in the GaN film in the present Example were small without penetrating the film.
Wafer A warped on the GaN film side in a recessed manner, an amount of warpage was 700 μm, and cracks counted in the GaN film were extremely many. Wafer B warped on the GaN film side in a recessed manner, an amount of warpage was 650 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 4×108 cm−2. Wafer C warped on the GaN film side in a recessed manner, an amount of warpage was 400 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 3×108 cm−2. Wafer D warped on the GaN film side in a recessed manner, an amount of warpage was 350 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2.5×108 cm−2. Wafer E warped on the GaN film side in a recessed manner, an amount of warpage was 320 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer F warped on the GaN film side in a recessed manner, an amount of warpage was 230 μm, cracks counted in the GaN film were extremely few, and dislocation density of the GaN film was 1×108 cm2. Wafer G warped on the GaN film side in a recessed manner, an amount of warpage was 200 μm, cracks counted in the GaN film were extremely few, and dislocation density of the GaN film was 1×108 cm−2. Wafer H warped on the GaN film side in a recessed manner, an amount of warpage was 10 μm, cracks counted in the GaN film were extremely few, and dislocation density of the GaN film was 1×108 cm−2. Wafer I warped on the GaN film side in a projecting manner, an amount of warpage was 20 μm, cracks counted in the GaN film were extremely few, and dislocation density of the GaN film was 1×108 cm−2. Wafer J warped on the GaN film side in a projecting manner, an amount of warpage was 110 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer K warped on the GaN film side in a projecting manner, an amount of warpage was 230 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 3×108 cm−2. Wafer L warped on the GaN film side in a projecting manner, an amount of warpage was 740 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 4×108 cm2. In wafer M, cracking occurred in support substrate M and a sufficient GaN film was not obtained. Wafer R warped on the GaN film side in a projecting manner, an amount of warpage was 750 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 4×108 cm2. Table 1 summarizes these results. In Table 1, “-” indicates that that physical property value was not measured.
Referring to Table 1, by employing a composite substrate (wafers B to K) having a support substrate in which coefficient of thermal expansion αS in a main surface was more than 0.8 time and less than 1.2 times (that is, 0.8<(αS/αGaN-a ratio)<1.2) as high as coefficient of thermal expansion αGaN-a of GaN crystal in the direction of a axis, as compared with a case where a sapphire substrate was employed (wafer R), a GaN film extremely less in warpage could be formed. In addition, from a point of view of further decrease in warpage and dislocation density of the GaN film in the wafer, coefficient of thermal expansion αS in a main surface of the support substrate of the composite substrate was preferably more than 0.9 time and less than 1.15 times (that is, 0.9<(αS/αGaN-a ratio)<1.15) as high as coefficient of thermal expansion αGaN-a of the GaN crystal in the direction of a axis (wafers D to J) and further preferably more than 0.95 time and less than 1.10 times (that is, 0.95<(αS/αGaN-a ratio)<1.10) as high as coefficient of thermal expansion αGaN-a of the GaN crystal in the direction of a axis (wafers F to I).
1. Measurement of Coefficient of Thermal Expansion of GaN Crystal
Measurement as in Example 1 was conducted and then average coefficient of thermal expansion αGaN-a of the GaN crystal in the direction of a axis from 25° C. to 800° C. was 5.84×10−6/° C.
2. Step of Preparing Composite Substrate
(1) Sub Step of Preparing Support Substrate
Referring to
YSZ-mullite-based sintered body A0 had a content of YSZ to the total of YSZ and mullite (hereinafter referred to as a YSZ content) of 0 mass %, average coefficient of thermal expansion αS from 25° C. to 800° C. (hereinafter simply referred to as average coefficient of thermal expansion αS) thereof was not measured, and a ratio of coefficient of thermal expansion αS of the sintered body to average coefficient of thermal expansion αGaN-a of the GaN crystal in the direction of a axis (hereinafter simply referred to as average coefficient of thermal expansion αGaN-a) (hereinafter denoted as the αS/αGaN-a ratio) was not calculated.
YSZ-mullite-based sintered body B1 had a YSZ content of 20 mass %, a content of Y2O3 (yttria) to YSZ (hereinafter referred to as a Y2O3 content) of 0 mol %, average coefficient of thermal expansion αS of 4.40×10−6/° C., and the αS/αGaN-a ratio of 0.753. YSZ-mullite-based sintered body B2 had a YSZ content of 20 mass %, a Y2O3 content of 3 mol %, average coefficient of thermal expansion αS of 4.58×10−6/° C., and the αS/αGaN-a ratio of 0.784. YSZ-mullite-based sintered body B3 had a YSZ content of 20 mass %, a Y2O3 content of 5 mol %, average coefficient of thermal expansion αS of 4.68×10−6/° C., and the αS/αGaN-a ratio of 0.801. YSZ-mullite-based sintered body B4 had a YSZ content of 20 mass %, a Y2O3 content of 6 mol %, average coefficient of thermal expansion αS of 4.69×10−6/° C., and the αS/αGaN-a ratio of 0.803. YSZ-mullite-based sintered body B5 had a YSZ content of 20 mass %, a Y2O3 content of 10 mol %, average coefficient of thermal expansion αS of 4.72×10−6/° C., and the αS/αGaN-a ratio of 0.808. YSZ-mullite-based sintered body B6 had a YSZ content of 20 mass %, a Y2O3 content of 20 mol %, average coefficient of thermal expansion αS of 4.81×10−6/° C., and the αS/αGaN-a ratio of 0.823. YSZ-mullite-based sintered body B7 had a YSZ content of 20 mass %, a Y2O3 content of 50 mol %, average coefficient of thermal expansion αS of 5.06×10−6/° C., and the αS/αGaN, ratio of 0.866. YSZ-mullite-based sintered body B8 had a YSZ content of 20 mass %, a Y2O3 content of 100 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN, ratio was not calculated.
YSZ-mullite-based sintered body C1 had a YSZ content of 25 mass %, a Y2O3 content of 0 mol %, average coefficient of thermal expansion αS of 4.48×10−6/° C., and the αS/αGaN, ratio of 0.767. YSZ-mullite-based sintered body C2 had a YSZ content of 25 mass %, a Y2O3 content of 3 mol %, average coefficient of thermal expansion αS of 4.62×10−6/° C., and the αS/αGaN, ratio of 0.791. YSZ-mullite-based sintered body C3 had a YSZ content of 25 mass %, a Y2O3 content of 5 mol %, average coefficient of thermal expansion αS of 5.26×10−6/° C., and the αS/αGaN, ratio of 0.901. YSZ-mullite-based sintered body C4 had a YSZ content of 25 mass %, a Y2O3 content of 6 mol %, average coefficient of thermal expansion αS of 5.27×10−6/° C., and the αS/αGaN-a ratio of 0.903. YSZ-mullite-based sintered body C5 had a YSZ content of 25 mass %, a Y2O3 content of 10 mol %, average coefficient of thermal expansion αS of 5.31×10−6/° C., and the αS/αGaN-a ratio of 0.909. YSZ-mullite-based sintered body C6 had a YSZ content of 25 mass %, a Y2O3 content of 20 mol %, average coefficient of thermal expansion αS of 5.40×10−6/° C., and the αS/αGaN-a ratio of 0.925. YSZ-mullite-based sintered body C7 had a YSZ content of 25 mass %, a Y2O3 content of 50 mol %, average coefficient of thermal expansion αS of 5.69×10−6/° C., and the αS/αGaN-a ratio of 0.974. YSZ-mullite-based sintered body C8 had a YSZ content of 25 mass %, a Y2O3 content of 100 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated.
YSZ-mullite-based sintered body D1 had a YSZ content of 30 mass %, a Y2O3 content of 0 mol %, average coefficient of thermal expansion αS of 4.56×10−6/° C., and the αS/αGaN, ratio of 0.781. YSZ-mullite-based sintered body D2 had a YSZ content of 30 mass %, a Y2O3 content of 3 mol %, average coefficient of thermal expansion αS of 4.65×10−6/° C., and the αS/αGaN, ratio of 0.796. YSZ-mullite-based sintered body D3 had a YSZ content of 30 mass %, a Y2O3 content of 5 mol %, average coefficient of thermal expansion αS of 5.55×10−6/° C., and the αS/αGaN-a ratio of 0.950. YSZ-mullite-based sintered body D4 had a YSZ content of 30 mass %, a Y2O3 content of 6 mol %, average coefficient of thermal expansion αS of 5.56×10−6/° C., and the αS/αGaN, ratio of 0.952. YSZ-mullite-based sintered body D5 had a YSZ content of 30 mass %, a Y2O3 content of 10 mol %, average coefficient of thermal expansion αS of 5.60×10−6/° C., and the αS/αGaN, ratio of 0.959. YSZ-mullite-based sintered body D6 had a YSZ content of 30 mass %, a Y2O3 content of 20 mol %, average coefficient of thermal expansion αS of 5.70×10−6/° C., and the αS/αGaN-a ratio of 0.976. YSZ-mullite-based sintered body D7 had a YSZ content of 30 mass %, a Y2O3 content of 50 mol %, average coefficient of thermal expansion αS of 6.00×10−6/° C., and the αS/αGaN-a ratio of 1.027. YSZ-mullite-based sintered body D8 had a YSZ content of 30 mass %, a Y2O3 content of 100 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated.
YSZ-mullite-based sintered body E1 had a YSZ content of 35 mass %, a Y2O3 content of 0 mol %, average coefficient of thermal expansion αS of 4.77×10−6/° C., and the αS/αGaN-a ratio of 0.816. YSZ-mullite-based sintered body E2 had a YSZ content of 35 mass %, a Y2O3 content of 3 mol %, average coefficient of thermal expansion αS of 4.86×10−6/° C., and the αS/αGaN-a ratio of 0.832. YSZ-mullite-based sintered body E3 had a YSZ content of 35 mass %, a Y2O3 content of 5 mol %, average coefficient of thermal expansion αS of 5.80×10−6/° C., and the αS/αGaN-a ratio of 0.993. YSZ-mullite-based sintered body E4 had a YSZ content of 35 mass %, a Y2O3 content of 6 mol %, average coefficient of thermal expansion αS of 5.81×10−6/° C., and the αS/αGaN-a ratio of 0.995. YSZ-mullite-based sintered body E5 had a YSZ content of 35 mass %, a Y2O3 content of 10 mol %, average coefficient of thermal expansion αS of 5.85×10−6/° C., and the αS/αGaN-a ratio of 1.002. YSZ-mullite-based sintered body E6 had a YSZ content of 35 mass %, a Y2O3 content of 20 mol %, average coefficient of thermal expansion αS of 5.96×10−6/° C., and the αS/αGaN-a ratio of 1.020. YSZ-mullite-based sintered body E7 had a YSZ content of 35 mass %, a Y2O3 content of 50 mol %, average coefficient of thermal expansion αS of 6.27×10−6/° C., and the αS/αGaN-a ratio of 1.074.
YSZ-mullite-based sintered body E8 had a YSZ content of 35 mass %, a Y2O3 content of 100 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated.
YSZ-mullite-based sintered body F2 had a YSZ content of 40 mass %, a Y2O3 content of 0 mol %, average coefficient of thermal expansion αS of 4.97×10−6/° C., and the αS/αGaN-a ratio of 0.851. YSZ-mullite-based sintered body F2 had a YSZ content of 40 mass %, a Y2O3 content of 3 mol %, average coefficient of thermal expansion αS of 5.07×10−6/° C., and the αS/αGaN, ratio of 0.868. YSZ-mullite-based sintered body F3 had a YSZ content of 40 mass %, a Y2O3 content of 5 mol %, average coefficient of thermal expansion αS of 6.05×10−6/° C., and the αS/αGaN-a ratio of 1.036. YSZ-mullite-based sintered body F4 had a YSZ content of 40 mass %, a Y2O3 content of 6 mol %, average coefficient of thermal expansion αS of 6.06×10−6/° C., and the αS/αGaN, ratio of 1.038. YSZ-mullite-based sintered body F5 had a YSZ content of 40 mass %, a Y2O3 content of 10 mol %, average coefficient of thermal expansion αS of 6.10×10−6/° C., and the αS/αGaN-a ratio of 1.045. YSZ-mullite-based sintered body F6 had a YSZ content of 40 mass %, a Y2O3 content of 20 mol %, average coefficient of thermal expansion αS of 6.21×10−6/° C., and the αS/αGaN-a ratio of 1.064. YSZ-mullite-based sintered body F7 had a YSZ content of 40 mass %, a Y2O3 content of 50 mol %, average coefficient of thermal expansion αS of 6.54×10−6/° C., and the αS/αGaN-a ratio of 1.120. YSZ-mullite-based sintered body F8 had a YSZ content of 40 mass %, a Y2O3 content of 100 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated.
YSZ-mullite-based sintered body G1 had a YSZ content of 70 mass %, a Y2O3 content of 0 mol %, average coefficient of thermal expansion αS of 4.99×10−6/° C., and the αS/αGaN-a ratio of 0.854. YSZ-mullite-based sintered body G2 had a YSZ content of 70 mass %, a Y2O3 content of 3 mol %, average coefficient of thermal expansion αS of 5.09×10−6/° C., and the αS/αGaN-a ratio of 0.872. YSZ-mullite-based sintered body G3 had a YSZ content of 70 mass %, a Y2O3 content of 5 mol %, average coefficient of thermal expansion αS of 6.07×10−6/° C., and the αS/αGaN-a ratio of 1.039. YSZ-mullite-based sintered body G4 had a YSZ content of 70 mass %, a Y2O3 content of 6 mol %, average coefficient of thermal expansion αS of 6.08×10−6/° C., and the αS/αGaN-a ratio of 1.041. YSZ-mullite-based sintered body G5 had a YSZ content of 70 mass %, a Y2O3 content of 10 mol %, average coefficient of thermal expansion αS of 6.12×10−6/° C., and the αS/αGaN-a ratio of 1.048. YSZ-mullite-based sintered body G6 had a YSZ content of 70 mass %, a Y2O3 content of 20 mol %, average coefficient of thermal expansion αS of 6.23×10−6/° C., and the αS/αGaN-a ratio of 1.067. YSZ-mullite-based sintered body G7 had a YSZ content of 70 mass %, a Y2O3 content of 50 mol %, average coefficient of thermal expansion αS of 6.56×10−6/° C., and the αS/αGaN-a ratio of 1.123.
YSZ-mullite-based sintered body G8 had a YSZ content of 70 mass %, a Y2O3 content of 100 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated.
YSZ-mullite-based sintered body H1 had a YSZ content of 100 mass %, a Y2O3 content of 0 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated. YSZ-mullite-based sintered body H2 had a YSZ content of 100 mass %, a Y2O3 content of 3 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated. YSZ-mullite-based sintered body H3 had a YSZ content of 100 mass %, a Y2O3 content of 5 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated. YSZ-mullite-based sintered body H4 had a YSZ content of 100 mass %, a Y2O3 content of 6 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated. YSZ-mullite-based sintered body H5 had a YSZ content of 100 mass %, a Y2O3 content of 10 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated. YSZ-mullite-based sintered body H6 had a YSZ content of 100 mass %, a Y2O3 content of 20 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated. YSZ-mullite-based sintered body H7 had a YSZ content of 100 mass %, a Y2O3 content of 50 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated. YSZ-mullite-based sintered body H8 had a YSZ content of 100 mass %, a Y2O3 content of 100 mol %, average coefficient of thermal expansion αS was not measured, and the αS/αGaN-a ratio was not calculated.
A support substrate having a diameter of 4 inches (101.6 mm) and a thickness of 1 mm was cut from each of the 57 types of the YSZ-mullite-based sintered bodies above, and opposing main surfaces of each support substrate were mirror-polished to thereby obtain 57 types of support substrates A0, B1 to B8, C1 to C8, D1 to D8, E1 to E8, F1 to F8, G1 to G8, and H1 to H8. Namely, a content of YSZ to the total of YSZ and mullite (a YSZ content), a content of Y2O3 (yttria) to YSZ (a Y2O3 content), and an average coefficient of thermal expansion from 25° C. to 800° C. of the 57 types of the support substrates above were equal to a YSZ content, a Y2O3 content, and an average coefficient of thermal expansion from 25° C. to 800° C. of the 57 types of the YSZ-mullite-based sintered bodies above, respectively. Tables 2 to 8 summarize these results. In Tables 2 to 8, “-” indicates that that physical property value was not measured or calculated.
(2) Sub Step of Forming Single Crystal Film on Underlying Substrate Referring to
A GaN film having a thickness of 0.4 μm was formed as single crystal film 13 on main surface 30n of the Si substrate (underlying substrate 30) above with the MOCVD method. Regarding film formation conditions, a TMG gas and an NH3 gas were used as source gases, an H2 gas was used as a carrier gas, a film formation temperature was set to 1000° C., and a film formation pressure was set to 1 atmosphere. Main surface 13m of the GaN film (single crystal film 13) thus obtained had a plane orientation having an off angle within ±1° with respect to the (0001) plane and it was a Ga atomic plane.
(3) Sub Step of Bonding Support Substrate and Single Crystal Film to Each Other
Referring to (C1) in
Referring also to (C2) in
Referring next to (C3) in
(4) Sub Step of Removing Underlying Substrate
Referring to
3. Step of Forming GaN-Based Film
Referring to
Regarding the 57 types of the wafers taken out of a film formation apparatus after cooling to a room temperature, warpage of the wafer and crack count and dislocation density of the GaN film were determined. Here, a shape of warpage and an amount of warpage of the wafer were determined based on interference fringes observed at the main surface of the GaN film with FM200EWafer of Corning Tropel. Regarding crack count in the GaN film, the number of cracks per unit length was counted with a Nomarski microscope, and evaluation as “extremely few”, “few”, “many”, and “extremely many” was made when the count was smaller than 1 count/mm, when the count was not smaller than 1 count/mm and smaller than 5 counts/mm, when the count was not smaller than 5 counts/mm and smaller than 10 counts/mm, and when the count was not smaller than 10 counts/mm, respectively. Dislocation density of the GaN film was measured with CL (cathode luminescence) based on the number of dark points per unit area. It is noted that cracks generated in the GaN film in the present Example were small without penetrating the film.
Wafer A0 was extremely many in cracks counted in the GaN film, and a shape of warpage, an amount of warpage, and dislocation density of the GaN film were not measured. Table 2 summarizes results.
Wafer B1 warped on the GaN film side in a recessed manner, an amount of warpage was 670 μm, cracks counted in the GaN film were many, and dislocation density of the GaN film was 5×108 cm−2. Wafer B2 warped on the GaN film side in a recessed manner, an amount of warpage was 660 μm, cracks counted in the GaN film were many, and dislocation density of the GaN film was 5×108 cm−2. Wafer B3 warped on the GaN film side in a recessed manner, an amount of warpage was 655 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer B4 warped on the GaN film side in a recessed manner, an amount of warpage was 650 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer B5 warped on the GaN film side in a recessed manner, an amount of warpage was 645 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer B6 warped on the GaN film side in a recessed manner, an amount of warpage was 610 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer B7 warped on the GaN film side in a recessed manner, an amount of warpage was 480 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer B8 was few in cracks counted in the GaN film, and a shape of warpage, an amount of warpage, and dislocation density of the GaN film were not measured. Table 2 summarizes results.
Wafer C1 warped on the GaN film side in a recessed manner, an amount of warpage was 665 μm, cracks counted in the GaN film were many, and dislocation density of the GaN film was 5×108 cm−2. Wafer C2 warped on the GaN film side in a recessed manner, an amount of warpage was 657 μm, cracks counted in the GaN film were many, and dislocation density of the GaN film was 5×108 cm−2. Wafer C3 warped on the GaN film side in a recessed manner, an amount of warpage was 390 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer C4 warped on the GaN film side in a recessed manner, an amount of warpage was 385 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer C5 warped on the GaN film side in a recessed manner, an amount of warpage was 380 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer C6 warped on the GaN film side in a recessed manner, an amount of warpage was 350 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer C7 warped on the GaN film side in a recessed manner, an amount of warpage was 180 μm, cracks counted in the GaN film were extremely few, and dislocation density of the GaN film was 1×108 cm−2. Wafer C8 was few in cracks counted in the GaN film, and a shape of warpage, an amount of warpage, and dislocation density of the GaN film were not measured. Table 3 summarizes results.
Wafer D1 warped on the GaN film side in a recessed manner, an amount of warpage was 660 μm, cracks counted in the GaN film were many, and dislocation density of the GaN film was 5×108 cm−2. Wafer D2 warped on the GaN film side in a recessed manner, an amount of warpage was 650 μm, cracks counted in the GaN film were many, and dislocation density of the GaN film was 5×108 cm−2. Wafer D3 warped on the GaN film side in a recessed manner, an amount of warpage was 250 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer D4 warped on the GaN film side in a recessed manner, an amount of warpage was 240 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer D5 warped on the GaN film side in a recessed manner, an amount of warpage was 230 μm, cracks counted in the GaN film were extremely few, and dislocation density of the GaN film was 1×108 cm−2. Wafer D6 warped on the GaN film side in a recessed manner, an amount of warpage was 180 μm, cracks counted in the GaN film were extremely few, and dislocation density of the GaN film was 1×108 cm−2. Wafer D7 warped on the GaN film side in a recessed manner, an amount of warpage was 10 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer D8 was few in cracks counted in the GaN film, and a shape of warpage, an amount of warpage, and dislocation density of the GaN film were not measured. Table 4 summarizes results.
Wafer E1 warped on the GaN film side in a recessed manner, an amount of warpage was 630 μm, cracks counted in the GaN film were many, and dislocation density of the GaN film was 5×108 cm−2. Wafer E2 warped on the GaN film side in a recessed manner, an amount of warpage was 520 μm, cracks counted in the GaN film were many, and dislocation density of the GaN film was 5×108 cm−2. Wafer E3 warped on the GaN film side in a recessed manner, an amount of warpage was 150 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm2. Wafer E4 warped on the GaN film side in a recessed manner, an amount of warpage was 120 μm, cracks counted in the GaN film were extremely few, and dislocation density of the GaN film was 1×108 cm2. Wafer E5 warped on the GaN film side in a recessed manner, an amount of warpage was 1 μm, cracks counted in the GaN film were extremely few, and dislocation density of the GaN film was 1×108 cm−2. Wafer E6 warped on the GaN film side in a projecting manner, an amount of warpage was 7 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer E7 warped on the GaN film side in a projecting manner, an amount of warpage was 12 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer E8 was few in cracks counted in the GaN film, and a shape of warpage, an amount of warpage, and dislocation density of the GaN film were not measured. Table 5 summarizes results. Wafer F1 warped on the GaN film side in a recessed manner, an amount of warpage was 500 μm, cracks counted in the GaN film were many, and dislocation density of the GaN film was 5×108 cm−2. Wafer F2 warped on the GaN film side in a recessed manner, an amount of warpage was 480 μm, cracks counted in the GaN film were many, and dislocation density of the GaN film was 5×108 cm2. Wafer F3 warped on the GaN film side in a projecting manner, an amount of warpage was 10 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer F4 warped on the GaN film side in a projecting manner, an amount of warpage was 10 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer F5 warped on the GaN film side in a projecting manner, an amount of warpage was 11 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer F6 warped on the GaN film side in a projecting manner, an amount of warpage was 12 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm−2. Wafer F7 warped on the GaN film side in a projecting manner, an amount of warpage was 110 μm, cracks counted in the GaN film were few, and dislocation density of the GaN film was 2×108 cm2. Wafer F8 was few in cracks counted in the GaN film, and a shape of warpage, an amount of warpage, and dislocation density of the GaN film were not measured. Table 6 summarizes results.
Wafer G1 warped on the GaN film side in a recessed manner, an amount of warpage was 510 μm, cracks counted in the GaN film were extremely many, and dislocation density of the GaN film was 5×108 cm−2. Wafer G2 warped on the GaN film side in a recessed manner, an amount of warpage was 490 μm, cracks counted in the GaN film were extremely many, and dislocation density of the GaN film was 5×108 cm−2. Wafer G3 warped on the GaN film side in a projecting manner, an amount of warpage was 10 μm, cracks counted in the GaN film were extremely many, and dislocation density of the GaN film was 2×108 cm−2. Wafer G4 warped on the GaN film side in a projecting manner, an amount of warpage was 11 μm, cracks counted in the GaN film were extremely many, and dislocation density of the GaN film was 2×108 cm−2. Wafer G5 warped on the GaN film side in a projecting manner, an amount of warpage was 11 μm, cracks counted in the GaN film were extremely many, and dislocation density of the GaN film was 2×108 cm−2. Wafer G6 warped on the GaN film side in a projecting manner, an amount of warpage was 12 μm, cracks counted in the GaN film were extremely many, and dislocation density of the GaN film was 2×108 cm−2. Wafer G7 warped on the GaN film side in a projecting manner, an amount of warpage was 110 μm, cracks counted in the GaN film were extremely many, and dislocation density of the GaN film was 2×108 cm−2. Wafer G8 was extremely many in cracks counted in the GaN film, and a shape of warpage, an amount of warpage, and dislocation density of the GaN film were not measured. Table 7 summarizes results.
All of wafers H1 to H8 were extremely many in cracks counted in the GaN film, and a shape of warpage, an amount of warpage, and dislocation density of the GaN film were not measured. Table 8 summarizes results.
Referring to Tables 2 to 8, by employing a composite substrate (wafers B3 to B7, C3 to C7, D3 to D7, E1 to E7, F1 to F7, and G1 to G7) having a support substrate in which coefficient of thermal expansion αS in a main surface was more than 0.8 time and less than 1.2 times (that is, 0.8<(αS/αGaN-a ratio)<1.2) as high as coefficient of thermal expansion αGaN-a of GaN crystal, a GaN film less in warpage, low in dislocation density, and excellent in crystallinity could be formed. In addition, from a point of view of further decrease in warpage and dislocation density of the GaN film, coefficient of thermal expansion αS in a main surface of the support substrate of the composite substrate was preferably more than 0.9 time and less than 1.15 times (that is, 0.9<(αS/αGaN-a ratio)<1.15) as high as coefficient of thermal expansion αGaN-a of the GaN crystal (wafers C3 to C7, D3 to D7, E3 to E7, F3 to F7, and G3 to G7) and further preferably more than 0.95 time and less than 1.1 times (that is, 0.95<(αS/αGaN-a ratio)<1.1) as high as coefficient of thermal expansion αGaN-a of the GaN crystal (wafers C7, D3 to D7, E3 to E7, F3 to F6, and G3 to G6).
In addition, Table 9 summarizes relation between cracks counted in a GaN film (GaN-based film 20) grown on a GaN single crystal layer (GaN-based single crystal layer 23) of each of the 57 types of the composite substrates and a YSZ content and a Y2O3 content of each of the 57 types of support substrate 11 of the 57 types of composite substrate 10 in Table 2 to Table 8.
Referring to Table 9, when a content of YSZ to the total of mullite (an Al2O3—SiO2 composite oxide) and YSZ (yttria stabilized zirconia) contained in the support substrate of the composite substrate was not lower than 20 mass % and not higher than 40 mass % and more preferably not lower than 25 mass % and not higher than 35 mass %, cracks counted in the GaN film (GaN-based film) formed on the GaN film (single crystal film) of the composite substrate significantly decreased. In addition, when a content of Y2O3 (yttria) to YSZ was not lower than 5 mol % and more preferably not lower than 6 mol % and not higher than 50 mol %, cracks counted in the GaN film (GaN-based film) formed on the GaN film (single crystal film) of the composite substrate extremely significantly decreased.
Though a case where a non-doped GaN film was formed on the composite substrate was shown in the example above, substantially the same results as in the example above were obtained also in a case where a GaN film provided with n- or p-type conductivity by doping was formed and in a case where a GaN film of which resistivity was raised by doping was formed.
Further, in a case of forming a GaN-based film such as a GaxInyAl1-x-yN film (x>0, y≧0, x+y≦1) instead of a GaN film as well, results as in the example above were obtained. In particular, in a case of forming a GaxInyAl1-x-yN film (x>0.5, y≧0, x+y≦1) instead of a GaN film, substantially the same results as in the example above were obtained.
Furthermore, a plurality of GaN-based films (specifically, GaxInyAl1-x-yN films (x>0, y≧0, x+y≦1) and the like)) can also be formed by varying a composition ratio of such a group III element as Ga, In and Al. Namely, a plurality of GaN-based films such as GaxInyAl1-x-yN films (x>0, y≧0, x+y≦1) and the like instead of a GaN film can be formed by varying a composition ratio of such a group III element as Ga, In and Al.
In carrying out the present invention, a known dislocation lowering technique such as an ELO (Epitaxial Lateral Overgrowth) technique is applicable in forming a GaN-based film.
In addition, after the GaN-based film is formed on the composite substrate, only the support substrate of the composite substrate or the entire composite substrate (the support substrate and the single crystal film) may be etched away. Here, the GaN-based film may be transferred to another support substrate.
It should be understood that the embodiments and the examples disclosed herein are illustrative and non-restrictive in every respect. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.
10 composite substrate; 11 support substrate; 11m, 12m, 12am, 12bn, 13m, 13n, 21m, 23m, 30n main surface; 12, 12a, 12b adhesive layer; 13 single crystal film; 20 GaN-based film; 21 GaN-based buffer layer; 23 GaN-based single crystal layer; 30 underlying substrate; and 40 wax.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2010-254529 | Nov 2010 | JP | national |
| 2010-255840 | Nov 2010 | JP | national |
| 2011-230682 | Oct 2011 | JP | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/JP2011/075961 | 11/10/2011 | WO | 00 | 10/24/2012 |
