JP4035136B2 - Method for fixing seed crystal and method for producing single crystal using the fixing method - Google Patents

Description

  The present invention relates to a fixing method for fixing a seed crystal for growing a single crystal such as silicon carbide to a graphite pedestal and a method for producing a single crystal using them.

  In recent years, a silicon carbide single crystal substrate has been developed as a semiconductor substrate of a semiconductor device for high withstand voltage and high power such as a high withstand voltage power transistor and a high withstand voltage diode. And as a manufacturing method of this silicon carbide single crystal substrate, a sublimation recrystallization method (an improved Rayleigh method) is mainly employed. FIG. 5 is a schematic view of an apparatus used for this sublimation recrystallization method. SiC powder 16 is contained as a raw material powder in the lower half of a graphite crucible composed of a container body 14 and a lid body 15 provided with a base. There is a seed crystal 17 on the lower surface of the lid 15 facing this. The inside of the crucible is held so that the SiC powder 16 side is at a high temperature and the seed crystal 17 side is at a low temperature, and the sublimation gas of the SiC powder 16 is recrystallized on the low temperature seed crystal 17 so that a single crystal 18 grows.

  In the above method, the seed crystal 17 is usually attached to a pedestal provided on the lid 15 using an adhesive. However, in this case, as shown in FIG. 6A, bubbles are likely to be generated in the adhesive layer 19 due to the heat treatment process for drying and curing the adhesive, and the voids 20 remain in the adhesive layer 19. If a single crystal is grown in the presence of this void 20, heat is conducted to the lid 15 at a location where there is no void 20 and is in close contact with the adhesive layer 19, so the temperature gradient between the seed crystal 17 and the lid 15 is In addition, since the seed crystal 17 does not release heat to the lid 15 in a place where the void 20 exists, a temperature gradient is locally generated between the seed crystal 17 and the lid 15.

  As a result, as shown in FIG. 6B, sublimation of the seed crystal 17 occurs from the back surface of the seed crystal 17 toward the lid 15 having a low temperature in the gap portion 20. Such back surface sublimation occurring on the back surface of the seed crystal 17 occurs at a plurality of locations on the bonding surface of the seed crystal 17 and the lid 15 and continuously occurs during single crystal growth. As shown in FIG. This causes a large defect (macro defect) 21 propagating in the growth direction from the interface between the base and the base. The presence of the macro defect 21 makes it difficult to cut out a large number of practical SiC wafers even when a long SiC single crystal is obtained. May induce flaws. For this reason, there is a problem that it is difficult to obtain a high-quality SiC single crystal in a large area.

In order to solve this problem, for example, a method of making the in-plane temperature distribution of the seed crystal uniform by interposing a carbonized layer between the seed crystal and the pedestal of the lid and bonding them together has been proposed ( Patent Document 1). This configuration is not shown because the adhesive layer 19 in FIGS. 6A-C is simply replaced by the carbonized layer 22. In addition, a surface other than the surface on which the single crystal of the seed crystal grows is coated with a protective layer made of a material (for example, tantalum) that is stable under the single crystal growth conditions, and then bonded to the base of the lid with an adhesive. Has proposed a method for suppressing macro defects (see Patent Document 2).
Japanese Patent Laid-Open No. 9-110584 Japanese Patent Laid-Open No. 9-268096

  However, in the method of Patent Document 1, a carbonized layer is formed between the seed crystal 17 and the base of the lid 15 by interposing a liquid adhesive containing a polymer material (for example, a resist) and then performing heat treatment at a high temperature. Since the seed crystal 17 is fixed by forming 22, voids 20 are easily generated due to bubbles generated by the heat treatment process, and the macro defects 21 are completely suppressed as shown in FIG. 6C. I can't.

  Also in the method of Patent Document 2, as shown in FIGS. 7A to 7C, local temperature distribution occurs due to the non-uniformity of the adhesive layer 19, and as a result, the crack 24 enters the protective layer 23, and the crack 24, there is a problem that the back surface sublimation occurs between the seed crystal 17 and the pedestal pasting surface of the lid 15 through 24, and the macro defect 21 occurs.

  The present invention solves the above-mentioned conventional problem, prevents back surface sublimation occurring between the back surface of the seed crystal and the pedestal, and more reliably suppresses the occurrence of macro defects extending in the grown crystal. A method for fixing a seed crystal for growing a single crystal and a method for producing a single crystal using the seed crystal obtained thereby are provided.

The seed crystal fixing method of the present invention is a seed crystal fixing method for growing a silicon carbide single crystal from the seed crystal by fixing the seed crystal to a graphite pedestal.
A pressure member for disposing a metal material having a melting point lower than the growth temperature of the single crystal on the graphite base, disposing the seed crystal on the metal material, and further applying a load on the seed crystal. To form a laminated body,
A container body that covers the laminate is disposed on the graphite pedestal, and while adjusting the difference between the lower temperature of the graphite pedestal and the upper temperature of the container body, the lower temperature of the graphite pedestal is equal to or higher than the melting point of the metal material. And the heat treatment is performed so that the temperature is not higher than the growth temperature of the single crystal, and the graphite pedestal, the metal material, and the seed crystal are fixed and integrated,
Cooling the container and the laminate,
The container body and the pressure member are removed from the laminate.

  In the method for producing a single crystal of the present invention, when the sublimation gas from the silicon carbide raw material is supplied onto the seed crystal held on the graphite pedestal and the silicon carbide single crystal is grown on the seed crystal, the seed crystal is A seed crystal fixed by the fixing method is used.

  According to the seed crystal fixing method and the single crystal manufacturing method using the fixing method of the present invention, an adhesive layer made of a metal carbide is formed between the seed crystal and the graphite pedestal for holding the seed crystal. As a result, no voids are formed between the seed crystal and the graphite pedestal at high temperatures during single crystal growth, so that the generation of macro defects inside the crystal extending from the back surface of the seed crystal can be suppressed, and high-quality carbonization can be achieved. A silicon single crystal can be obtained.

  In the present invention, the metal material is preferably at least one material selected from titanium, vanadium and zirconium. These metal materials make it possible to form a metal carbide layer that closely adheres the seed crystal without voids and the graphite pedestal, prevent sublimation from the back surface of the seed crystal, and prevent macro defects extending into the growing single crystal. Generation can be suppressed.

  The thickness of the metal material is preferably in the range of 20 μm to 200 μm. If the thickness is within this range, sufficient adhesive strength can be obtained, and wraparound of the molten metal to the surface of the seed crystal can be prevented. Metal contamination in a single crystal to be grown can be suppressed.

  The heating temperature in the heating step when titanium is used as the metal material is preferably in the range of 1700 ° C. or more and 2000 ° C. or less. If it is this temperature range, titanium will fuse | melt and a graphite base and a seed crystal can be fixedly integrated.

  The heating temperature in the heating step when vanadium is used for the metal material is preferably in the range of 1900 ° C. or higher and 2200 ° C. or lower. If it is this temperature range, vanadium will fuse | melt and a graphite base and a seed crystal can be fixed-integrated.

  The heating temperature in the heating step when zirconium is used as the metal material is preferably in the range of 1900 ° C. or higher and 2200 ° C. or lower. Within this temperature range, the zirconium melts, and the graphite pedestal and the seed crystal can be fixed and integrated.

  The heating time in the heating step is preferably in the range of 1 hour to 6 hours. If it is the said reaction time, the graphite and metal material of a graphite base surface will react, and the metal carbide layer which has melting | fusing point more than the growth temperature of a single crystal can be formed efficiently.

  The pressurizing member preferably has a weight capable of applying a pressure of 7.84 kPa to 87.5 kPa to the seed crystal. The graphite on the surface of the graphite pedestal and the metal material are brought into close contact with each other and reacted uniformly to form a metal carbide layer efficiently.

  The pressure member is preferably made of at least one material selected from graphite, tantalum, niobium, molybdenum, tantalum carbide, niobium carbide, and molybdenum carbide. This is because the pressure member does not melt at the heating temperature and does not contaminate the surface of the seed crystal.

  The cooling step is preferably performed at a rate of 5 to 15 ° C./min while the temperature falls to 1100 ° C. from the heating temperature in the heating step, and then left at room temperature until it reaches room temperature in the apparatus. If it is gradually cooled as described above, it is possible to prevent the metal carbide layer from cracking due to the stress caused by the difference in thermal expansion among the graphite pedestal, the metal carbide layer, and the seed crystal. Here, room temperature refers to a range of 10 ° C. or higher and 30 ° C. or lower.

  In the present invention, when a container is used, the area of the seed crystal is larger than the area of the portion of the graphite pedestal where the seed crystal is fixed, and the pressure member may contact only the periphery of the seed crystal and apply a load. A shape that can be formed is preferable. For example, when using a graphite pedestal having a diameter of 10 mm for fixing the seed crystal and a seed crystal having a diameter of 10.5 to 12 mm, the area ratio is 1.1 to 1.44 times the seed crystal area. Moreover, when the part which fixes the seed crystal of a graphite base is 16 mm in diameter, about 16.5-18 mm in diameter is preferable for a seed crystal, and area ratio becomes 1.06-1.27 times. In other words, it is preferable that the size of the seed crystal protrudes about 0.5 to 1 mm from the portion where the seed crystal is fixed.

  In the above description, the shape in which the pressure member can apply a load by contacting only the peripheral portion of the seed crystal is, for example, a hollow cylindrical shape.

  The peripheral portion of the seed crystal is an area of the seed crystal that is larger than the area of the portion that fixes the seed crystal of the graphite pedestal, and at least a part of the seed crystal region that protrudes from the area of the portion that fixes the seed crystal of the graphite pedestal. Preferably there is.

  The protruding member is in contact with the protruding portion, and the surface of the seed crystal of this portion is roughened. However, in the single crystal growth atmosphere, the roughened region of the protruding surface becomes high temperature and is etched away. Accordingly, the region where the surface roughness has occurred is eliminated, and only a good surface which does not cause the surface roughness remains. This smooth surface that is not roughened is effective for producing a single crystal of good quality.

  Embodiments of a seed crystal fixing method for growing a silicon carbide single crystal of the present invention and a method of manufacturing a single crystal using them will be described below in detail with reference to the drawings.

Example 1
A specific method corresponding to the seed crystal fixing method of the present invention will be described with reference to FIGS.

  In this example, a Rayleigh seed crystal having a thickness of about 0.3 to 0.5 mm manufactured by the Rayleigh method is used as a seed crystal. A sublimation recrystallization method (an improved Rayleigh method) is applied to the Rayleigh seed crystal. A substrate obtained from an ingot grown by using a single crystal produced by the Atchison method or the like can also be used as a seed crystal.

  First, as shown in FIG. 1A, a metal material 2 made of titanium having a thickness of 50 μm, a Rayleigh seed crystal 3 and a pressure member 4 for applying a load made of graphite are arranged in this order on the graphite pedestal 1. A laminated body 25 was formed. As a load, 33.5 kPa was applied to the seed crystal.

  Next, as shown in FIG. 1B, a container body 5 made of graphite is covered with a laminated body 25, and further covered with a heat insulating material 6 having a hole in the upper and lower sides, and a water-cooled duplex capable of adjusting the internal pressure. It was installed in a quartz container 7. An RF coil 8 for heating a graphite container 5 around the double quartz container 7 and a pyrometer for monitoring the lower temperature of the graphite pedestal 1 and the upper temperature of the graphite crucible container 5 at the upper and lower parts of the double quartz container 7. 9 and 10 are provided. Actually, there are a fixing jig for holding the laminated body, a vacuum pump for evacuating the double quartz container 7, a pressure adjusting valve for adjusting the atmosphere in the double quartz container 7, and the like are omitted.

Next, the double quartz container was evacuated to 5 × 10 ⁻⁴ Pa or less, and argon gas was introduced to adjust the pressure to 79.8 kPa (600 Torr). In this state, the lower temperature of the graphite pedestal 1 was heated to 1700 ° C. as measured by the pyrometer 10 and held for about 3 hours. Thereafter, the temperature was lowered to 1100 ° C. at a rate of 10 ° C./min, and then naturally cooled to around room temperature (25 ° C.) over 4 hours. The laminated body 25 was taken out from the double quartz container 7 and the pressure member 4 for applying a load made of graphite was removed.

  Thus, the Rayleigh seed crystal 3 shown in FIG. 1C was fixed and integrated on the graphite pedestal 1 by the metal carbide layer 11 made of titanium carbide. Although argon gas was used as the heating atmosphere, it may be any material as long as it does not react with pedestal graphite, seed crystal silicon carbide, or the metal material used, and helium or a mixed gas of argon and helium may be used.

  The metal carbide layer 11 made of titanium carbide formed by such a fixing method becomes an adhesive layer having a melting point of about 3000 ° C. or more for closely adhering the Rayleigh seed crystal 3 and the graphite pedestal 1 without voids. At a temperature of about 2300 ° C. at which the crystal is grown, no temperature gradient is generated between the Rayleigh seed crystal 3 and the graphite pedestal 1, sublimation from the back surface of the seed crystal 3 can be prevented, and the occurrence of macro defects can be suppressed. . Titanium carbide has a melting point that varies depending on the carbon content, but has a melting point of 2500 ° C. or more when the carbon content is 18 atom% or more. Therefore, the seed crystal can be held in a solid phase even at the silicon carbide single crystal growth temperature. Is possible. Actually, the SEM (scanning electron microscope) observation and EDX (Energy Dispersive X-ray Fluorescence Spectrometer) analysis of the seed crystal and the graphite pedestal interface showed that there was a void of 45atom% titanium and 55atom% carbon. A titanium carbide layer was not formed.

  By the way, although titanium was used as a specific example of the metal material 2, even if vanadium or zirconium is used, it becomes possible to form the metal carbide layer 11 that makes the seed crystal 3 and the graphite pedestal 1 in close contact with each other, and the back surface of the seed crystal. Sublimation can be prevented and the occurrence of macro defects extending into the grown single crystal can be suppressed. Vanadium has a melting point of about 1890 ° C., a vanadium carbide having a carbon content of 35 atom% and a melting point of 2500 ° C. or more, and zirconium has a melting point of 1850 ° C. and a carbon content of 15 atom% and a zirconium carbide having a melting point of 2500 ° C. or more.

  In the same way as when using titanium, SEM observation and EDX analysis of the seed crystal of the seed crystal fixed and integrated with vanadium and zirconium and the graphite pedestal interface and EDX analysis are performed. When vanadium is used, vanadium When a vanadium carbide layer having a melting point of about 2700 ° C. without voids of 51 atom% and carbon 49 atom% is used with zirconium, a zirconium carbide layer with a melting point of about 2900 ° C. without voids of 40 atom percent of carbon and 60 atoms of carbon is formed. It had been. However, considering the melting point, the heating temperature was 1900 ° C. for vanadium, 1900 ° C. for zirconium, and the other conditions were the same as those used for titanium.

  Moreover, although the thickness of the metal material 2 made of titanium is 50 μm, sufficient adhesive strength can be obtained by using a material between 20 and 200 μm. In addition, it is possible to prevent the molten metal from wrapping around the surface of the seed crystal, and it is possible to suppress the occurrence of micropipe defects centered on the metal adhering to the surface of the seed crystal and the metal contamination in the growing single crystal. Table 1 shows the results when heat treatment is performed at a heating temperature of 1700 ° C. for 3 hours in an argon gas atmosphere with a load of 33.5 kPa and a pressure of 79.8 kPa (600 Torr) while changing the thickness of titanium.

  The evaluation items were determined by the adhesive strength and the presence or absence of metal wrapping around the crystal surface. That is, a circle having no wraparound was marked with ◯, and a circle with wraparound was marked with ×.

Regarding the adhesive strength, those that can withstand a load greater than 100 g / cm ² were rated as “◯”, and those that could only withstand a load of 100 g / cm ² or less were marked as “X”. Since this is a value related to the size of the crystal that can be produced by the sublimation method, a larger value is better, but a value larger than necessary is unnecessary. For example, in the case of producing a single crystal ingot having a length of about 100 mm using a 2 inch diameter seed crystal, it is sufficient that it can withstand a load of at least 32 g / cm ^2. Inadequate. Considering practicality, the criterion for determining the adhesive strength was whether or not the load could withstand a load of 100 g / cm ² .

  The above evaluation criteria are the same in Table 2 and below.

  As is clear from the results in Table 1, the thinner the metal material, the lower the adhesive strength. When the thickness was 20 μm, the adhesive strength was sufficient, but when the thickness was 5 μm, the adhesive strength decreased, and the Rayleigh seed crystal was not adhered to the graphite pedestal. On the other hand, as the thickness of the metal material increases, the metal material wraps around the seed crystal. When the metal material had a thickness of 200 μm, the adhesive strength was sufficient, but a titanium melt mark in which titanium slightly turned around the side surface of the seed crystal remained. However, since there is no wraparound of titanium on the Rayleigh seed crystal surface, this is considered to be the limit of wraparound. Further, when the thickness was increased to 250 μm, the trace of melting of titanium could be confirmed in the peripheral portion of the surface of the seed crystal and the pressure member made of graphite. For these reasons, the thickness of titanium is preferably 20 to 200 μm.

  Similarly, Table 2 shows the results when heat treatment was performed at a heating temperature of 1900 ° C. for 3 hours in an argon gas atmosphere with a load of 33.5 kPa and a pressure of 79.8 kPa (600 Torr) while changing the thickness of vanadium and zirconium. 3 shows. Evaluation items are the same as in Table 1.

  In the case of vanadium and zirconium, the result was the same as when titanium was used, and the result that 20 to 200 μm was optimum was obtained. From these things, when vanadium or zirconium is used, the thickness is preferably 20 to 200 μm.

  From the above, the thickness of the metal material 2 is preferably 20 to 200 μm in the above three types of materials.

  Next, specific conditions for the heating temperature will be described. In this example, the heating temperature of the heating process was 1700 ° C. when titanium was used as the metal material, 1900 ° C. when vanadium was used, and 1900 ° C. when zirconium was used. Indicates the heating temperature range. Within this appropriate heating temperature range, the seed crystal 3 and the graphite pedestal 1 can be firmly bonded to the metal carbide layer 11 and macro defects can be suppressed.

  Table 3 shows the results of heat treatment for 3 hours in an argon gas atmosphere with a load of 33.5 kPa and a pressure of 79.8 kPa (600 Torr) using titanium, vanadium, and zirconium with a thickness of 50 μm as the metal material while changing the heating temperature. Shown in 4-6.

As evaluation items, judgment was made based on adhesive strength and macro defect density. Macro defects were evaluated by growing a single crystal of about 8 mm under the same growth condition on a seed crystal fixed under predetermined conditions, slicing the crystal parallel to the growth direction, polishing and observing a cross section. By utilizing the transmission mode of the optical microscope, the number of macro defects is counted while shifting the focus in the thickness direction of the cross-sectional observation sample, and then divided by the width of the seed crystal and the thickness of the cross-section observation sample. Density was calculated. Macro defect density 0 to 100 / cm ² to ○, a 100~500 / cm ² △, and as × 500 / cm ² or more. This evaluation criterion is the same in Table 4 and below.

  The single crystal growth procedure for evaluating macro defects is as follows. As shown in FIG. 2, SiC powder 12 is housed as a raw material powder in the lower half of a graphite crucible made of graphite pedestal 1 in which seed crystal 3 is fixed by container body 5 made of graphite and metal carbide layer 11, This is covered with a heat insulating material 6 with a hole in the top and bottom, and installed in a water-cooled double quartz container 7 capable of adjusting the internal pressure. An RF coil 8 for heating the graphite crucible 7 is provided around the double quartz container 7, and pyrometers 9, 10 for monitoring the temperature of the graphite pedestal 1 and the container body 5 made of graphite are provided at the upper and lower parts of the double quartz container 7. Yes. Actually, there are a fixing jig for holding the crucible, a vacuum pump for evacuating the double quartz container 7, a pressure adjusting valve for adjusting the atmosphere in the double quartz container 7, and the like are omitted. Subsequently, the SiC powder 12 side is set to a high temperature, the seed crystal 3 side is set to a low temperature, and a sublimation gas is recrystallized on the seed crystal to grow a silicon carbide single crystal 13. As the growth conditions of the silicon carbide single crystal for observing the macro defect, the SiC powder side temperature is 2350 ° C., the seed crystal side temperature is 2200 ° C., and the growth time is 30 hours in an argon gas atmosphere at a pressure of 3.99 kPa (30 Torr). did. The growth amount is about 8 mm.

  Table 4 shows the relationship between the heating temperature, adhesive strength, and macro defects when titanium is used as the metal material.

  As apparent from Table 4, when the heating temperature is 1500 ° C. to 1600 ° C., the adhesive strength is sufficient, but the macro defect density tends to increase as the heating temperature decreases. This is because the heating temperature is lower than about 1675 ° C., which is the melting point of titanium, and titanium does not completely melt and the reaction with the graphite of the pedestal does not proceed, so the carbon content is lower than 18 atom% and the melting point is This is probably because a titanium carbide layer lower than 2500 ° C. is formed. On the other hand, when the heating temperature was in the range of 1700 to 2000 ° C., good results were obtained for both adhesive strength and macro defect density. This is presumably because the titanium was completely melted and the reaction with the graphite of the pedestal proceeded sufficiently, so that a titanium carbide layer having a carbon content higher than 18 atom% and a melting point higher than 2500 ° C. was formed. When the heating temperature is further increased to 2050 ° C., the seed crystal is not bonded to the pedestal. This is presumably because, due to the high heating temperature, the molten titanium and pedestal graphite reacted excessively, and the titanium carbide layer became carbon-rich, resulting in a porous layer structure. Therefore, when titanium is used as the metal material, the heating temperature is desirably 1700 to 2000 ° C.

  Next, Table 5 shows the relationship between the heating temperature, adhesive strength, and macro defects when vanadium is used as the metal material.

  As is apparent from Table 5, the seed crystal was not bonded to the graphite pedestal at the heating temperature of 1700 ° C. At a heating temperature of 1800 ° C., macro defects were confirmed although the adhesive strength was sufficient. This is because the heating temperature is lower than the melting point of vanadium of about 1890 ° C., so the vanadium is not completely melted and the reaction with the graphite of the pedestal does not proceed, so the carbon content is lower than 35 atom% and the melting point is 2500 ° C. This is probably because a lower vanadium carbide layer is formed. When the heating temperature was in the range of 1900 to 2200 ° C., good results were obtained for both adhesive strength and macro defect density. This is probably because vanadium was completely melted and the reaction with the base graphite sufficiently progressed, so that a vanadium carbide layer having a carbon content higher than 35 atom% and a melting point higher than 2500 ° C. was formed. . At the heating temperature of 2300 ° C., the seed crystal was not adhered to the pedestal. As with the case where titanium is used as the metal material, the heating temperature is high, so the molten vanadium and the base graphite react excessively, and the vanadium carbide layer becomes carbon-rich and has a porous layer structure. It is presumed that. Therefore, the heating temperature when vanadium is used as the metal material is desirably 1900 to 2200 ° C.

  Table 6 shows the relationship between the heating temperature, adhesion strength, and macro defects when zirconium is used as the metal material.

  As apparent from Table 6, the seed crystal was not adhered to the pedestal at the heating temperature of 1700 ° C. At a heating temperature of 1800 ° C., macro defects were confirmed although the adhesive strength was sufficient. This is because the heating temperature is lower than the melting point of about 1850 ° C. of zirconium, the zirconium is not completely melted and the reaction with the graphite of the pedestal does not proceed, so the carbon content is lower than 15 atom% and the melting point is 2500 ° C. This is probably because a lower zirconium carbide layer is formed. When the heating temperature was in the range of 1900 to 2200 ° C., good results were obtained for both adhesive strength and macro defect density. This seems to be because the zirconium was completely melted and the reaction with the graphite of the pedestal proceeded sufficiently, so that a zirconium carbide layer having a carbon content higher than 15 atom% and a melting point higher than 2500 ° C. was formed. It is. At the heating temperature of 2300 ° C., the seed crystal was not adhered to the pedestal. In this regard, as with the case of using titanium and vanadium as the metal material, the heating temperature is high, so the molten zirconium and the pedestal graphite react excessively, and the zirconium carbide layer becomes carbon-rich, resulting in a porous layer structure. Presumed to be because. Therefore, the heating temperature when zirconium is used as the metal material is desirably 1900 to 2200 ° C.

  Next, specific conditions for the heating time will be described. In this example, as a specific example of the heating time of the heating step, it was set to 3 hours when titanium, vanadium and zirconium were used as the metal material 2, but in the range of 1 to 6 hours in all metal materials, The Rayleigh seed crystal 3 and the graphite pedestal 1 can be firmly bonded by the metal carbide layer 11 and macro defects can be suppressed. Results of adhesion experiment when heat treatment was performed in an argon gas atmosphere with a load of 33.5 kPa and a pressure of 79.8 kPa (600 Torr) using titanium, vanadium and zirconium with a thickness of 50 μm as the metal material while changing the heating time Are shown in Tables 7-9. From the above results, the heating temperature was in the range of 1700 ° C. to 2000 ° C. when titanium was used as the metal material, the range of 1900 ° C. to 2200 ° C. was used when vanadium was used as the metal material, and zirconium was used as the metal material. In this case, the temperature range was 1900 ° C to 2200 ° C. Evaluation items were adhesive strength and macro defect density.

  Table 7 shows the relationship between the heating time, adhesive strength, and macro defects when titanium is used as the metal material.

  As apparent from Table 7, the seed crystal was not adhered to the graphite pedestal at the heating temperatures of 1700 ° C., 1800 ° C., and 1900 ° C. when the heating time was 0.5 hours. This is considered to be because the heating time of 0.5 hours is not sufficient for the titanium to melt in the heating temperature range of 1700-1900 ° C. At 2000 ° C., macro defects were confirmed although the adhesive strength was sufficient. This is because when the heating temperature is 2000 ° C. and the heating time is 0.5 hour, the titanium melts and reacts with the graphite pedestal, but because the time is short, the reaction time is short, the carbon content is lower than 18 atom%, and the melting point is 2500 ° C. This is probably because a lower titanium carbide layer is formed. In the heating time of 1 to 6 hours, good results were obtained for both the adhesive strength and the macro defect density in the heating temperature range of 1700 to 2000 ° C. This is because, under this condition range, titanium was completely melted and sufficiently reacted with the graphite of the pedestal, so that a titanium carbide layer having a carbon content higher than 18 atom% and a melting point higher than 2500 ° C. was formed. It appears to be. When the heating time was 8 hours or more, there was a tendency for the adhesive strength to decrease and the macro defect density to increase at all heating temperatures. In this regard, it is presumed that because the heating time is long, the molten titanium reacts excessively with the graphite of the pedestal, the titanium carbide layer becomes carbon-rich and has a porous layer structure. Therefore, it is desirable that the heating time when titanium is used be in the range of 1 to 6 hours.

  Table 8 shows the relationship between the heating time, adhesive strength, and macro defects when vanadium is used as the metal material.

  As shown in Table 8, the seed crystal was not adhered to the graphite pedestal at the heating temperatures of 1900 ° C. and 2000 ° C. when the heating time was 0.5 hours. This is considered to be caused by the fact that in the range of the heating temperature of 1900 to 2000 ° C., the heating time of 0.5 hour is not sufficient time for the vanadium to melt. At 2200 ° C., macro defects were confirmed although the adhesive strength was sufficient. This is because when the heating temperature is 2200 ° C. and the heating time is 0.5 hour, the vanadium melts and reacts with the graphite pedestal, but because the time is short, the reaction time is short, the carbon content is lower than 35 atom%, and the melting point is 2500 ° C. This is probably because a lower vanadium carbide layer is formed. In the heating time of 1 to 6 hours, good results were obtained for both the adhesive strength and the macro defect density in the heating temperature range of 1900 to 2200 ° C. This is because in this condition range, vanadium was completely melted and sufficiently reacted with the graphite of the pedestal, so that a vanadium carbide layer having a carbon content higher than 35 atom% and a melting point higher than 2500 ° C. was formed. It seems to be. When the heating time was 8 hours or more, there was a tendency for the adhesive strength to decrease and the macro defect density to increase at all heating temperatures. Regarding this, it is presumed that because the heating time is long, the melted vanadium reacts excessively with the graphite of the pedestal, the vanadium carbide layer becomes carbon-rich and has a porous layer structure. Therefore, the heating time when vanadium is used is desirably in the range of 1 to 6 hours.

  Table 9 shows the relationship between the heating time, adhesion strength, and macro defects when zirconium is used as the metal material.

  As is apparent from Table 9, the seed crystal was not adhered to the graphite pedestal at the heating temperatures of 1900 ° C. and 2000 ° C. when the heating time was 0.5 hours. This is considered to be because the heating time of 0.5 hours is not sufficient for the zirconium to melt in the heating temperature range of 1900 to 2000 ° C. At 2200 ° C., macro defects were confirmed although the adhesive strength was sufficient. This is because when the heating temperature is 2200 ° C. and the heating time is 0.5 hour, the zirconium melts and reacts with the graphite pedestal, but because the time is short, the reaction time is short, the carbon content is lower than 15 atom%, and the melting point is 2500 ° C. This is probably because a lower zirconium carbide layer is formed. In the heating time of 1 to 6 hours, good results were obtained for both the adhesive strength and the macro defect density in the heating temperature range of 1900 to 2200 ° C. This is because, in this condition range, the zirconium was completely melted and reacted sufficiently with the graphite of the pedestal, so that a zirconium carbide layer having a carbon content higher than 15 atom% and a melting point higher than 2500 ° C. was formed. It appears to be. When the heating time was 8 hours or more, there was a tendency for the adhesive strength to decrease and the macro defect density to increase at all heating temperatures. In this regard, it is assumed that because the heating time is long, the molten zirconium reacts excessively with the pedestal graphite, and the zirconium carbide layer becomes carbon-rich and has a porous layer structure. Therefore, it is desirable that the heating time when zirconium is used be in the range of 1 to 6 hours.

  Next, pressurizing conditions at the time of bonding will be described. In this embodiment, the weight of the pressure member 4 for applying a load at the time of adhesion is 33.5 kPa for a graphite member, but it is 7.84 when any metal material such as titanium, vanadium and zirconium is used. Within the range of ˜87.5 kPa, the seed crystal 3 and the graphite pedestal 1 can be firmly bonded to the metal carbide layer 11 and macro defects can be suppressed. Tables 10 to 12 show the results when heat treatment was performed for 3 hours in an argon gas atmosphere at a pressure of 79.8 kPa (600 Torr) using titanium, vanadium, and zirconium having a thickness of 50 μm as the metal material 2 while changing the pressure. Shown in The heating temperature ranges from 1700 ° C. to 2000 ° C. when titanium is used as the metal material, 1900 ° C. to 2200 ° C. when vanadium is used as the metal material, and 1900 ° C. when zirconium is used as the metal material. The range was 2200 ° C. As evaluation items, judgment was made based on adhesive strength and macro defect density.

  Table 10 shows the relationship between pressure, adhesion strength, and macro defects when titanium is used as the metal material.

  As is clear from Table 10, in all temperature ranges, a macro defect could be confirmed although the adhesive strength was sufficient at a pressure of 4.23 kPa, though it was sufficient. This is because the pressure is weak, so there is a part that is not in contact with the part that is in local contact at the interface between the seed crystal and titanium, or the pedestal interface between titanium and graphite. Furthermore, it seems that this is because there are a part where the reaction between titanium and graphite is sufficiently performed and a part where the reaction is insufficient. As a result, the structure of the adhesive layer is a layer in which a titanium carbide layer having a carbon content higher than 18 atom% and a titanium carbide layer having a carbon content lower than 18 atom% are mixed, and although there is adhesive strength, macro defects can be suppressed. I guess it is not. In the pressure range of 7.84 to 87.5 kPa, good results were obtained for both adhesive strength and macro defect density in the entire temperature range. This is because good contact is obtained over the entire surface at the seed crystal / titanium interface or at the titanium / graphite pedestal interface, so that the reaction between titanium and graphite occurs uniformly throughout the heat treatment. It seems that it is because of At a pressure of 95.3 kPa, the adhesive strength was sufficient, but since the pressure was high, the periphery of the seed crystal was chipped or cracked from the periphery after bonding. The x (*) in the table indicates a state that is adhered but influences the subsequent single crystal growth. Therefore, it is desirable that the applied pressure when titanium is used in the range of the heating temperature of 1700 to 2000 ° C. is 7.84 to 87.5 kPa.

  Table 11 shows the relationship between the applied pressure, the adhesive strength, and the macro defects when vanadium is used as the metal material.

  As is apparent from Table 11, good results were obtained for both the adhesive strength and the macro defect density in the range of pressure from 7.84 to 87.5 kPa. The reason is the same as in the experiment using titanium. Therefore, it is desirable that the applied pressure when vanadium is used in the heating temperature range of 1900 to 2200 ° C. is 7.84 to 87.5 kPa.

  Table 12 shows the relationship between pressure, adhesion strength, and macro defects when zirconium is used as the metal material.

  As is clear from Table 12, good results were obtained for both the adhesive strength and the macro defect density in the pressure range of 7.84 to 87.5 kPa. The reason is the same as in the experiment using titanium. Accordingly, it is desirable that the pressure applied when zirconium is used in the heating temperature range of 1900 to 2200 ° C. is 7.84 to 87.5 kPa.

  Further, although graphite is used as a specific example of the pressure member 4 for applying a load, it may be made of at least one of tantalum, niobium, molybdenum, tantalum carbide, niobium carbide, and molybdenum carbide. During the heating process, since the pressure member 4 is not melted or the molten pressure member 4 does not react with the seed crystal 3, the seed crystal surface is not contaminated, and impurities are taken into the single crystal to be grown. Can be suppressed.

  As described above, under optimum heat treatment conditions, the metal material is melted and reacted with graphite, and the seed crystal and the graphite pedestal are firmly fixed with a void-free metal carbide layer having a melting point equal to or higher than the single crystal growth temperature. As a result, it is possible to greatly suppress macro defects that extend into the single crystal grown on the seed crystal.

(Example 2)
A specific method corresponding to another seed crystal fixing method of the present invention will be described with reference to FIGS. 3A to 3C.

  First, as shown in FIG. 3A, on a graphite pedestal 1, a metal material 2 made of titanium having a thickness of 50 μm, a Rayleigh seed crystal 3 having an area larger than the area of the region supporting the seed crystal of the graphite pedestal 1, and the graphite pedestal 1. The pressurizing member 4 for applying a load made of graphite so that at least a part of the seed crystal region protruding from the region supporting the seed crystal was in contact with each other formed a laminate 25 arranged in this order. As a load, 33.5 kPa is applied to the seed crystal. Since FIG. 3A is represented by a cross section, the shape of the pressure member is difficult to understand, but when viewed from above, it has a donut shape with a hole at the center where the surface of the seed crystal other than the region in contact with the pressure member can be confirmed. ing.

Next, as shown in FIG. 3B, a container body 5 made of graphite is covered with a laminated body 25, and further covered with a heat insulating material 6 having a hole in the top and bottom, and a water-cooled double body capable of adjusting the internal pressure. It was installed in a quartz container 7. An RF coil 8 for heating a laminated body 25 covered with a container body 5 made of graphite around the double quartz container 7, and a lower temperature of the graphite pedestal 1 and an upper part of the container body 5 made of graphite at the upper and lower parts of the double quartz container 7. Pyrometers 9 and 10 are provided for monitoring the temperature. Actually, there are a fixing jig for holding the crucible, a vacuum pump for evacuating the double quartz container 7, a pressure adjusting valve for adjusting the atmosphere in the double quartz container 7, etc., which are omitted. Subsequently, the double quartz container is evacuated to 5 × 10 ⁻⁴ Pa or less, argon gas is introduced to adjust the pressure to 93.1 kPa (700 Torr), and the lower temperature of the graphite pedestal 1 is set to the pyrometer 10. The measured value of 1700 ° C. was heated until the upper temperature of the vessel 5 made of graphite reached 1700 ° C. or higher as measured by the pyrometer 9, and the temperature was maintained for about 3 hours in the state of (lower temperature−upper temperature) ≦ 0. Thereafter, the temperature was lowered to 1100 ° C. at a rate of 10 ° C./min, and then naturally cooled to around room temperature (25 ° C.) over 4 hours. Thereafter, the laminated body 25 covered with the container body 5 made of graphite was taken out of the double quartz container 7, and the pressure member 4 for applying a load made of graphite and the container body 5 made of graphite was removed. As shown in FIG. 3C, the Rayleigh seed crystal 3 was fixed and integrated on the graphite pedestal 1 by a metal carbide layer 11 made of titanium carbide. The adjustment of the difference between the lower temperature of the graphite pedestal 1 and the upper temperature of the container body 5 made of graphite is controlled so as to be constant at a predetermined temperature on the lower temperature side of the graphite pedestal 1, and is made of the RF coil 8 and graphite. The relative positional relationship of the laminated body 25 covered with the container body 5 was changed. This changed the top temperature.

  The metal carbide layer 11 made of titanium carbide formed by such a fixing method becomes an adhesive layer having a melting point of about 3000 ° C. or more for closely adhering the Rayleigh seed crystal 3 and the graphite pedestal 1 without voids. At a temperature of about 2300 ° C. at which the crystal is grown, no temperature gradient is generated between the Rayleigh seed crystal 3 and the graphite pedestal 1, sublimation from the back surface of the seed crystal 3 can be prevented, and generation of macro defects can be suppressed. It was. In addition, since the surface of the seed crystal 3 is not roughened after the seed crystal 3 is fixed to the graphite pedestal 1, when a single crystal is grown on the seed crystal 3, micropipe defects generated from the interface between the seed crystal surface and the growth layer And a high-quality silicon carbide single crystal could be obtained.

  The difference between the second embodiment of the present invention and the first embodiment is that the area of the seed crystal 3 is larger than the area of the graphite pedestal 1 supporting the seed crystal, and the seed protruding beyond the area of the graphite pedestal 1 supporting the seed crystal. Two points are that at least a part of the crystal region is in contact with the pressing member 4 and that the difference between the lower temperature of the graphite pedestal 1 and the upper temperature of the container body 5 made of graphite is adjusted. Since the other types of the metal material, the thickness of the metal material, the heating temperature range, the heating time range, the pressurizing range, and the like are the same as those in Example 1, the description thereof is omitted. In this embodiment, the case where titanium is used as the metal material will be described, but the same applies to vanadium and zircodium.

  The major feature of the present embodiment is that the area of the seed crystal 3 is larger than the area of the area supporting the seed crystal of the graphite pedestal 1, and the seed crystal area protruding from the area where the pressure member 4 supports the seed crystal of the graphite pedestal 1. It is made to contact only at least one part. This effect will be described below.

  The seed crystal was fixed under the conditions of a metal material of 50 μm titanium, a lower temperature of 1700 ° C., an upper temperature of 1665 ° C., a heating time of 3 hours, and a load of 33.5 kPa. The size of the area for supporting the seed crystal of the graphite pedestal is 10 mm in diameter, the Rayleigh seed crystal is approximately circular and 11 mm in diameter, and the outer periphery of the seed crystal is made of graphite at about 0.5 mm. The pressure member was brought into contact. As evaluation items, the initial surface roughness Ra of the seed crystal and the surface roughness Ra after fixing the seed crystal are measured, and a single crystal is grown using this seed crystal to observe a cross section. Pipe density was measured. The single crystal growth conditions were the same as in Example 1. Ra was measured using New View 5032 manufactured by Zygo. The measurement area is vertical: 0.14 mm and horizontal: 0.11 mm. As a comparison, the same experiment was performed when the pressure member made of graphite was in contact with the entire surface of the Rayleigh seed crystal. The results are shown in Table 13.

As apparent from Table 13, the surface roughness Ra after fixation of the seed crystal is 1.596 nm when the pressure member is in contact with only the peripheral portion, and after fixation of the seed crystal when it is in contact with the entire surface. The surface roughness Ra is 8.365 nm, and the surface roughness after fixing the seed crystal is obviously small in the method in which the pressure member is brought into contact with only the peripheral portion. This is because heat treatment is performed at 1700 ° C., and silicon, which is a constituent element of silicon carbide and has a melting point as low as about 1450 ° C., is easily removed from the surface of the seed crystal. This is probably because the contact between silicon and graphite is promoted when they are in contact. Even when only the peripheral part is in contact, the peripheral region where the pressure member is in contact is deteriorated to Ra = 7.963 nm as in the case where the pressure member is in contact with the entire surface. Since the seed crystal region that protrudes beyond the region supporting the seed crystal becomes high temperature and is lost by thermal etching, the quality of the single crystal is not affected. Regarding the macro defect density, both are 0, and it is considered that good adhesion is obtained by the titanium carbide layer. Regarding the micropipe density, it was 3300 / cm ² when the pressurizing member was in contact with only the peripheral portion, and 10,000 / cm ² when the pressurizing member was in contact with the entire surface. Since the seed crystal fixing conditions and the single crystal growth conditions are the same, the micropipe density is presumed to depend on the surface roughness after fixing the seed crystals.

  From the above, the pressure member for applying the load at the time of bonding is optimally shaped so as to contact only at least a part of the seed crystal region protruding from the region supporting the seed crystal of the graphite pedestal. I know that there is.

  Next, the area of the seed crystal 3 is larger than the area of the region supporting the seed crystal of the graphite pedestal 1, and only the seed crystal region protruding from the region where the pressure member 4 supports the seed crystal of the graphite pedestal 1 is in contact. The effect of adjusting the difference between the lower temperature of the graphite pedestal 1 and the upper temperature of the container body 5 made of graphite will be described together with the results. The conditions for fixing the seed crystal were as follows: the metal material was titanium 50 μm, the lower temperature was 1700 ° C., the heating time was 3 hours, the load was 33.5 kPa, and the upper temperature was a parameter. The size of the area for supporting the seed crystal of the graphite pedestal is 10 mm in diameter, the Rayleigh seed crystal is approximately circular and 11 mm in diameter, and the outer periphery of the seed crystal is made of graphite at about 0.5 mm. The pressure member was brought into contact. The evaluation items were measurement of the surface roughness Ra after fixing the seed crystal, and measurement of the macro defect density and the micropipe density. The initial surface roughness Ra of the seed crystal was adjusted to about 0.5 nm. The single crystal growth conditions were the same as in Example 1.

The results are shown in FIGS. 4A-B. FIG. 4A shows the result of the surface roughness after fixing, and FIG. 4B shows the result of the micropipe density generated from the interface between the seed crystal and the growth layer. Under the condition of (lower temperature−upper temperature) ≦ 0, Ra after fixing the seed crystal is about 0.7 nm, which shows that surface roughness is suppressed. Correspondingly, the micropipe density is drastically reduced under the condition of (lower temperature−upper temperature) ≦ 0, and 50 / cm ² or less can be realized. This is because under the condition of (lower temperature−upper temperature)> 0, the upper temperature is low, so that a kind of sublimation method using seed crystals as a raw material occurs, the seed crystals are sublimated, and the surface becomes rough ( Under the condition of (lower temperature−upper temperature) ≦ 0, it is assumed that the reverse occurs and the seed crystal is not sublimated. Macro defects were 0 under all conditions.

  From the above, the pressure member for applying a load at the time of adhesion is shaped so as to contact only at least a part of the seed crystal region protruding from the region supporting the seed crystal of the graphite pedestal, and the lower temperature of the graphite pedestal It can be seen that it is optimal to set the relationship of the upper temperature of the graphite container to (lower temperature−upper temperature) ≦ 0.

  As described above, the area of the seed crystal is made larger than the area of the region supporting the seed crystal of the graphite pedestal, and at least a part of the seed crystal region protruding from the region supporting the seed crystal of the graphite pedestal is in contact with the pressure member. By making the relationship between the lower temperature of the graphite pedestal and the upper temperature of the graphite container (lower temperature−upper temperature) ≦ 0, macro defects extending into the single crystal grown on the seed crystal are significantly suppressed. In addition, micropipes generated from the interface between the seed crystal and the growth layer can also be suppressed.

(Example 3)
The method for producing a single crystal in the present invention will be specifically described. A silicon carbide single crystal is grown using the seed crystal fixed in Example 1.

  As shown in FIG. 2, in the lower half of the graphite crucible made of the graphite pedestal 1 fixed with the metal carbide layer 11 made of titanium carbide or the like with the container body 5 made of graphite and the seed crystal fixed in Example 1, SiC powder 12 is accommodated as a raw material powder, covered with a heat insulating material 6 having holes in the upper and lower sides, and placed in a water-cooled double quartz container 7 capable of adjusting the internal pressure. An RF coil 8 for heating the graphite crucible 7 is provided around the double quartz container 7, and pyrometers 9, 10 for monitoring the temperature of the graphite pedestal 1 and the container body 5 made of graphite are provided at the upper and lower parts of the double quartz container 7. ing. Actually, there are a fixing jig for holding the crucible, a vacuum pump for evacuating the double quartz container 7, a pressure adjusting valve for adjusting the atmosphere in the double quartz container 7, and the like are omitted.

  Subsequently, the SiC powder 12 side is set to a high temperature, the seed crystal 3 side is set to a low temperature, and a sublimation gas is recrystallized on the seed crystal to grow a silicon carbide single crystal 13. By performing growth using the seed crystal 3 fixed and integrated on the graphite pedestal 1 with the metal carbide layer 11 having no voids, a high-quality single crystal 13 having no macro defects can be obtained.

  Actually, the metal material was heat-treated in an argon gas atmosphere at a pressure of 93.1 kPa (700 Torr) under the conditions of 50 μm titanium, heating temperature 1700 ° C., heating time 3 hours, and load 33.5 kPa, and fixed to the graphite pedestal. Using a Rayleigh seed crystal, growth was performed on this seed crystal in an argon gas atmosphere at a pressure of 3.99 kPa (30 Torr) at a SiC powder side temperature of 2350 ° C. and a seed crystal side temperature of 2200 ° C. for 30 hours. As a result, a single crystal having a height of about 8 mm was obtained. As a result of observing a cross section by slicing the crystal parallel to the growth direction, no macro defect extending from the interface between the seed crystal and the graphite pedestal into the grown single crystal was observed. In order to confirm the reproducibility, the same experiment was performed 10 times. As a result, the macro defect was completely suppressed 9 times.

As a comparative sample of the conventional example, a single crystal was grown using a seed crystal in which a Rayleigh seed crystal was attached to a graphite pedestal with a resist and the resist was carbonized and fixed by a heat treatment at about 500 ° C. for 3 hours. The growth conditions were the same as described above, and the obtained crystal height was almost the same as about 8 mm. As a result of observing the cross section of this crystal, macro defects extending from the interface between the seed crystal and the graphite pedestal to the single crystal portion through the seed crystal were observed. When the macro defect density was measured by observing the transmission mode of the microscope while gradually shifting the focus in the thickness direction of the sliced crystal, it was about 970 cm ⁻² .

  As described above, a silicon carbide single crystal is formed by using a seed crystal in which a seed crystal and a graphite pedestal are fixed with a metal carbide having a melting point equal to or higher than a single crystal growth temperature without voids and melted with a metal material. By growing it, macro defects can be suppressed and a high quality single crystal can be obtained.

Example 4
The manufacturing method of the other single crystal in this invention is demonstrated concretely. A silicon carbide single crystal is grown using the seed crystal fixed in Example 2. Since the single crystal growth procedure is the same as that of Example 3, the description thereof is omitted.

Actually, the metal material was heat-treated in an argon gas atmosphere with a pressure of 93.1 kPa (700 Torr) under the conditions of 50 μm titanium, lower temperature 1700 ° C., upper temperature 1740 ° C., heating time 3 hours, and load 33.5 kPa. Using a Rayleigh seed crystal fixed and integrated on a graphite pedestal, an SiC gas side temperature of 2350 ° C. and a seed crystal side temperature of 2200 ° C. were applied on the seed crystal in an argon gas atmosphere at a pressure of 3.99 kPa (30 Torr) for 30 hours. Growing up. As a result, a single crystal having a height of about 8 mm was obtained. Ra of the surface of the seed crystal after fixation was 0.706 nm. As a result of observing a cross section by slicing the crystal parallel to the growth direction, no macro defect extending from the interface between the seed crystal and the graphite pedestal into the grown single crystal was observed. Further, the micropipe density in the grown crystal is about 80 / cm ² because there are no seeds generated by macro defects and no seed crystal and grown layer interface due to surface roughness. I was able to.

  As described above, using a seed crystal without surface roughness in which a metal material is melted and reacted with graphite, the seed crystal and the graphite pedestal are fixed with a metal carbide having a melting point equal to or higher than the single crystal growth temperature. By growing the silicon carbide single crystal, macro defects can be suppressed and micropipes can also be suppressed, and a high-quality single crystal can be obtained.

[Industrial applicability]
A method for fixing a seed crystal for growing a silicon carbide single crystal according to the present invention and a method for producing a single crystal using the same prevent sublimation from the back surface of the seed crystal and introduce macro defects into the grown crystal. Therefore, it can be applied to cadmium sulfide (CdS), cadmium selenide (CdSe), zinc sulfide (ZnS), aluminum nitride (AlN), boron nitride (BN), and the like, which are single crystals that can be grown by the sublimation method.

1A to 1C are cross-sectional views showing a method for fixing a seed crystal for growing a single crystal in Example 1 of the present invention. FIG. 2 is a cross-sectional view showing a method for producing a silicon carbide single crystal in Example 1 of the present invention. 3A to 3C are cross-sectional views showing a method for fixing a seed crystal for single crystal growth in Example 2 of the present invention. 4A-B are graphs showing the results of the surface roughness after fixation of the seed crystal and the density of micropipes generated from the interface between the seed crystal and the growth layer in Example 2 of the present invention. FIG. 5 is a schematic cross-sectional view of a conventional sublimation apparatus for growing a single crystal. 6A to 6C are cross-sectional views showing a seed crystal structure and a macro defect generation mechanism in a conventional method. 7A to 7C are cross-sectional views showing the structure of a seed crystal coated with a protective layer and a macro defect generation mechanism in a conventional method.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Graphite base 2 Metal material 3,17 Seed crystal 4 Pressure member 5,14 Container body 6 Heat insulating material 7 Double quartz container 8 RF coil 9,10 Pyrometer 11 Metal carbide layer 12,16 SiC powder 13,18 Single crystal 15 Lid 19 Adhesive layer 20 Void 21 Macro defect 22 Carbonized layer 23 Protective layer 24 Crack 25 Laminate

Claims (14)

In the seed crystal fixing method for growing the silicon carbide single crystal from the seed crystal by fixing the seed crystal to the graphite pedestal,
A pressure member for disposing a metal material having a melting point lower than the growth temperature of the single crystal on the graphite base, disposing the seed crystal on the metal material, and further applying a load on the seed crystal. To form a laminated body,
A container body that covers the laminate is disposed on the graphite pedestal, and while adjusting the difference between the lower temperature of the graphite pedestal and the upper temperature of the container body, the lower temperature of the graphite pedestal is equal to or higher than the melting point of the metal material. And the heat treatment is performed so that the temperature is not higher than the growth temperature of the single crystal, and the graphite pedestal, the metal material, and the seed crystal are fixed and integrated,
Cooling the container and the laminate,
A seed crystal fixing method, wherein the container body and the pressure member are removed from the laminate. The area of the seed crystal is larger than the area of the portion for fixing the seed crystal of the graphite pedestal, and the pressure member has a shape that can apply a load by contacting only the peripheral portion of the seed crystal. The seed crystal fixing method according to claim 1 . The peripheral portion of the seed crystal has an area of the seed crystal that is larger than an area of the portion that fixes the seed crystal of the graphite pedestal and protrudes beyond an area of the portion that fixes the seed crystal of the graphite pedestal. The seed crystal fixing method according to claim 2 , which is at least a part. The relationship between the lower temperature of the graphite pedestal and the upper temperature of the container body,
(Lower temperature-Upper temperature) ≤ 0
The seed crystal fixing method according to claim 1 , wherein the relationship is satisfied. 2. The seed crystal fixing method according to claim 1 , wherein the metal material is at least one material selected from titanium, vanadium, and zirconium. The method for fixing a seed crystal according to claim 1 , wherein the thickness of the metal material is in a range of 20 μm to 200 μm. The method for fixing a seed crystal according to claim 5 , wherein the heating temperature in the heating step is such that the lower temperature of the graphite pedestal is in the range of 1700 ° C to 2000 ° C when the metal material is titanium. 6. The seed crystal fixing method according to claim 5 , wherein the heating temperature in the heating step is such that the lower temperature of the graphite pedestal is in the range of 1900 ° C. to 2200 ° C. when the metal material is vanadium. The method for fixing a seed crystal according to claim 5 , wherein the heating temperature in the heating step is such that the lower temperature of the graphite pedestal is in the range of 1900 ° C to 2200 ° C when the metal material is zirconium. The method for fixing a seed crystal according to claim 1 , wherein the heating time in the heating step is in the range of 1 hour to 6 hours. Said pressure member, the method of fixing the seed crystal according to claim 1 having a weight applied pressure 87.5kPa following range of 7.84kPa to the seed crystal. 2. The seed crystal fixing method according to claim 1 , wherein the pressing member is made of at least one material selected from graphite, tantalum, niobium, molybdenum, tantalum carbide, niobium carbide, and molybdenum carbide. 2. The seed crystal according to claim 1 , wherein the cooling step is performed at a rate of 5 to 15 ° C./min while the temperature drops to 1100 ° C. from the heating temperature of the heating step, and then left at room temperature until reaching room temperature. Fixing method. In the manufacturing method which supplies the sublimation gas from a silicon carbide raw material on the seed crystal hold | maintained at the graphite base, and grows a silicon carbide single crystal on the said seed crystal, By the fixing method in any one of Claims 1 thru | or 13 A method for producing a silicon carbide single crystal, comprising using a fixed seed crystal.

Images