SILICON CARBIDE SINGLE CRYSTAL SUBSTRATE AND PROCESS FOR PRODUCING SAME
20190024257 ยท 2019-01-24
Assignee
Inventors
- Shinya Sato (Tokyo, JP)
- Tatsuo Fujimoto (Tokyo, JP)
- Hiroshi TSUGE (Tokyo, JP)
- Masakazu KATSUNO (Tokyo, JP)
Cpc classification
International classification
Abstract
Provided are: a silicon carbide single crystal substrate which is cut out from a silicon carbide bulk single crystal grown by the Physical Vapor Transport method; and a process for producing the same. The number of screw dislocations in one of the semicircle areas of the substrate is smaller than that in the other thereof, namely, the number of screw dislocations in a given area of the substrate is reduced. The semicircle areas of the substrate correspond respectively to the halves of the substrate. The present invention pertains to: a silicon carbide single crystal substrate which is cut out from a silicon carbide bulk single crystal grown by the Physical Vapor Transport method and which is characterized in that the average value of the screw-dislocation densities observed at multiple measurement points in one of the semicircle areas, which correspond respectively to the halves of the substrate, is 80% or less of the average value of screw-dislocation densities observed at multiple measurement points in the other of the semicircle areas; and a process for producing the same.
Claims
1. A silicon carbide single crystal substrate cut from a bulk silicon carbide single crystal grown by a physical vapor transport method, wherein an average value of screw dislocation densities observed at a plurality of measurement points in one semicircular region which is one-half of the substrate is not more than 80% of an average value of screw dislocation densities observed at a plurality of measurement points in one semicircular region which is the other one-half of the substrate.
2. The silicon carbide single crystal substrate according to claim 1, wherein the substrate has a main surface having an angle .sub.w of more than 0 and not more than 12, the angle .sub.w being formed by the normal penetrating center point O of the substrate and the [0001] direction; when two semicircular regions bounded by a diameter R.sub. of the substrate are defined, the diameter R.sub. being perpendicular to the virtual direction V obtained by projecting the [0001] direction axis on the main surface from the center point O, the average value of screw dislocation densities observed at a plurality of measurement points in a first semicircular region which is one of the semicircular regions is not more than 80% of the average value of screw dislocation densities observed at a plurality of measurement points in a second semicircular region which is the other semicircular region.
3. The silicon carbide single crystal substrate according to claim 2, wherein the average value of screw dislocation densities observed at a plurality of measurement points in the first semicircular region is not more than 60% of the average value of screw dislocation densities observed at a plurality of measurement points in the second semicircular region.
4. The silicon carbide single crystal substrate according to claim 2, wherein the average value of screw dislocation densities observed at a plurality of measurement points in the first semicircular region is not more than 50% of the average value of screw dislocation densities observed at a plurality of measurement points in the second semicircular region.
5. The silicon carbide single crystal substrate according to claim 2, wherein when a sectoral region having a central angle of 45 from a radius r.sub. which equally divides the first semicircular region is defined, the average value of screw dislocation densities observed at a plurality of measurement points in the sectoral region is not more than 40% of the average value of screw dislocation densities observed at a plurality of measurement points in the second semicircular region.
6. The silicon carbide single crystal substrate according to claim 2, wherein when a sectoral region having a central angle of 45 from a radius r.sub. which equally divides the first semicircular region is defined, the average value of screw dislocation densities observed at a plurality of measurement points in the sectoral region is not more than 30% of the average value of screw dislocation densities observed at a plurality of measurement points in the second semicircular region.
7. The silicon carbide single crystal substrate according to claim 2, wherein the substrate has an off angle corresponding to the angle .sub.w of which off direction d.sub.w is <11-20> direction.
8. The silicon carbide single crystal substrate according to claim 2, wherein 24 radii r.sub.1 to r.sub.24 extending radially from the center point O of the substrate set to 0 are regarded as axes, each scaled from 0 to 1, the 24 radii r.sub.1 to r.sub.24 existing in 12 diameters dividing the circumference of the substrate into 24 equal parts, respectively, the average value of the screw dislocation densities observed in the first semicircular region is an average of values measured at all of 27 measurement points of i) to iii) below and the average value of the screw dislocation densities observed in the second semicircular region is an average of values measured at all of 22 measurement points of iv) to v) below: i) center point O ii) a.sub.1 to a.sub.13 iii) b.sub.1 to b.sub.13 iv) a.sub.14 to a.sub.24 v) b.sub.14 to b.sub.24 (wherein numerals appended to symbols a and b correspond to numerals of the radii r.sub.1 to r.sub.24 and, for example, a.sub.1 and b.sub.1 represent measurement points present on the radius r.sub.1; among the measurement points, a represents measurement points present within a range of from more than 0 and not more than 0.5 in each radius, and b represents measurement points present within a range of from more than 0.5 and not more than 1; and 24 measurement points having the same symbol are present on the same circle of each of the symbols a and b; one of the two radii existing in the diameter R.sub. is r.sub.1 and the radius adjacent to r.sub.1 and located in the first semicircular region is defined as r.sub.2, and the reference numbers of radii are serially allocated along the direction of the circumference).
9. The silicon carbide single crystal substrate according to claim 8, wherein the average value of screw dislocation densities observed in the first semicircular region is not more than 600/cm.sup.2.
10. The silicon carbide single crystal substrate according to claim 5, wherein 24 radii r.sub.1 to r.sub.24 extending radially from the center point O of the substrate set to 0 are regarded as axes, each scaled from 0 to 1, the 24 radii r.sub.1 to r.sub.24 existing in 12 diameters dividing the circumference of the substrate into 24 equal parts, respectively, the average value of the screw dislocation densities observed in the sectoral region is an average of values measured at all of 15 measurement points of i), vi) and vii) below and the average value of the screw dislocation densities observed in the second semicircular region is an average of values measured at all of 22 measurement points of iv) to v) below: i) center point O vi) a.sub.4 to a.sub.10 vii) b.sub.4 to b.sub.10 iv) a.sub.14 to a.sub.24 v) b.sub.14 to b.sub.24 (wherein numerals appended to symbols a and b correspond to numerals of the radii r.sub.1 to r.sub.24 and, for example, a.sub.1 and b.sub.1 represent measurement points present on the radius r.sub.1; among the measurement points, a represents measurement points present within a range of from more than 0 and not more than 0.5 in each radius, and b represents measurement points present within a range of from more than 0.5 and not more than 1; and 24 measurement points having the same symbol are present on the same circle of each of the symbols a and b; one of the two radii existing in the diameter R.sub. is r.sub.1 and the radius adjacent to r.sub.1 and located in the first semicircular region is defined as r.sub.2, and the reference numbers of radii are serially allocated along the direction of the circumference).
11. The silicon carbide single crystal substrate according to claim 10, wherein the average value of screw dislocation densities observed in the sectoral region is not more than 400/cm.sup.2.
12. The silicon carbide single crystal substrate according to claim 3, wherein when a sectoral region having a central angle of 45 from a radius r.sub. which equally divides the first semicircular region is defined, the average value of screw dislocation densities observed at a plurality of measurement points in the sectoral region is not more than 40% of the average value of screw dislocation densities observed at a plurality of measurement points in the second semicircular region.
13. The silicon carbide single crystal substrate according to claim 4, wherein when a sectoral region having a central angle of 45 from a radius r.sub. which equally divides the first semicircular region is defined, the average value of screw dislocation densities observed at a plurality of measurement points in the sectoral region is not more than 40% of the average value of screw dislocation densities observed at a plurality of measurement points in the second semicircular region.
14. The silicon carbide single crystal substrate according to claim 3, wherein when a sectoral region having a central angle of 45 from a radius r.sub. which equally divides the first semicircular region is defined, the average value of screw dislocation densities observed at a plurality of measurement points in the sectoral region is not more than 30% of the average value of screw dislocation densities observed at a plurality of measurement points in the second semicircular region.
15. The silicon carbide single crystal substrate according to claim 4, wherein when a sectoral region having a central angle of 45 from a radius r.sub. which equally divides the first semicircular region is defined, the average value of screw dislocation densities observed at a plurality of measurement points in the sectoral region is not more than 30% of the average value of screw dislocation densities observed at a plurality of measurement points in the second semicircular region.
Description
BRIEF DESCRIPTION OF DRAWINGS
[0030]
[0031]
[0032]
[0033]
[0034]
[0035]
[0036] Hereinafter, the present invention will be described in more detail.
[0037] The SiC single crystal substrate according to the present invention is cut from a bulk SiC single crystal grown by the PVT method. As described above, there are reported cases in which screw dislocations were successfully reduced in crystal growth in the MSE method and epitaxial growth of SiC in the CVD method. However, the epitaxial growth of SiC by the MSE method or the CVD method is performed at a growth rate of 1/10 or less of the growth rate of an ordinary SiC single crystal grown by the PVT method and is completely different, in terms of the level of productivity, from a manner of manufacturing by cutting from a bulk SiC single crystal grown by the PVT method, as in the SiC single crystal substrate of the present invention. Further, by the MSE method or the CVD method, a SiC epitaxial film is usually grown on the base SiC substrate to obtain a SiC epitaxial substrate. Therefore, the existence form of the screw dislocations in the substrate is different from that in the SiC single crystal substrate cut from a bulk SiC single crystal grown by the PVT method as in the present invention. More specifically, while screw dislocations exist as they are in the base SiC substrate in a SiC epitaxial substrate obtained by the MSE method or the like, in the SiC single crystal substrate of the present invention, the screw dislocations are decreased at least in a semicircular region through the direction of thickness of the substrate.
[0038] The present inventors conducted intensive and extensive studies on means for obtaining a SiC single crystal substrate having less screw dislocations by using the PVT method and, as a result, found that, when growing a bulk SiC single crystal, screw dislocation or the threading mixed dislocation (in the present specification, these are together referred to as screw dislocation) can be structurally converted by employing specific conditions during the growth.
[0039] More specifically, in a bulk SiC single crystal formed by the PVT method, screw dislocations generated by succession from the seed crystal or generated at the interface with the seed crystal exist in substantially parallel to the c axis. In more detail, screw dislocations have a property to extend in the direction of the normal of the isothermal line during the growth. In the PVT method, generally, in order to obtain a stable polytype SiC single crystal, there is provided a temperature gradient lower in the center region of the grown crystal than in the peripheral region thereof (usually, the temperature in the center region is lower by about more than 0 C. and about not more than 20 C.) to form a bulk SiC single crystal while maintaining a protruded crystal growth surface. Thus, since the isothermal line during the growth has a protruded shape, strictly speaking, the directions of extension of the screw dislocations differ from surface to surface of the growing crystal.
[0040] On the other hand, since screw dislocations in the crystal have a Burgers vector in the <0001> direction, screw dislocations most stably exist when they extend in this direction. Therefore, when the direction of extension of screw dislocations is shifted from the <0001> direction, such screw dislocations are in an unstable state because the distortion of the crystal lattice is increased so that the energy state seems to be high. In view of this, in the present invention, by creating an equilibrium state or a quasi-equilibrium state in the PVT method during the growth of the crystal, the screw dislocations in an unstable state are structurally converted so as to obtain a SiC single crystal substrate in which the screw dislocations are partially decreased.
[0041] More specifically, as depicted in
[0042] Since the direction of extension of the stacking faults and basal plane dislocations is substantially perpendicular to the growing direction of the crystal, they are expelled outside from the side surface of the crystal as the growth proceeds. Accordingly, by using such a structural conversion in the PVT method, in the (iii) subsequently grown crystal 4, the screw dislocations 6 are decreased (without including stacking faults and the like), so that a bulk SiC single crystal 5 in which the screw dislocations are decreased in one-half region in the longitudinal cross-section can be obtained. The schematic cross-sectional view depicted in
[0043] In obtaining such a bulk SiC single crystal 5, when the initially grown layer 2 is grown in the step (i), usual growth conditions in the PVT method can be employed. In the present invention, converting the screw dislocations in the structural conversion layer 3 during the step (ii) has been taken into consideration. Since the screw dislocations 6 in the bulk SiC single crystal 5 are mainly those generated by succession from the seed crystal 1 or generated at the interface with the seed crystal 1, it is acceptable to generate these screw dislocations in the step (i) in the same manner as in the conventional method.
[0044] Specifically, the temperature (growth temperature) of the seed crystal 1 is desirably not lower than 2100 C. and not higher than 2400 C., preferably not lower than 2200 C. and not higher than 2300 C. The growth atmosphere pressure is desirably not lower than 0.13 kPa and not higher than 2.6 kPa, preferably not lower than 0.65 kPa and not higher than 1.95 kPa. The initially grown layer 2 is grown under the conditions combining these conditions at a growth rate of, for example, not less than 100 m/h, preferably not less than 300 m/h. In view of sufficiently growing the screw dislocations 6 along the growth direction, the thickness of the initially grown layer 2 grown on the seed crystal 1 is desirably not less than 1 mm, preferably not less than 2 mm. However, in view of the fact that its effect is saturated and in view of the productivity and so on of the SiC single crystal substrate finally cut from the bulk SiC single crystal 5, the thickness of the initially grown layer 2 is preferably not more than 10 mm.
[0045] Regarding the growth conditions of the structural conversion layer 3 in the step (ii), the growth conditions are acceptable as long as the equilibrium state or the quasi-equilibrium state in the PVT method is created so that the screw dislocations in the unstable state are structurally converted. Usually, by increasing the growth atmosphere pressure, the amount of the atoms reaching the crystal growth surface is decreased because the diffusion of the raw material gas of the SiC sublimation raw material is slowed. On the other hand, the amount of the atoms leaving from the crystal surface is determined by the growth surface temperature. Accordingly, when the structural conversion layer 3 is grown, specifically, the temperature of the seed crystal 1 is desirably not lower than 2100 C. and not higher than 2400 C., preferably not lower than 2200 C. and not higher than 2300 C., and the growth atmosphere pressure is desirably higher than 2.6 kPa and not higher than 65 kPa, preferably not lower than 6.5 kPa and not higher than 39 kPa. In contrast to the fact that the usual growth rate of a SiC single crystal by the PVT method is about 100 to about 1000 m/h, when the structural conversion layer 3 is grown in (ii), by the combination of the above-mentioned growth temperature and the growth atmosphere pressure, the desired conversion of the screw dislocations is sought at a low growth rate of less than 100 m/h, preferably not more than 50 m/h, more preferably not more than 25 m/h. In view of the growth efficiency and so on, it is desired to set the growth rate of the structural conversion layer 3 to not less than 1 m/h.
[0046] The thickness of the structural conversion layer 3 is desirably at least 0.5 mm, preferably not less than 1 mm. When the thickness of the structural conversion layer 3 is less than 0.5 mm, the structural conversion from the screw dislocations to the stacking faults or the like may not be attained sufficiently. On the other hand, although the structural conversion of the screw dislocations is enhanced with the increase in the thickness of the structural conversion layer 3, in view of the fact that the effect is saturated and in view of the productivity and so on, it is sufficient to set the upper limit of the thickness of the structural conversion layer 3 to 10 mm.
[0047] After growing the structural conversion layer 3 in (ii), the main grown crystal 4 may be grown at an increased growth rate in the step (iii). As mentioned above, since the screw dislocations 6 in an unstable state are structurally converted to stacking faults, basal plane dislocations or the like in the structural conversion layer 3, the screw dislocations are decreased in the specific region in the crystal grown thereafter. Therefore, normal growth conditions in the PVT method can be employed, specifically, as the conditions for growing the main grown crystal 4, the growth temperature and the growth atmosphere pressure mentioned in the description of the initially grown layer 2 in (i) can be employed. Under these conditions, in view of the productivity and so on, the growth rate is desirably not less than 100 m/h, preferably not less than 300 m/h. In view of the fact that the main grown crystal 4 is the main grown part in the bulk SiC single crystal 5, and that the SiC single crystal substrate 8 is cut from the obtained bulk SiC single crystal 5, the thickness of the main grown crystal 4 grown after forming the structural conversion layer 3 is desirably not less than 10 mm. In view of using an existing equipment, the upper limit of the growth rate of the crystal is about 2000 m/h (including the case of (i)), and the upper limit of the thickness of the main grown crystal 4 is about 200 mm.
[0048] As described above, the bulk SiC single crystal 5 is obtained through at least three growth steps (i) to (iii). Regarding the growth atmosphere pressure, among these steps, the growth atmosphere pressure is set to be highest in the step (ii). On the other hand, regarding the growth temperature (seed crystal temperature), in view of the fact that a higher growth temperature results in higher temperature of the SiC sublimation raw material and, in turn, increasing the sublimated raw material to increase the growth rate, the temperature of the step (iii) may be set to be the highest. Alternatively, a constant growth temperature may be maintained throughout the steps (i) to (iii). When changing the growth temperature or the growth atmosphere pressure between the steps, since larger amplitude of the change per unit of time results in the larger change per time in growth rate, the crystal growth during the change is unstable and crystal defects such as the coexistence of a different poly type can occur. Accordingly, it is desired to change the pressure preferably at a rate of 13.3 kPa or less per hour, more preferably at a rate of 1.33 kPa or less per hour. For the same reason, it is desired to change the temperature preferably at a rate of 40 C. or less per hour, more preferably at a rate of 10 C. or less per hour.
[0049] In the present invention, as the seed crystal 1 used for obtaining the bulk SiC single crystal 5, it is preferred to use a SiC single crystal composed of SiC single crystal and having an off angle .sub.s relative to the (0001) plane. In other words, it is preferred that the normal of the SiC seed crystal substrate have an off angle .sub.s relative to the (0001) plane, and that the off direction d.sub.s be the <11-20> direction. To attain the desired structural change of the screw dislocations described above, it is important that the seed crystal 1 have an off angle .sub.s of more than 0. By this, among the screw dislocations generated in the initially grown layer 2, those in an energetically unstable state are concentrated to the region opposite to the off direction d.sub.s, and structurally converted in the structural conversion layer 3. Although the off angle .sub.s is not limited, a larger off angle is preferred because the effect of concentrating those in the energetically unstable state in a specific region is promoted. For example, the off angle .sub.s may be more than 4 or more than 8. On the other hand, since the present inventors have confirmed that when the off angle is too large, the probability that the coexistence of a different poly type occurs is increased, the off angle is desirably not more than 16, preferably not more than 12.
[0050] As depicted in
[0051] The SiC single crystal substrate 8 can be taken out from the bulk SiC single crystal 5 at a highest efficiency when the substrate is cut such that it has a cross section perpendicular to the direction of the crystal growth. In this case, in the obtained SiC single crystal substrate 8, the off angle .sub.w formed by the normal n penetrating the center point O of the substrate and [0001] is identical to the off angle .sub.s of the seed crystal 1, and its off direction d.sub.w is also identical to the off direction d.sub.s of the seed crystal 1. Thus, the SiC single crystal 8 in this case has an off angle .sub.w (=off angle .sub.s) between the normal n penetrating the center point O of the substrate and [0001], and has a main surface of which off direction d.sub.w (=off direction d.sub.s) is [11-20] direction. As depicted in
[0052] For example, in the SiC single crystal substrate 8 cut from the bulk SiC single crystal 5 depicted in
[0053] On the other hand, in view of the use, the post treatment such as forming an epitaxial film, and so on, the off angle .sub.w of the SiC single crystal substrate 8 may be arbitrarily selected. Specifically, a SiC single crystal substrate 8 may be obtained, which substrate has an off angle .sub.w different from the off angle .sub.s of the seed crystal 1, of which off direction d.sub.w is <11-20> direction, by appropriately selecting the direction when the substrate is cut from the main grown crystal 4. Here, although the off angle .sub.w is not limited, the off angle .sub.w is desirably not more than 12 in order to obtain a smooth growth surface, and preferably not more than 8 in view of obtaining a larger number of substrates from the bulk SiC single crystal (ingot). As depicted in
[0054] The SiC single crystal substrate 8 of the present invention is a silicon carbide single crystal substrate having partially decreased screw dislocations, wherein the number of screw dislocations generated in a semicircular region which is one-half of the substrate is smaller than that in the other semicircular region. More specifically, the average value of screw dislocation densities observed at a plurality of measurement points in the first semicircular region 8a is not more than 80%, preferably not more than 60%, more preferably as low as not more than 50%, of the average value of screw dislocation densities observed at a plurality of measurement points in the second semicircular region 8b. Further, when a sectoral region 8a having a central angle of 45 from a radius r.sub.1 which equally divides the first semicircular region is defined (
[0055] The method for measuring the respective average values of the screw dislocations in the first semicircular region 8a and in the second semicircular region 8b is not particularly limited. However, the most common method employed can be etching the substrate surface by immersion in molten KOH at approximately 500 C. and then observing the shape of an etched pit through an optical microscope to measure a screw dislocation density. Then, through the optical microscope, screw dislocation density can be measured at plural measurement points in the respective regions to obtain respective average values.
[0056] In this case, in order to suitably determine whether the substrate has lower screw dislocation density in the first semicircular region 8a than the second semicircular region 8b bounded by the diameter R.sub., for example, it is preferable to select measurement points in the respective regions in a manner as described below to measure screw dislocation densities so as to obtain the respective average values. However, the selection of the measurement points described below is merely one example and obviously, the present invention is not limited thereto.
[0057] More specifically, 24 radii r.sub.1 to r.sub.24 extending radially from the center point O of the substrate set to 0 are selected in the substrate and regarded as axes, each scaled from 0 to 1, the 24 radii r.sub.1 to r.sub.24 existing in 12 diameters dividing the circumference of the substrate into 24 equal parts, respectively, as depicted in
[0058] i) center point O
[0059] ii) a.sub.1 to a.sub.13
[0060] iii) b.sub.1 to b.sub.13
[0061] iv) a.sub.14 to a.sub.24
[0062] v) b.sub.14 to b.sub.24
[0063] The numerals appended to symbols a and b correspond to numerals of the radii r.sub.1 to r.sub.24 and, for example, a.sub.1 and b.sub.1 represent measurement points present on the radius r.sub.1. Among the measurement points, a represents measurement points present within a range of from more than 0 and not more than 0.5 in each radius, and b represents measurement points present within a range of from more than 0.5 and not more than 1. Here, the scale 0 corresponds to the center point O of the substrate, and the scale 1 represents the position corresponds to the point on the circumference of the substrate. The 24 measurement points having the same symbol are present on the same circle of each of the symbols a and b. One of the two radii existing in the diameter R.sub. is r.sub.1 and the radius adjacent to r.sub.1 and located in the first semicircular region is defined as r.sub.2, and the numerals of radii are serially allocated along the direction of the circumference. Thus, the radii r.sub.1 to r.sub.13 exist in the first semicircular region 8a, and radii r.sub.14 to r.sub.24 exist in the second semicircular region 8b.
[0064] In the SiC single crystal substrate 8 of the present invention, for example, the average value of the screw dislocation densities observed in the second semicircular region 8b is about 1250/cm.sup.2, while the average value of the screw dislocation densities observed in the first semicircular region 8a is not more than 950/cm.sup.2. More specifically, the average value of the screw dislocation densities observed in the second semicircular region 8b is mainly within the range of 800 to 1500/cm.sup.2, while the average value of the screw dislocation densities observed in the first semicircular region 8a is not more than 600/cm.sup.2, preferably not more than 400/cm.sup.2. In other words, the average value of the screw dislocation densities observed in the first semicircular region 8a is not more than 80%, preferably not more than 60%, more preferably not more than 50%, of the average value of the screw dislocation densities observed in the second semicircular region 8b. The lower limit of the average value of the screw dislocation densities observed in the first semicircular region 8a is, theoretically, 0.1 dislocation/cm.sup.2. However, in view of the fact that screw dislocations are inevitably generated by factors disturbing the growth, such as impurities contained in the raw material and attachment of the graphite from the wall surface of a graphite crucible, the practical lower limit at present is thought to be 1 dislocation/cm.sup.2.
[0065] In order to suitably determine whether the substrate has a lower screw dislocation density in the sectoral region 8a than in the second semicircular region 8b, the average value of the screw dislocation density in the sectoral region 8a is preferably determined from the measurement values at the total 15 measurement points i), vi) and vii) among the measurement points depicted in
[0066] i) center point O
[0067] vi) a.sub.4 to a.sub.10
[0068] vii) b.sub.4 to b.sub.10
[0069] The average value of the screw dislocation densities observed in the sectoral region 8a is not more than 400/cm.sup.2, preferably not more than 250/cm.sup.2. Thus, the average value of the screw dislocation densities observed in the sectoral region 8a is not more than 70%, preferably not more than 40%, more preferably not more than 30%, of the average value of the screw dislocation densities observed in the second semicircular region 8b. Regarding the lower limit of the average value of the screw dislocation densities observed in the sectoral region 8a, the same explanation as for the first semicircular region 8a is applied.
[0070] Since the structural conversion described above is utilized in the present invention, there is no limitation by the polytype of the bulk SiC single crystal 5. Therefore, it is applicable for obtaining a bulk SiC single crystal such as 4H type as a typical polytype, 6H type, and 3C type. In particular, the present invention is advantageous because it is applicable to the 4H type SiC single crystal substrate 8 expected to be more promising as applications to power devices. Furthermore, an apparatus for manufacturing a silicon carbide single crystal using a commonly used the PVT method can be used. Accordingly, for example, nitrogen doping or the like can be appropriately performed in crystal according to the purpose of use while controlling the amount of nitrogen gas or the like supplied into a growth atmosphere through a high purity gas pipe or a mass flow controller. Furthermore, regarding the crystal diameter of the obtained bulk SiC single crystal substrate 8, although there is a limitation posed by the apparatus for producing the bulk SiC single crystal 5, there is no theoretical limitation in applying the principle of the present invention.
[0071] The SiC single crystal substrate 8 of the present invention has partially decreased screw dislocations, and specifically, the screw dislocation density observed in the first semicircular region 8a is not more than 80%, or less, of the screw dislocation density observed in the conventional substrates. Thus, by using the semicircular region in which the screw dislocations are decreased, high performance devices in which the leakage current and decrease in the life time of the oxide film caused by the screw dislocations are small can be produced, so that the production of high performance SiC devices can be easily realized. For example, it is suitable for producing MOSFET, JFET and the like. In addition, it is advantageous that devices of high quality can be obtained, since, for example, the reliability of oxide film in transistors can be greatly improved. Further, since the production of devices depending on the distribution of the screw dislocations can be easily attained, the overall yield in the device production can also be improved.
EXAMPLES
[0072] Hereinafter, the present invention will be described in more detail with reference to Examples and the like. The present invention is not limited the content of the Examples below.
[0073]
Example 1
[0074] First, a SiC single crystal substrate having a diameter of 75 mm was cut from a bulk SiC single crystal obtained in advance, subjected to mirror surface polishing to prepare a seed crystal 1 having an off angle of 4 degrees in a (0001) plane. The normal of the SiC seed crystal substrate has an off angle s=4 with respect to the [0001] direction, and the off direction d.sub.s is the [11-20] direction. The seed crystal 1 was attached to the inner surface of the graphite lid 12 of the apparatus for growing a single crystal described above and placed in the graphite crucible 11 filled with the sublimation raw material 10, and the crucible was covered with a graphite felt 15. The seed crystal 1 was mounted on the graphite supporting rod 14 and installed inside the double quartz tube 13.
[0075] Then, after vacuum evacuating the inside of the double quartz tube 13, high purity Ar gas as an atmosphere gas was flown therein, and the pressure in the quartz tube was set to be 80 kPa. Under this pressure, a high frequency current is applied to a work coil 16 to increase the temperature of the seed crystal 1 up to 2200 C. Then, the pressure was reduced down to 1.3 kPa for 30 minutes, and 20-hour crystal growth was performed in which the crystal growth surface was the (000-1) plane of the seed crystal 1 [growth process (i)]. An estimate of the measurement results from another crystal growth performed under the same condition indicates that a SiC single crystal (initially grown layer 2) having a thickness of 6 mm was grown on the seed crystal by the 20-hour crystal growth (growth rate: 300 m/h). In the crystal growth, an appropriate amount of nitrogen was introduced so as to have a nitrogen concentration of approximately 110.sup.19 cm.sup.3 in the grown crystal, and the flow rate of nitrogen was maintained until the end of the crystal growth through all growth steps including the subsequent crystal growth.
[0076] After growing the SiC single crystal having the thickness of 6 mm in the manner as above, subsequently, pressure was reduced at a pressure changing rate of 1.3 kPa/h and a temperature was increased at a changing rate of 0 C./h for 4 hours to change the growth condition, and 20-hour crystal growth was performed keeping a growth atmosphere pressure of 6.5 kPa and a seed crystal temperature of 2200 C. (growth process (ii)). From an estimate of a result of separately performed crystal growth, as a result of the 20-hour crystal growth, a SiC single crystal having a thickness of 1.6 mm (structural conversion layer 3) is assumed to have grown (growth rate 80 m/h).
[0077] Then, at a pressure changing rate of 1.3 kPa/h and a temperature changing rate 5 C./h, the growth condition was changed for 4 hours, the growth condition was changed for 4 hours, and 30-hour crystal growth was performed keeping growth atmosphere pressure at 1.3 kPa and seed crystal temperature at 2220 C. (growth process (iii)). From an estimate of a result of crystal growth separately performed in the same manner, as a result of the 30-hour crystal growth, a SiC single crystal having a thickness of 9 mm (main grown crystal 4) is assumed to have grown (growth rate 300 m/h). Thus, a bulk SiC single crystal 5 (ingot) having a diameter of approximately 75 mm and a height of 16.6 mm was obtained.
[0078] In the bulk SiC single crystal 5 obtained above, a SiC single crystal was cut approximately perpendicularly to the crystal growth direction, by confirming the orientation with an X-ray crystal orientation measurement device, from a portion of the main grown crystal 4 grown in the growth process (iii), and subjected to mirror surface polishing to obtain a SiC single crystal substrate 8 having a thickness of 500 m according to Example 1. This SiC single crystal substrate 8 had a main surface 9 (substrate surface) of which the off angle .sub.W was 4, and the off direction d.sub.W was the [11-20] direction.
[0079] The resultant SiC single crystal substrate 8 was immersed in a molten KOH at 520 C. for 5 minutes such that the entire surface of the substrate was immersed, so as to perform etching with the molten KOH. The main plane 9 of the substrate subjected to the etching was observed through an optical microscope (magnification: 80 times) to measure a screw dislocation density. Herein, in accordance with a method described in J. Takahashi et al., Journal of Crystal Growth, 135, (1994), 61-70, a shell-shaped pit was defined as a basal plane dislocation, a small-sized hexagonal pit was defined as a threading edge dislocation, and middle-sized and large-sized hexagonal pits were defined as screw dislocations to classify dislocation defects by etch-pit configurations and obtain a screw dislocation density. In addition, in the observation through the optical microscope, as explained in
[0080] Table 1 indicates the results.
TABLE-US-00001 TABLE 1 Measurement points (unit: /cm.sup.2) Example 1 Example 2 i) Center point O 1224 846 ii) a.sub.1, a.sub.2, a.sub.3, a.sub.4, 1148, 1182, 1046, 788, 694, 548, 376, a.sub.5, a.sub.6, a.sub.7, a.sub.8, 1072, 766, 782, 744, 394, 218, 286, 388, a.sub.9, a.sub.10, a.sub.11, a.sub.12, 752, 956, 1082, 1012, 214, 288, 846, 684, a.sub.13 1256, 1174 798 iii) b.sub.1, b.sub.2, b.sub.3, b.sub.4, 1132, 1016, 902, 784, 856, 796, 694, 442, b.sub.5, b.sub.6, b.sub.7, b.sub.8, 712, 854, 722, 788, 266, 244, 488, 482, b.sub.9, b.sub.10, b.sub.11, b.sub.12, 768, 698.966, 904, 356, 484, 824, 784, b.sub.13 858 856 iv) a.sub.14, a.sub.15, a.sub.16, a.sub.17, 1344, 1182, 1588, 1034, 946, 1058, a.sub.18, a.sub.19, a.sub.20, a.sub.21, 1278, 1644, 1078, 1044, 844, 954, 842, a.sub.22, a.sub.23, a.sub.24 1366, 1378, 1274, 1068, 1074, 1018, 1172, 1346 1236 v) b.sub.14, b.sub.15, b.sub.16, b.sub.17, 1396, 1002, 1228, 1046, 852, 862, 958, b.sub.18, b.sub.19, b.sub.20, b.sub.21, 1204 1114, 1194, 758, 1324, 1024, b.sub.22, b.sub.23, b.sub.24 1106, 1218 1178, 1026, 1058, 1268, 1234, 1004 954 vi) a.sub.4, a.sub.5, a.sub.6, a.sub.7, 1072, 766, 782, 744, 376, 394, 218, 286, a.sub.8, a.sub.9, a.sub.10 752, 956, 1082 388, 214, 288 vii) b.sub.4, b.sub.5, b.sub.6, b.sub.7, 784, 712, 854, 722, 442, 266, 244, 488, b.sub.8, b.sub.9, b.sub.10 788, 768, 698 482, 356, 484, average of total 27 937 553 measurement points of i) to iii) average of total 15 847 385 measurement points of i), vi), vii) average of total 22 1251 1011 measurement points of iv) to v)
[0081] As indicated in
[0082] In addition, in the bulk of the SiC single crystal remaining after cutting the SiC single crystal substrate 8, a (1-100) plane substrate was cut from the crystal including the seed crystal 1. The (1-100) plane substrate was immersed in a molten KOH at 520 C. for 5 minutes to perform etching with the molten KOH, and then the surface of the etched substrate was observed through an optical microscope (magnification: 100 times). As a result, in a portion corresponding to the structural conversion layer 3 of the growth process (ii), linear etching marks extending substantially perpendicularly to the growing direction were observed and the occurrence of a large number of stacking faults was found. In particular, the stacking faults concentrated in the region on the side of the off direction d.sub.s of the seed crystal 1.
[0083] Then, the region where the stacking faults occurred was observed by high resolution x-ray topography. In the observation, the (0004) plane was used as a diffraction plane. As a result, the X-ray topographic image showed defects extending in parallel to the growing direction of the SiC single crystal. From the relationship with the observation condition (transmission (0004)), the defects can be identified as crystal defects that include a screw dislocation component of which Burgers vector is <0001>. In other words, screw dislocations were confirmed to have been converted to the defects extending substantially perpendicularly to the growing direction. The points of occurrence of the defects match with the positions of stacking faults observed in the molten KOH etching, indicating that the SiC single crystal having the thickness of 1.6 mm corresponding to the structural conversion layer 3 of the growth process (ii) served as the structural conversion layer so that the screw dislocations were structurally converted to the stacking faults.
Example 2
[0084] First, a SiC single crystal substrate having a diameter of 75 mm was cut from a bulk SiC single crystal obtained in advance, subjected to mirror surface polishing to prepare a seed crystal 1 having an off angle of 4 degrees in a (0001) plane. The normal of the SiC seed crystal substrate has an off angle s=4 with respect to the [0001] direction, and the off direction d.sub.s is the [11-20] direction. The seed crystal 1 was attached to the inner surface of the graphite lid 12 of the apparatus for growing a single crystal described above and placed in the graphite crucible 11 filled with the sublimation raw material 10, and the crucible was covered with a graphite felt 15, and then, the seed crystal 1 was mounted on the graphite supporting rod 14 and installed inside the double quartz tube 13.
[0085] Then, after vacuum evacuating the inside of the double quartz tube 13, high purity Ar gas as an atmosphere gas was flown therein, and the pressure in the quartz tube was set to be 80 kPa. Under this pressure, a high frequency current is applied to a work coil 16 to increase the temperature of the seed crystal 1 up to 2200 C. Then, the pressure was reduced down to 1.3 kPa, and 20-hour crystal growth was performed in which the crystal growth surface was the (000-1) plane of the seed crystal 1 [growth process (i)]. An estimate of the measurement results from another crystal growth performed under the same condition indicates that a SiC single crystal (initially grown layer 2) having a thickness of 6 mm was grown on the seed crystal by the 20-hour crystal growth (growth rate: 300 m/h). In the crystal growth, an appropriate amount of nitrogen was introduced so as to have a nitrogen concentration of approximately 110.sup.19 cm.sup.3 in the grown crystal. In addition, the flow rate of nitrogen was maintained until the end of the crystal growth through all growth steps including the subsequent crystal growth.
[0086] After growing the SiC single crystal having the thickness of 6 mm in the manner as above, subsequently, pressure was reduced at a pressure changing rate of 1.2 kPa/h and a temperature was increased at a changing rate of 0 C./h for 10 hours to change the growth condition, and 20-hour crystal growth was performed keeping a growth atmosphere pressure of 13.3 kPa and a seed crystal temperature of 2200 C. [growth process (ii)]. From an estimate of a result of separately performed crystal growth, as a result of the 20-hour crystal growth, a SiC single crystal having a thickness of 1 mm (structural conversion layer 3) is assumed to have grown (growth rate 50 m/h).
[0087] Then, at a pressure changing rate of 1.2 kPa/h and a temperature changing rate 5 C./h, the growth condition was changed for 10 hours, and 30-hour crystal growth was performed keeping growth atmosphere pressure at 1.3 kPa and seed crystal temperature at 2250 C. [growth process (iii)]. From an estimate of a result of crystal growth separately performed in the same manner, as a result of the 30-hour crystal growth, a SiC single crystal having a thickness of 9 mm (main grown crystal 4) is assumed to have grown (growth rate 300 m/h). Thus, a bulk SiC single crystal 5 (ingot) having a diameter of approximately 75 mm and a height of 16 mm was obtained.
[0088] In the bulk SiC single crystal obtained above, a SiC single crystal was cut from a portion of the main grown crystal 4 grown in the growth process (iii) near the center in the height direction in the same manner as Example 1, and subjected to mirror surface polishing to obtain a SiC single crystal substrate having a thickness of 500 m according to Comparative example 1. This SiC single crystal substrate had a main surface (substrate surface) of which the off angle .sub.W was 4, and the off direction d.sub.W was the [11-20] direction.
[0089] The SiC single crystal substrate obtained above was subjected to molten KOH etching in the same manner as Example 1 and the (0001) plane of the etched substrate was observed through an optical microscope to obtain a screw dislocation density. Selection of measurement points and the measurement of the screw dislocation density were all performed in the same manner as Example 1.
[0090] Table 1 indicates the results.
[0091] As indicated in Table 1, when values measured at 27 total measurement points of i) to iii) included in the first semicircular region 8a were averaged, the average value of the screw dislocation densities observed in the first semicircular region 8a of the SiC single crystal substrate according to Example 2 was 553/cm.sup.2. On the other hand, when values measured in 22 total measurement points of iv) to v) included in the second semicircular region 8b were averaged, the average value of screw dislocation densities observed in the second semicircular region 8b was 1011/cm.sup.2, and the average value of screw dislocation densities observed in a plurality of measurement points within the first semicircular region 8a was 54.7% of the average value of screw dislocation densities observed in a plurality of measurement points within the second semicircular region 8b. Moreover, values measured in 15 total measurement points of i), vi), and vii) included in the sectoral region 8a were averaged, the average value of screw dislocation densities observed in the sectoral region 8a was 385/cm.sup.2, and 38.1% of the average value in the case of the second semicircular region 8b. When the first semicircular region 8a is defined without the measurement points of diameter R.sub., the average value of screw dislocation densities observed in the first semicircular region 8a is 491/cm.sup.2, and 48.6% of the average value of screw dislocation densities observed in a plurality of measurement points within the second semicircular region 8b.
Example 3
[0092] First, a SiC single crystal substrate having a diameter of 75 mm was cut from a bulk SiC single crystal obtained in advance, subjected to mirror surface polishing to prepare a seed crystal 1 having an off angle of 5 degrees in a (0001) plane. The normal of the SiC seed crystal substrate has an off angle s=5 with respect to the [0001] direction, and the off direction d.sub.s is the [11-20] direction. The seed crystal 1 was attached to the inner surface of the graphite lid 12 of the apparatus for growing a single crystal described above and placed in the graphite crucible 11 filled with the sublimation raw material 10, and the crucible was covered with a graphite felt 15, and then, the seed crystal 1 was mounted on the graphite supporting rod 14 and installed inside the double quartz tube 13.
[0093] Then, after vacuum evacuating the inside of the double quartz tube 13, high purity Ar gas as an atmosphere gas was flown therein, and the pressure in the quartz tube was to be 80 kPa. Under this pressure, a high frequency current is applied to a work coil 16 to increase the temperature of the seed crystal 1 up to 2200 C. Then, the pressure was reduced down to 1.3 kPa, and 20-hour crystal growth was performed in which the crystal growth surface was the (000-1) plane of the seed crystal 1 [growth process (i)]. An estimate of the measurement results from another crystal growth performed under the same condition indicates that a SiC single crystal (initially grown layer 2) having a thickness of 6 mm was grown on the seed crystal by the 20-hour crystal growth (growth rate: 300 m/h). In the crystal growth, an appropriate amount of nitrogen was introduced so as to have a nitrogen concentration of approximately 110.sup.19 cm.sup.3 in the grown crystal. In addition, the flow rate of nitrogen was maintained until the end of the crystal growth through all growth steps including the subsequent crystal growth.
[0094] After growing the SiC single crystal having the thickness of 6 mm in the manner as above, subsequently, pressure was reduced at a pressure changing rate of 1.2 kPa/h and a temperature was increased at a changing rate of 0 C./h for 10 hours to change the growth condition, and 20-hour crystal growth was performed keeping a growth atmosphere pressure of 13.3 kPa and a seed crystal temperature of 2200 C. [growth process (ii)]. From an estimate of a result of separately performed crystal growth, as a result of the 20-hour crystal growth, a SiC single crystal having a thickness of 1 mm (structural conversion layer 3) is assumed to have grown (growth rate 50 m/h).
[0095] Then, at a pressure changing rate of 1.2 kPa/h and a temperature changing rate 5 C./h, the growth condition was changed for 10 hours, and 30-hour crystal growth was performed keeping growth atmosphere pressure at 1.3 kPa and seed crystal temperature at 2250 C. [growth process (iii)]. From an estimate of a result of separately performed crystal growth, as a result of the 30-hour crystal growth, a SiC single crystal having a thickness of 9 mm (main grown crystal 4) is assumed to have grown (growth rate 300 m/h). Thus, a bulk SiC single crystal 5 (ingot) having a diameter of approximately 75 mm and a height of 16 mm was obtained.
[0096] In the bulk SiC single crystal obtained above, a SiC single crystal was cut from a portion of the main grown crystal 4 grown in the growth process (iii) near the center in the height direction in the same manner as Example 3, and subjected to mirror surface polishing to obtain a SiC single crystal substrate having a thickness of 500 m according to Comparative example 1. This SiC single crystal substrate had a main surface (substrate surface) of which the off angle .sub.W was 5, and the off direction d.sub.W was the [11-20] direction.
[0097] The SiC single crystal substrate obtained above was subjected to molten KOH etching in the same manner as Example 1 and the (0001) plane of the etched substrate was observed through an optical microscope to obtain a screw dislocation density. Selection of measurement points and the measurement of the screw dislocation density were all performed in the same manner as Example 1.
[0098] Table 1 indicates the results.
[0099] As indicated in Table 2, when values measured at 27 total measurement points of i) to iii) included in the first semicircular region 8a were averaged, the average value of the screw dislocation densities observed in the first semicircular region 8a of the SiC single crystal substrate according to Example 3 was 494/cm.sup.2. On the other hand, values measured in total measurement points of iv) to v) included in the second semicircular region 8b were 1001/cm.sup.2, and the average value of screw dislocation densities observed in a plurality of measurement points within the first semicircular region 8a was 49.4% of the average value of screw dislocation densities observed in a plurality of measurement points within the second semicircular region 8b. Moreover, values measured in 15 total measurement points of i), vi), and vii) included in the sectoral region 8a were averaged, the average value of screw dislocation densities observed in the sectoral region 8a was 350/cm.sup.2, and 35.0% of the average value in the case of the second semicircular region 8b. When the first semicircular region 8a is defined without the measurement points of diameter R.sub., the average value of screw dislocation densities observed in the first semicircular region 8a is 426/cm.sup.2, and 42.6% of the average value of screw dislocation densities observed in a plurality of measurement points within the second semicircular region 8b.
TABLE-US-00002 TABLE 2 Measurement points (unit: /cm.sup.2) Example 3 i) Center point O 924 ii) a.sub.1, a.sub.2, a.sub.3, a.sub.4, a.sub.5, a.sub.6, a.sub.7, a.sub.8, a.sub.9, a.sub.10, 726, 688, 528, 456, a.sub.11, a.sub.12, a.sub.13 332, 260, 298, 236 382, 376, 562, 674, 792 iii) b.sub.1, b.sub.2, b.sub.3, b.sub.4, b.sub.5, b.sub.6, b.sub.7, b.sub.8, b.sub.9, b.sub.10, 778, 682, 594, 266, b.sub.11, b.sub.12, b.sub.13 242, 398, 274, 226 284, 294, 622, 696, 742 iv) a.sub.14, a.sub.15, a.sub.16, a.sub.17, a.sub.18, a.sub.19, a.sub.20, a.sub.21, 968, 924, 1046, 1002, a.sub.22, a.sub.23, a.sub.24 1132, 1032, 1082, 1016 1282, 1074, 832 v) b.sub.14, b.sub.15, b.sub.16, b.sub.17, b.sub.18, b.sub.19, b.sub.20, b.sub.21, 846, 976, 922, 1048, b.sub.22, b.sub.23, b.sub.24 1062, 1024, 1066, 1096, 868, 864, 852 vi) a.sub.4, a.sub.5, a.sub.6, a.sub.7, a.sub.8, a.sub.9, a.sub.10 456, 332, 260, 298, 236, 382, 376 vii) b.sub.4, b.sub.5, b.sub.6, b.sub.7, b.sub.8, b.sub.9, b.sub.10 266, 242, 398, 274, 226, 284, 294 average of total 27 measurement points of 494 i) to iii) average of total 15 measurement points of 350 i), vi), vii) average of total 22 measurement points of 1001 iv) to v)
Example 4
[0100] First, a SiC single crystal substrate having a diameter of 75 mm was cut from a bulk SiC single crystal obtained in advance, subjected to mirror surface polishing to prepare a seed crystal 1 having an off angle of 8 degrees in a (0001) plane. The normal of the SiC seed crystal substrate has an off angle s=8 with respect to the [0001] direction, and the off direction d.sub.s is the [11-20] direction. The seed crystal 1 was attached to the inner surface of the graphite lid 12 of the apparatus for growing a single crystal described above and placed in the graphite crucible 11 filled with the sublimation raw material 10, and the crucible was covered with graphite felt 15, and then, the seed crystal 1 was mounted on the graphite supporting rod 14 and installed inside the double quartz tube 13.
[0101] Then, after vacuum evacuating the inside of the double quartz tube 13, high purity Ar gas as an atmosphere gas was flown therein, and the pressure in the quartz tube was to be 80 kPa. Under this pressure, a high frequency current is applied to a work coil 16 to increase the temperature of the seed crystal 1 up to 2200 C. Then, the pressure was reduced down to 1.3 kPa, and 20-hour crystal growth was performed in which the crystal growth surface was the (000-1) plane of the seed crystal 1 [growth process (i)]. An estimate of the measurement results from another crystal growth performed under the same condition indicates that a SiC single crystal having a thickness of 6 mm (initially grown layer 2) was grown on the seed crystal by the 20-hour crystal growth (growth rate: 300 m/h). In the crystal growth, an appropriate amount of nitrogen was introduced so as to have a nitrogen concentration of approximately 110.sup.19 cm.sup.3 in the grown crystal. In addition, the flow rate of nitrogen was maintained until the end of the crystal growth through all growth steps including the subsequent crystal growth.
[0102] After growing the SiC single crystal having the thickness of 6 mm in the manner as above, subsequently, pressure was reduced at a pressure changing rate of 1.265 kPa/h and a temperature was increased at a changing rate of 0 12/h for 20 hours to change the growth condition, and 40-hour crystal growth was performed keeping a growth atmosphere pressure of 26.6 kPa and a seed crystal temperature of 2200 C. [growth process (ii)]. From an estimate of a result of a separately performed crystal growth, as a result of the 40-hour crystal growth, a SiC single crystal having a thickness of 1 mm (structural conversion layer 3) is assumed to have grown (growth rate 25 m/h).
[0103] Then, at a pressure changing rate of 1.265 kPa/h and a temperature changing rate 2 C./h, the growth condition was changed for 30 hours, and 30-hour crystal growth was performed keeping growth atmosphere pressure at 1.3 kPa and seed crystal temperature at 2260 C. (growth process (iii)). From an estimate of a result of separately performed crystal growth, as a result of the 30-hour crystal growth, a SiC single crystal having a thickness of 9 mm (main grown crystal 4) is assumed to have grown (growth rate 300 m/h). Thus, a bulk SiC single crystal 5 (ingot) having a diameter of approximately 75 mm in and a height of 16 mm was obtained.
[0104] In the bulk SiC single crystal obtained above, a SiC single crystal was cut from a portion of the main grown crystal 4 grown in the growth process (iii) near the center in the height direction in the same manner as Example 1, and subjected to mirror surface polishing to obtain a SiC single crystal substrate having a thickness of 500 m according to Example 4. This SiC single crystal substrate had a main surface (substrate surface) of which the off angle .sub.W was 8, and the off direction d.sub.W was the [11-20] direction.
[0105] The SiC single crystal substrate obtained above was subjected to molten KOH etching in the same manner as Example 1 and the (0001) plane of the etched substrate was observed through an optical microscope to obtain a screw dislocation density. Selection of measurement points and the measurement of the screw dislocation density were all performed in the same manner as Example 1.
[0106] Table 3 indicates the results.
TABLE-US-00003 TABLE 3 Measurement points (unit: /cm.sup.2) Example 4 Comparative example 1 i) Center point O 618 1028 ii) a.sub.1, a.sub.2, a.sub.3, a.sub.4, a.sub.5, 566, 558, 448, 254, 1296, 1120, 952, 1084, a.sub.6, a.sub.7, a.sub.8, a.sub.9, a.sub.10, 176, 148, 206, 144, 980, 868, 1020, 912, a.sub.11, a.sub.12, a.sub.13 224, 256, 584, 522, 1044, 1068, 920, 944, 592 1022 iii) b.sub.1, b.sub.2, b.sub.3, b.sub.4, b.sub.5, 632, 598, 478, 248, 1092, 1028, 936, 992, b.sub.6, b.sub.7, b.sub.8, b.sub.9, b.sub.10, 178, 204, 168, 216, 884, 792, 784, 900, b.sub.11, b.sub.12, b.sub.13 248, 218, 388, 628, 932, 976, 1052, 900, 552 1036 iv) a.sub.14, a.sub.15, a.sub.16, a.sub.17, 884, 852, 758, 1046, 1084, 1252, 958, 1046, a.sub.18, a.sub.19, a.sub.20, a.sub.21, 958, 862, 844, 852, 1158, 962, 1044, 1052, a.sub.22, a.sub.23, a.sub.24 954, 866, 946 1154, 1066, 1146 v) b.sub.14, b.sub.15, b.sub.16, b.sub.17, 946, 752, 844, 764, 924, 860, 868, 908, b.sub.18, b.sub.19, b.sub.20, b.sub.21, 862, 654, 696, 842, 808, 900, 952, 908, b.sub.22, b.sub.23, b.sub.24 684, 980, 622 1008, 1128, 1244 vi) a.sub.4, a.sub.5, a.sub.6, a.sub.7, a.sub.8, 254, 176, 148, 206 1084, 980, 868, 1020, a.sub.9, a.sub.10 144, 224, 256 912, 1044, 1068 vii) b.sub.4, b.sub.5, b.sub.6, b.sub.7, b.sub.8, 248, 178, 204, 168, 992, 884, 792, 784, b.sub.9, b.sub.10 216, 248, 218 900, 932, 976 average of total 27 372 984 measurement points of i) to iii) average of total 15 234 951 measurement points of i), vi), vii) average of total 22 839 1020 measurement points of iv) to v)
[0107] As indicated in Table 3, when values measured at 27 total measurement points of i) to iii) included in the first semicircular region 8a were averaged, the average value of the screw dislocation densities observed in the first semicircular region 8a of the SiC single crystal substrate according to Example 3 was 372/cm.sup.2. On the other hand, when values measured in 22 total measurement points of iv) to v) included in the second semicircular region 8b were averaged, the average value of screw dislocation densities observed in the second semicircular region 8b was 839/cm.sup.2, and the average value of screw dislocation densities observed in a plurality of measurement points within the first semicircular region 8a was 44.3% of the average value of screw dislocation densities observed in a plurality of measurement points within the second semicircular region 8b. Moreover, values measured in 15 total measurement points of i), vi), and vii) included in the sectoral region 8a were averaged, the average value of screw dislocation densities observed in the sectoral region 8a was 234/cm.sup.2, and 27.9% of the average value in the case of the second semicircular region 8b. When the first semicircular region 8a is defined without the measurement points of diameter R.sub., the average value of screw dislocation densities observed in the first semicircular region 8a is 322/cm.sup.2, and 38.4% of the average value of screw dislocation densities observed in a plurality of measurement points within the second semicircular region 8b.
Comparative Example 1
[0108] First, a SiC single crystal substrate having a diameter of 75 mm was cut from a bulk SiC single crystal obtained in advance, subjected to mirror surface polishing to prepare a seed crystal 1 having an off angle of 4 degrees in a (0001) plane. The normal of the SiC seed crystal substrate has an off angle s=4 with respect to the [0001] direction, and the off direction d.sub.s is the [11-20] direction. The seed crystal 1 was attached to the inner surface of the graphite lid 12 of the apparatus for growing a single crystal described above and placed in the graphite crucible 11 filled with the sublimation raw material 10, and the crucible was covered with a graphite felt 15, and then, the seed crystal 1 was mounted on the graphite supporting rod 14 and installed inside the double quartz tube 13.
[0109] Then, after vacuum evacuating the inside of the double quartz tube 13, high purity Ar gas as an atmosphere gas was flown therein, and the pressure in the quartz tube was to be 80 kPa. Under this pressure, a high frequency current is applied to a work coil 16 to increase the temperature of the seed crystal 1 up to 2200 C. Then, the pressure was reduced down to 1.3 kPa, and 60-hour crystal growth in which the crystal growth surface was the (000-1) plane of the seed crystal 1 was performed to obtain a bulk SiC single crystal having a diameter of approximately 75 mm and a height of 18 mm (growth rate: 300 m/h). In the crystal growth, an appropriate amount of nitrogen was introduced so as to have a nitrogen concentration of approximately 110.sup.19 cm.sup.3 in the grown crystal.
[0110] In the bulk SiC single crystal obtained above, a SiC single crystal was cut from near the center in the height direction in the same manner as Example 1, and subjected to mirror surface polishing to obtain a SiC single crystal substrate having a thickness of 500 m according to Comparative example 1. This SiC single crystal substrate had a main surface (substrate surface) of which the off angle .sub.W was 4, and the off direction d.sub.W was the [11-20] direction.
[0111] The SiC single crystal substrate obtained above was subjected to molten KOH etching in the same manner as Example 1 and the (0001) plane of the etched substrate was observed through an optical microscope to obtain a screw dislocation density. Selection of measurement points and the measurement of the screw dislocation density were all performed in the same manner as Example 1.
[0112] Table 3 indicates the results.
[0113] As indicated in Table 3, when values measured at 27 total measurement points of i) to iii) included in the first semicircular region 8a were averaged, the average value of the screw dislocation densities observed in the first semicircular region 8a of the SiC single crystal substrate according to Comparative example 1 was 984/cm.sup.2. On the other hand, values measured in 22 total measurement points of iv) to v) included in the second semicircular region 8b were 1020/cm.sup.2, and the value for the first semicircular region 8a was comparable to that for the second semicircular region 8b (the average value of the screw dislocation density observed in the first semicircular region was 96.5% of that in the second semicircular region). Moreover, the average of values measured in 15 total measurement points of i), vi), and vii) included in the sectoral region 8a of the SiC single crystal substrate according to the Comparative example 1 was 951/cm.sup.2, and was also comparable. When the first semicircular region 8a is defined without the measurement points of diameter R.sub., the average value of screw dislocation densities observed in the first semicircular region 8a is 959/cm.sup.2, and 94.0% of the average value of screw dislocation densities observed in a plurality of measurement points within the second semicircular region 8b.
[0114] Then, in the bulk of the SiC single crystal remaining after cutting the SiC single crystal substrate of Comparative example 1, a (1-100) plane substrate was cut from the crystal including the seed crystal. The substrate was subjected to molten KOH etching in the same manner as Example 1, and the surface of the etched substrate was observed through an optical microscope. As a result, no stacking faults as observed in Example 1 were particularly found and thus the structural conversion of screw dislocations was not observed.
LIST OF REFERENCE SIGN
[0115] 1: seed crystal [0116] 2: initially grown layer [0117] 3: structural conversion layer [0118] 4: main grown crystal [0119] 5: bulk SiC single crystal [0120] 6: screw dislocation [0121] 7: stacking fault [0122] 8: SiC single crystal substrate [0123] 8a: first semicircular region [0124] 8a: sectoral region [0125] 8b: the second semicircular region [0126] 9: main surface of SiC single crystal substrate [0127] 10: SiC sublimation raw material [0128] 11: graphite crucible [0129] 12: graphite lid [0130] 13: double quartz tube [0131] 14: graphite supporting rod [0132] 15: graphite felt [0133] 16: work coil [0134] 17: pipe [0135] 18: mass flow controller [0136] 19: vacuum pumping device