Silicon carbide epitaxial wafer and process for producing same

Abstract

A subject of present invention is to enable reducing, even in growth at a high C/Si ratio, contamination by different polytypes with respect to a silicon carbide epitaxial wafer having a low off-angle, and to provide the silicon carbide epitaxial wafer which enables forming a reliable high voltage silicon carbide semiconductor element. The silicon carbide epitaxial wafer of the present invention is a silicon carbide epitaxial wafer comprising an epitaxially grown layer disposed on a silicon carbide substrate having an -type crystal structure and an off-angle tilted at an angle of more than 0 and less than 4 from a (0001) Si plane or a (000-1) C plane, wherein a region of a step bunching including five to ten bunched steps of 1 nm in height occupies 90% or more of the surface of the silicon carbide substrate.

Claims

1. A silicon carbide epitaxial wafer comprising an epitaxially grown layer disposed on a silicon carbide substrate having an -type crystal structure and an off-angle tilted at an angle larger than 0 and 1 or less from a (0001) Si plane, wherein a step bunching region comprising bunched steps of 1 nm unit height for each step and a total step height from 5 nm to 10 nm occupies 90% or more of a surface of the silicon carbide substrate along step terraces of the bunched steps at a boundary between the silicon carbide substrate and the epitaxially grown layer.

2. A silicon carbide epitaxial wafer comprising an epitaxially grown layer disposed on a silicon carbide substrate comprising an -type crystal structure and an off-angle tilted at an angle larger than 0 and less than 4 from a (0001) Si plane or a (000-1) C plane, wherein a step bunching region comprising bunched steps of 1 nm unit height for each step and a total step height from 5 nm to 10 nm occupies 90% or more of a surface of the silicon carbide substrate along step terraces of the bunched steps at the surface of the silicon carbide substrate.

3. The silicon carbide epitaxial wafer according to claim 1, wherein a surface roughness of the epitaxially grown layer is more than 0 nm and 0.1 nm or less.

4. The silicon carbide epitaxial wafer according to claim 1, wherein a stacking fault density of a different polytype is more than 0/cm.sup.2 and 0.2/cm.sup.2 or less.

5. A method for producing the silicon carbide epitaxial wafer of claim 2, the method comprising: preparing a silicon carbide substrate having an -type crystal structure and an off-angle tilted at an angle of more than 0 and less than 4 from a (0001) Si plane or a (000-1) C plane; forming a region of a step bunching on a surface of the silicon carbide substrate; fabricating, under such a condition that a C/Si ratio is 2 or more, an epitaxially grown silicon carbide layer on the surface of the silicon carbide substrate comprising the step bunching formed thereon; and planarizing the surface of the formed epitaxially grown layer by carrying out chemical mechanical polishing on the surface, to fabricate the silicon carbide epitaxial wafer, wherein in the fabricated silicon carbide epitaxial wafer, the region of the step bunching comprising five to ten bunched steps of 1 nm in height occupies 90% or more in a surface of the silicon carbide substrate along the step terraces of the step bunching of a boundary between the silicon carbide substrate and the epitaxially grown layer.

6. The silicon carbide epitaxial wafer according to claim 2, wherein a surface roughness of the epitaxially grown layer is more than 0 nm and 0.1 nm or less.

7. The silicon carbide epitaxial wafer according to claim 2, wherein a stacking fault density of a different polytype is more than 0/cm.sup.2 and 0.2/cm.sup.2 or less.

8. The silicon carbide epitaxial wafer according to claim 6, wherein a stacking fault density of a different polytype is more than 0/cm.sup.2 and 0.2/cm.sup.2 or less.

9. The silicon carbide epitaxial wafer according to claim 3, wherein a stacking fault density of a different polytype is more than 0/cm.sup.2 and 0.2/cm.sup.2 or less.

10. The silicon carbide epitaxial wafer according to claim 2, wherein the off-angle of the silicon carbide substrate is larger than 0 and less than 1 from a (0001) Si plane.

11. A silicon carbide epitaxial wafer comprising: a silicon carbide substrate having an -type crystal structure and an off-angle tilted at an angle larger than 0 and less than 4 from a (0001) Si plane; and an epitaxially grown layer disposed on the silicon carbide substrate, wherein at a boundary between the silicon carbide substrate and the epitaxially grown layer, 90% or more of a surface of the silicon carbide substrate is occupied by a step bunching region comprising bunched steps of 1 nm unit height for each step and total step height from 5 nm to 10 nm.

12. The silicon carbide epitaxial wafer according to claim 11, wherein a surface roughness of the epitaxially grown layer is more than 0 nm and 0.1 nm or less.

13. The silicon carbide epitaxial wafer according to claim 11, wherein a stacking fault density of a different polytype is more than 0/cm.sup.2 and 0.2/cm.sup.2 or less.

14. The silicon carbide epitaxial wafer according to claim 11, wherein the off-angle of the silicon carbide substrate is larger than 0 and 1 or less from a (0001) Si plane.

15. The silicon carbide epitaxial wafer according to claim 1, wherein the epitaxially grown layer substantially has an -type crystal structure.

16. The silicon carbide epitaxial wafer according to claim 2, wherein the epitaxially grown layer substantially has an -type crystal structure.

17. The silicon carbide epitaxial wafer according to claim 11, wherein the epitaxially grown layer substantially has an -type crystal structure.

18. The silicon carbide epitaxial wafer according to claim 1, wherein the epitaxially grown layer substantially has a 4H-type crystal structure, and the silicon carbide substrate substantially has a 4H-type crystal structure.

19. The silicon carbide epitaxial wafer according to claim 2, wherein the epitaxially grown layer substantially has a 4H-type crystal structure, and the silicon carbide substrate substantially has a 4H-type crystal structure.

20. The silicon carbide epitaxial wafer according to claim 11, wherein the epitaxially grown layer substantially has a 4H-type crystal structure, and the silicon carbide substrate substantially has a 4H-type crystal structure.

Description

BRIEF DESCRIPTION OF DRAWINGS

(1) FIG. 1 is an explanatory drawing of a process of the present invention.

(2) FIG. 2 is an AFM image with a cross-sectional profile of an epitaxial wafer fabricated in Example 1.

(3) FIG. 3 is an AFM image with a cross-sectional profile orthogonal to step terraces, of a substrate for observational use fabricated under a condition of a hydrogen etching step in Example 1.

(4) FIG. 4 is an AFM image with a cross-sectional profile orthogonal to step terraces, of a substrate for observational use fabricated under a condition of a hydrogen etching step in Comparative example 1.

(5) FIG. 5 is a cross-sectional profile orthogonal to step terraces, obtained by the AFM measurement in a region where step bunching occurs.

(6) FIG. 6 is a graph showing a relation of the region occupied by step bunching in the surface of the silicon carbide substrates according to Examples 1 and 2 and Comparative examples 1 to 3 with stacking fault density owing to 3C contamination of the substrates.

EMBODIMENTS FOR CARRYING OUT THE INVENTION

(7) (Silicon Carbide Epitaxial Wafer)

(8) The silicon carbide epitaxial wafer of the present invention is a silicon carbide epitaxial wafer including an epitaxially grown layer disposed on a silicon carbide substrate having an -type crystal structure and e an off-angle tilted at an angle of more than 0 and less than 4 from a (0001) Si plane or a (000-1) C plane, wherein the region of step bunching including five to ten bunched steps of 1 nm in height occupies 90% or more of the surface of the silicon carbide substrate.

(9) FIG. 1 illustrates an explanatory drawing of a process of the present invention. Hydrogen etching is carried out on a silicon carbide substrate to create step bunching, and then epitaxially grown layer is formed under a condition of a high C/Si ratio. After the formation of the epitaxial growth layer, the surface of the epitaxially grown layer is planarized by using chemical mechanical polishing and dry etching.

(10) Determination of whether step bunching have been created or not on the silicon carbide substrate can be carried out by a cross-sectional measurement using SEM (Scanning Electron Microscopy) or TEM (Transmission Electron microscopy) which observes the boundary between the silicon carbide substrate and the epitaxially grown layer of a specimen cut from the silicon carbide epitaxial wafer. The silicon carbide substrate and the epitaxially grown layer exhibit different contrasts in the SEM and TEM measurements, depending on difference in the species and concentration of impurities contained therein, and thus the boundary can be identified. Consequently, the geometry of the surface of the silicon carbide substrate can be observed.

EXAMPLES

Example 1

(11) A silicon carbide substrate having the 4H crystal structure was prepared which was a representative silicon carbide substrate having an -type crystal structure. This silicon carbide substrate had an off-angle tilted by an angle of 0.9 in the <11-20> direction from the (0001) Si plan.

(12) This silicon carbide substrate was placed in a reactor of a hot-wall CVD apparatus to produce a silicon carbide epitaxial wafer according to Example 1, as will be described below.

(13) <Hydrogen Etching Step>

(14) The pressure in the reactor was maintained to be 6 kPa, with hydrogen gas kept introduced into the reactor at a flow rate of 100 slm, and the silicon carbide substrate was heated to 1,725 C. by radio frequency induction heating. Hydrogen etching was carried out for 40 minutes under this condition to create step bunching on the surface of the silicon carbide substrate.

(15) Although the temperature condition in the hydrogen etching step is not particularly limited and can be appropriately selected depending on a purpose, it is preferably from 1,500 C. to 1,800 C. When the temperature condition is lower than 1,500 C., a longer time may be required for the hydrogen etching step, causing a higher production cost, and when the condition exceeds 1,800 C., it may be difficult to maintain the thermal resistance of the apparatus producing the silicon carbide epitaxial wafer.

(16) <Formation Step by Epitaxial Growth>

(17) Then, silane and propane were introduced into the reactor at flow rates of 50 sccm and 33 sccm, respectively, and an epitaxially grown layer of 10 m in thickness was formed on the surface of the silicon carbide substrate under such a condition that the C/Si ratio was 2.

(18) <Step of Planarizing Surface of Epitaxial Wafer>

(19) Chemical mechanical polishing was carried out on the surface of the produced epitaxial wafer to planarize the surface.

(20) As described above, the silicon carbide epitaxial wafer according to Example 1 was produced.

(21) An AFM (Atomic Force Microscope) measurement was carried out on the epitaxial wafer fabricated in Example 1.

(22) FIG. 2 illustrates an AFM image and a cross-sectional profile orthogonal to step terraces, of the epitaxial wafer fabricated in Example 1. The surface roughness of the wafer was 0.08 nm.

Comparative Example 1

(23) Further, a silicon carbide epitaxial wafer according to Comparative example 1 was produced in the same manner as that in Example 1 except that the etching time was changed from 40 minutes to 0 minutes in the hydrogen etching step.

(24) <Calculation of Density of Contamination owing to Different Polytypes>

(25) A measurement of light emission spectra was carried out by PL (Photoluminescence) measurement and contamination density owing to different polytypes was calculated based on the count of faults exhibiting light emission other than light emission owing to the band edge of 4H-SiC and owing to impurities in the silicon carbide substrate. As a result, for the epitaxial wafer fabricated in Example 1, stacking fault density owing to 3C contamination exhibiting light emission from the 3C bulk was 0.2/cm.sup.2, and no other stacking faults were detected. On the other hands, for the epitaxial wafer fabricated in Comparative example 1, the stacking fault density owing to 3C contamination was 3.1/cm.sup.2, and no other stacking faults were detected.

(26) <Measurement>

(27) A substrate for observational use was fabricated via the hydrogen etching step under the same condition as that of the hydrogen etching step in each of Example 1 and Comparative example 1.

(28) An AFM measurement was carried out on the substrate for observational use fabricated under the condition of the hydrogen etching step in each of Example 1 and Comparative example 1.

(29) FIG. 3 and FIG. 4 show an AFM image with a cross-sectional profile of the substrate for observational use fabricated under the condition of the hydrogen etching step in each of Example 1 and Comparative example 1.

(30) FIG. 5 shows a cross-sectional profile obtained by the AFM measurement of a region where step bunching occurred. For better understanding of the bunched number of the steps, the (0001) Si plane is taken parallel to the horizontal axis. As can be seen in FIG. 5, step bunching including five to ten bunched steps of 1 nm in height occurred on the surface of the silicon carbide substrate.

(31) <Calculation of Proportion of Region Occupied by Step Bunching in Surface of Silicon Carbide Substrate>

(32) In a cross-section orthogonal to step terraces from 10 nm in altitude (lowest altitude of roughness generated by hydrogen etching) to +10 nm in altitude (highest altitude of roughness generated by hydrogen etching) on the substrate surface, three AFM measurement images of a rectangular area of 10 m10 m parallel to the etched boundary were arbitrarily selected, and the area where step bunching occurred was divided by the total measurement area to obtain the proportion of the region occupied by the step bunching in the surface of the silicon carbide substrate.

(33) As can be seen in FIG. 3, in the vicinity of the surface of the substrate for observational use fabricated on the basis of the condition of the hydrogen etching step in Example 1, a large proportion of the region was occupied by the step bunching including five to ten bunched steps of 1 nm in height. Calculation by the above-mentioned method gave 93% for the proportion of the region occupied by the step bunching in the surface of silicon carbide substrate. On the other hand, the proportion was 44% with respect to the surface of the substrate for observational use fabricated on the basis of the condition of the hydrogen etching step of Comparative example 1. When step bunching of less than 5 nm in height occurred, their area to be excluded was determined to be the value of a width corresponding to a step height of less than 5 nm multiplied by 10 m which was a value of the depth of the cross-sectional profile orthogonal to step terraces.

(34) This Example was achieved on a surface having an off-angle tilted by an angle of 0.9 degree in the <11-20> direction from the (0001) Si plane, and a smaller tilt angle is more likely to create step bunching, leading to a similar result even for shorter etching time than that in this example. A larger tilt angle is less likely to create step bunching, but a similar result can be obtained by setting a longer etching time than that in this example.

Example 2

Example 2, and Comparative Examples 2 and 3

(35) Silicon carbide epitaxial wafers according to Example 2 and Comparative examples 2 and 3 were produced in the same manner as that in Example 1 except that the hydrogen etching temperature and the hydrogen etching time in the hydrogen etching step in Example 1 were changed to values indicated in Table 1 below.

(36) The proportion of the region occupied by step bunching in the surface of the silicon carbide substrate and the stacking fault density owing to the 3C contamination, measured in the same manner as that in Example 1, were also shown in Table 1 below.

(37) TABLE-US-00001 TABLE 1 proportion stacking fault of region density owing Example/ etching etching occupied to 3C Comparative temperature time by step contamination Example ( C.) (minutes) bunching (%) (/cm.sup.2) Example 1 1725 40 93 0.2 Example 2 1725 60 95 0.2 Comparative 1725 0 44 3.1 Example 1 Comparative 1725 20 88 0.7 Example 2 Comparative 1660 1 0 5.9 Example 3

(38) As can be seen in Table 1, when the hydrogen etching step was carried out under such a condition that the region occupied by step bunching was 90% or more of the surface of the silicon carbide substrate, the stacking fault density owing to the 3C contamination was able to be reduced 0.2/cm.sup.2 or less.

(39) On the other hand, when the hydrogen etching step was carried out under such a condition that the region occupied by step bunching was less than 90% of the surface of the silicon carbide substrate, the stacking fault density owing to the 3C contamination exhibited a value significantly higher than 0.2/cm.sup.2.

(40) Now, an explanation is made for a relation of the proportion of the region occupied by step bunching in the surface of the silicon carbide substrate with the stacking fault density owing to the 3C contamination. FIG. 6 shows a graph indicating a relation of the proportion of the region occupied by step bunching in the surface of the silicon carbide substrate with the stacking fault density owing to the 3C contamination, in the silicon carbide epitaxial wafer according to each of Examples 1, 2, and Comparative examples 1 to 3. It was confirmed that the region occupied by step bunching must be 90% or more of the surface of the silicon carbide substrate in order to reduce the stacking fault density owing to the 3C contamination down to 0.2/cm.sup.2 or less.