Method for manufacturing sic wafer

Abstract

To provide a method for manufacturing an SiC wafer from an SiC ingot with which it is possible to reduce the difference in level of a cut surface and yet suppress a decrease in throughput.SOLUTION: The method for manufacturing SiC wafers comprises: a step (a) for preparing an SiC ingot having off-angles; a step (b) for applying a short pulse laser with a pulse width of less than 1 ns along a scanning line substantially parallel to an off-angle formed direction; a step (c) for moving the SiC ingot relatively in an index feeding direction intersecting the scanning line at right angle; and a step (d) for applying stress to a target region which is a section of the SiC ingot located on the side opposite the index feeding direction with reference to the scanning line. The step (b) and the step (c) are repeatedly executed, and at least the step (b) is executed together with the step (d), with a cleavage extending in a direction orthogonal to the off-angle formed direction from a region where the short pulse laser is applied by the execution of the step (b).SELECTED DRAWING: Figure 11

Description

[Technical field]

[0001] The present invention relates to a method for producing SiC wafers from a SiC ingot. [Background technology]

[0002] SiC semiconductors using SiC single crystals are expected to be a material for next-generation power devices that will replace Si semiconductors. This type of semiconductor device is typically formed on wafers (SiC wafers) that are thinly sliced ​​from cylindrical SiC ingots.

[0003] Conventionally, a method has been known in which an ingot is sliced ​​into thin wafers using a wire saw, but in the case of SiC, the hardness is higher than that of silicon, making slicing with a wire saw relatively difficult.

[0004] Also, when cutting and polishing the front and back sides, some parts of the ingot are unavoidably produced that are not used for the wafers. Unlike Si, SiC is generally manufactured using the sublimation method, which has a slow growth rate, so its manufacturing costs are higher than those of Si. For this reason, when manufacturing SiC wafers, it is desirable to minimize the amount of ingot parts that are not used for the SiC wafers.

[0005] From this perspective, a method for manufacturing SiC wafers with relatively little SiC loss is known in which a modified layer is formed inside a SiC single crystal ingot by focusing laser light of a wavelength that is transparent to SiC, and then cutting along the intended cutting plane obtained by extending a crack from this modified layer (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Patent No. 6698468 [Patent Document 2] Patent Publication No. 2021-34680 Summary of the Invention [Problem to be solved by the invention]

[0007] There are many polytypes of SiC crystals, but most, with some exceptions, have a hexagonal crystal structure as shown in Figure 1. In device development, 4H-SiC, which is known to have a relatively high carrier mobility, is used, and this 4H-SiC has a hexagonal crystal structure. SiC crystals with a hexagonal structure, including 4H-SiC, have a (0001) face terminated by Si and a (000-1) face terminated by C. These faces are collectively called the {0001} face. The {0001} face is a comprehensive expression that includes crystallographically equivalent faces such as the (0001) face and the (000-1) face. In this specification, the symbol (bar) that indicates the plane orientation or direction is indicated with a number immediately following it.

[0008] The {0001} plane is generally referred to as the c-plane, and the plane perpendicular to this c-plane <0001> The direction is denoted as the c-axis. <0001> The orientation is a generic term that includes the

[0001] orientation and orientations that are crystallographically equivalent to the

[0001] orientation.

[0009] When forming a SiC ingot, a technology is generally used in which a plane inclined at a small angle of less than 10° (typically about 4°) with respect to the c-plane is used as the growth plane, and a single crystal is grown on this growth plane. This inclination angle is called the "off-angle."

[0010] FIG. 2 is a schematic cross-sectional view of a SiC ingot. The SiC ingot 10 has a pair of main surfaces (11, 12). As described above, the {0001} plane is inclined with respect to the main surfaces (11, 12). More specifically, the {0001} plane is inclined at an off angle θ1 with respect to the first main surface 11. This off angle θ1 is determined by a relationship between a normal 14 of the first main surface 11 and <0001> The {0001} plane also corresponds to the inclination angle with respect to the direction.

[0011] 2, the direction Df in which the off angle θ1 is formed is substantially parallel to the Y direction and substantially perpendicular to the X direction. In the following drawings, the Z direction corresponds to the direction of a normal 14 to the main surface (11, 12) of the SiC ingot 10.

[0012] 3 to 5 are schematic diagrams for explaining the method described in Patent Document 1 (hereinafter referred to as "Conventional Method 1"). In Fig. 3, a plan view of a SiC ingot 10 is referenced, in Fig. 4, a cross-sectional view of the SiC ingot 10 in the XZ plane is referenced, and in Fig. 5, a cross-sectional view of the SiC ingot 10 in the YZ plane is referenced.

[0013] As shown in Fig. 4, a pulsed laser L70 is irradiated onto a SiC ingot 10 by a laser processing device 70. The laser processing device 70 has a built-in focusing optical system, and a depth position 52 of a focusing point f70 of the pulsed laser L70 is set according to the thickness of a product SiC wafer. More specifically, the depth position 52 is set at a depth that takes into consideration a cutting allowance for planarization and the like in addition to the thickness of the product SiC wafer.

[0014] In conventional method 1, the focal point f70 of the pulsed laser L70 is scanned in a direction substantially perpendicular to the off-angle formation direction Df. That is, as shown in Fig. 3, the scan line 51 is parallel to the X direction and perpendicular to the Y direction. That is, it can be said that the direction of the scan line 51 is substantially perpendicular to the off-angle formation direction Df.

[0015] The SiC ingot 10 is placed on an X stage 61 and a Y stage 62, and the position of the focal point f70 of the pulsed laser L70 moves as each stage (61, 62) moves. More specifically, the pulsed laser L70 is irradiated while the focal point f70 is relatively moved in the X direction (processing feed) by moving the X stage 61. When the irradiation position of the pulsed laser L70 reaches the X side end of the SiC ingot 10, the Y stage 62 is moved to relatively move the focal point f70 in the Y direction (index feed). Thereafter, the pulsed laser L70 is irradiated again along the X direction. That is, in the conventional method 1, the Y direction substantially parallel to the off-angle formation direction Df corresponds to the index feed direction.

[0016] Extremely high energy is input to the vicinity of the focal point f70 of the pulsed laser L70 in the SiC ingot 10. This modifies the SiC ingot 10 near the focal point f70. That is, when the SiC ingot 10 is irradiated with the pulsed laser L70 while being scanned in the X direction, modified regions 53 are formed at a predetermined depth position 52 and arranged intermittently in the X direction. The modified regions 53 formed in the X direction have intervals that depend on the repetition frequency of the pulsed laser L70.

[0017] When the modified region 53 is formed at a predetermined location of the SiC ingot 10, cleavage occurs starting from the modified region 53. The cleavage extends in all circumferential directions starting from the modified region 53, but is relatively more likely to extend in the off-angle formation direction Df.

[0018] The X direction, which is the scanning direction (also called the "processing feed direction"), is substantially perpendicular to the formation direction Df of the off-angle. However, in the X direction, multiple modified regions 53 are formed at intervals that depend on the repetition frequency of the pulsed laser L70, so the intervals between adjacent modified regions 53 in the X direction are extremely narrow. In addition, since the X direction does not cross the inclined surface resulting from the off-angle, the cleavage planes exist within the same plane. Therefore, the modified regions 53 adjacent in the X direction are easily connected to each other by cleavage starting from the modified regions 53.

[0019] 5, the cleavage 54 tends to extend in the Y direction, which is the index feed direction. As a result, a cleavage 54 originating from a modified region 53 formed on one scan line and a cleavage 54 originating from a modified region 53 on an adjacent scan line, i.e., adjacent in the Y direction, are connected in the Y direction.

[0020] As a result, a large number of modified regions 53 and cleavages 54 are connected in a planar manner at a predetermined depth position in SiC ingot 10. Therefore, by cutting SiC ingot 10 using this surface as a cutting surface, a thickness equivalent to depth position 52 is cut from SiC ingot 10, and a wafer region 56 as shown in FIG.

[0021] However, in the case of the conventional method 1, as shown in FIG. 7, a large step 58 occurs on the cut surface 57 of the wafer region 56. FIG. 7 is a partially enlarged view of FIG. 6. This is due to the fact that the cleavage 54 is likely to extend along the off-angle formation direction Df. As shown in FIG. 6, even if the cleavage extends along the off-angle formation direction Df, the respective cleavage planes are on different planes, so in order to connect adjacent cleavage planes, it is necessary to generate a crack in the Z direction while extending the cleavage so that the cleavage planes overlap in the Z direction. In the case of the conventional method 1, specifically, during the peeling process performed after the laser processing process, a crack is generated by applying ultrasonic vibration or mechanical impact, and the cleavage planes adjacent in the Z direction are connected through the crack.

[0022] In the conventional method 1, the Y direction, which is substantially parallel to the direction in which the cleavage 54 is likely to extend, is set as the index feed direction, and the index distance (index amount) is widened, so that it is considered that a high throughput can be obtained in the irradiation process of the pulse laser L70. However, since the index distance is long, the interval between adjacent cleavage planes in the Z direction is widened, and it is necessary to apply a relatively large amount of energy to connect the cleavage planes in the Z direction by cracks. In addition, as shown in FIG. 7, a large step 58 is generated on the cross section, and a large amount of polishing waste is generated when this step 58 is flattened. This places a restriction on the number of wafers that can be obtained from an ingot, and leads to an increase in the number of processing steps required for flattening, which results in an increase in the manufacturing cost of wafers.

[0023] In addition, since the step 58 becomes large, it is necessary to set the depth position 52 of the focal point f70 of the pulse laser L70 at a position sufficiently deeper than the depth equivalent to the thickness of the product SiC wafer, taking into consideration the thickness polished in the polishing process performed for flattening. This reduces the number of SiC wafers obtainable from a single SiC ingot 10.

[0024] 8, in the method described in Patent Document 2 (hereinafter referred to as "Conventional Method 2"), the scanning line 21 is set in a direction (Y direction) that is substantially parallel to the off-angle formation direction Df. That is, in Conventional Method 2, the index feed direction is the X direction that is substantially perpendicular to the off-angle formation direction Df.

[0025] In this method, in the Y direction, which is the scanning direction, a plurality of modified regions 53 are formed at intervals that depend on the repetition frequency of the pulsed laser L70, so that the interval between adjacent modified regions 53 in the Y direction is extremely narrow. This means that the interval between the modified regions 53 in FIG. 6 is narrow. Therefore, as described above, although the cleavage extending in the Y direction substantially parallel to the formation direction Df of the off-angle extends along different planes, the interval in the Y direction is narrower than that in the conventional method 1, so that the interval in the Z direction between adjacent cleavage planes is shorter. As a result, the cleavage planes can be connected in the Z direction by cracks with less energy than that in the conventional method 1.

[0026] The cleavage 54 originating from the modified region 53 also extends in the X direction, which is substantially perpendicular to the formation direction Df of the off angle. This direction does not cross the inclined surface caused by the off angle, so the cleavage 54 tends to extend along the same surface. Therefore, by appropriately setting the separation distance (index distance) between the scan lines 21 adjacent in the X direction, it is possible to connect the cleavage 54 originating from the modified region 53 formed on one scan line and the cleavage 54 originating from the modified region 53 on the adjacent scan line, i.e., adjacent in the X direction, with a step that is smaller than that of the conventional method 1.

[0027] 9 is a plan view showing a modified region 53 produced by conventional method 2 and the spread of a cleavage 54 originating from the modified region 53. As described above, it is considered that the cleavage 54 originating from the modified region 53 spreads relatively more easily in the Y direction substantially parallel to the off-angle formation direction Df than in the X direction substantially perpendicular to the off-angle formation direction Df. Therefore, when viewed in the Z direction, the shape of the cleavage 54 has a length in the Y direction that is relatively longer than the length in the X direction.

[0028] 9, if the distance (index distance Ix) between adjacent scan lines 21 in the X direction in conventional method 2 is set to be approximately the same as the distance between adjacent scan lines 51 in the Y direction in conventional method 1, it is expected that cleavage 54 adjacent in the index feed direction (X direction) will not be connected to each other. In this case, a cutting surface is not formed, and therefore a SiC wafer cannot be obtained from the SiC ingot.

[0029] Therefore, in order to connect the cleavage 54 adjacent in the index feed direction (X direction), it is necessary to shorten the separation distance (index distance Ix) between the scan lines 21 adjacent in the X direction, as shown in Fig. 10. However, this leads to an increase in the number of times that the SiC ingot 10 is irradiated with the pulse laser L70, resulting in a decrease in throughput.

[0030] In view of the above, an object of the present invention is to provide a method for producing SiC wafers from a SiC ingot, which can suppress a decrease in throughput while reducing the step difference at the cut surface. [Means for solving the problem]

[0031] The method for producing a SiC wafer according to the present invention includes the steps of: A step (a) of preparing a SiC ingot in which a c-plane has an off angle with respect to a first main surface; a step (b) of irradiating the first main surface with a short pulse laser having a pulse width of less than 1 ns from outside the first main surface along a scanning line substantially parallel to a direction in which the off-angle is formed, while adjusting a focusing position to a predetermined depth from the first main surface; (c) relatively moving the SiC ingot by a predetermined index distance in an index feed direction perpendicular to the scan line; and (d) applying stress to a target area that is a part of the SiC ingot located on the opposite side of the index feed direction with respect to the scan line; The steps (b) and (c) are repeatedly performed, and at least the step (b) is performed together with the step (d); The method is characterized in that, by carrying out the step (b), cleavage extends from the region irradiated with the short-pulse laser in a direction perpendicular to the direction in which the off-angle is formed.

[0032] In this specification, the off angle direction being "substantially parallel" to the scan line means that the angle between the off angle direction and the extension direction of the scan line is an angle equivalent to the off angle. The angle equivalent to the off angle is an angle having a fluctuation range of the allowable error (±20%) with respect to the off angle.

[0033] According to the above method, the SiC ingot is irradiated with a short pulse laser along a scanning line substantially parallel to the off-angle direction, and index feed is performed in a direction perpendicular to the scanning line. Therefore, as described above, the cleavage extending from the modified region generated by the irradiation of the short pulse laser in the index feed direction extends substantially planarly. Therefore, steps are less likely to occur on the cut surface compared to conventional method 1.

[0034] According to the above method, during irradiation with a short pulse laser, stress is applied to a target region, which is a part of the SiC ingot located on the opposite side of the index feed direction with respect to the scanning line. For convenience of explanation, the index feed direction is the +X direction, and the opposite direction is the -X direction, and both are collectively referred to as the X direction. By pressing the -X side of the modified region, the +X side of the modified region is opened. As a result, cleavage starting from the modified region is particularly likely to extend to the +X side. This does not simply mean that the cleavage is more likely to extend to the +X side than to the -X side, but means that the proportion of extension to the +X side is increased among all circumferential directions in which the cleavage can extend.

[0035] As a result, according to the above method, the cleavage can be extended further in the index feed direction compared to the case where a short pulse laser is irradiated without applying stress, and therefore the index distance set in step (c) can be set longer than in the conventional case, thereby improving throughput.

[0036] By setting the pulse width of the short-pulse laser irradiated in step (b) to less than 1 ns, it is possible to transmit shock waves before thermal vibration is transmitted to the irradiation target area. This makes it possible to locally break the bond between Si and C, resulting in a modified region. From this point of view, it is preferable that the pulse width of the short-pulse laser is 100 ps or less, and more preferably 10 ps or less. It is also possible to use a femtosecond laser as a generator of the short-pulse laser.

[0037] The stress applied to the target region in step (d) is appropriately set according to the thickness of the SiC ingot, and is preferably on the order of kPa or MPa, which allows the target region in the SiC ingot to be deformed by a deformation equal to or greater than the lattice constant, enabling cleavage to extend through gaps equal to or greater than one atomic layer.

[0038] The step (d) may be a step of physically pressing the target area.

[0039] This makes it possible to apply stress to the target area with a simple configuration, specifically, for example, a predetermined pressing mechanism such as a roller.

[0040] The target area may be at least a portion of the area of ​​the SiC ingot that is located on the opposite side of the index feed direction relative to the scan line, and may be the entire width of the SiC ingot in the direction parallel to the scan line.

[0041] According to the above method, it is not necessary to change the pressed location along the scan line at the timing when the short pulse laser is irradiated along the scan line, and control is simplified.

[0042] The step (d) is a step of physically pressing the target area using a pressing mechanism; The step (c) may include a step of relatively moving the pressing mechanism by the index distance in the index feed direction.

[0043] For example, by using a rotatable roller as the pressing mechanism, when the SiC ingot is moved in the index feed direction, the pressing mechanism can rotate while being in contact with the SiC ingot, which allows the target area to move in the index feed direction without impeding the movement of the SiC ingot in the index feed direction.

[0044] The pressing mechanism may be configured to physically press the target area while applying ultrasonic vibrations thereto.

[0045] According to the above method, the stress applied to the target region can be further increased, making it easier to extend the cleavage further in the index feed direction.

[0046] Step (d) may comprise heating the target area.

[0047] According to the above method, it is possible to apply a stress associated with thermal expansion to the target area. As a specific example, the step (d) may include a step of irradiating the target area with heating light.

[0048] The method for producing a SiC wafer includes the steps of: and (e) detecting the extent of the cleavage extension after the step (b) and before the step (c), When the extension index indicating the degree of extension of the crack exceeds a predetermined threshold in the step (e), the step (b) is executed again along the same scanning line as the scanning line on which the step (b) was executed immediately before; If the extension index falls below the threshold value in step (e), step (c) may be executed.

[0049] According to the above method, when it is detected that the cleavage has not been sufficiently extended in the index feed direction, the short pulse laser is irradiated again along the same scanning line, thereby forming a cutting surface on the SiC ingot in which the cleavage is connected without returning the scanning line in the opposite direction to the index feed direction.

[0050] The step (e) may include a step of imaging an area including the scanning line of the object on which the step (b) was previously performed, and comparing information based on the obtained captured image with the threshold value.

[0051] When the cleavage is sufficiently extended, the light for imaging is reflected specularly from the cleavage surface. On the other hand, when the cleavage is not sufficiently extended, the light for imaging is scattered at the unconnected part of the cleavage, increasing the brightness. Therefore, for example, the brightness of the captured image is binarized for each pixel, and the total value is compared with a threshold value, so that the extent of the cleavage extension can be evaluated. Effect of the Invention

[0052] According to the method of the present invention, it is possible to obtain a SiC wafer region having a small step difference in the cut surface from a SiC ingot while suppressing a decrease in throughput. [Brief description of the drawings]

[0053] [Figure 1] 1 is a drawing showing a typical crystal structure of SiC. [Diagram 2] FIG. 2 is a schematic cross-sectional view of a SiC ingot. [Diagram 3] FIG. 1 is a schematic diagram for explaining the method described in Patent Document 1, and an XY plan view of a SiC ingot is referenced. [Figure 4]FIG. 1 is a schematic diagram for explaining the method described in Patent Document 1, and a cross-sectional view of an SiC ingot in the XZ plane is referenced. [Diagram 5] FIG. 1 is a schematic diagram for explaining the method described in Patent Document 1, and a cross-sectional view of a SiC ingot in the YZ plane is shown. [Figure 6] FIG. 2 is a cross-sectional view showing a schematic diagram of a wafer region obtained by cutting a SiC ingot along a cutting surface. [Figure 7] FIG. 7 is a partially enlarged view of FIG. [Figure 8] FIG. 1 is a schematic diagram for explaining the method described in Patent Document 2, in which an XY plan view of a SiC ingot is referenced. [Figure 9] FIG. 1 is a schematic diagram for explaining the method described in Patent Document 2, which diagrammatically shows the spread of cleavage on the XY plane. [Figure 10] The state of cleavage spreading when the index distance is shorter than that in FIG. 9 is shown diagrammatically on the XY plane. [Figure 11] FIG. 1 is a schematic diagram for explaining one embodiment of a method for producing a SiC wafer according to the present invention, and an XY plan view of a SiC ingot is referenced. [Figure 12] FIG. 1 is a schematic diagram for explaining one embodiment of a method for producing a SiC wafer according to the present invention, and a cross-sectional view of a SiC ingot in a YZ plane is shown. [Figure 13] FIG. 1 is a schematic diagram for explaining one embodiment of a method for producing a SiC wafer according to the present invention, and a cross-sectional view of an SiC ingot in an XZ plane is shown. [Figure 14] 1 is a graph showing the relationship between the travel distance from the laser irradiation position and the temperature change depending on the pulse width of the pulse laser irradiated onto the SiC ingot. [Figure 15] 1 is a schematic diagram for explaining one embodiment of a method for producing a SiC wafer according to the present invention, and shows a schematic view of cleavage expansion on an XY plane. [Figure 16]This is the result of simulating the distribution of stress generated in a target area due to an external force from a roller. [Figure 17] FIG. 1 is a schematic diagram for explaining another embodiment of a method for producing a SiC wafer according to the present invention, in which a cross-sectional view of a SiC ingot in a YZ plane is referenced. [Figure 18] FIG. 2 is a schematic diagram for explaining another embodiment of the method for producing a SiC wafer according to the present invention, and an XY plan view of a SiC ingot is referenced. [Figure 19] FIG. 2 is a schematic diagram for explaining another embodiment of the method for producing a SiC wafer according to the present invention, in which a cross-sectional view of an SiC ingot in an XZ plane is referenced. [Figure 20] FIG. 2 is a schematic diagram for explaining another embodiment of the method for producing a SiC wafer according to the present invention, and an XY plan view of a SiC ingot is referenced. [Figure 21] FIG. 2 is a schematic diagram for explaining another embodiment of the method for producing a SiC wafer according to the present invention, in which a cross-sectional view of an SiC ingot in an XZ plane is referenced. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0054] Hereinafter, an embodiment of the SiC wafer manufacturing method according to the present invention will be described with reference to the drawings as appropriate. Note that the drawings are all schematic illustrations, and the dimensional ratios in the drawings do not necessarily match the actual dimensional ratios, and the dimensional ratios between the drawings do not necessarily match.

[0055] The method for producing a SiC wafer in this embodiment aims to obtain a SiC wafer by cutting a SiC ingot along a cutting surface.

[0056] Fig. 11 to Fig. 13 are schematic diagrams for explaining a method for producing a SiC wafer in this embodiment. Fig. 11 shows an XY plan view of a SiC ingot, Fig. 12 shows a cross-sectional view of the SiC ingot in a YZ plane, and Fig. 13 shows a cross-sectional view of the SiC ingot in an XZ plane.

[0057] In the following description, the XYZ coordinate system shown in Figs. 11 to 13 etc. will be referred to as appropriate. Here, when a positive or negative direction is to be distinguished when expressing a direction, it is described with a positive or negative sign, such as "+X direction" and "-X direction". When a direction is to be expressed without distinguishing between positive and negative directions, it is simply described as "X direction". In other words, in this specification, when it is simply described as "X direction", both the "+X direction" and the "-X direction" are included.

[0058] 2 to 10, in the following drawings, an XYZ coordinate system is referenced in which the normal direction to main surface 11 of SiC ingot 10 is defined as the Z direction, and a plane parallel to main surface 11 is defined as the XY plane. Main surface 11 corresponds to a "first main surface."

[0059] 2, SiC ingot 10 has a c-plane inclined at an off angle θ1 with respect to first main surface 11. The off angle θ1 is 10° or less, and typically about 3° to 5°.

[0060] In the following drawings including FIG. 11, as in FIGS. 3 and 8, the direction substantially parallel to the formation direction Df of the off angle is represented as the "Y direction", and the direction substantially perpendicular to the formation direction Df of the off angle is represented as the "X direction".

[0061] 12, a short-pulse laser L30 is irradiated onto a SiC ingot 10 from a laser processing device 30. The laser processing device 30 has a built-in focusing optical system, and a depth position 22 of a focusing point f30 of the short-pulse laser L30 is adjusted to the thickness of a SiC wafer to be obtained.

[0062] The short-pulse laser L30 exhibits a pulse waveform with a pulse width of less than 1 ns. Preferably, the pulse width of the short-pulse laser L30 is 100 ps or less, and more preferably, the pulse width is 10 ps or less.

[0063] FIG. 14 is a graph showing the results of calculations of temperature changes according to the distance (travel distance) from the laser irradiation point when the pulse width of the pulse laser irradiated onto SiC ingot 10 is changed.

[0064] The results in Fig. 14 are based on the physical properties of SiC: thermal conductivity λ = 264 [W / (m K)], density ρ = 3210 [kg / cm 3 ], specific heat c = 670 [J / (kg K)], thermal diffusivity σ [m 2 / s]=λ / (ρ·c).

[0065] 14, when the pulse width of the pulse laser is 100 ps, ​​the temperature at the irradiation point is less than 1% when moving 500 nm from the irradiation point. Also, when the pulse width of the pulse laser is 10 ps, ​​the temperature at the irradiation point is less than 10% when moving 100 nm from the irradiation point, and the temperature at the irradiation point is less than 0.1% when moving 200 nm from the irradiation point.

[0066] That is, the shorter the pulse width of the short-pulse laser L30, the faster the shock wave can be transmitted locally to the target location than the heat can be transmitted, and as a result, the bond between Si and C can be broken to form a local modified region. Considering that the thickness of a SiC wafer is generally 100 μm or more, by setting the pulse width of the short-pulse laser L30 to less than 1 ns, a modified region 23 can be formed locally at a depth position 22 of the SiC ingot 10. In particular, since the lattice constant in the C-axis direction of 4H-SiC is 1.01 nm, by setting the pulse width to less than 1 ns, high energy can be concentrated on a region at the level of several tens of atomic layers, and the material can be modified locally.

[0067] In the method of the present embodiment, the focal point f30 of the short-pulse laser L30 is scanned (processed and fed) in the Y direction, which is a direction substantially parallel to the off-angle formation direction Df. In other words, the scanning direction is the same as the method described above with reference to FIG.

[0068] For example, the SiC ingot 10 is placed on an X-stage 61 and a Y-stage 62, and the position of the focal point f30 of the short pulse laser L30 moves as each stage (61, 62) moves. More specifically, the focal point f30 is moved relatively in the Y direction by moving the Y-stage 62, and the short pulse laser L30 is irradiated. Note that, although FIG. 13 shows the X-stage 61 placed on the Y-stage 62, this is merely an example, and the positions of the two may be reversed. Also, the SiC ingot 10 may be placed on a single stage that can be displaced in both the X and Y directions.

[0069] For example, when the scanning direction 30s is the +Y direction shown in Fig. 11, when the irradiation position of the short pulse laser L30 reaches the +Y side end of the SiC ingot 10, the X stage 61 is moved to relatively move the focal point f30 in the X direction (+X direction in the example of Fig. 11). After that, the short pulse laser L30 is irradiated again along the Y direction. That is, in the method of this embodiment, the scanning direction 30s (extension direction of the scan line 21) corresponds to the Y direction substantially parallel to the off-angle formation direction Df, and the index feed direction corresponds to the X direction substantially perpendicular to the off-angle formation direction Df.

[0070] In addition, when the scanning direction 30s is fixed in the +Y direction, after the irradiation position of the short pulse laser L30 reaches the +Y side end of the SiC ingot 10, the Y stage 62 may be moved so that the position of the focal point f30 corresponds to the -Y side end of the SiC ingot 10, and then the X stage 61 may be moved to index-feed the SiC ingot 10. On the other hand, when both the +Y direction and the -Y direction can be set alternately as the scanning direction 30s, after the irradiation position of the short pulse laser L30 reaches the Y-direction end of the SiC ingot 10, the X stage 61 may be moved without moving the Y stage 62 to index-feed the SiC ingot 10, and then the short pulse laser L30 may be irradiated while moving the Y stage 62 in the direction opposite to the previous scanning direction 30s.

[0071] In the method of this embodiment, during irradiation with the short-pulse laser L30, stress is applied to a region of the SiC ingot 10 located on the -X side of the currently scanned scanning line 21, i.e., on the opposite side to the index feed direction (hereinafter referred to as "target region 5"). In the example of Fig. 11 and Fig. 13, the short-pulse laser L30 is irradiated while an external force 32a is applied to the SiC ingot 10 in the thickness direction (Z direction) by a roller 32, which is an example of a pressing mechanism.

[0072] Since extremely high energy is input near the focal point f30 of the short-pulse laser L30 in the SiC ingot 10, the SiC ingot 10 present in this region is modified to form a modified region 23 (see FIG. 15). When the modified region 23 is formed in a predetermined location of the SiC ingot 10, a cleavage 24 occurs starting from the modified region 23. FIG. 15 is a plan view that diagrammatically illustrates, following FIGS. 9 and 10, the modified region 23 generated by executing the method of the present embodiment and the spread of the cleavage 24 starting from the modified region 23.

[0073] When no stress is applied to the target region 5 of the SiC ingot 10, the cleavage 24 originating from the modified region 23 is considered to extend in the same manner as the cleavage 54 described above with reference to Figs. 9 and 10. However, in this embodiment, stress is applied to the SiC ingot 10 located on the -X side (opposite the index feed direction) of the modified region 23. As a result, the SiC ingot 10 located on the +X side (the same side as the index feed direction) as viewed from the modified region 23 is in a relatively open state. Therefore, the cleavage 24 originating from the modified region 23 has a higher proportion of extension in the +X direction among all circumferential directions in which it may extend. As a result, as shown diagrammatically in Fig. 15, the cleavage 24 is more likely to extend in the X direction. Fig. 15 diagrammatically shows that the cleavage 24 obtained by the method of this embodiment extends in the X direction by a distance 24x compared to the cleavage 54 shown in Fig. 9.

[0074] As a result, the index distance Ix that allows adjacent cleavages 24 in the index feed direction (X direction) to be connected can be set to a longer distance than in the conventional method 2 described above with reference to Fig. 10. This allows the number of scan lines 21 to be reduced, thereby reducing the number of times the SiC ingot 10 is irradiated with the short pulse laser L30, and improving throughput.

[0075] From the above perspective, if the magnitude of stress applied to the target region 5 is equal to or greater than the magnitude at which the crystal of the SiC ingot 10 is deformed at the lattice constant level, it is believed that a gap larger than one atomic layer will be created, making it easier to extend the cleavage 24.

[0076] 16 is a diagram showing the results of a simulation of the distribution of stress generated in the target region 5 when a stress of 1 MPa is applied to the target region 5, assuming that the roller 32 applies an external force 32a to the SiC ingot 10 in the thickness direction (Z direction) and the tip of the roller 32 is deformed to a width of 1 mm. According to FIG. 16, it is confirmed that a tensile stress is generated near the end of the roller 32 in the target region 5. It is believed that the cleavage 24 extends in the X direction due to this tensile stress.

[0077] Taking into consideration the simulation results shown in FIG. 16, it can be seen that the stress applied to the target region 5 of the SiC ingot 10 is appropriately set according to the thickness of the SiC ingot 10, but may be on the order of kPa or MPa.

[0078] 13, it is preferable that the rollers 32 have a circular shape when viewed from the Y direction. This makes it possible to prevent the rollers 32 from rotating and interfering with the index feed of the SiC ingot 10 when the SiC ingot 10 is index-fed in the X direction. Accordingly, the target area 5 to which stress is applied by the rollers 32 also moves relatively in the index feed direction.

[0079] According to the above method, even when the SiC ingot 10 is irradiated with the short pulse laser L30 while being scanned in the Y direction substantially parallel to the off-angle formation direction Df, and index-fed in the X direction substantially perpendicular to the off-angle formation direction Df, it is possible to set the index distance relatively long.

[0080] When the short-pulse laser L30 is irradiated onto the entire area of ​​the SiC ingot 10, a surface to be cut that is substantially parallel to the XY plane is formed in the vicinity of a depth position 22 of the SiC ingot 10 that is set according to the thickness of the SiC wafer to be obtained. Thereafter, a part of the SiC ingot 10 is peeled off along the surface to be cut by a known method to obtain a SiC wafer.

[0081] An example of the irradiation conditions of the short pulse laser L30 is shown below. By applying stress, it is possible to increase the index distance in the X direction by about 100 μm or more. Wavelength: 1064nm Repetition frequency: 500kHz Average output: 0.2W~10W Pulse width: 10ps Index distance in X direction: 200μm Y-direction machining feed rate: 200~1000mm / s

[0082] For example, if the processing feed speed is 500 mm / s, the short pulse laser L30 is irradiated at a pitch of 1 μm in the Y direction along the scan line 21.

[0083] 17, ultrasonic vibrations 32b may be applied from roller 32 to target area 5 of SiC ingot 10 in addition to external force 32a. This increases the stress applied to target area 5 of SiC ingot 10. The mechanism for applying ultrasonic vibrations 32b may be configured as a mechanism separate from roller 32, which serves as a mechanism for applying external force 32a.

[0084] From the above viewpoint, it is understood that there is no limitation on the way of applying stress as long as stress can be applied to the SiC ingot 10 located on the opposite side of the index feed direction (-X side) from the irradiated region at the time of irradiating with the short pulse laser L30. For example, as shown in Figs. 18 and 19, a heating light source 35 may be used to irradiate the target region 5 of the SiC ingot 10 with heating light L35 to thermally expand the target region 5, thereby generating stress.

[0085] 18, it is assumed that a heating light source 35 having multiple light sources arranged along the Y direction is used. However, the heating light source 35 is not limited to this embodiment, and multiple light sources may be arranged in a matrix in the X and Y directions. As another example, the heating light source 35 may be configured to irradiate the heating light L35 only to a limited area, and as the focal point f30 of the short pulse laser L30 from the laser processing device 30 is scanned in the Y direction, the irradiation area of ​​the heating light L35 may also be scanned in the Y direction.

[0086] 20, a detection mechanism 40 for detecting whether the cleavage 24 is sufficiently extended in the X direction may be provided. For convenience of illustration, a mechanism for applying stress to the target region 5 is omitted in FIGS. 20 and 21.

[0087] FIG. 21 is a diagram illustrating the details of the detection mechanism 40 of FIG. 20, and a cross-sectional view in the XZ plane of the SiC ingot 10 is referenced following FIG. 17. The detection mechanism 40 includes an imaging light source 41 and a light receiving unit 42. For example, when the focal point f30 position of the short pulse laser L30 reaches an end of the SiC ingot 10 in the +Y direction, that is, when the irradiation of the short pulse laser L30 along one scanning line 21 is completed, the imaging light source 41 irradiates a region 45 including the scanning line 21 and showing a predetermined width in the X direction with visible light L41. The light receiving unit 42 receives the return light L10 from the region 45 of the SiC ingot 10 and analyzes the received light intensity.

[0088] When the cleavage 24 is sufficiently extended in the X direction, which is the index feed direction, a cutting target surface is obtained in which the modified region 23 and the cleavage 24 are connected, and the visible light L41 is specularly reflected by this cutting target surface. In contrast, when the cleavage 24 is not sufficiently extended, there are portions where the cleavage 24 are not connected to each other, and the visible light L41 is scattered at these portions. In other words, it is expected that the intensity of the return light L10 will be higher in the latter case than in the former case.

[0089] Therefore, the light receiving unit 42 checks the received light intensity of the return light L10 and compares it with a reference threshold value. If the received light intensity exceeds the threshold value, it can be determined that the cleavage 24 has not sufficiently extended in the X direction. In this case, the short pulse laser L30 may be irradiated again along the same scan line 21. This allows the SiC ingot 10 to be index-fed toward the next scan line 21 with the cleavage 24 reliably extended in the X direction.

[0090] When the extension of the cleavage 24 is detected by the above method, an index showing the degree of extension of the cleavage 24 (extension index) corresponds to the received light intensity of the return light L10 emitted from the SiC ingot 10.

[0091] [Another embodiment] Further embodiments will be described below.

[0092] <1> As described above, the Y direction, which is the extension direction of the scanning line 21, is substantially parallel to the off-angle formation direction Df. Therefore, the depth position 22 of the light-focusing point f30 may be changed according to the moving distance in the Y direction depending on the inclination of the off-angle. On the other hand, according to the above-mentioned method, the step of the cutting target surface can be reduced, so the depth position 22 of the light-focusing point f30 may be the same regardless of the position in the Y direction.

[0093] <2> In the above embodiment, the position of the focal point f30 is moved by moving the X-stage 61 and the Y-stage 62 on which the SiC ingot 10 is placed, but any method may be used to move the relative position of the focal point f30 with respect to the SiC ingot 10. For example, the laser processing apparatus 30 itself may be moved in the X-direction or Y-direction while the SiC ingot 10 is fixed.

[0094] In this case, the mechanism for applying stress to the target region 5 may also be one that moves in the X direction or the Y direction, similar to the laser processing device 30.

[0095] 11 illustrates an embodiment in which the roller 32 presses the entire area of ​​the SiC ingot 10 in the Y direction. However, the pressing mechanism represented by the roller 32 may be configured to press a region of the SiC ingot 10 located on the −X side of the focal point f30 at least in the vicinity of the focal point f30. [Explanation of symbols]

[0096] 5:Target area 10: SiC ingot 11: First principal surface 14: Normal to the first principal surface 21: Scan line 22: Depth position 23: Modified area 24: Cleavage 30: Laser processing equipment 30s: Scanning direction 32: Laura 32a: External force 32b: Ultrasonic vibration 35: Heating light source 40: Detection mechanism 41: Light source for imaging 42: Light receiving part 45: Imaging area 51: Scan line 52: Depth position 53: Modified area 54 :Open 56: Wafer area 57: Cut surface 58: Step 61: X Stage 62: Y Stage 70: Laser processing equipment Df: Off-angle formation direction L10: Return light L30: Short pulse laser L35: Heating light L41: Visible light L70: Pulsed laser f30: Focus point f70: Focus point

Claims

1. A step (a) of preparing a SiC ingot having a c-plane having an off angle with respect to a first main surface; a step (b) of irradiating the first main surface with a short pulse laser having a pulse width of less than 1 ns from outside the first main surface along a scanning line substantially parallel to a direction in which the off-angle is formed, while adjusting a focusing position to a predetermined depth from the first main surface; (c) relatively moving the SiC ingot by a predetermined index distance in an index feed direction perpendicular to the scan line; and (d) applying stress to a target area that is a part of the SiC ingot located on an opposite side of the index feed direction with respect to the scan line; The steps (b) and (c) are repeatedly performed, and at least the step (b) is performed together with the step (d); a cleavage extending from the region irradiated with the short pulse laser in a direction perpendicular to a direction in which the off-angle is formed by performing the step (b).

2. 2. The method for producing a SiC wafer according to claim 1, wherein the step (d) is a step of physically pressing the target area.

3. 3. The method for manufacturing a SiC wafer according to claim 2, wherein the target area is at least a portion of an area of ​​the SiC ingot that is located on the opposite side of the index feed direction with respect to the scanning line, and is the entire width of the SiC ingot in a direction parallel to the scanning line.

4. The step (d) is a step of physically pressing the target area using a pressing mechanism; 4. The method for producing a SiC wafer according to claim 3, wherein the step (c) includes the step of relatively moving the pressing mechanism by the index distance in the index feed direction.

5. The method for producing a SiC wafer according to claim 4 , wherein the pressing mechanism is configured to be capable of physically pressing the target region while applying ultrasonic vibrations thereto.

6. 2. The method for producing a SiC wafer according to claim 1, wherein the step (d) is a step of heating the target area.

7. 7. The method for producing a SiC wafer according to claim 6, wherein the step (d) includes the step of irradiating the target area with heating light.

8. and (e) detecting the extent of extension of the cleavage after the step (b) and before the step (c), When an extension index indicating the degree of extension of the cleavage exceeds a predetermined threshold in the step (e), the step (b) is executed again along the same scanning line as the scanning line on which the step (b) was executed immediately before; 8. The method for producing a SiC wafer according to claim 1, wherein, when the extension index falls below the threshold value in the step (e), the step (c) is performed.

9. 9. The method for producing a SiC wafer according to claim 8, wherein the step (e) includes a step of imaging an area including the scanning line of the object on which the step (b) was previously executed, and comparing information based on the obtained captured image with the threshold value.

Images