Laser-assisted method for cutting crystalline materials

Abstract

This invention provides a crystalline material processing method that produces a large number of wafers in a single slicing procedure through highly efficient parallel slicing. [Solution] The crystalline material processing method includes forming a subsurface laser damage 434 at a first mean depth position to form a crack propagating outward from at least one subsurface laser damage pattern inside the substrate 430, then imaging the top surface 433 of the substrate with an imaging device 4, analyzing the image to identify conditions indicating the presence of crack-free areas within the substrate, and performing one or more procedures in response to the analysis.

Description

Detailed Description of the Invention

[0001] [Description of Related Applications] This application claims priority to U.S. Patent Application No. 16 / 410,487, filed May 13, 2019; U.S. Provisional Patent Application No. 62 / 803,340, filed February 8, 2019; U.S. Patent Application No. 16 / 274,064, filed February 12, 2019; and U.S. Provisional Patent Application No. 62 / 786,333, filed December 29, 2018, the entire disclosures of which are hereby incorporated herein by reference. This application also incorporates by reference the entire disclosure of U.S. Patent Application No. 16 / 274,045, filed February 12, 2019.

[0002] [Technical Field] The present disclosure relates to methods for processing crystalline materials, and more particularly, to laser-assisted methods for separating or removing relatively thin layers of crystalline material from a substrate such as a boule or wafer.

[0003] [Background] Thin layers of crystalline materials are required as starting structures for fabricating various useful systems in a wide range of applications in microelectronics, optoelectronics, and microfabrication. Conventional methods for cutting thin layers (e.g., wafers) from large-diameter crystalline ingots of crystalline materials include the use of wire saws. Wire sawing techniques have been applied to various crystalline materials such as silicon, sapphire, and silicon carbide. Wire saw tools consist of extremely fine steel wires (typically less than 0.2 mm in diameter) that are passed through grooves in one or more guide rollers. Two slicing methods exist: free abrasive slicing and fixed abrasive slicing. Free abrasive slicing involves attaching a slurry (typically abrasive particles suspended in oil) to a steel wire moving at high speed, and the ingot is cut as the abrasive particles roll between the wire and the workpiece. Unfortunately, the environmental impact of the slurry cannot be ignored. To mitigate such effects, a wire with fixed diamond abrasive grains is sometimes used as a fixed abrasive slicing method, which requires only a water-soluble coolant (not a slurry). High-efficiency parallel slicing makes it possible to produce a large number of wafers in a single slicing procedure. Figure 1 shows a conventional wire saw tool 1, which includes a parallel wire section 3 that extends between rollers 4A-4C and is positioned to simultaneously saw the ingot 2 into multiple thin sections (e.g., wafers 8A-8G), each having a face substantially parallel to the end face 6 of the ingot 2. During the sawing process, the wire section 3, supported by rollers 4A-4C, can be pushed downward 5 toward a holder 7 located beneath the ingot 2. If the end face 6 is parallel to the crystallographic c-plane of the ingot 2, and the wire section 3 saws the ingot 2 parallel to the end face 6, then each resulting wafer 8A-8G will have an "on-axis" end face 6' parallel to the crystallographic c-plane.

[0004] It is also possible to produce micro-slope wafers (also known as off-cut or "off-axis") that have end faces that are not parallel to the crystallographic c-plane. Micro-slope wafers with a 4-degree offcut (e.g., of SiC) are often used as growth substrates for high-quality physical vapor transport growth and epitaxial growth of other materials (e.g., AlN and other Group III nitrides). Micro-slope wafer production is done by growing an ingot away from the c-axis (e.g., growing on a micro-slope species material) and sawing this ingot perpendicular to the ingot sidewall, or by starting the ingot from an on-axis species material and growing this ingot perpendicular to the ingot sidewall. This can be done by sewing at an angle that moves away from the straight direction.

[0005] Wire sawing of semiconductor materials has several limitations. Kerf loss, based on the width of material removed with each cut, is unavoidable in saw cutting and represents a significant loss of semiconductor material. The stress applied to the wafer by wire saw cutting is relatively high, resulting in non-zero bow and warp characteristics. Processing time for a single boule (or ingot) is very long, and events such as wire breakage can extend processing time and lead to unwanted material loss. Chipping and cracking on the cut surface of the wafer can reduce wafer strength. At the end of the wire sawing process, debris must be removed from the resulting wafer.

[0006] For silicon carbide (SiC), which has high wear resistance (and hardness comparable to diamond and boron nitride), wire sawing can require considerable time and resources, resulting in significant manufacturing costs. SiC substrates enable the fabrication of desirable power electronics devices, radio frequency devices, and optoelectronic devices. SiC appears in many different crystalline structures called polytypes, and certain polytypes (e.g., 4H-SiC and 6H-SiC) have a hexagonal structure.

[0007] Figure 2 is an oblique crystal plane view showing the coordinate system for hexagonal crystals such as 4H-SiC, where the c-plane (the (0001) plane corresponding to the

[0001] (perpendicular) direction of crystal growth) is perpendicular to both the m-plane ((1-100) plane) and the a-plane ((11-20) plane), the (1-100) plane is perpendicular to the [1-100] direction, and the (11-20) plane is perpendicular to the [11-20] direction. Figure 3 is a second oblique crystal plane view of a hexagonal crystal showing a microslope 9 that is not parallel to the c-plane, where the vector 10 (orthogonal to the microslope 9) is tilted by an angle β away from the

[0001] direction, and the angle β is tilted (slightly) toward the [11-20] direction. Figure 4A is a perspective wafer orientation diagram showing the orientation of the beveled wafer 11A with respect to the c-plane ((0001) plane), where the vector 10A (orthogonal to the wafer plane 9A) is tilted by an angle β away from the

[0001] direction. This angle β is equal to the orthogonal tilt (or orientation shift angle) β that spans between the (0001) plane and the projection 12A of the wafer plane 9A. Figure 4B is a simplified cross-sectional view of the beveled wafer 11A superimposed on a portion of the ingot 14A (for example, an on-axis ingot having an end face 6A parallel to the (0001) plane) from which the beveled wafer 11A was defined. Figure 4B shows that the alignment of the wafer plane 9A of the beveled wafer 11A is shifted by an angle β from the (0001) plane.

[0008] Figure 5 includes an upper surface 26 (for example, parallel to the (0001) plane (c plane) and perpendicular to the

[0001] direction), and perpendicular to the (11-20) plane and parallel to the [11-20] direction (length L F This is a top plan view of an exemplary SiC wafer 25, with a substantially circular edge 27 (having a diameter D) as its lateral boundary, including a primary flat 28 (having a c-plane). The SiC wafer may include an outer surface that is misaligned from the c-plane (e.g., oblique and off-axis with respect to the c-plane).

[0009] Due to the difficulties associated with the fabrication and processing of SiC, the cost of SiC device wafers is higher compared to various other semiconductor material wafers. Typical kerf loss resulting from SiC wire sawing can exceed approximately 250 microns per wafer, and the resulting wafer thickness from the wire sawing process is approximately 350 microns, which is not negligible considering that it is subsequently thinned (by grinding) to a final thickness of approximately 100 to 180 microns depending on the end application. Slicing wafers thinner than approximately 350 microns involves considering the problems of wire sawing and device fabrication. That had previously seemed unrealistic.

[0010] In an attempt to address the limitations associated with wire sawing, alternative techniques have been developed for removing thin layers of semiconductor material from bulk crystals. One such technique, involving the removal of silicon carbide layers from larger crystals, is described by Kim et al., "4H-SiC wafer slicing by using femtosecond laser double pulses," Optical Materials Express, p. 2450, vol. This is described in 7, no. 7 (2017). Such techniques involve forming a laser writing track by impacting silicon carbide with laser pulses to induce subsurface damage, then bonding the crystal to a locking jig and applying tensile force to cause fracturing along the subsurface damage zone. By using the laser to weaken specific areas of the material and then causing fracturing between these areas, the laser scanning time is reduced.

[0011] Additional separation techniques, including the formation of subsurface laser damage on a SiC ingot using a pulsed laser beam and subsequent induction of fragmentation by application of ultrasonic vibrations, are disclosed in U.S. Patents 9,925,619 and 10,155,323 of DISCO Corporation. Additional techniques for removing thin layers of semiconductor material from bulk crystals are disclosed in U.S. Patent Application Publication 2018 / 0126484A1 of Siltectra GmbH.

[0012] In this technical field, tools for forming laser subsurface damage in semiconductor materials are known and commercially available from various suppliers, such as Disco Corporation (Tokyo, Japan). Such tools allow laser radiation to be focused into the interior of a crystalline substrate, enabling lateral movement of the laser relative to the substrate. Typical laser damage patterns involve the formation of parallel lines spaced laterally apart from one another at a certain depth within the crystalline material substrate. Parameters such as focusing depth, laser power, and translational velocity can be adjusted to impart laser damage, but adjusting certain factors involves trade-offs. Increasing the laser power tends to impart greater subsurface damage, which can increase the ease of fracturing (for example, by reducing the stress required to complete the fracturing). However, greater subsurface damage leads to greater surface irregularity along the surface exposed by the fracturing, which may require additional processing to smooth such surfaces sufficiently for subsequent processing (for example, for integration into electronic devices). While reducing the lateral spacing between subsurface laser damage lines can improve the ease of fracturing, this also increases the number of parallel paths between the substrate and the laser, thereby reducing tool throughput. Furthermore, the results obtained by laser processing may vary within the substrate depending on the lateral or radial position at a particular vertical depth, and / or the perpendicular position of the substrate surface as part of the ingot relative to its original growth position.

[0013] Due to variations in material and / or optical properties within thick substrates such as SiC ingots, and also between different ingots of the same composition, it is difficult to easily produce wafers of uniform thickness that can be reproduced by laser processing and subsequent crushing while avoiding unwanted material loss.

[0014] Therefore, in this field, improved laser-assisted methods for separating or removing relatively thin layers of crystalline (e.g., semiconductor) material from a substrate are still being sought to address the problems associated with conventional methods.

[0015] [Overview of the prefecture] This disclosure relates to a method and apparatus for processing crystalline material substrates, and takes various forms. When additional laser substrate damage is required at a first depth position, and The imaging and / or analysis of crack-free areas after the formation of subsurface laser damage in the substrate is used as an indicator for determining when to change the instruction set for forming subsurface laser damage at the next depth position, thereby addressing the variability of laser damage formation requirements (e.g., laser power, laser focusing depth, number of damage formation passes) per substrate and at different depth positions within a single substrate. The crystalline material processing method includes generating a subsurface laser damage area in an area at a first mean depth position of the crystalline material to facilitate the formation of cracks in the substrate that propagate outward from the subsurface laser damage pattern; imaging the top surface of the substrate; analyzing the image to identify conditions indicating the presence of crack-free areas within the substrate; and performing one or more actions in response to the analysis (e.g., when appropriate conditions are achieved). One possible action includes forming auxiliary subsurface laser damage at the first mean depth position to facilitate the formation of additional cracks in the crack-free area for the purpose of forming a first thickness reduction portion of the substrate (e.g., a first wafer). Possible alternative procedures include modifying the instruction set to cause subsequent subsurface laser damage formation (at a second or subsequent mean depth position to form a second and any thickness reduction portion of the substrate) without necessarily forming additional damage at the first mean depth position. The laser damage facilitates subsequent fracture of the substrate to produce multiple thickness reduction portions of the substrate. The material processing apparatus includes a laser processing station having a laser, at least one translational stage, a scattering light source positioned to be located on the first lateral side of the substrate, and an imaging device positioned on the second lateral side opposite the substrate. The light source may be positioned substantially perpendicular to the primary flat of the substrate and / or within ±5 degrees from the direction perpendicular to the <11-20> direction of the hexagonal structure of the substrate to enhance visibility of crack-free areas through the top surface of the substrate.

[0016] In one embodiment, the Disclosure provides: supplying and executing laser radiation focused along a first mean depth position within the crystalline material of a substrate to form a subsurface laser damage having at least one subsurface laser damage pattern, wherein the at least one subsurface laser damage pattern is configured to facilitate the formation of at least one group of multiple cracks within the substrate that propagate substantially outward from the at least one subsurface laser damage pattern; generating at least one image of the top surface of the substrate after the formation of the at least one subsurface laser damage pattern; analyzing at least one image to identify conditions indicating the presence of crack-free regions within the substrate; and, in response to the analysis, the following steps The present invention relates to a crystal material processing method, comprising: (i) or (ii): (i) performing relative movement between a laser and a substrate while supplying laser radiation focused into the interior of the substrate in at least a crack-free region, in order to facilitate the formation of additional cracks in a crack-free region along or adjacent to a first thickness reduction portion of the substrate; or (ii) changing the instruction set associated with the substrate for forming subsurface laser damage when creating subsurface laser damage patterns at a second mean depth position and any subsequent mean depth position in the substrate, in order to form at least one additional thickness reduction portion of the substrate.

[0017] In certain embodiments, the analysis includes quantifying the top area characteristics of one or more crack-free regions within the substrate and comparing the top area characteristics to at least one predetermined threshold area characteristic.

[0018] In certain embodiments, at least one predetermined threshold area characteristic is a first predetermined threshold area characteristic The method includes a top area characteristic and a second predetermined threshold area characteristic, wherein the second predetermined threshold area characteristic is greater than the first predetermined threshold area characteristic, and the method includes performing step (ii) if the top area characteristic is at least the same size as the first predetermined threshold area characteristic, and performing step (i) if the top area characteristic is at least the same size as the second predetermined threshold area characteristic.

[0019] In certain embodiments, the method includes performing both steps (i) and (ii) in response to analysis.

[0020] In certain embodiments, step (ii) includes adjusting at least one of (a) the average laser power, (b) the laser focusing depth relative to the exposed surface of the substrate, or (c) the number of laser damage formation steps when creating subsurface laser damage patterns at a second average depth position and any subsequent average depth position in the substrate.

[0021] In certain embodiments, modifying the instruction set according to step (ii) includes increasing the average laser power by a value within the range of 0.15 to 0.35 watts.

[0022] In certain embodiments, step (i) includes adjusting at least one of (a) the average laser power or (b) the laser focusing depth relative to the exposed surface of the substrate when creating auxiliary subsurface laser damage to assist at least one subsurface laser damage pattern and to facilitate the formation of additional cracks in crack-free regions along or adjacent to a first average depth position.

[0023] In certain embodiments, the substrate has a generally circular edge with a primary flat, and generating at least one image includes (a) irradiating the top surface with scattered light generated by a scattered light source disposed on a first lateral side of the substrate and disposed substantially perpendicular to the primary flat, and (b) capturing at least one image using an imaging device disposed on a second lateral side of the substrate opposite the first lateral side.

[0024] In certain embodiments, the crystal material has a hexagonal structure, and generating at least one image includes (a) irradiating the top surface with scattered light generated by a scattered light source disposed on a first lateral side of the substrate and disposed within ±5 degrees from a direction perpendicular to the <112-0> direction of the hexagonal structure, and (b) capturing at least one image using an imaging device disposed on a second lateral side of the substrate opposite the first lateral side.

[0025] In certain embodiments, at least one subsurface laser damage pattern includes a first subsurface laser damage pattern and a second subsurface laser damage pattern formed after the first subsurface laser damage pattern, the first subsurface laser damage pattern includes a first plurality of substantially parallel lines and the second subsurface laser damage pattern, the lines of the second plurality of substantially parallel lines are dispersed among the lines of the first plurality of substantially parallel lines, and at least some of the lines of the second plurality of substantially parallel lines do not intersect any of the lines of the first plurality of substantially parallel lines.

[0026] In certain embodiments, each line of the second plurality of substantially parallel lines is disposed between different pairs of adjacent lines of the first plurality of substantially parallel lines.

[0027] In certain embodiments, each line of the first plurality of substantially parallel lines and each line of the second plurality of substantially parallel lines are within ±5 degrees from a direction perpendicular to the <11-20> direction of the hexagonal structure of the crystal material and are substantially parallel to the surface of the substrate.

[0028] In certain embodiments, at least one subsurface laser damage pattern includes a first subsurface laser damage pattern and a second subsurface laser damage pattern formed after the first subsurface laser damage pattern. At least one group of a plurality of substantially parallel lines includes a first plurality of substantially parallel lines and a second plurality of substantially parallel lines. The lines of the first plurality of substantially parallel lines are non-parallel to the lines of the second plurality of substantially parallel lines. The angular direction of the lines of the second plurality of substantially parallel lines differs from the angular direction of the lines of the first plurality of substantially parallel lines by 10 degrees or less. At least some of the lines of the second plurality of substantially parallel lines do not intersect any of the lines of the first plurality of substantially parallel lines.

[0029] In certain embodiments, at least one subsurface laser damage pattern further includes a third subsurface laser damage pattern formed after the second subsurface laser damage pattern. At least one group of a plurality of substantially parallel lines further includes a third plurality of substantially parallel lines. At least one group of cracks includes first, second, and third pluralities of cracks. The first subsurface laser damage pattern forms a first plurality of cracks that propagate laterally outward from the lines of the first plurality of substantially parallel lines within the substrate. The second subsurface laser damage pattern forms a second plurality of cracks that propagate laterally outward from the lines of the second plurality of substantially parallel lines within the substrate. The second plurality of cracks do not connect with the first plurality of cracks. The third subsurface laser damage pattern forms a third plurality of cracks that propagate laterally outward from the lines of the third plurality of substantially parallel lines within the substrate. At least some of the third plurality of cracks connect with at least some of the first plurality of cracks and at least some of the second plurality of cracks.

[0030] In certain embodiments, the method further includes detecting conditions indicating non-uniform doping of a crystalline material over at least a portion of the surface of a substrate, wherein the non-uniform doping includes a first doping region and a second doping region, and in response to the detection of conditions indicating non-uniform doping of a crystalline material, performing at least one of the following steps: (A) modifying the laser output so as to provide a first power level of laser radiation when forming a subsurface laser damage in the first doping region and a second power level of laser radiation when forming a subsurface laser damage in the second doping region, or (B) changing the average depth of subsurface laser damage formation on the substrate when forming a subsurface laser damage in either the first doping region or the second doping region.

[0031] In certain embodiments, the method further includes fragmenting the crystalline material substantially along at least one subsurface laser damage pattern such that first and second crystalline material portions are produced, each having a reduced thickness compared to the substrate but substantially the same length and width as the substrate.

[0032] In certain embodiments, the substrate comprises silicon carbide. In certain embodiments, the substrate comprises an ingot having a diameter of at least 150 mm.

[0033] In another aspect, the present disclosure relates to a material processing apparatus comprising a laser processing station configured to process a substrate of a crystalline material, the laser processing station comprising: a laser configured to form a subsurface laser damage region inside the substrate; at least one translational stage configured to perform relative movement between the laser and the substrate; a scattering light source configured to illuminate the top surface of the substrate, the scattering light source being positioned on a first lateral side of the substrate; and at least one of the top surfaces of the substrate The imaging device is configured to generate an image and is configured to be positioned on the second lateral side of a substrate opposite to the first lateral side.

[0034] In certain embodiments, the substrate has a substantially circular edge with a primary flat, and the scattering light source is positioned substantially perpendicular to the primary flat on the first lateral side of the substrate.

[0035] In certain embodiments, the crystalline material has a hexagonal structure, and the scattering light source is positioned on the first transverse side of the substrate, within ±5 degrees from a direction perpendicular to the <11-20> direction of the hexagonal structure.

[0036] In certain embodiments, the material processing apparatus further comprises a computing device configured to analyze at least one image to identify conditions indicating the presence of crack-free areas within the substrate.

[0037] In certain embodiments, the computing device is further configured to perform at least one of the following steps in response to analysis by the computing device: (i) causing relative movement between a laser and the substrate while supplying laser radiation focused into the interior of the substrate at least in crack-free areas to form auxiliary subsurface laser damage on the substrate for the purpose of forming a first thickness reduction portion of the substrate and to facilitate the formation of additional cracks in crack-free areas along or adjacent to a first mean depth position; or (ii) modifying the instruction set associated with the substrate for forming subsurface laser damage when creating subsurface laser damage patterns on the substrate at a second mean depth position and any subsequent mean depth positions for the purpose of forming a second and any subsequent thickness reduction portion of the substrate.

[0038] In certain embodiments, the analysis performed by the computing device includes quantifying the top area characteristics of one or more crack-free regions within the substrate and comparing the top area characteristics to at least one predetermined threshold area characteristic.

[0039] In a particular embodiment, at least one predetermined threshold area characteristic includes a first predetermined threshold area characteristic and a second predetermined threshold area characteristic, wherein the second predetermined threshold area characteristic is greater than the first predetermined threshold area characteristic, and the computing device is configured to control the material processing apparatus to perform step (ii) if the top area characteristic is at least the same size as the first predetermined threshold area characteristic, and the computing device is configured to control the material processing apparatus to perform step (i) if the top area characteristic is at least the same size as the second predetermined threshold area characteristic.

[0040] In certain embodiments, the material processing apparatus further comprises a memory configured to store a set of instructions associated with a substrate for forming subsurface laser damage on the substrate, and the memory is accessible by a computing device.

[0041] In certain embodiments, the material processing apparatus further comprises a crushing station configured to receive a substrate from a laser processing station.

[0042] In other embodiments, any of the embodiments described above and / or any of the various distinct embodiments and features described herein may be combined to obtain additional advantages. Unless otherwise indicated herein, any of the various features and elements disclosed herein may be combined. These may be combined with one or more other disclosed features and elements.

[0043] Other aspects, features, and embodiments of this disclosure will be more readily apparent from the subsequent disclosure and the accompanying claims.

[0044] The accompanying drawings incorporated herein and forming part thereof illustrate several aspects of this disclosure and serve to illustrate the principles of this disclosure together with this description. [Brief explanation of the drawing]

[0045] [Figure 1] The figure includes a first frame providing a perspective view of an ingot being received by a conventional wire saw tool and subjected to a wire sawing process, and a second frame providing perspective views of a plurality of wafers obtained by the wire sawing process. [Figure 2] This is the first perspective crystal plane view showing the coordinate system for hexagonal crystals such as 4H-SiC. [Figure 3] This is a second perspective crystal plane view for hexagonal crystals, showing a microslope nonparallel to the c-plane. [Figure 4A] This is an oblique wafer orientation diagram showing the orientation of a slightly sloped wafer with respect to the c-plane. [Figure 4B] Figure 4A is a simplified cross-sectional view of a slightly sloped wafer superimposed on a portion of an ingot. [Figure 5] This is a top view of an example SiC wafer, with the superimposed arrows indicating the crystal orientation. [Figure 6A] This is a schematic side elevation view of an on-axis ingot of crystalline material. [Figure 6B] Figure 6A is a schematic side elevation view of the ingot rotated 4 degrees, with an overlapping pattern for cutting the end of the ingot. [Figure 6C] This is a schematic side elevation view of the ingot after the end has been removed to provide an end face that is not perpendicular to the c-direction. [Figure 7] This is a schematic perspective view of a movable laser tool configured to focus laser radiation into the interior of a crystalline material to create subsurface damage. [Figure 8A] This figure provides an exemplary laser tool movement path for a crystalline material to form subsurface damage within the crystalline material. [Figure 8B]Includes superimposed arrows indicating the orientation of subsurface damage lines with respect to the [11-20] direction in the hexagonal structure of the crystalline material. [Figure 9] This is a schematic perspective view of the surface structure of a 4H-SiC crystal after crushing but before smoothing, showing off-axis or slightly sloped surfaces (relative to the c-axis). The crushed surface exhibits plateau and stepped sections. [Figure 10A-10D] This is a schematic cross-sectional view of the formation of subsurface laser damage on a crystalline material substrate by focusing laser radiation into the bare substrate, through the surface of a substrate supported by a carrier, through the carrier and adhesive layer into the substrate, and through the carrier into the substrate. [Figure 11A] The figure provides a top plan view of a crystalline material substrate, including dispersed first, second, and third subsurface laser damage patterns defined according to one embodiment, wherein each damage pattern includes a plurality of substantially parallel lines perpendicular to the [11-20] direction (and substantially perpendicular to the primary substrate flat), and the laser damage patterns are combined to form a plurality of 3-line groups, which are separated from each other by inter-group spacings that exceed the spacing between adjacent lines in each 3-line group. [Figure 11B] Figure 11A is a schematic top view of the crystalline substrate under fabrication after the formation of the first subsurface laser damage pattern, illustrating the first set of cracks inside the substrate propagating transversely outward from the first set of substantially parallel lines. [Figure 11C] Figure 11B is a top plan view of the crystalline substrate during the formation of a second subsurface laser damage pattern following a first subsurface laser damage pattern, illustrating a second set of cracks inside the substrate that propagate laterally outward from a second set of substantially parallel lines but do not contact the first set of cracks. [Figure 11D]Figure 11C is a top plan view of the crystalline substrate during the formation of a third subsurface laser damage pattern following the first and second subsurface laser damage patterns, illustrating a third set of cracks within the substrate that propagate laterally outward from a third set of substantially parallel lines, connecting the first and second sets of cracks. [Figure 12] This is a schematic top plan view of a crystalline substrate, including dispersed first to third subsurface laser damage patterns defined according to an embodiment similar to the embodiment shown in Figure 11A, wherein each damage pattern includes a plurality of substantially parallel lines deflected by 3 degrees with respect to the direction perpendicular to the [11-20] direction along the substrate surface (and substantially perpendicular to the primary substrate flat), and the laser damage patterns are combined to form a plurality of 3-line groups, which are separated from each other by inter-group spacings that exceed the spacing between adjacent lines in each 3-line group. [Figure 13] This is a schematic top plan view of a crystalline material substrate containing dispersed first to fourth laser damage patterns, where all lines are parallel to each other and perpendicular to the [11-20] direction along the substrate surface (and substantially perpendicular to the primary substrate flat). [Figure 14] A schematic top plan view of a crystalline material substrate, comprising dispersed first, second, and third subsurface laser damage patterns defined according to one embodiment, wherein the first and second groups of lines are parallel to each other and perpendicular to the [11-20] direction along the substrate surface (and substantially perpendicular to the primary substrate flat), and the third group of lines is non-parallel to the first and second groups of lines but does not intersect with the lines of the first and second groups of lines within the substrate. [Figure 15]This is a schematic top plan view of a crystalline material substrate, including dispersed first, second, and third subsurface laser damage patterns defined according to one embodiment, wherein the first and second groups of lines are each parallel to each other and deflected by 3 degrees from a direction perpendicular to the [11-20] direction along the substrate surface (and substantially perpendicular to the primary substrate flat), and the third group of lines is perpendicular to the primary substrate flat but does not intersect with the lines of the first and second groups of lines within the substrate. [Figure 16] This is a schematic top plan view of a crystalline material substrate containing dispersed first, second, and third subsurface laser damage patterns defined according to one embodiment, wherein all laser damage lines are parallel to each other, and the spacing between groups of laser damage lines is not uniform in at least part of the substrate. [Figure 17] This is a schematic top view of a crystalline material substrate containing dispersed first, second, and third subsurface laser damage patterns drawn according to one embodiment, where all laser damage lines are parallel to each other, and the laser damage lines exhibit variability in group spacing, inter-group spacing, and group configuration. [Figure 18] This is a schematic top plan view of a crystalline material substrate including continuously formed first, second, and third subsurface laser damage patterns, drawn according to one embodiment, wherein the first and second groups of laser damage lines are parallel to each other, but the third group of laser damage lines is non-parallel to and intersects with the first and second groups of laser damage lines. [Figure 19] This is a schematic top view of a crystalline material substrate containing continuously formed first, second, and third subsurface laser damage patterns, where each group of laser damage lines contains parallel lines and each group of laser damage lines is non-parallel to each other group of laser damage lines. [Figure 20A] This is a top plan view of a crystalline substrate showing non-overlapping first, second, and third areas where laser damage regions may be formed. [Figure 20B]Figure 20A is a top view of the crystalline substrate after the first multiple subsurface laser damage regions have been formed within the first to third areas. [Figure 20C] Figure 20B is a top view of the crystalline substrate after forming a second set of subsurface laser-damaged regions within the first to third areas. [Figure 20D] Figure 20C is a top view of the crystalline substrate after forming a third set of subsurface laser-damaged regions within the first to third areas. [Figure 21] This is a schematic top view of a holder for a laser processing apparatus, which is arranged to hold four substrates on which subsurface laser damage can be formed using one or more lasers. [Figure 22A] This is a schematic top view of a single substrate being processed with split laser beams to simultaneously form subsurface laser damage following a first subsurface laser damage pattern in two portions of the substrate. [Figure 22B] This is a schematic top view of two substrates being processed with split laser beams to simultaneously form subsurface laser damage following a first subsurface laser damage pattern on both substrates. [Figure 23A] This is a schematic cross-sectional view of a crystalline material substrate, including a first subsurface laser damage pattern centered on a first depth. [Figure 23B] Figure 23A is a schematic cross-sectional view of the substrate after the formation of a second subsurface laser damage pattern, which is centered on a second depth and aligned with the first subsurface laser damage pattern, and the vertical extensions of the first and second damage patterns overlap. [Figure 24A] This is a perspective view of a SiC wafer after it has been separated from a sapphire support bonded with a thermoplastic adhesive, according to the method described herein. [Figure 24B] Figure 24A is an oblique view of the sapphire support from which the SiC wafer has been separated. [Figure 24C]This is a partially inverted version of the SiC wafer photograph in Figure 24A, with the colors partially reversed to enhance the contrast between the central doping ring and the outer annular portion of the wafer. [Figure 24D] Figure 24C shows an image where a dotted oval is used to indicate the boundary between the central doping ring and the outer annular portion of the wafer. [Figure 25] This is a schematic lateral cross-sectional view of a SiC ingot grown on a seed crystal, showing a cylindrical doping region that extends upward from the seed crystal along the central part of the ingot across its entire thickness. [Figure 26] This is a schematic top view of a SiC wafer obtained from the SiC ingot shown in Figure 25, along the thin cross-sectional portion shown. [Figure 27] This is a schematic lateral cross-sectional view of a SiC ingot grown on a seed crystal, showing a frustoconical doping region that extends upward from the seed crystal along the central part of the ingot across its entire thickness. [Figure 28] This is a schematic lateral cross-sectional view of a SiC ingot grown on a slightly inclined (e.g., off-cut) seed crystal, showing a frustoconical doping region extending upward from the seed crystal at a point off-center from the seed crystal, across the entire thickness of the ingot. [Figure 29] This is an oblique view of the Si surface of a SiC wafer separated from an ingot by a process including the formation of subsurface laser damage and subsequent separation. The inset area (upper right) depicts a fragment of the SiC wafer, including the edge depicted in the subsequent scanning electron microscope (SEM) image. [Figure 30A] Figure 29 is a 45x magnified SEM image of a portion of the SiC wafer fragment, taken at a tilt angle of 15 degrees. The superimposed arrows indicate the orientation of the [1-100] and [11-20] crystal planes. [Figure 30B] This is a 1,300x magnified SEM image of a portion of the SiC wafer fragment shown in Figure 29, taken at a tilt angle of 15 degrees. [Figure 30C] This is a 350x magnified SEM image of a portion of the SiC wafer fragment shown in Figure 29, taken at a tilt angle of 15 degrees. [Figure 30D] This is a 100x magnified SEM image of a portion of the SiC wafer fragment shown in Figure 29, taken at a tilt angle of 2 degrees. [Figure 30E] This is a 1,000x magnified SEM image of a portion of the SiC wafer fragment shown in Figure 29, taken at a tilt angle of 2 degrees. [Figure 31A] Figure 29 is a confocal laser scanning microscope image of a small central portion of a SiC wafer, where the locations of "trenches" formed by laser scanning are marked by superimposed crosshairs. [Figure 31B] Figure 31A shows a plot of a portion of the surface profile of the SiC wafer. [Figure 32A] Figure 29 is a confocal laser scanning microscope image of a larger portion of the SiC wafer near the top (at the time of imaging), with the locations of the "trenches" formed by the laser scanning marked by superimposed crosshairs. [Figure 32B] Figure 32A is a plot of the surface profile of the portion of the SiC wafer closest to the top. [Figure 33A] Figure 29 is a confocal laser scanning microscope image of a larger portion of the SiC wafer near the bottom (at the time of imaging), with the locations of the "trenches" formed by the laser scanning marked by superimposed crosshairs. [Figure 33B] Figure 33A is a plot of the surface profile of the portion of the SiC wafer closest to the bottom. [Figure 34A] This is a schematic lateral cross-sectional view of a robust carrier to which an adhesive material is bonded to the surface. [Figure 34B] Figure 34A is a schematic cross-sectional view of an assembly including a robust carrier and adhesive material, bonded to a crystalline material substrate having a subsurface laser-damaged region adjacent to the adhesive material lip. [Figure 34C] Figure 34B is a schematic cross-sectional view of the assembly, in which the surface of the robust carrier is positioned on a cooling device in the form of a water-cooled chuck. [Figure 34D]This is a schematic cross-sectional view of the majority of a crystalline substrate separated from a bonded assembly (on a water-cooled chuck), including the portion of crystalline material removed from the robust support and substrate after fragmentation of the crystalline material along the subsurface laser-damaged area. [Figure 34E] Figure 34D is a schematic cross-sectional view of the joined assembly after removal from the water-cooled chuck, showing residual laser damage along the upper-facing surface. [Figure 34F] This is a schematic cross-sectional view of a portion of a crystalline material supported by a heated vacuum chuck, where the rigid carrier and adhesive material are laterally translated away from the crystalline material portion after the adhesive material has been thermally softened and released. [Figure 35] This is a schematic cross-sectional view of a subsurface laser-damaged crystalline material bonded to a rigid support, with the crystalline material and support located in a liquid tank of an ultrasonic generator. [Figures 36A-36C] This is a schematic cross-sectional view illustrating a step for fracturing a subsurface laser-damaged crystalline material, which includes applying a mechanical force close to one edge of the support to impart a bending moment to at least a portion of the support. [Figure 37A-37O] This is a schematic cross-sectional diagram illustrating the steps of a device wafer splitting process, according to which a thick wafer is crushed from a crystalline material, at least one epitaxial layer is grown on the thick wafer, and the thick wafer is crushed to form first and second bonded assemblies, each containing a carrier and a thin wafer separated from the thick wafer, the first bonded assembly containing at least one epitaxial layer as part of a functional semiconductor-based device. [Figure 38] This flowchart schematically illustrates each step of the process, from creating subsurface laser damage to bonding a rigid support to an ingot of crystalline material (e.g., SiC), then laser-cutting the bonded assembly containing the support and a portion of the crystalline material, to further processing the bonded assembly to form an epitaxial layer on a device wafer, and finally returning the ingot and rigid support to the beginning of the process. [Figure 39]Figure 38 is a schematic cross-sectional view of a portion of a crystalline material substrate showing subsurface laser damage, where superimposed dotted lines identify the expected kerf loss material areas that may result from laser damage and subsequent surface processing (e.g., grinding and planarization). [Figure 40] This is a schematic diagram of a material processing apparatus according to one embodiment, which includes a laser processing station, a material crushing station, a plurality of parallel-arranged rough grinding stations, a precision grinding station, and a CMP station. [Figure 41] This is a schematic diagram of a material processing apparatus according to one embodiment, which is similar to the embodiment shown in Figure 40, but in which the edge grinding station is located between the precision grinding station and the rough grinding station. [Figure 42] This is a schematic diagram of a material processing apparatus according to one embodiment, which includes a laser processing station, a material crushing station, a plurality of parallel-arranged rough grinding stations, a precision grinding station, a surface coating station, an edge grinding station, a coating removal station, and a CMP station. [Figure 43A] This is a schematic lateral cross-sectional view of a first apparatus for holding an ingot having an end face not perpendicular to a side wall, according to one embodiment. [Figure 43B] This is a schematic lateral cross-sectional view of a second device for holding an ingot having an end face not perpendicular to a side wall, according to one embodiment. [Figure 44] This is a schematic side cross-sectional view of a conventional laser focusing device, which focuses an incoming horizontal beam with a lens to form an outgoing beam having a beam waist pattern with minimum width at a downstream position corresponding to the focal length of the lens. [Figure 45] This is a schematic lateral cross-sectional view of a vertically oriented focused laser beam exhibiting a beam waist within a crystalline material, illustrating resolution threshold points located at different perpendicular positions to the beam waist. [Figure 46A] This is a plot of laser power versus the sequential wafer identification of wafers obtained from SiC ingots, showing the increase in laser power along with the wafer identification number. [Figure 46B] This is a plot of laser power versus the sequential wafer identification of wafers obtained from SiC ingots, showing the increase in laser power along with the wafer identification number. [Figure 46C] This is a plot of laser power versus the sequential wafer identification of wafers obtained from SiC ingots, showing the increase in laser power along with the wafer identification number. [Figure 47] This plot shows resistivity (Ohm-cm) versus slice numbers of 50 wafers produced from a SiC ingot. The superimposed polynomial fit shows that resistivity decreases with slice number, and a higher slice number indicates that the ingot is closer to the seed crystal from which it was grown (e.g., via a physical vapor transport (PVT) process). [Figure 48] Figure 47 plots laser power (watts) against the resistivity of wafers produced from SiC ingots. The superimposed polynomial fit shows that as the resistivity value increases, the laser power required to achieve it decreases. [Figure 49A] This is a schematic side cross-sectional view of a scattering light source and imaging device positioned in close proximity to the substrate within a laser processing station. [Figure 49B] This is a top view of the scattering light source and imaging device positioned in close proximity to the substrate within the laser processing station. [Figure 50A] This is an image of the top surface of a crystalline SiC substrate with subsurface laser damage, showing areas of different colors and irregularly shaped dark areas corresponding to crack-free areas within the substrate. [Figure 50B] This is a schematic diagram of the substrate representation, showing the irregularly shaped dark areas of Figure 50A within the dotted line area that substantially corresponds to the boundary between areas of different colors on the top surface of the substrate in Figure 50A. [Figure 50C] These are enlarged views of the irregularly shaped dark areas in Figures 50A and 50B, with rectangular boxes added around each individual region. [Figure 51]This is a schematic diagram of a material processing apparatus according to one embodiment, which includes a laser processing station comprising a laser, at least one translational stage, a scattering light source configured to illuminate the top surface of a substrate, and an imaging device configured to generate at least one image of the top surface of a substrate. [Figure 52] This is a flowchart illustrating the steps of a first crystalline material processing method, which includes generating an image of the top surface of a substrate with subsurface laser damage; analyzing the image to identify the presence of conditions indicating one or more crack-free regions; comparing one or more properties of the crack-free regions to first and second thresholds; and taking action in response to the comparison to increase the reliability of producing a substrate portion (e.g., a wafer) from the substrate (i.e., (A) performing an additional laser pass at substantially the same depth position by optionally adjusting one or more laser parameters to form auxiliary laser damage, and / or (B) adjusting one or more laser parameters to form subsurface laser damage at second and subsequent depth positions). [Figure 53] This is a flowchart illustrating the steps of a second crystalline material processing method, which includes generating an image of the top surface of a substrate with subsurface laser damage; analyzing the image to quantify the top area characteristics of one or more crack-free regions; comparing the top area characteristics with first and second threshold area characteristics; and taking action in response to the comparison (i.e., performing an additional laser pass at the same depth position and / or adjusting the output for subsurface laser damage at the next depth position) to improve the reliability of producing a substrate portion (e.g., a wafer) from the substrate. [Figure 54] This is a generalized schematic diagram of a computer system that may be included in any component of the system or method disclosed herein. [Modes for carrying out the invention]

[0046] This disclosure relates to a method and apparatus for processing a crystalline material substrate, and takes various forms. The crystalline material processing method includes generating a subsurface laser damage area in an area at a first mean depth position of the crystalline material to facilitate the formation of cracks in the substrate that propagate outward from a subsurface laser damage pattern; imaging the top surface of the substrate; analyzing the image to identify conditions indicating the presence of crack-free areas within the substrate; and performing one or more actions in response to the analysis (e.g., when appropriate conditions are met). One possible action includes forming auxiliary subsurface laser damage at the first mean depth position to facilitate the formation of additional cracks in crack-free areas for the purpose of forming a first thickness reduction portion of the substrate (e.g., a first wafer). Another possible action includes modifying the set of instructions to perform subsequent laser damage formation (at a second or subsequent mean depth position to form a second and any thickness reduction portion of the substrate) without necessarily forming additional damage at the first mean depth position. The laser damage facilitates subsequent fracture of the substrate to produce multiple thickness reduction portions of the substrate.

[0047] In certain embodiments, the analysis includes quantifying the top area characteristics of one or more crack-free regions within the substrate and comparing the top area characteristics to at least one predetermined threshold area characteristic. In certain embodiments, if the first threshold area characteristic is exceeded, the average laser power is progressively increased in the next laser damage formation step (i.e., at the second or next average depth position for forming second and subsequent reduced-thickness substrate portions), without necessarily involving the formation of additional damage at the first average depth position. Instead of increasing the laser power, or in addition to this, the laser focusing depth relative to the top surface may be changed and / or several laser damage formation steps may be modified in the instruction set for performing the subsequent second laser damage formation step. If a larger second threshold area characteristic is exceeded (suggesting that the crack-free region may be large enough to prevent fracture), an auxiliary subsurface laser damage pattern is formed at the first mean depth position to assist at least one subsurface laser damage pattern and to promote the formation of additional cracks in the crack-free region along or adjacent to the first mean depth position, for the purpose of forming a substrate thickness reduction region. Damage can be formed before the substrate is removed from the laser processing station, thereby avoiding the extra step of removing and reinstalling the substrate and improving the throughput of the laser processing station.

[0048] In additional embodiments, the Disclosure relates to a material processing apparatus comprising a laser processing station configured to process a substrate of crystalline material, the laser processing station comprising: a laser configured to form subsurface laser-damaged areas within the substrate; at least one translational stage configured to perform relative movement between the laser and the substrate; a scattering light source configured to illuminate the top surface of the substrate, the scattering light source being positioned to be located on a first transverse side of the substrate; and an imaging device configured to produce at least one image of the top surface of the substrate, the imaging device being positioned on a second transverse side of the substrate opposite to the first transverse side. Such an apparatus makes crack-free areas adjacent to subsurface laser damage within the substrate visible on its surface as dark (e.g., black or nearly black) spots on the top surface of the substrate. Such an apparatus also makes areas with different degrees of crack formation between subsurface laser-damaged areas exhibit different colors on the top surface of the substrate. The dark spots typically first appear within the area of ​​a facet (corresponding to a doping ring), so in certain embodiments, the area of ​​a facet can be distinguished.

[0049] As already mentioned, variations in material and / or optical properties within thick substrates such as SiC ingots, and between different ingots of the same composition, make it difficult to easily and reproducibly produce wafers of uniform thickness by laser processing while avoiding unwanted material loss. The applicant has found that when wafers are continuously formed from a SiC ingot by subsurface laser damage formation and subsequent fracture, the laser power needs to be increased as damage formation progresses at deeper levels to enable successful fracture. (Again, when forming multiple wafers from a SiC ingot, the initial wafers distal to the seed crystal can be successfully cut after laser damage formation produced at a lower average laser power, but as the growth position of the wafers to be cut approaches the seed crystal, progressively higher laser power levels are required for the laser damage used to cut the next wafer.) This behavior is thought to be primarily driven by changes in bulk light absorption, but may also be influenced by other changes in the crystal lattice. One theoretical solution to this problem is to simply use high laser power at each successive depth position when forming subsurface damage. However, in this case, unwanted material loss is thought to occur when damage is created "early" in the ingot (e.g., at the first few depth positions distal to the seed crystal). Furthermore, significant variations in wafer thickness are thought to occur due to variability in both the damage depth and the resolution point relative to the laser beam waist (as a result of the focal length of the beam focusing optical system). Constantly trying to adjust the wafer thickness is not only impractical but also lacks accuracy due to measurement inaccuracies caused by the rough surface created by the laser separation process and the relationship between laser depth and required laser power.

[0050] Before detailing the specific features of the above-described method and apparatus (specific embodiments are shown in relation to Figures 45 to 51), we will introduce the apparatus and method for processing crystalline material substrates.

[0051] [Technical Terms and Definitions] Terms such as "first," "second," etc., may be used in this specification to describe various elements, but it should be understood that these elements are not limited by those terms. These terms are used solely to distinguish one element from another. For example, without departing from the scope of this disclosure, the first element may be referred to as the second element, and similarly, the second element may be referred to as the first element. As used herein, the term "and / or" includes any combination of one or more of the related enumerated things.

[0052] When an element, such as a layer, region, or substrate, is described as existing "on top of" or extending "upwards" from another element, it will be understood that the element may exist directly on or extend directly onto that other element, or there may be an intervening element. In contrast, when an element is described as existing "directly on top of" another element or extending "directly upwards" from another element, there is no intervening element. Again, when an element, such as a layer, region, or substrate, is described as existing "on top of" or extending "upwards" from another element, it will be understood that the element may exist directly on or extend directly onto that other element, or there may be an intervening element. In contrast, when an element is described as existing "directly on top of" another element or extending "directly onto" another element, there is no intervening element. Furthermore, when an element is described as being "connected" or "bonded" to another element, it will be understood that the element may be directly connected to or bonded to that other element, or there may be an intervening element. In contrast, when one element is described as being "directly connected" or "directly coupled" to another element, there is no intervening element.

[0053] In this specification, relative terms such as “below,” “up,” “top,” “upper,” “lower,” “horizontal,” or “vertical” may be used to describe the relationship between one element, layer, or region and another, as shown in the figures. It will be understood that these terms, and the terms discussed above, are intended to encompass various orientations of the device in addition to the orientation depicted in the figures.

[0054] The technical terms used herein are intended solely to describe specific embodiments and are not intended to limit the disclosure. Where used herein, the singular forms “a,” “an,” and “the” are intended to include the plural form unless the context explicitly indicates otherwise. It will be further understood that, where used herein, the terms “comprise,” “comprising,” “include,” and / or “including” indicate the presence of a mentioned feature, integer, step, operation, element, and / or component, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0055] Unless otherwise defined, all terms used herein (including technical and / or scientific terms) have the same meaning as those generally understood by those skilled in the art to which this disclosure belongs. It will be further understood that terms used herein should be interpreted in a way that is consistent with their meaning in the context of this specification and the related art, and not in an idealized or overly formal sense unless expressly defined herein.

[0056] As used herein, “substrate” means a crystalline material, such as a single-crystal semiconductor material, which is divisible into at least two thinner portions having substantially the same lateral dimensions (e.g., diameter, or length and width) as the substrate, and which has sufficient thickness to (i) be surface-processed (e.g., lapping and polishing) to support the epitaxial growth of one or more semiconductor material layers, and optionally (ii) stand upright when separated from a rigid support, and optionally includes an ingot or wafer. In certain embodiments, the substrate may have a substantially cylindrical shape and / or at least one or more of the following thicknesses They may have the same thickness: 300 μm, 350 μm, 500 μm, 750 μm, 1 mm, 2 mm, 3 mm, 5 mm, 1 cm, 2 cm, 5 cm, 10 cm, 20 cm, 30 cm, or more. In certain embodiments, the substrate may include a thicker wafer that can be divided into two thinner wafers. In certain embodiments, the substrate may be part of a thicker wafer on which one or more epitaxial layers (optionally coupled with one or more metal contacts) are disposed as part of a device wafer having a plurality of electrically operating devices. The device wafer may be divided according to aspects of the present disclosure to obtain a thinner device wafer and a second thinner wafer on which one or more epitaxial layers (optionally coupled with one or more metal contacts) may be subsequently formed. In certain embodiments, the substrate may have a diameter of 150 mm or more, or 200 mm or more. In certain embodiments, the substrate may include 4H-SiC having a diameter of 150 mm, 200 mm, or greater, and a thickness in the range of 100 to 1000 microns, or 100 to 800 microns, or 100 to 600 microns, or 150 to 500 microns, or 150 to 400 microns, or 200 to 500 microns, or any other thickness range, or any other thickness value specified herein.

[0057] The terms “first mean depth position,” “second mean depth position,” and “next mean depth position,” as used herein, refer to depth positions within the substrate (e.g., horizontal plane) measured from the first top surface of the substrate for forming the thickness reduction portion of the substrate. For example, the first mean depth position may correspond to the subsurface laser damage position for forming the first wafer from the ingot, the second mean depth position may correspond to the subsurface laser damage position for forming the second wafer from the ingot, and so on. In certain embodiments, each thickness reduction portion obtained from the substrate has the same or substantially the same thickness. The term “mean depth position” is used instead of depth position because it is recognized that in certain embodiments, there may be small differences in laser focusing depth between processes for forming a laser damage pattern intended for forming a single thickness reduction portion of a substrate (e.g., a single wafer), or even within a single process, and such small differences are preferably in the range of 1 to 10 microns, or 2 to 8 microns, or 2 to 6 microns. This should be distinguished from the much larger difference between the first average depth position and the second average depth position, which is typically within the range of at least 100 microns (or at least 150 microns, 200 microns, 300 microns, 400 microns, 500 microns, or more).

[0058] The embodiments described below represent the information necessary to enable a person skilled in the art to carry out the embodiments and represent the best mode of carrying out the embodiments. A person skilled in the art will understand the concepts of this disclosure and recognize uses of these concepts that are not detailed herein by reading the following description in conjunction with the accompanying drawings. It should be understood that these concepts and uses are within the scope of this disclosure and the accompanying claims.

[0059] [material] The methods disclosed herein can be applied to substrates of various crystalline materials, both single-crystal and polycrystalline. In certain embodiments, the methods disclosed herein can utilize cubic, hexagonal, and other crystalline structures and may apply to crystalline materials having on-axis and off-axis crystal orientations. In certain embodiments, the methods disclosed herein can be applied to semiconductor materials and / or wide-bandgap materials. Examples of materials, but not limited to, include Si, GaAs, and diamond. In certain embodiments, such methods may utilize single-crystal semiconductor materials having a hexagonal structure, such as 4H-SiC, 6H-SiC, or Group III nitride materials (e.g., GaN, AlN, InN, InGaN, AlGaN, or AlInGaN). In the various exemplary embodiments described below, we generally refer to SiC or specifically to 4H-SiC, but it will be understood that any suitable crystalline material can be used. Among the various SiC polytypes, the 4H-SiC polytype is particularly attractive for power electronics devices due to its high thermal conductivity, wide bandgap, and isotropic electron mobility. Bulk SiC can be grown on-axis (i.e., without intentional angular deflection from its c-plane, suitable for forming undoped or semi-insulating materials) or off-axis (typically away from the growth axis, such as the c-axis, at a non-zero angle typically within the range of 0.5 to 10 degrees (or a lower range such as 2 to 6 degrees or another lower range), which may be suitable for forming N-doped or highly conductive materials). The embodiments disclosed herein are applicable to on-axis and off-axis crystalline materials, as well as doped and unintentionally doped crystalline semiconductor materials. Doped semiconductor materials (e.g., N-doped SiC) exhibit some infrared absorption and therefore require the use of higher laser power than undoped materials to impart subsurface laser damage. In certain embodiments, the crystalline material may include single-crystal materials and further include single-crystal semiconductor materials. Specific embodiments disclosed herein can utilize on-axis 4H-SiC or micro-slope (off-axis) 4H-SiC having an offcut of 1 to 10 degrees, or in the range of 2 to 6 degrees, or about 4 degrees.

[0060] In certain embodiments of this specification, doped or undoped SiC substrates, such as SiC ingots (also known as booleans), can be used, which can be grown by physical vapor transport (PVT) or other conventional ingot fabrication methods. When doped SiC is used, such doping makes the SiC properties N-type or semi-insulating. In certain embodiments, N-type SiC ingots are intentionally doped with nitrogen. In certain embodiments, N-type SiC ingots have resistivity values ​​in the range of 0.015 to 0.028 Ohm-cm. In certain embodiments, SiC ingots may have resistivity values ​​that vary with vertical position, in which case different substrate portions (e.g., wafers) will have different resistivity values, which may be due to variations in bulk doping levels during ingot growth. In certain embodiments, SiC ingots may have doping levels that vary horizontally, from higher doping areas closer to the center of the ingot to lower doping levels closer to its lateral edges. Due to variations in ingot doping and resistivity with respect to horizontal and vertical positions, it may be necessary to adjust the laser damage formation parameters to form various thickness reduction portions (e.g., wafers) of a substrate (e.g., ingot) and / or during the formation of a single thickness reduction portion of the substrate. In certain embodiments, resistivity is highest near the exposed surface of the ingot and lowest near the growth species. The decrease in resistivity corresponds to increased doping and increased laser absorption.

[0061] Figures 6A and 6C schematically show on-axis and off-axis crystalline substrates in the form of ingots that may be used with the methods disclosed herein. Figure 6A is a schematic side elevation view of an on-axis ingot 15 of a crystalline material having a first end face 16 and a second end face 17 perpendicular to the c direction (i.e., the

[0001] direction of a hexagonal structure material such as 4H-SiC). Figure 6B is a schematic side elevation view of the ingot 15 of Figure 6A rotated by 4 degrees, with an overlapping pattern 18 (shown by dotted lines) for cutting and removing the ends of the ingot 15 adjacent to the end faces 16,17. Figure 6C is a schematic side elevation view of an off-axis ingot 15A formed from the ingot 15 of Figure 6B after the ends have been removed to provide new end faces 16A,17A that are not perpendicular to the c direction. When laser radiation to a first depth is supplied through the end face 16 of the ingot 15 to form subsurface laser damage, a carrier (not shown) is bonded to the end face 16, the ingot 15 is fractured along the subsurface laser damage, and an on-axis wafer can then be formed. Conversely, if laser radiation to a first depth is supplied through the end face 16A of an off-axis ingot 15A to form subsurface laser damage, a support (not shown) may be bonded to the end face 16A, and the ingot 15A may be fractured along the subsurface laser damage, thereby forming an off-axis wafer.

[0062] [Subsurface laser damage formation] By processing a crystalline material substrate to form multiple patterns of subsurface laser damage, subsequent fracture of the substrate is facilitated to produce first and second crystalline material portions of the substrate with reduced thickness. A particular method involves dispersing multiple groups of substantially parallel lines of multiple subsurface laser damage patterns, wherein at least some of the lines of the second (e.g., next formed) multiple lines do not intersect with the lines of the first multiple lines. A particular method involves forming first and subsequent subsurface laser damage patterns in a crystalline material substrate, each containing another plurality of substantially parallel lines, wherein the lines of the first and subsequent plurality of substantially parallel lines are non-parallel to each other, the difference in the angular direction of the lines of the subsequent plurality of substantially parallel lines from the angular direction of the lines of the first plurality of substantially parallel lines is 10 degrees or less, and at least some of the lines of the subsequent plurality of substantially parallel lines do not intersect with any of the lines of the first plurality of substantially parallel lines. A particular method includes forming a first subsurface laser damage pattern substantially centered at an initial depth within the crystalline material of the substrate, and forming a second subsurface laser damage pattern substantially centered at a subsequent depth (different from the first depth) within the substrate, wherein the second subsurface laser damage pattern is substantially aligned with the first subsurface laser damage pattern, and the vertical extensions of at least a portion of the first and subsurface laser damage patterns overlap.

[0063] The continuous formation of subsurface laser damage patterns distributed, dispersed, or alternating across the crystalline material is considered to beneficially maintain sufficient stress within the crystalline material to facilitate subsequent material fracturing using the methods herein, while enabling high throughput of the laser tool associated with moderate material damage and correspondingly low kerf loss. To facilitate fracturing along the laser damage lines, it would be in principle simple to scan almost the entire crystalline material using high laser power. While such a technique can reliably separate thin layers of crystalline material from a bulk substrate (e.g., an ingot), high laser power increases material damage, and considerable surface processing (e.g., grinding and planarization) tends to be required to remove the damage. Closer spacing between laser damage lines helps to facilitate fracturing, but at the cost of significantly reducing the throughput of the laser processing tool. Conventional methods for forming subsurface laser damage include forming subsurface laser damage lines in an advancing direction across the crystalline material, then performing lateral relative indexing between the material and the laser, then forming subsurface laser damage lines in a receding direction, then performing lateral indexing in the same lateral direction, and so on. Such methods generally require higher laser power or closer spacing between the continuously formed laser damage lines, which tends to reduce throughput or impart greater damage, resulting in increased kerf loss due to the need to remove additional material from the laser-processed surface to remove the laser damage. This conventional method does not involve the formation of a first distributed subsurface laser damage pattern (e.g., including the formation of a first plurality of laser damage regions across multiple non-overlapping areas of the substrate) and the subsequent formation of a second distributed subsurface laser damage pattern (e.g., including the formation of a second plurality of laser damage regions across the same plurality of non-overlapping areas of the substrate), wherein the second subsurface laser damage pattern is alternately arranged or dispersed between the first subsurface laser damage pattern.

[0064] Various embodiments disclosed herein enable high throughput for laser tools. The concern is to facilitate reliable separation of a thin layer of crystalline material (e.g., a wafer) from a substrate without excessively high laser power, while achieving low kerf loss. Certain embodiments of this specification include forming an initial distributed subsurface laser damage pattern on a crystalline material substrate (e.g., across each of several non-overlapping areas of the substrate), and then forming at least one subsequent distributed subsurface laser damage pattern on the same substrate (e.g., across each of the same several non-overlapping areas), wherein at least a portion (e.g., lines) of the at least one subsequent laser damage pattern is located within the gaps between the laser damage lines of the initial laser damage pattern, thereby providing a dispersed or alternating subsurface laser damage pattern. In certain embodiments, at least some (or all) of the laser damage lines of the at least one subsequently formed laser damage pattern do not intersect with the laser damage lines of the initial subsurface laser damage pattern. It is believed that the non-intersection of the laser damage patterns can beneficially avoid the diffusion of localized stress. In certain embodiments, the formation of first and second dispersed subsurface laser damage patterns is carried out in such a manner that localized subsurface cracks do not propagate between them, but the application of a third (or subsequent) dispersed subsurface laser damage pattern allows localized subsurface cracks to propagate and bond in a substantially continuous manner across the entire internal plane of the crystalline material substrate, thereby facilitating subsequent fracturing along the laser-damaged area using the technique disclosed herein. The formation of dispersed subsurface laser damage according to the method described herein has been shown to enable reliable separation of thin layers of crystalline material from the substrate with fewer laser damage lines per layer to be removed, and to beneficially achieve increased laser tool throughput while achieving low levels of laser damage (enabling low kerf loss).

[0065] In various embodiments, we refer to laser subsurface damage including lines oriented with respect to the crystalline structure of the substrate. In certain embodiments, the substrate comprises a crystalline material having a hexagonal structure, and the laser damage lines are oriented perpendicular to the <11-20> direction of the hexagonal structure, or within ±5 degrees from the direction perpendicular thereto, and parallel or substantially parallel to the surface of the substrate (e.g., within ±5 degrees, ±3 degrees, or ±1 degree from thereto). While primary flats on conventional 4H-SiC wafers are intended to be oriented parallel to the <11-20> direction of the hexagonal structure, primary flats may not be truly parallel to such a direction due to manufacturing variability. Various SiC wafer manufacturers provide published specifications regarding the orientation of primary flats within ±5 degrees from the direction parallel to the <11-20> direction of the hexagonal structure. Therefore, it is preferable to use X-ray diffraction (XRD) data rather than wafer flat alignment to determine the appropriate laser orientation for the formation of subsurface laser damage.

[0066] In this art, tools for forming laser subsurface damage in crystalline materials are known and commercially available from various suppliers, such as DISCO Corporation (Tokyo, Japan). Such tools allow laser radiation to be focused into the interior of a crystalline material substrate, enabling lateral movement of the laser relative to the substrate. Typical laser damage patterns in this art involve the formation of parallel lines spaced laterally apart from one another at a certain depth within the crystalline substrate. Parameters such as focusing depth, laser power, translational velocity, and subsurface damage line spacing can be adjusted to impart laser damage, but adjusting certain factors involves trade-offs. Increasing laser power tends to impart greater subsurface damage, which can improve the ease of fracturing (for example, by reducing the stress required to complete the fracturing). However, greater subsurface damage leads to greater surface irregularity along the surface exposed by the fracturing. As a result, additional processing may be required to smooth such surfaces sufficiently for subsequent processing (e.g., for integration into electronic devices), and this additional processing results in additional kerf loss. While reducing the lateral spacing between damage lines can improve the ease of shattering, doing so increases the number of parallel paths between the substrate and the laser, which reduces the tool's throughput.

[0067] Figure 7 is a schematic perspective view of an example of a laser tool 29 configured to focus laser radiation into the interior of a crystalline material 30 to form subsurface damage 40. The crystalline material 30 includes an upper surface 32 and an opposite lower surface 34, and the subsurface damage 40 is formed inside the crystalline material 30 between the upper surface 32 and the lower surface 34. The laser radiation 36 is focused using a lens assembly 35 to form a focused beam 38, whose focal point is inside the crystalline material 30. Such laser radiation 36 can be pulsed at any preferred frequency and beam intensity (typically in the range of nanoseconds, picoseconds, or femtoseconds) and has a wavelength less than the band gap of the crystalline material 30 so that the laser radiation 36 can be focused to a target depth below the surface of the crystalline material 30. At the focal point, the beam size and short pulse width result in an energy density high enough to produce a highly localized absorption that forms the subsurface damage. One or more properties of the lens assembly 35 can be modified to adjust the focal point of the focused beam 38 to a desired depth within the crystalline material 30. As schematically shown by the dotted line 44, a relative lateral movement (e.g., lateral translation) can be made between the lens assembly 35 and the crystalline material 30 to propagate the subsurface damage 40 in a desired direction. Such lateral movement may be repeated in various patterns, including those described below.

[0068] Figures 8A and 8B provide exemplary laser tool migration paths to a crystalline material for forming subsurface damage within the crystalline material. In certain embodiments, the laser tool portion (including, for example, a lens assembly) can be configured to move while the crystalline material remains stationary, whereas in other embodiments, the laser tool portion can be held stationary while the crystalline material moves relative to the tool portion. Figure 8A shows a reversing linear scanning motion 46 in the y-direction suitable for forming a pattern of transversely spaced parallel lines of subsurface damage within a first crystalline material 45A. Figure 8B shows a linear scanning motion 48 in the y-direction across (and beyond) the entire surface of the crystalline material 45B (advancing slightly in the x-direction for each reversal in the y-direction) sufficient to form parallel subsurface laser damage lines distributed across the crystalline material 45B. As shown, the laser damage lines are perpendicular to the [11-20] direction of the hexagonal structure of the crystalline material 45B along the surface of the crystalline material 45B and substantially parallel to the surface of the crystalline material 45B.

[0069] Covering the entire surface of a crystalline material with a laser beam formed in the y-direction, while advancing in one direction in the x-direction each time there is a reversal in the y-direction, is sometimes referred to as a single pass of laser damage formation. In certain embodiments, laser processing of the crystalline material to form subsurface damage may be performed in two, three, four, five, six, seven, or eight passes, or any other suitable number of passes. Kerf loss can be reduced by increasing the number of passes with lower laser power. To achieve a desirable balance between material loss and process rate, a desirable number of laser subsurface damage formation passes has been found to be two to five passes, or three to four passes, before performing the fracturing step.

[0070] In certain embodiments, the lateral spacing between adjacent laser subsurface damage lines (whether formed in a single or multiple pass) may be in the range of 80 to 400 microns, 100 to 300 microns, or 125 to 250 microns. The lateral spacing between adjacent laser subsurface damage lines affects the laser processing time, ease of shattering, and effective laser damage depth (depending on c-plane orientation or misalignment).

[0071] It has been observed that forming subsurface laser damage lines in crystalline materials results in the formation of small cracks within the material that propagate outward from the laser damage lines (e.g., transversely outward). Such cracks appear to extend substantially or primarily along the C-plane. The length of such cracks appears to be functionally related to the laser power level (which can be calculated as the product of the energy per pulse multiplied by the pulse frequency). It has been observed that for adjacent laser subsurface damage lines separated by a certain distance, increasing the laser power when forming such laser subsurface damage lines tends to increase the ability of the cracks to connect or bond between them, which is desirable for promoting easy fracture.

[0072] If the crystalline material subjected to laser damage formation includes off-axis (i.e., non-c-plane) orientation (e.g., within the range of 0.5–10 degrees, 1–5 degrees, or other orientation deviations), such orientation deviations may affect the desired laser damage line spacing.

[0073] SiC substrates may include surfaces that are misaligned from the c-plane (for example, off-axis at an oblique angle to the c-plane). Off-axis substrates are sometimes called micro-slope substrates. After crushing such a substrate, the surface of the crushed material may include plateaus and steps (these can then be smoothed by surface processing such as grinding and polishing). Figure 9 is a schematic perspective view of the surface structure of an off-axis 4H-SiC crystal 50 (with angle A to the c-axis base plane) after crushing but before smoothing. The crushed surface exhibits steps 52 and plateaus 54 with respect to the c-axis base plane 56. In the case of a 4-degree off-axis surface, the steps theoretically have a height of approximately 17 microns relative to a plateau width of 250 microns. In the case of a 4H-SiC crystal with subsurface laser damage, the 250-micron gap between laser rays forms a plateau with a width of 250 microns. After crushing, the stepped surface is smoothed, planarized, and polished in preparation for epitaxial growth of one or more layers thereon.

[0074] When subsurface laser damage is formed in a crystalline material (e.g., SiC) and the subsurface laser damage lines are oriented away from the direction perpendicular to the substrate flat (i.e., not perpendicular to the [11-20] direction), such laser damage lines extend through multiple steps and plateaus in a manner similar to that of off-axis semiconductor materials. For the purposes of the following discussion, the term "off-axis laser subsurface damage lines" will be used to refer to laser subsurface damage lines not perpendicular to the [11-20] direction.

[0075] If the spacing between adjacent subsurface laser damage lines is too large, the fragmentation of the crystalline material will be hindered. If the spacing between adjacent subsurface laser damage lines is too small, the height of the steps tends to decrease, but the number of vertical steps increases, and an increase in the number of vertical steps usually requires a greater separation force to complete the fragmentation.

[0076] Reducing the distance between adjacent laser damage lines to an excessively small distance results in a decrease in yield and substantially increases processing time and cost. SiC decomposition requires a minimum laser energy threshold. If, at this minimum energy level, connected cracks are created between two laser lines spaced approximately 100 microns apart, then the benefit of narrowing the laser line spacing below this threshold is negligible in terms of reducing kerf loss.

[0077] The surface roughness of the crystalline material exposed by crushing can affect not only subsequent handling such as robotic vacuuming, but also the wear of grinding wheels, which are a major consumable expense. Roughness is influenced by both the spacing of subsurface laser damage lines and the orientation of such subsurface damage lines relative to the crystalline structure of the semiconductor material. Reducing the gap between subsurface damage lines simply reduces the height of possible steps. Off-axis laser subsurface damage lines By providing this feature, long, parallel steps that would otherwise be present in the laser-damaged area tend to break, which also helps mitigate at least some of the effects from the slope or curvature of the C-plane. When the laser beam is perpendicular to the flat surface of the substrate, the cleavage plane parallel to the laser beam along the C-plane extends approximately 150 mm from the flat to the curved end on the opposite side of the wafer. A slight deflection of the slope or curvature of the C-plane (which is common for SiC substrates) causes the plane to bounce as the fracturing propagates (forces plane Because it jumps, it can result in significant variability in the fractured surface. A drawback of providing off-axis laser subsurface damage lines is that such subsurface damage lines generally require increased laser power to form connected cracks between adjacent laser lines. For this reason, in certain embodiments, a good balance is provided by forming a combination of on-axis (perpendicular to the primary flat) and off-axis laser subsurface damage lines, which avoids excessive variability in the fractured surface without requiring excessively high laser power to form connected cracks between adjacent laser lines.

[0078] In certain embodiments, a laser having a wavelength of 1064 nm can be used to carry out the method disclosed herein, and the inventors have experience processing 4H-SiC. In certain embodiments, a wide range of pulse frequencies may be used, but pulse frequencies from 120 kHz to 150 kHz have been successfully utilized. A translational stage speed of 936 mm / s between the laser and the substrate being processed has been successfully utilized, but in certain embodiments, higher or lower translational stage speeds may be used by appropriately adjusting the laser frequency to maintain the desired overlap of laser pulses. The average laser power range for forming subsurface laser damage in doped SiC material is in the range of 3 W to 8 W, and 1 W to 4 W for undoped SiC material. The energy of the laser pulse can be calculated by dividing the power by the frequency. Laser pulse widths of 3 ns to 4 ns can be used, but other pulse widths can be used in other embodiments. In certain embodiments, numerical apertures (NA) of laser lenses in the range of 0.3 to 0.8 may be used. In embodiments aimed at processing SiC, considering the change in refractive index from air (approximately 1) to SiC (approximately 2.6), a large change in the refraction angle occurs inside the SiC material to be processed, making it important to correct the NA and aberrations of the laser lens to achieve the desired results.

[0079] One of the main factors contributing to kerf loss is subsurface laser damage located beneath the primary fracture region on the ingot surface. Generally, increased subsurface laser damage leads to increased kerf loss. One possible reason for increased subsurface laser damage is the inability to adequately compensate for the optical properties of the crystalline material. In certain embodiments, optical parameter optimization can be performed periodically (e.g., each time a crystalline material substrate (e.g., an ingot) is supplied to the laser tool) before the formation of subsurface laser damage on the substrate. Such optimization can utilize variable height adjustment to achieve an initial state in which the best focus point of the laser beam is formed on the upper surface of the crystalline material substrate, and thereafter, the aperture and / or compensation ring adjustment ring of the laser tool can be adjusted to correspond to the desired depth of subsurface laser damage formation in the crystalline material depending on the next state.

[0080] In certain embodiments, a crystalline material substrate may exhibit doping that varies in position (e.g., transversely and / or in diameter) across the main surface (e.g., plane) of the substrate. Dopant concentrations are typically higher in the central region of the SiC{0001} wafer, which is observable by the darker color of such regions. This increase in dopant concentration is due to increased impurity incorporation during facet growth. During the growth of a SiC{0001} ingot, {0001} facets appear near the center of the ingot. Rapid helical growth occurs on the {0001} facets, <0001> The crystal growth rate along the direction is relatively slow. Therefore, along the {0001} facet region The impurity concentration increases. The dopant concentration in the center of the SiC wafer (i.e., the facet region) may be 20% to 50% higher than the dopant concentration outside this region. The formation of doping ring regions with increased dopant concentration in SiC is shown in Figures 14A, 14C, and 14D. Such regions have higher laser absorptivity and a slightly altered refractive index, both of which affect the depth of focus of the laser radiation in the substrate. Various properties of the doping ring region can be compensated for by increasing the laser power used when focusing the laser radiation into the doping ring region compared to the power used when focusing the laser radiation into the material outside the doping ring region. In certain embodiments, the presence of conditions indicating heterogeneous doping of the crystalline material over at least a portion of the substrate surface can be detected to determine the presence of at least one first doping region and at least one second doping region. (Methods for detecting various doping conditions include, but are not limited to, interferometry, resistivity measurement, absorptiometry or reflectivity measurement, and other techniques known to those skilled in the art.) Subsequently, in response to the detection of conditions indicating heterogeneous doping of the crystalline material, the laser output can be modified during the formation of a subsurface laser damage pattern to provide a first average power laser emission when forming subsurface laser damage in a first doping region, and a second average power laser emission when forming subsurface laser damage in a second doping region, with the first and second average power levels being different from each other. Alternatively, or additionally, the depth of subsurface laser damage formation in the substrate (i.e., relative to the exposed surface of the substrate) can be modified when forming subsurface laser damage in either the first or second doping region. In certain embodiments, the difference in laser focusing depth between the first and second doping regions for forming a single thickness reduction portion of the substrate (e.g., one wafer) may be in the range of 1 to 15 microns, or 1 to 10 microns, or 2 to 8 microns, or 4 to 6 microns.

[0081] In certain embodiments, a crystalline material substrate may exhibit different laser absorption levels with respect to the vertical position in the substrate (e.g., within an ingot), especially in the case of intentionally doped materials. Laser absorption levels may also vary depending on the substrate (e.g., the ingot). Such variations are thought to be due to variations in doping. In certain embodiments, a lower average laser power (e.g., 3W) can be used to form subsurface laser damage in the substrate region distal to the growth species, and a higher average laser power (e.g., 5.5W) can be used to form subsurface laser damage in the substrate region proximal to the growth species.

[0082] In certain embodiments, optical measurements of the laser focal depth in a semiconductor material can be performed (e.g., taking into account the refractive index changes of the semiconductor material / air) to initialize the laser subsurface damage to an appropriate depth relative to the surface of the crystalline material substrate, and the laser damage settings (e.g., laser power, laser focal, and / or the number of laser damage formation steps) can be adjusted in response to such measurements before scanning the entire surface of the substrate. In certain embodiments, optical measurements of the laser focal depth can be performed once per ingot, or once each time after a portion of the ingot has been crushed and removed (i.e., before the formation of a subsurface laser damage pattern for each substrate layer that will be removed in the next crushing).

[0083] In certain embodiments, the semiconductor material processing method disclosed herein may include some or all of the following items and / or steps: A second carrier wafer can be mounted on the bottom surface of a crystalline material substrate (e.g., an ingot). The top surface of the crystalline material substrate is then ground or polished to an average surface roughness R of less than about 5 nanometers, for example, as preparation for a surface for transmitting laser energy. a This can be achieved. Then, laser damage can be applied to one or more desired depths within the crystalline material substrate. The spacing and direction of the laser damage marks optionally depend on the crystal orientation of the crystalline substrate. A first support can be bonded to the top surface of the crystalline substrate. The wafer obtained from the crystalline substrate can be associated with an identification code or other information linked to the first support. Alternatively, laser marking can be applied to the wafer (not the support) before separation to facilitate tracking of the wafer during and after fabrication. The crystalline substrate is then crushed along the subsurface laser damage areas (using one or more methods disclosed herein) to obtain a portion of the semiconductor substrate bonded to the first support and the remaining portion of the crystalline substrate bonded to the second support. Both the removed portion and the remaining portion of the semiconductor substrate are smoothed and cleaned as necessary to remove any residual subsurface laser damage. The removed portion of the semiconductor substrate can be separated from the support. The process can then be repeated using the remaining portion of the semiconductor substrate.

[0084] Wire sawing of SiC wafers typically results in a kerf loss of at least about 250 microns per wafer, but the laser-assisted and carrier-assisted separation methods for SiC disclosed herein can achieve a kerf loss in the range of 80 to 140 microns per wafer.

[0085] In certain embodiments, laser subsurface damage can be formed on a crystalline material substrate before the substrate is bonded to a rigid support. In certain embodiments, a rigid support that transmits laser radiation of a desired wavelength can be bonded to the crystalline material substrate before subsurface laser damage formation. In such embodiments, laser radiation can be selectively transmitted through the rigid support into the interior of the crystalline material substrate. Different support-substrate subsurface laser formation configurations are shown in Figures 10A to 10D. Figure 10A is a schematic diagram of laser radiation 61 focused through the surface of a bare substrate 62 to form subsurface laser damage 63 within the substrate 62, to which a rigid support may be attached after the subsurface laser damage is formed. Figure 10B is a schematic diagram of laser radiation 61 focused through the surface of a substrate 62 to form subsurface laser damage 63 within the substrate 62, to which the substrate 62 is previously bonded to a rigid support 66 using an adhesive material 64. Figure 10C is a schematic diagram of laser radiation 61 focused through the rigid carrier 66 and adhesive 64 to form subsurface laser damage 63 within a substrate 62 pre-bonded to the rigid carrier 66. In certain embodiments, the surface of the substrate 62 distal to the rigid carrier 66 may include one or more epitaxial layers and / or metallized layers, and the substrate 62 embodies an operable electrical device before the formation of subsurface laser damage 63. Figure 10D is a schematic diagram of laser radiation 61 focused through the rigid carrier 66 into the substrate 62 (without intervening layers) to form subsurface laser damage 63 within a substrate 62 pre-bonded to the rigid carrier 66 (e.g., via anodic bonding or other adhesive-free means).

[0086] [Dispersed subsurface laser damage] In certain embodiments, subsurface laser damage can be formed in a crystalline material by the continuous formation of multiple dispersed laser damage patterns, each subsurface laser damage pattern containing multiple substantially parallel lines. In certain embodiments, each subsurface laser damage pattern may extend substantially along the entire length (e.g., perpendicular to the substrate flat) with respect to the substrate of the crystalline material and may contain spaced lines distributed substantially across the entire width. In certain embodiments, the dispersed damage pattern may include continuously formed first and second, or first to third, or first to fourth subsurface laser damage, each subsurface laser damage pattern containing multiple parallel lines. To increase the ease of fracturing of the crystalline material along or adjacent to the subsurface laser damage region, multiple subsurface laser damage patterns may be formed continuously in a dispersed manner (e.g., forming a first subsurface damage pattern, then It is considered preferable to form a second subsurface damage pattern, and then any subsequent subsurface damage pattern, distributing the various lines of each damage pattern among the other damage patterns, rather than forming the same traces without dispersion. Regarding the reason for the improvement in fracture results obtained by dispersing subsurface laser damage patterns in crystalline materials, we do not wish to be bound by any particular theory, but it is thought that the continuous formation of dispersed subsurface laser damage patterns increases the degree of conservation of internal stress within the semiconductor material, and as a result, the lateral propagation of cracks originating from different subsurface laser damage lines is promoted.

[0087] In certain embodiments, a first subsurface laser damage pattern in a crystalline material includes a first plurality of parallel lines within the crystalline material and a first plurality of cracks propagating transversely outward from the first plurality of substantially parallel lines (e.g., primarily or substantially along the c-plane), wherein the cracks originating from each line are not connected to the cracks originating from adjacent lines. In certain embodiments, after the formation of the first subsurface laser damage pattern, a second subsurface laser damage pattern is formed in the crystalline material, including a second plurality of parallel lines, wherein the second subsurface laser damage pattern includes a second plurality of cracks propagating transversely outward from the second plurality of substantially parallel lines within the crystalline material, wherein at least some of the second plurality of cracks are connected to cracks originating from two adjacent lines of the first plurality of lines (e.g., resulting in the formation of continuous cracks).

[0088] In certain embodiments, first, second, and third subsurface laser damage patterns are formed sequentially within the crystalline material, each subsurface laser damage pattern containing a plurality of parallel lines, the lines of each subsurface laser damage pattern distributed among the lines of each other subsurface laser damage pattern. In certain embodiments, the first subsurface laser damage pattern comprises a first plurality of cracks propagating transversely outward from a plurality of substantially parallel lines within the crystalline material, the second subsurface laser damage pattern comprises a second plurality of cracks propagating transversely outward from a second plurality of substantially parallel lines within the crystalline material, the second plurality of cracks not connected to the first plurality of cracks, and the third subsurface laser damage pattern comprises a third plurality of cracks propagating transversely outward from a third plurality of substantially parallel lines within the crystalline material. In such embodiments, at least some of the third plurality of cracks connect with (i) at least some of the first plurality of cracks and (ii) at least some of the second plurality of cracks (for example, resulting in the formation of a continuous crack). In certain embodiments, a fourth subsurface laser damage pattern can be formed after the first to third subsurface laser damage, and the fourth subsurface laser damage pattern serves to further connect cracks originating from any two or more of the first, second, or third lines. In certain embodiments, three, four, five, or more dispersed patterns of subsurface laser damage can be provided.

[0089] In certain embodiments, one or more portions of the substrate may include dispersed subsurface laser damage patterns, while other portions of the substrate may include non-dispersed laser damage patterns. In certain embodiments, subsurface laser damage with different dispersion patterns can be provided on the same substrate. For example, a dispersion pattern of subsurface laser damage on a single substrate may include a damage pattern of five damage lines in a first region, four damage lines in a second region, three damage lines in a third region, two damage lines in a fourth region, one damage line in a fifth region (i.e., non-dispersed), zero damage lines in a sixth region, or any combination of two or three of the above, each of which optionally has substantially the same unit area. In certain embodiments, within at least one region of the substrate, there may be a regular (e.g., regularly repeating) pattern of dispersed damage lines. (i) Patterns may exist, and irregular (e.g., without regular repetition) patterns of dispersed or non-dispersed damage lines may exist within at least one other region of the substrate.

[0090] Figure 11A provides a top plan view of a crystalline material substrate 70, including dispersed first, second, and third subsurface laser damage patterns defined according to one embodiment. The first, second, and third subsurface damage patterns each individually include a first plurality of parallel lines 71, a second plurality of parallel lines 72, and a third plurality of parallel lines 73, extending perpendicular to the primary substrate flat 78 (and perpendicular to the [11-20] direction). The three laser damage patterns are combined to form a plurality of three-line groups 74, which are separated from each other by inter-group spacings 75 that are greater than the spacings 76, 77 between adjacent lines in each three-line group 74. For clarity, cracks formed by the first plurality of parallel lines 71, the second plurality of parallel lines 72, and the third plurality of parallel lines 73 are not shown in Figure 11A. In certain embodiments, a first plurality of parallel lines 71 are formed in a first stroke, a second plurality of parallel lines 72 are formed in a second stroke, and a third plurality of parallel lines 73 are formed in a third stroke. The third stroke serves to connect any cracks that initially originate from either the first parallel lines 71 or the second parallel lines 72.

[0091] Continuing to refer to Figure 11A, in one embodiment, a first set of parallel lines 71 can be formed at a pitch of 500 microns (i.e., the spacing between lines), and a second set of parallel lines 72 can be formed at a pitch of 500 microns, offset by 250 microns from the first set of parallel lines 71. Subsequently, a third set of parallel lines 73 can be formed at a pitch of 500 microns, offset by 125 microns from the first set of parallel lines 71. This configuration creates a set of three-line groups 74 separated from each other by a gap of 250 microns, where adjacent lines within each three-line group are separated from each other by a gap of 125 microns.

[0092] The inventors found that the order of the three-step laser damage formation process, as described in relation to Figure 11A, is important. When the order of the steps is changed to form the first, third, and second subsurface laser damage lines in succession, a higher laser power is required to complete the crack initiation across the 250-micron inter-group spacing 75. This is thought to be due to the crack initiation occurring between lines spaced 125 microns apart in the second step when using the original (first, second, third) continuous order, in which case the crack formed in the third step is simply large enough to connect to the crack originating from the second subsurface damage line that crosses the second 125-micron gap 77. When the order of the steps is first, third, second, crack initiation across the inter-group spacing 75 is not observed unless the laser power is increased, but increasing the laser power usually increases kerf loss. Therefore, according to a particular embodiment in which the order of the steps is first, second, and third, it may be desirable that the cracks formed in the first and second steps are not connected to each other, and that the cracks formed in the third step then create connected cracks that traverse both the 125 micron gaps 76, 77 and the 250 micron intergroup spacings 75.

[0093] In certain embodiments, the boundaries of each three-line group 74 may be considered to demarcate the damage-bearing areas of the substrate 70, and the damage-bearing areas of each three-line group 74 are separated from the damage-bearing areas of other three-line groups (i.e., by the inter-group spacing 75). In particular, as will be shown later in Figure 11D, cracks formed by subsurface laser damage may propagate across the inter-group spacing 75 between adjacent three-line groups 74.

[0094] Figures 11B to 11D show the fabrication of the crystalline material substrate 70 shown in Figure 11A. Figure 11B shows the substrate 70 after the formation of a first plurality of subsurface laser damage lines 71 (perpendicular to the flat 78 of the substrate 70) which have a pitch (or inter-line spacing) 71B and form a first subsurface laser damage pattern 71A. Cracks 71C propagate laterally outward from the first plurality of subsurface laser damage lines 71, but cracks originating from different subsurface laser damage lines 71 do not connect with each other.

[0095] Figure 11B shows the substrate 70 after the formation of a second plurality of subsurface laser damage lines 72 (perpendicular to the flat 78 of the substrate 70) which have a pitch (or line spacing) 72B and form a second subsurface laser damage pattern 72A. Cracks 72C propagate laterally outward from the second plurality of subsurface laser damage lines 71, but cracks originating from different subsurface laser damage lines 71 do not connect with each other.

[0096] Figure 11C shows the substrate 70 after the formation of a third plurality of subsurface laser damage lines 73 (perpendicular to the flat 78 of the substrate 70) which have a pitch (or inter-line spacing) 73B and form a third subsurface laser damage pattern 73A. Cracks 73C propagate laterally outward from the third plurality of subsurface laser damage lines 73, and such cracks 73C are sufficient to connect cracks 71C, 72C formed by the first subsurface laser damage line 71 and the second subsurface laser damage line 72. As shown, the connections of cracks between the first, second, and third plurality of subsurface damage lines are also sufficient to cause the cracks to propagate further and connect across the inter-group spacing 75.

[0097] In certain embodiments, the third laser stroke, which forms a third subsurface damage pattern, is performed at a higher laser power level than the first two strokes to help extend the cracks to connect across inter-group spacings 75 that are wider than the inter-line spacings 76, 77 within each three-line group 74. The inventors found that increasing the laser power during the third stroke to a degree sufficient to connect cracks not only between laser subsurface damage lines spaced 125 μm apart (as shown in Figure 11D) but also between laser subsurface damage lines positioned 250 μm apart. This resulted in an approximately 25% increase in tool throughput, with a small kerf loss as a penalty (e.g., approximately 110 μm kerf loss instead of 100 μm).

[0098] In certain embodiments, all laser subsurface damage lines may be non-perpendicular to the primary substrate flat (and with respect to the [11-20] direction) within a range of about 1 to 5 degrees from the perpendicular direction. For example, Figure 12 is a schematic top plan view of a crystalline material substrate 80, including a substrate flat 88 and a plurality of substantially parallel subsurface laser damage lines 81-83, first, second, and third, dispersed or dispersed between each other to form first to third subsurface laser damage patterns. Each plurality of substantially parallel subsurface laser damage lines 81-83 is deflected by 3 degrees with respect to the perpendicular direction (and with respect to the [11-20] direction), and the laser damage patterns combine to form a plurality of 3-line groups 89, which are separated from each other by inter-group spacings 85 that are greater than the spacing (or gaps) 86, 87 between adjacent lines in each 3-line group 89. In one embodiment, a first plurality of parallel lines 81 can be formed with a pitch of 500 microns (i.e., the spacing between lines), and a second plurality of parallel lines 82 can be formed with a pitch of 500 microns, offset by 250 microns from the first plurality of parallel lines 81. Subsequently, a third plurality of parallel lines 83 can be formed with a pitch of 500 microns, offset by 125 microns from the first plurality of parallel lines 81. This configuration creates a plurality of three-line groups 89 separated from each other three-line group by a gap of 250 microns, and in this case, adjacent lines within each three-line group are separated from each other by a gap of 125 microns. As shown, the parallel subsurface laser damage lines 81-83 in each group are parallel to each other.

[0099] Figure 13 is a schematic top plan view of a crystalline material substrate 90, which includes a substrate flat 98 and a plurality of substantially parallel subsurface laser damage lines 91-94, dispersed or alternating between each other to form a first to fourth subsurface laser damage pattern, all of which are parallel to each other and perpendicular to the substrate flat 98 (and to the [11-20] direction). In certain embodiments, each of the first to fourth plurality of subsurface laser damage lines 91-94 may include lines having a pitch of 500 nm, a second plurality of lines 92 offset by 250 microns from the first plurality of lines 91, a third plurality of lines offset by 125 microns from the first plurality of lines 91, and a fourth plurality of lines offset by 375 microns from the first plurality of lines 91. Ultimately, there will be a gap of 125 microns between each of the first to fourth plurality of lines 91-94. The first line 91 to the fourth line 94 form a four-line repeating group 95.

[0100] An alternative method for forming a crystalline material substrate similar to the substrate 90 shown in Figure 13 involves using four laser subsurface damage formation steps, each forming a line with a pitch of 500 microns. After the first step, the line formed by the second step is shifted 125 microns from the line formed by the first step, then the line formed by the third step is shifted 250 microns from the line formed by the first step, then the line formed by the fourth step is shifted 375 microns from the line formed by the first step.

[0101] Figure 14 is a schematic top plan view of a crystalline material substrate 100, which includes a substrate flat 108 and includes dispersed first to third plurality of subsurface laser damage lines 101 to 103 that form first, second, and third subsurface laser damage patterns. The first plurality of lines 101 and the second plurality of lines 102 are each parallel to each other and perpendicular to the primary substrate flat 108 (and to the [11-20] direction), while the third plurality of lines 103 are nonparallel to the first plurality of lines 101 and the second plurality of lines 102 (for example, with an angular difference in the range of 1 to 5 degrees), but do not intersect with either the first or second line 101 or the second line 102 within the substrate 100. In certain embodiments, the first plurality of parallel lines 101 and the second plurality of parallel lines 102 are formed first, and then the third plurality of parallel lines 103 are formed thereafter. In a particular embodiment, the first set of parallel lines 101 and the second set of parallel lines 102 each have a pitch of 500 microns, and the second set of parallel lines 102 are offset by 250 microns from the first set of parallel lines 101. The first line 101 to the third line 103 constitute a set of repeating lines group 104.

[0102] In Figure 14, the subsurface laser damage lines do not intersect, but in certain embodiments, one or more subsurface laser damage lines (formed, for example, in the next laser damage formation step) may intersect with one or more other subsurface damage lines (formed, for example, in a preceding or first laser damage formation step). In certain embodiments, the relative angle between intersecting subsurface laser damage lines may be in the range of 4 to 30 degrees, or 5 to 20 degrees, or 5 to 15 degrees, or 5 to 10 degrees.

[0103] Figure 15 is a schematic top plan view of a crystalline material substrate 110, which includes a substrate flat 118 and includes dispersed first to third subsurface laser damage lines 111-113 that form first to third subsurface laser damage patterns. The first group of lines 111 and the second group of lines 112 are each parallel to each other and non-perpendicular to the primary substrate flat 108 (for example, with an angular difference in the range of 1 to 5 degrees), while the third group of lines 113 are perpendicular to the primary substrate flat 118, but at least some (or all) of its lines are parallel to the first group of lines 111 and lines 112 in the substrate 110. The lines do not intersect with the lines of the second group. In a particular embodiment, the first set of parallel lines 111 and the second set of parallel lines 112 each have a pitch of 510 microns, and the second set of parallel lines 112 are offset by 250 microns from the first set of parallel lines 111. The first line 111 to the third line 113 constitute a three-line repeating group 114.

[0104] Figure 16 is a schematic top plan view of a crystalline material substrate containing dispersed first, second, and third subsurface laser damage patterns defined according to one embodiment, wherein all laser damage lines are parallel to each other, and the spacing between groups of laser damage lines is not uniform in at least part of the substrate.

[0105] Figure 17 is a schematic top plan view of a crystalline material substrate containing dispersed first, second, and third subsurface laser damage patterns drawn according to one embodiment, where all laser damage lines are parallel to each other, and the laser damage lines exhibit variability in group spacing, inter-group spacing, and group configuration.

[0106] Figure 18 is a schematic top plan view of a crystalline material substrate including continuously formed first, second, and third subsurface laser damage patterns, drawn according to one embodiment, wherein the first and second groups of laser damage lines are parallel to each other, but the third group of laser damage lines is non-parallel to and intersects with the first and second groups of laser damage lines.

[0107] Figure 19 is a schematic top plan view of a crystalline material substrate containing continuously formed first, second, and third subsurface laser damage patterns, where each group of laser damage lines contains parallel lines and each group of laser damage lines is non-parallel to each other group of laser damage lines. Figures 11A to 19 show embodiments containing three or four groups of multiple subsurface laser damage lines, but it will be understood that any preferred number of groups of subsurface laser damage lines can be provided. For example, in certain embodiments, the first and second multiple subsurface laser damage lines may be dispersed in the absence of a third and / or fourth multiple subsurface laser damage line. In certain embodiments, the first and second multiple subsurface laser damage lines can be formed in first and second strokes, respectively, where each multiple laser damage line has a pitch of 250 microns, and the second multiple laser damage line is offset by 125 microns from the first multiple laser damage line.

[0108] In certain embodiments, subsurface laser damage is distributed among multiple non-overlapping areas of the crystalline material by forming a first group of subsurface laser damage sites within non-overlapping first and second areas of the crystalline material, followed by the formation of a second group of subsurface laser damage sites within the first and second areas, where at least some (or all) of the second group of subsurface laser damage sites do not intersect with the sites of the first group of subsurface laser damage sites formed within the non-overlapping areas. Subsequently, one or more additional groups of subsurface laser damage sites may be formed and distributed among the same non-overlapping first and second areas of the crystalline material. While first and second areas have been described, it will be understood that any preferred number of non-overlapping areas (e.g., three, four, five, six, or more areas) may be defined. In certain embodiments, such areas may not only be non-overlapping at all, but may also be spaced apart from each other in a non-contact relationship (e.g., spaced laterally).

[0109] Figure 20A is a top plan view of the crystalline substrate 150, showing non-overlapping first, second, and third areas 150A-150C where laser damage regions may be formed. The first area 150A is shown to highlight the boundary between the first to third areas 150A-150C. Although shading has been added to the third area 150C for illustrative purposes, it should be understood that the actual crystalline material substrate 150 is usually a uniform color. Each area 150A to 150C is in contact with a portion of the primary flat 150' of the substrate 150. Figures 20A to 20D show three areas 150A to 150C, but any preferred number of areas such as two, three, four, five, six, or more are contemplated, and such areas may be arranged in any preferred structure, for example, in a one-dimensional array, in a two-dimensional array, as a plurality of sectors (e.g., wedge-shaped sectors) extending from a central point, etc.

[0110] Figure 20B is a top plan view of the crystalline substrate 150 of Figure 20A after the formation of a first plurality of subsurface laser damage regions 151 within the first area 150A to the third area 150C. As shown, the laser damage regions 151 are provided as substantially perpendicular and substantially parallel lines to the primary flat 150' of the substrate 150. A plurality of laser damage regions 151 are provided within each of the first to third areas 150A to 150C. Although not shown in Figure 20B, lateral cracks (as shown in Figure 11B) may originate from the laser damage regions 151, but it will be understood that it may be preferable that these do not connect adjacent laser damage regions 151. In certain embodiments, a plurality of subsurface laser damage regions 151 can be formed within the first area 150A, then within the second area 150B, and finally within the third area 150C.

[0111] Figure 20C is a schematic top plan view of the crystalline material substrate 150 of Figure 20B after the formation of a second plurality of subsurface laser damage regions 152 within the first to third areas 150A to 150C. As shown, the laser damage regions 152 of the second plurality of subsurface laser damage regions 152 are provided as substantially parallel lines substantially perpendicular to the primary flat 150', with a plurality of laser damage regions 152 provided within each of the first to third areas 150A to 150C. Furthermore, each laser damage region 152 of the second plurality of subsurface laser damage regions 152 is substantially parallel to the first plurality of subsurface laser damage regions 151. Although not shown in Figure 20C, it will be understood that while cracks extending laterally from each laser damage region 151, 152 may occur, it is preferable that such cracks do not connect adjacent laser damage regions 151, 152. In certain embodiments, the subsurface laser damage regions 152 can be formed in the same order as the first subsurface laser damage regions 151 (for example, the subsurface laser damage regions 152 can be formed in the first area 150A, then in the second area 150B, and finally in the third area 150C). In this way, the laser damage regions 152 of the second plurality of subsurface laser damage regions 152 are dispersed among the laser damage regions 151 of the first plurality of subsurface laser damage regions 151.

[0112] Figure 20D is a top plan view of the crystalline substrate of Figure 20C after a third plurality of subsurface laser damage regions 153 have been formed within the first to third areas 150A to 150C. As shown, the laser damage regions 153 of the third plurality of subsurface laser damage regions 153 are provided as substantially parallel lines substantially perpendicular to the primary flat 150', and the plurality of laser damage regions 153 of the third plurality of subsurface laser damage regions 153 are located within each of the first to third areas 150A to 150C. Each laser damage region 153 of the third plurality of subsurface laser damage regions 153 may be substantially parallel to the first plurality of subsurface laser damage regions 151 and the second plurality of subsurface laser damage regions 152. Multiple three-line groups 154 are formed by subsurface laser damage patterns provided by the first to third subsurface laser damage regions 151-153, and these are separated by an inter-group spacing 154' that exceeds the spacing between adjacent laser damage regions 151-153 in each three-line group 154. , they are spaced apart from each other. Although not shown in Figure 20C, cracks extending laterally may occur from each laser-damaged region 151-153, and it will be understood that these cracks (as shown in Figure 11D) extend laterally between all the laser-damaged regions 151-153, thereby facilitating the subsequent fracture of the upper portion of the substrate 150 from the rest of the substrate 150. In certain embodiments, the subsurface laser-damaged regions 153 of the multiple subsurface laser-damaged regions 152 can be formed in the same order as the first subsurface laser-damaged region 151 and the second subsurface laser-damaged region 152 (for example, the subsurface laser-damaged regions 153 can be formed in the first area 150A, then in the second area 150B, and finally in the third area 150C). In this way, the laser damage regions 153 of the third plurality of subsurface laser damage regions 153 are dispersed among the laser damage regions 151 and 152 of the first plurality of subsurface laser damage regions 151 and the second plurality of subsurface laser damage regions 152.

[0113] [Parallel machining and / or laser beam splitting] In certain embodiments, to improve tool throughput, multiple areas of a single substrate can be processed simultaneously to form subsurface laser damage within multiple substrate regions, and / or multiple substrates can be arranged within a single tool for simultaneous or substantially simultaneous laser processing. In certain embodiments, the output beam of a single laser can be split into multiple beams using one or more beam splitters, and individual beams of these beams can be supplied to either different substrates or different areas of a single substrate to form subsurface laser damage therein using the methods disclosed herein. In certain embodiments, multiple lasers can be used to simultaneously supply beams to multiple substrates or multiple areas of a single substrate to form subsurface laser damage therein using the methods disclosed herein.

[0114] Figure 21 is a schematic top plan view of a laser processing apparatus holder 163, which is arranged to hold four substrates 155A-155D on which subsurface laser damage can be formed using one or more lasers. As shown, each substrate 155A-155D includes a subsurface laser damage pattern thereon, such a pattern including a plurality of substantially parallel lines 156-158 of the first, second, and third types. The three laser damage patterns are combined to form a plurality of three-line groups 156, which are separated from each other by inter-group spacings 160 that are greater than the spacings 161, 162 between adjacent lines in each three-line group 159. In certain embodiments, laser damage patterns can be formed on the first substrate 155A and the third substrate 155C by a first laser or a first segmented laser beam portion, and on the second substrate 155B and the fourth substrate 155D by a second laser or a second segmented laser beam portion. In certain embodiments, the holder 163 supporting the substrates 155A to 155D is configured to move (for example, in two (x, y) lateral directions), while one or more lasers and / or their focusing optical systems are restricted from moving laterally (but may move vertically (z direction)).

[0115] Figure 22A is a schematic top view of a single substrate 164 being processed with a laser beam divided into multiple parts to simultaneously form subsurface laser damage regions following a first subsurface laser damage pattern in multiple areas of the substrate 164. As shown, the substrate 164 includes multiple areas 164A to 164C (similar to areas 150A to 150C depicted in Figures 20A to 20C, for example). The first laser damage formation step includes directing two divided laser beam portions to simultaneously form laser damage regions 165' in the first area 164A and the second area 164B. The substrate 164 can be indexed laterally with respect to the laser (for example, in the opposite direction to the rightward arrow), and the next laser damage formation step is to direct the first area 1 The process includes applying two divided laser beam portions to simultaneously form laser damage areas 165'' in areas 64A and the second area 164B. This process is repeated to form additional laser damage areas 165''', 165'''' in the first area 164A and the second area 164B, and to finally cover the first, second, and third areas 164A-164C to form the first subsurface laser damage pattern. The process may then be repeated to form second and third subsurface laser damage patterns, respectively, dispersed together with the first subsurface laser damage pattern. Using the first and second divided laser beam portions, a subsurface laser damage pattern distributed throughout the substrate 164 can be formed in half the time it would take to form the pattern with a single, undivided laser beam. Figure 22B is a schematic top plan view of two substrates 166A, 166B supported by a holder 168, being processed with a laser beam split into two portions to simultaneously form subsurface laser damage on both substrates 166A, 166B according to at least one subsurface laser damage pattern. The first laser damage formation step includes applying two split laser beam portions to simultaneously form laser damage areas 167' on the first substrate 166A and the second substrate 166B. The holder supporting the substrates 166A, 166B can be indexed laterally with respect to the laser (for example, in the opposite direction to the rightward arrow), and the next laser damage formation step includes applying two split laser beam portions to simultaneously form laser damage areas 167'' on the first substrate 166A and the second substrate 166B. This process is repeated to form additional laser-damaged areas 167'''', 167'''' on the first substrate 166A and the second substrate 166B, and finally to cover the first substrate 164A and the second substrate 166B and form a first subsurface laser-damaged pattern thereon.The process may then be repeated to form second and third subsurface laser damage patterns, respectively, on substrates 166A and 166B, which are dispersed together with the first subsurface laser damage pattern.

[0116] [Formation of overlapping subsurface laser damage at different depths] In certain embodiments, a first subsurface laser damage centered on a first depth can be formed within a crystalline material substrate, and an additional subsurface laser damage centered on a second depth can be formed within the substrate, where the additional subsurface laser damage is substantially aligned with the first subsurface laser damage, and at least a portion of the vertical extension of the additional subsurface laser damage overlaps with at least a portion of the vertical extension of the first laser damage. Again, to provide subsurface laser damage with overlapping vertical extensions, one or more subsequent steps configured to impart laser damage of different depths can be added on top of one or more preceding steps. In certain embodiments, the addition of overlapping subsurface damage may be performed in response to a pre-fragmentation (e.g., by optical analysis) determination that one or more preceding subsurface laser damage formation steps are incomplete. In certain embodiments, the difference in laser focusing depth between the first and second laser damage patterns for forming a single thickness reduction portion of the substrate (e.g., one wafer) may be in the range of 1 to 10 microns, or 2 to 8 microns, or 2 to 6 microns. The formation of overlapping subsurface laser damage at different depths may be carried out in conjunction with any other method step herein, including (but not limited to) the formation of multiple dispersed subsurface laser damage patterns.

[0117] Figure 23A is a schematic cross-sectional view of a crystalline material substrate 170, including a first subsurface laser damage pattern 173 centered on a first depth relative to the first surface 171 of the substrate 1770, the subsurface damage pattern 173 being created by focused laser radiation 179. The first subsurface laser damage pattern 173 is opposite the first surface 171. It has a vertical extension 174 that remains inside the substrate 170 between the second surface 172 on the side. Figure 23B is a schematic cross-sectional view of the substrate of Figure 23A after the formation of the second subsurface laser damage pattern 175, which is centered on a second depth and aligned with the first subsurface laser damage pattern 173, and the vertical extension 176 of the second damage pattern 175 overlaps with the vertical extension 174 of the first damage pattern 173 in the damage overlap region 177. In certain embodiments, subsequent fracture of the crystalline material 170 may be performed along or through the damage overlap region 177.

[0118] [Formation of non-overlapping subsurface laser damage at different depths] In certain embodiments, subsurface laser damage lines can be formed at various depths of the substrate without alignment with other (e.g., previously formed) subsurface laser damage lines, and / or without overlapping vertical extension features of the first and subsequent laser damages. In certain embodiments, the dispersed pattern of subsurface laser damage may include groups of laser lines, different groups of which are focused at different depths relative to the substrate surface. In certain embodiments, the focusing depth of the laser radiation inside the substrate differs between different groups of laser lines (e.g., at least two different groups from a first and second group, a first to a third group, a first to a fourth group, etc.) by a distance ranging from about 2 microns to about 5 microns (i.e., about 2 μm to about 5 μm).

[0119] [Laser Tool Calibration] One of the main causes of kerf loss is subsurface laser damage located beneath the primary fracture region on the ingot surface. Generally, increased subsurface laser damage leads to increased kerf loss. One possible reason for increased subsurface laser damage is the inability to adequately compensate for the optical properties of the crystalline material.

[0120] In certain embodiments, laser calibration may be performed each time a crystalline material substrate (e.g., an ingot) is supplied to the laser tool, prior to the formation of subsurface laser damage in the crystalline material substrate. Such calibration can utilize variable height adjustment to achieve an initial state in which the best focusing point of the laser beam is formed on the upper surface of the crystalline material substrate, and thereafter, the aperture or compensating ring of the laser tool can be adjusted to correspond to the desired depth of subsurface laser damage formation in the crystalline material, depending on the next state.

[0121] [Wafer photograph showing the doping area (also known as the doping ring)] Figure 24A is a perspective photograph of the SiC wafer 180 after separation from the carrier (i.e., the sapphire carrier 181 bonded with the thermoplastic adhesive shown in Figure 24B) using the thermal-induced fracturing method described herein. Both the wafer 180 and the carrier 181 have a diameter of 150 mm. No wafer breakage was observed after thermal-induced fracturing. Figure 24C is a partially inverted version of the SiC wafer photograph in Figure 24A to emphasize the contrast between the central doping ring 182 and the outer annular portion 183 of the SiC wafer 180. Figure 24D shows the image of Figure 24C with a dotted oval notation indicating the boundary between the central doping ring 182 and the outer annular portion 183 of the SiC wafer 180. The doping ring 182 represents the region of high doping concentration relative to the outer annular portion 183 of the SiC wafer. Since doped semiconductor materials such as SiC exhibit improved absorption at IR wavelengths, it may be beneficial to increase the laser output in the doping ring 182 compared to the outer annular portion 183 when attempting to form subsurface laser damage on a SiC wafer. In certain embodiments, the presence of conditions indicating non-uniform doping of the crystalline material over at least a portion of the substrate surface can be detected, for example, by detecting changes in the reflectance or absorptance of light by optical means, in order to determine the presence of at least one first doping region and at least one second doping region (e.g., the doping ring 182 and the outer annular portion 183). Subsequently, the non-uniform doping of the crystalline material In response to the detection of conditions indicating a subsurface laser damage pattern, the laser output can be modified during the formation of the subsurface laser damage pattern to provide a first average power laser emission when forming subsurface laser damage within a first doping region (e.g., doping ring 182) and a second average power laser emission when forming subsurface laser damage within a second doping region (e.g., outer annular portion 183), wherein the first and second average power levels are different from each other.

[0122] [Schematic diagram of an ingot exhibiting a doping ring] Figure 25 is a schematic lateral cross-sectional view of a SiC ingot 184 grown on a seed crystal 185, showing a substantially cylindrical doping region 187 extending upward from the seed crystal 185 (located on the first surface or bottom surface 185' of the ingot 184) along the central portion of the ingot 184 over its entire thickness, with the doping region 187 located on the second surface or top surface 186' of the ingot 184. The side of the doping region 187 is surrounded by a substantially annular undoped (e.g., low-concentration doped or unintentionally doped) region 186. A thin cross-sectional portion 189 of the ingot 184 taken between the first surface 185' and the second surface 186' may define a wafer 189A as shown in Figure 26. The wafer 189A includes a central doping region 187 and a substantially ring-shaped undoped region 186, partly bordered by a primary flat 189'. In certain embodiments, the wafer 189A can be produced from an ingot 184 using a laser-assisted cutting method as described herein.

[0123] Figure 26 shows the size (e.g., width or diameter) of the doping region 187 as being substantially constant throughout the thickness of the ingot 184, but the inventors have confirmed that the size of the doping region can vary with its vertical position in the ingot (e.g., typically the width or diameter is larger closer to the seed crystal and smaller further away from the seed crystal). It has also been confirmed that the amount of doping within the doping region can vary with its vertical position in the ingot.

[0124] Figure 27 is a schematic lateral cross-sectional view of a SiC ingot 184A grown on a seed crystal 185A, showing a frustoconical doping region 187A extending upward from the seed crystal 185A (located on the first surface or bottom surface 185A' of the ingot 184A) along the central portion of the ingot 184A over its entire thickness. As shown, the doping region 187A is present on the second surface or top surface 186A' of the ingot 184A, but the width or diameter of the doping region 187A is smaller on the second surface 186A' than on the first surface 185A'. The lateral side of the doping region 187A is surrounded by an undoped (e.g., low-concentration doped or unintentionally doped) region 186A, which is substantially annular in shape. In certain embodiments, the doping region 187A may have a width and doping level that varies with its perpendicular position to the seed crystal 185A.

[0125] The inventors also confirmed that when a micro-slope seed crystal (e.g., an offcut at an angle nonparallel to the c-plane) is used for growing the SiC ingot, the lateral position and shape of the doping region may differ compared to the configuration shown in Figure 27. For example, when a micro-slope seed crystal is used, the doping region may be oval rather than circular, and / or may be shifted laterally relative to the center of the ingot.

[0126] Figure 28 is a schematic lateral cross-sectional view of a SiC ingot 184B grown on a slightly sloped (e.g., off-cut) seed crystal 185B, showing a frustoconical doping region 187B extending upward from the seed crystal 185B at a point off-center from the seed crystal 185B and across the entire thickness of the ingot 184B. As shown, the doping region 187B is located on the second face or top face 186B' of the ingot 184B, but the doping region 1 The width or diameter of 87B is smaller on the second surface 186B' than on the first surface 185B'. The doped region 187B may have a substantially oval shape when viewed from above. The sides of the doped region 187B are surrounded by an undoped (e.g., low-concentration doped or unintentionally doped) region 186B. In certain embodiments, the doped region 187B may have a shape, width, and / or doping level that varies with its perpendicular position to the seed crystal 185B.

[0127] [Wafer enlargement photo] Figure 29 is an oblique view of the Si surface of a SiC wafer separated from an ingot by a process including the formation of subsurface laser damage and subsequent separation, where the inset area (upper right) depicts a deliberately separated fragment of the SiC wafer, including the edge depicted in the subsequent scanning electron microscope (SEM) image.

[0128] Figure 30A is a 45x magnified SEM image of a portion of the SiC wafer fragment from Figure 29, taken at a 15-degree tilt angle, with superimposed arrows indicating the directions of the [1-100] and [11-20] crystal planes. The laser beams are perpendicular to the [11-20] direction and spaced approximately 250 microns apart. Figure 30B is a 1,300x magnified SEM image of a portion of the SiC wafer fragment from Figure 29, taken at a 15-degree tilt angle. Figure 30C is a 350x magnified SEM image of a portion of the SiC wafer fragment from Figure 29, taken at a 15-degree tilt angle. As shown in Figure 30C, the off-axis cleavage plane correlates to some extent with the laser spacing, but is not consistent across the wafer surface. This can be at least partially due to variations in the laser beam position relative to the cleavage plane. In this wafer, fragmentation began at the polycrystalline introduction site.

[0129] Figure 30D is a 100x magnified SEM image of a portion of the SiC wafer fragment from Figure 29, taken at a tilt angle of 2 degrees. Figure 30E is a 1,000x magnified SEM image of a portion of the SiC wafer fragment from Figure 29, taken at a tilt angle of 2 degrees. Figures 30D and 30E show that the laser damage is considerably shallower compared to the surface features along the fractured area. In particular, the variability of the resulting fracture damage can be seen in the central part of Figure 30E.

[0130] Figure 31A is a confocal laser scanning microscope image of a small central portion of the SiC wafer shown in Figure 29, where the positions of "trenches" formed by laser scanning are marked by superimposed crosshairs. Figure 31B is a plot of the surface profile of a portion of the SiC wafer shown in Figure 31A. Referring to Figure 31B, the variability of the laser beam position relative to the SiC cleavage plane can be observed.

[0131] Figure 32A is a confocal laser scanning microscope image of a larger portion of the SiC wafer near the top (at the time of imaging) shown in Figure 29, with superimposed crosshairs marking the locations of "trenches" or lines formed by the laser scanning. Figure 32B is a plot of the surface profile of the portion of the SiC wafer near the top in Figure 32A. In Figure 32B, the first pair of lines corresponding to laser damage (represented as crosshairs in ellipse 200) are separated by a depth of more than 30 microns, and the second pair of lines corresponding to laser damage (represented as crosshairs in ellipse 201) are separated by a depth of more than 20 microns. Figures 32A and 32B show irregular spacing between laser lines, where the individual lines within the first pair of lines (in ellipse 200) are closer to each other and the individual lines within the second pair of lines (in ellipse 201) are closer to each other compared to other laser damage lines drawn.

[0132] Figure 33A is a confocal laser scanning microscope image of a larger portion of the SiC wafer shown in Figure 29, near the bottom (at the time of imaging), with the locations of the "trenches" formed by the laser scanning marked by superimposed crosshairs. Figure 33B shows the SiC of Figure 33A. This is a plot of the surface profile of the portion of the wafer near the bottom. Figure 33B shows the lateral distance variation between adjacent pairs of laser damage lines, with some pairs separated by only 334 microns and others by only 196 microns, but the maximum variation in depth is 13 microns.

[0133] [Substrate fragmentation after subsurface laser damage formation] As already discussed herein, subsurface laser damage can be formed within a crystalline material substrate to prepare the substrate for fracturing to remove at least one thin layer (e.g., a wafer) of crystalline material from the substrate. Examples of specific fracturing techniques (e.g., cooling a CTE mismatched support bonded to the substrate, applying ultrasound to the substrate, or applying a bending moment to a support mounted on the substrate) are described below herein, and it will be understood that the various subsurface laser damage formation techniques described herein can be used within any suitable fracturing technique, including fracturing techniques already known to those skilled in the art.

[0134] [Fracture of rigid supports with CTE mismatch between support / substrate due to cooling] Figures 34A to 34F illustrate steps of a carrier-assisted method for crushing a crystalline material according to one embodiment of the present disclosure, utilizing a rigid carrier having a CTE greater than that of the crystalline material bonded to the crystalline material. Figure 34A is a schematic lateral cross-sectional view of a rigid carrier 202 having a layer of adhesive material 198 bonded to a first surface 203 of the rigid carrier 202, and a second surface 204 opposite to the first surface 203.

[0135] Figure 34B is a schematic cross-sectional view of an assembly 188 comprising a rigid support 202 and an adhesive material 198 of Figure 34A, bonded to a crystalline material substrate 190 having a subsurface laser-damaged region 196 within it. The rigid support 202 has a larger diameter or lateral elongation than the substrate 190. The substrate 190 includes a first surface 192 adjacent to the adhesive material 198 and a second surface 194 on the opposite side, with the subsurface laser damage 196 being closer to the first surface 192 than to the second surface 194. The adhesive material 198 extends between the first surface 192 of the crystalline substrate 190 and the first surface 203 of the rigid support 202. The adhesive material 198 can be cured according to the requirements of the selected bonding method (e.g., thermal compression adhesive bonding, compression UV bonding, chemical reactive bonding, etc.). In certain embodiments, a second carrier (not shown) may be bonded to the second surface 194 of the substrate 190, in which case the second carrier will not extend beyond the substrate 190 and / or the substrate 190 and the CTE will coincide, but these are optional.

[0136] Figure 34C is a schematic cross-sectional view of the assembly of Figure 34B after the second surface 204 of the rigid support 202 is positioned on the support surface 208 of a cooling device in the form of a cooled chuck 206 configured to receive a coolant. Heat is transferred from the rigid support 202 to the cooled chuck 206 by contact between the rigid support 202 and the cooled chuck 206. During the cooling process, the rigid support 202 shrinks laterally more than the crystalline material substrate 190 due to the CTE of the support 202 being greater than that of the substrate 190, and as a result, the support 202 exerts shear stress on the substrate 190. Due to the presence of subsurface laser damage 196 near the adhesive layer 198 that bonds the rigid support 202 to the substrate 190, the crystalline material is fractured along or near the subsurface laser damage region 196 by the shear stress exerted on the substrate 190.

[0137] In certain embodiments, the cooled chuck 206 has a diameter smaller than the diameter of the rigid support 202. Coolant can be supplied to the cooled chuck 206, but the rigid support 202 does not need to reach liquid nitrogen temperature (-160°C) in order to successfully complete the thermally induced fracturing of the crystalline material substrate 190. Using a cooled chuck maintained at -70°C is preferred for the fracturing of single-crystal SiC material supported by a single-crystal sapphire substrate. Results have been obtained. Such temperatures can be maintained using various coolants, such as liquid methanol (which maintains fluidity above its freezing point of -97°C) supplied from a two-phase pump-type vaporization cooling system. Favorable separation results were also obtained by cooling the carrier, adhesive, and substrate in a freezer maintained at -20°C, in which case such temperatures can be maintained using a single-phase vaporization cooling system. The ability to use a single-phase vaporization cooling system or a two-phase pump-type vaporization cooling system instead of liquid nitrogen significantly reduces operating costs.

[0138] Figure 34D is a schematic cross-sectional view of the remaining portion of the crystalline material substrate 190A separated from the bonded assembly, which includes the rigid support 202, the adhesive material 198, and the portion 210 of crystalline material removed from the substrate 190A, after the crystalline material has been fractured along the subsurface laser-damaged region. The remaining portion of the crystalline material substrate 190A is bounded by a new first surface 193 (with residual laser damage 196A) opposite the second surface 194. Correspondingly, the removed portion 210 of crystalline material is bounded by a new second surface 212 (with residual laser damage 196B) opposite the first surface 192. The bonded assembly 215, including the rigid support 202, the adhesive material 198, and the removed portion 160 of crystalline material, can then be recovered from the cooled chuck 206.

[0139] Figure 34E is a schematic cross-sectional view of the joined assembly 215 of Figure 34D after recovery from the water-cooled chuck 206. By maintaining the portion of the removed crystalline material 210 attached to the rigid support 202, mechanical support for the portion of the removed crystalline material 210 is achieved, which is advantageous as it allows for the performance of one or more surface processing steps (e.g., grinding, polishing, etc.) on the new surface 212 to remove residual laser damage 196B and achieve a desired thickness of the crystalline material 210 (e.g., via grinding, and subsequent optional chemical mechanical planarization and / or polishing steps). In certain embodiments, laser damage removal and thinning may include a continuous grinding / polishing process, as well as any preferred polishing and cleaning steps to prepare the new surface 212 for the next process (e.g., surface injection, laser marking (e.g., along a wafer flat), epitaxial layer formation, metallization, etc.).

[0140] Figure 34F is a schematic cross-sectional view of the removed portion 210 of the crystalline material supported by the upper surface 218 of a heated vacuum chuck 216, where the rigid carrier 202 and adhesive material 198 are laterally translated away from the removed portion 212 of the crystalline material after the high-temperature softening and release of the adhesive material 198. That is, the heated vacuum chuck 216 can be heated to a temperature sufficient to soften and / or flow the adhesive material 198 so that when an external shear stress is applied to the second surface 204 of the rigid carrier 202, the rigid carrier 202 can be laterally translated away from the removed portion 212 of the crystalline material that is temporarily held in place by the heated vacuum chuck 216. The operation of the heated vacuum chuck 216 may then be stopped, and the removed portion 212 of the crystalline material will embody a self-supporting material. If desired, all adhesive residue 198 can be removed and cleaned from the first surface 203 of the rigid support 202, and the rigid support 202 can be optionally reused for another crushing process. The removed crystalline material can then be used as a growth substrate for growing one or more epitaxial layers and conductive metal layers to form device wafers, which are then pulverized to form individual semiconductor devices.

[0141] [Fracture induced by ultrasonic energy] Another method for causing laser-induced fracture along a subsurface damage zone of a crystalline material bonded to a rigid support involves applying ultrasonic energy to the crystalline material while it is bonded. Figure 35 shows a rigid support 20 using an intervening adhesive material 198A. This is a schematic cross-sectional view of assembly 188A, which includes a crystalline material 190A with subsurface laser damage 196A bonded to 2A, and is located in a liquid tank 225 of an ultrasonic generator device 220. The device 220 further includes a container 222 located in contact with an ultrasonic generating element 224, and the container 222 houses the liquid tank 225. The presence of a rigid support 202A can reduce or eliminate the fracture of the crystalline material 190A when subjected to ultrasonic energy, especially if residual stress remains between the rigid support 202A and the crystalline material 190A before separation (e.g., due to CTE mismatch). Such residual stress can reduce the amount of ultrasonic energy required to initiate the fracture of the crystalline material, thereby reducing the likelihood of material fracture.

[0142] [Crushing induced by mechanical force] In certain embodiments, the fracture of a crystalline material bonded to a rigid support can be facilitated by (i) the application of a mechanical force (e.g., optionally localized to one or more locations) adjacent to at least one edge of the support. Such a force can impart a bending moment to at least a portion of the support, and such a bending moment is transmitted to a subsurface laser-damaged region to initiate fracture. Exemplary embodiments are shown in Figures 36A to 36C.

[0143] Figures 36A to 36C are schematic cross-sectional views illustrating the steps for fracturing a crystalline material substrate 236 with subsurface laser damage 233 by applying a mechanical force close to one edge of a support 238 to which the substrate 236 is bonded. The bonded assembly includes a crystalline material substrate 236 having a subsurface laser damage region 233, bonded between rigid supports 238, 238'. Each rigid support 238, 238' includes laterally projecting tab portions 239, 239' aligned with a flat 235 of the substrate 236, providing a locally enlarged boundary region that defines a recess 231 into which a tool 219 can be inserted. Figure 36A shows the state before the tool 219 is inserted into the recess 191. Figure 36B shows the state after the tool 219 has been inserted into the recess, at which point the tool 216 is tilted upward, which causes a twisting force to act in a direction that promotes separation between the rigid supports 238 and 238', thereby causing a bending moment M to act on at least one support 238. In certain embodiments, the substrate 236 comprises a material having a hexagonal structure (e.g., 4H-SiC), and the bending moment M is oriented within ±5 degrees from the direction perpendicular to the [11-20] direction (or, equivalently, within ±5 degrees from the direction parallel to the [1-100] direction). Figure 36C shows the state of the crystalline substrate 236 after the initial fracturing along the subsurface laser-damaged region 233, in which case the upper portion 236 of the crystalline material remains bonded to the upper support 238, the lower portion 236B of the crystalline material remains bonded to the lower support 238', and the upper support 238 is tilted upward relative to the lower support 238'. Such fracture yields the first bonded assembly 229A (including the upper carrier 238 and the upper portion 236A of the crystalline material), separated from the second bonded assembly 229B (including the lower carrier 238' and the lower portion 236B of the crystalline material). In certain embodiments, mechanical forces can be applied in close proximity to the edges on both sides of the rigid carrier to which the substrate is bonded in order to facilitate the fracture of the subsurface laser-damaged crystalline material bonded to the carrier.

[0144] It should be noted that the combination of two or more fracturing techniques (e.g., CTE mismatch and ultrasonic-induced fracturing, or CTE mismatch and mechanical action-induced fracturing, or ultrasonic-induced and mechanical action-induced fracturing) is particularly intended. In certain embodiments, the liquid in the ultrasonic chamber can be cooled before or during the application of ultrasonic energy. The amount of mechanical force required to complete fracturing may be affected by the CTE difference between the substrate and the support. In certain embodiments, the CTE difference and mechanical force can be combined. If the CTE difference between the support and the substrate is small or nonexistent (i.e., CTEs are matched), a greater mechanical force may be required to complete fracturing. Conversely, if the CTE mismatch When the material density is large, the mechanical force required to complete the crushing can be reduced, or mechanical force may not be required at all.

[0145] [Device / wafer splitting process] In certain embodiments, laser-assisted and carrier-assisted separation methods can be applied to a crystalline material after forming at least one epitaxial layer (and optionally at least one metallic layer) on it as part of a operable semiconductor-based device. Such device wafer splitting processes are particularly advantageous because they can increase the yield of crystalline material (and reduce waste) by greatly reducing the need to grind and remove the substrate material after device formation.

[0146] Figures 37A to 37O are schematic cross-sectional diagrams illustrating the steps of the device wafer splitting process, according to which a thick wafer is crushed from a crystalline material, at least one epitaxial layer is grown on the thick wafer, and the thick wafer is crushed to form first and second bonded assemblies, each containing a carrier and a thin wafer separated from the thick wafer, the first bonded assembly containing at least one epitaxial layer as part of a functional semiconductor-based device.

[0147] Figure 37A shows a crystalline material substrate 240 having a first surface 241 and a subsurface laser damage 243 positioned to a certain depth relative to the first surface. Figure 37B shows the substrate 240 of Figure 37A after adhesive material 244 has been added to cover the first surface 241. Figure 37C shows the item depicted in Figure 37B after a rigid carrier 246 has been bonded to the substrate 240 using the adhesive material 244. Figure 37D shows the item of Figure 37D after the substrate 240 has been fractured along the subsurface laser damage 243 (for example, using one or more methods disclosed herein), resulting in the remaining portion of the substrate 240 separated from the bonded assembly, including the carrier 246, the adhesive material 244, and the crystalline material portion (e.g., a thick wafer) 242 removed from the substrate 240. In certain embodiments, the thick wafer 242 may have a thickness in the range of approximately 350 to 750 microns. The exposed surface 243A of the thick wafer 242 and the exposed surface 243B of the rest of the substrate 240 may exhibit surface irregularities, which can be reduced by surface processing steps such as grinding, CMP, polishing, etc. Figure 37E shows the thick wafer 242 after debonding and removal from the carrier 246, and the thick wafer 242 has a vertical edge profile. The vertical edge of the wafer is prone to fracturing during wafer handling, resulting in unacceptable edge chips and particles. To reduce the risk of breakage, the wafer edge can be edge-ground to create a non-vertical wafer edge with a chamfered or rounded edge. Figure 37F shows the thick wafer 242 supported between opposing upper gripping portions 248A and lower gripping portions 248B of a turntable adjacent to a rotary profile grinding tool 249 having a concave cutting surface 249A configured to impart a rounded edge profile 247 to the thick wafer 242 (e.g., impregnated with diamond particles). Figure 37G shows a thick wafer 242 after edge grinding (also known as edge profiling), which includes a rounded edge 247 that provides a boundary between the first wafer surface 251 and the second wafer surface 252.

[0148] Figure 37H shows the thick wafer 242 in Figure 37G after one or more epitaxial layers 253 have been grown on or covering the first surface 251 of the thick wafer 251. Due to the incompatibility of adhesives and epitaxy with high temperatures, the support shown in Figure 37D is not present. Figure 37I shows the structure in Figure 37H after conductive (e.g., metallic) contacts 254 have been formed on the epitaxial layer 253 to form at least one operable semiconductor device, and the thick wafer 242 still has rounded edges 247. Conventionally, grinding is performed to make the thick wafer 242 suitable for the resulting device relative to the second surface 252. This is done to thin the wafer to a very thin thickness (for example, 100 to 200 microns for Schottky diodes or MOSFETs). The technique disclosed herein utilizes laser-assisted and carrier-assisted separation to reduce the need for wafer grinding and instead remove a portion of the thick wafer, which can then be surface-finished and used to fabricate another operational semiconductor device.

[0149] The inventors found that the presence of rounded edges 247 on the thick wafer 242 hinders the control of the formation of subsurface laser damage adjacent to the edges 247, because the rounded profile negatively affects the control of laser focus and depth. To address this problem, the rounded edges 247 of the thick wafer 242 can be removed before further laser processing. Figure 37J shows the structure of Figure 37I being ground with an edge grinder 256 to remove the rounded edges 247 and to impart a substantially perpendicular edge 255 extending between the first surface 251 and the second surface 252 of the thick wafer 242, with the epitaxial layer 253 and contact 254 positioned on the first surface 251.

[0150] Figure 37K shows the structure of Figure 37J after adding a temporary adhesive material 257 covering the first surface 251, epitaxial layer 253, and contact 254 of the thick wafer 242 in preparation for receiving and bonding the first carrier. Figure 37L shows the structure of Figure 37K after adding the first carrier 258 to cover the temporary adhesive material 257 and after forming subsurface laser damage 259 within the thick wafer 242 by irradiating focused laser radiation through the second surface 252 of the thick wafer 242. Figure 37M shows the structure of Figure 37L after bonding the second rigid carrier 260 to the second surface 252 of the thick wafer 242 in proximity to the subsurface laser damage 259. For separation purposes, the second rigid carrier 260 acts as a front carrier intended to remove a portion (i.e., a layer) of the thick wafer 242.

[0151] In certain embodiments, laser radiation can be applied to a self-supporting thick device wafer, and the first and second carriers can be bonded substantially simultaneously to the front and back surfaces of the thick wafer. In certain embodiments, an adhesive material can be applied to one or both of the front and back surfaces of the carrier or wafer.

[0152] Figure 37N shows the items of Figure 37M after the thick wafer 242 has been fractured along subsurface laser damage 259 by applying at least one of the fracturing steps disclosed herein to obtain a first bonded secondary assembly 262A and a second bonded secondary assembly 262B. The first bonded secondary assembly 262A includes a first thin wafer portion 242A (separated from the thick wafer 242 in Figure 37M), an epitaxial layer 253, contacts 254, a temporary adhesive material, and a first carrier 258. The second bonded secondary assembly 262B includes a second thin wafer portion 242B (separated from the thick wafer 242 in Figure 37M), and a second carrier 260. The exposed surfaces 259A of the thin wafer portion 242A and 259B of the thin wafer portion 242B may exhibit surface irregularities due to laser damage and / or fracturing, but these can be reduced by conventional surface processing steps (e.g., grinding, CMP, and / or polishing). Figure 37O shows an operable semiconductor device 264 obtained from the first bonded subassembly 262A by removing the temporary adhesive 257 and the first carrier 258. Such a figure also shows the second thin wafer portion 242B after the second carrier 260 has been removed to prepare the second thin wafer portion 242B for further processing (e.g., epitaxial growth).

[0153] [Examples of methods including the reuse of carrier wafers] Figure 38 is a flowchart illustrating the steps of the method relating to this disclosure. Starting from the top, the laser 266 can focus its laser radiation below the first surface 272 of a thick crystalline material substrate 270 (e.g., a SiC ingot) to create a subsurface laser-damaged region 268. A carrier wafer 224 can then be bonded to the first surface 272 of the crystalline material substrate 270, the carrier wafer 274 comprising a first surface 276 (proximal to the first surface 272 of the substrate 270) and a second surface 278 opposite the first surface 276 of the carrier wafer 274. Such bonding between the carrier wafer 278 and the crystalline material substrate 270 can be performed by any of the methods disclosed herein, such as adhesive bonding or anodic bonding. Details relating to anodic bonding between a crystalline material substrate and a carrier are disclosed in U.S. Patent Application Publication No. 2016 / 0189954, the contents of which are incorporated herein by reference for all purposes. Subsequently, the crushing steps disclosed herein (e.g., cooling of the CTE mismatched carrier, application of ultrasonic energy, and / or application of mechanical force) are applied to crush the crystalline material 270 along the subsurface laser-damaged region 218 to separate the crystalline material portion 280, which is fixed to the carrier wafer 278, from the rest of the crystalline material substrate 270A. The newly exposed surface 282A of the rest of the crystalline material substrate 270A with residual laser damage is smoothed, cleaned, and returned to the beginning of the process (upper left in Figure 38). The newly exposed surface 284 of the removed crystalline material 280 is also smoothed while still attached to the carrier 274. Subsequently, the carrier wafer 274 can be separated from the removed portion of the crystalline material 280, and one or more layers can be epitaxially grown on the crystalline material 280 to form an epitaxial device 280', while the carrier wafer 274 is cleaned and returned to the beginning of the process (upper left of Figure 38) to allow for the removal of another relatively thin section of the crystalline material substrate 270.

[0154] Figure 39 is a schematic cross-sectional view of a portion of the crystalline material substrate (e.g., SiC ingot) 270 of Figure 38 showing subsurface laser damage 268, with the expected kerf-loss material region identified by superimposed dotted lines. The expected kerf-loss material region 290 includes the laser damage 268, and further material 284 that will be mechanically removed (e.g., by grinding and polishing) from the underside 288 (e.g., Si-terminated surface) of the crystalline material portion 280 (e.g., SiC wafer) that will be separated from the substrate 270, and further material 286 that will be mechanically removed (e.g., by grinding and polishing) from the upper surface 282A (e.g., C-terminated surface) of the remaining portion 270A of the substrate 270. The underside 288 of the crystalline material portion 280 is opposite to its upper surface 272. In certain embodiments, the entire kerf-loss material region may have a thickness in the range of 80 to 120 microns so that the SiC provides a substrate top surface 282A and a wafer bottom surface 288 sufficient for further processing.

[0155] [Material processing using multiple grinding stations / steps] In certain embodiments, a crystalline material subjected to laser processing and crushing may be further processed by a plurality of surface grinding steps to remove subsurface damage and edge grinding to impart a chamfered or rounded edge profile, in which case the order of the grinding steps is selected to minimize the likelihood of imparting additional surface damage and to prepare the crystalline material wafer for chemical mechanical planarization, and / or a protective surface coating is employed. Such steps can be performed, for example, using a material processing apparatus according to the embodiments disclosed herein, in which case the exemplary apparatus includes a laser processing station, a crushing station, a plurality of rough grinding stations arranged in parallel downstream of the crushing station, and at least one fine grinding station located downstream of the rough grinding stations. When processing a wire-sawed wafer, it is common to perform edge grinding before surface grinding or polishing to remove wire-sawed surface damage. However, the inventors have found that combining edge grinding of a substrate portion (e.g., wafer) with laser damage with crushing damage increases the likelihood of cracking in the substrate portion. I don't want to be bound by any particular theory regarding the reason for this phenomenon, but at least When edge grinding is performed before any surface treatment (grinding and / or polishing), the exposed cleavage planes resulting from surface fracture are thought to make the surface more susceptible to cracking. For this reason, it has been found that performing at least some surface treatment (e.g., grinding and / or polishing) before edge grinding is beneficial.

[0156] The rough grinding step (i.e., for removing laser and crushing damage along the crushed surface of the substrate portion and bulk substrate) has been found to take significantly longer to complete than the preceding laser processing and crushing steps, and significantly longer than the subsequent precision grinding step. For this reason, multiple rough grinding stations are provided in parallel to eliminate the bottleneck in the production of multiple wafers from bulk crystalline material (e.g., ingots). In certain embodiments, robotic handlers for controlling the loading and unloading of substrate portions may be located upstream and downstream of the multiple rough grinding stations. In certain embodiments, a carrier bonding station may be provided between the laser processing station and the crushing station, and a carrier removal station may be provided (either directly or indirectly) upstream of the edge grinding station. The carrier may preferably remain bonded to the substrate portion for at least some surface grinding steps to reduce the possibility of breakage, especially for thin substrate portions (e.g., wafers), but it is preferable that the carrier be removed before edge grinding (or before coating the wafer with a protective coating before edge grinding).

[0157] In certain embodiments, the carrier bonding station may use a carrier pre-coated with a temporary bonding medium, align and press the carrier against a substrate surface, and expose the bonding medium to the necessary conditions (e.g., heat and pressure) for bonding between the carrier and the substrate. Alternatively, the carrier bonding station may include a coating station that can be used to coat the carrier or substrate as required.

[0158] Figure 40 is a schematic diagram of a material processing apparatus 300 according to one embodiment, which includes a laser processing station 302, a carrier bonding station 303, a material crushing station 304, a plurality of parallel-arranged rough grinding stations 308A, 308B, a precision grinding station 312, a carrier removal station 313, and a CMP station 314. The laser processing station 302 includes at least one laser and a holder for at least one substrate arranged to receive at least one laser beam for forming subsurface laser damage in a crystalline material (e.g., an ingot). The carrier bonding station 303 is configured to bond the crystalline material (having subsurface laser damage) to at least one rigid carrier. The crushing station 304 is arranged to receive one or more assemblies (each containing a substrate bonded to a rigid carrier) from the carrier bonding station 303 and to crush at least one substrate along the subsurface laser damage area to remove substrate portions (which may be similar to wafers bonded to a carrier). Downstream of the crushing station 304, a first rough grinding station 308A and a second rough grinding station 308B are arranged in parallel, and a first robotic handler 306 is provided to alternately feed the substrate portion (as part of a joined assembly) received from the crushing station 304 to either the first rough grinding station 308A or the second rough grinding station 308B. Downstream of the first rough grinding station 308A and the second rough grinding station 308B, a second robotic handler 310 is provided to feed the roughly ground substrate portion (as part of a joined assembly) to the precision grinding station 312. Downstream of the precision grinding station 312, a carrier removal station 313 is provided, which serves to separate the ground substrate portion from the carrier. Downstream of the carrier removal station 313, a chemical mechanical planarization (CMP) station 314 is provided to prepare the substrate portion for further processing such as washing and epitaxial growth. CMP Station 314 removes damage remaining after precision grinding. While functioning to remove grits, precision grinding itself removes damage remaining after rough grinding. In certain embodiments, each rough grinding station 308A, 308B comprises at least one grinding wheel having a grinding surface of less than 5000 grits, and the precision grinding station 312 comprises at least one grinding wheel having a grinding surface of at least 5000 grits. In certain embodiments, each rough grinding station 308A, 308B is configured to remove crystalline material with a thickness of 20 to 100 microns from a crystalline material portion (e.g., a wafer), and the precision grinding station 312 is configured to remove crystalline material with a thickness of 3 to 15 microns. In certain embodiments, each rough grinding station 308A, 308B and / or the precision grinding station 312 may include multiple grinding substations, each substation comprising grinding wheels with different grits.

[0159] The apparatus according to the embodiment shown in Figure 40 can be modified to accommodate edge grinding for imparting rounded or chamfered edge profiles to crystalline substrate portions such as wafers. Such edge profiles reduce the risk of wafer edge breakage. Edge grinding may not be performed if the substrate portion is bonded to a carrier, and therefore the carrier removal station can be located (either directly or indirectly) upstream of the edge grinding station.

[0160] Figure 41 shows a material processing apparatus 320 according to one embodiment, similar to the embodiment in Figure 40, but incorporating an edge grinding station 332. The material processing apparatus 320 includes a laser processing station 322, a carrier bonding station 323, a material crushing station 324, a first robotic handler 326, a plurality of parallel-arranged rough grinding stations 328A, 328B, a second robotic handler 328, a carrier removal station 331, an edge grinding station 332, a precision grinding station 334, and a CMP station 336. The exemplary edge grinding station 332 may be positioned to grip the wafer between the upper and lower gripping portions of a turntable located in close proximity to a rotary grinding tool having a concave dueling surface (for example, as shown in Figure 37G). Gripping the wafer in this manner may cause undesirable damage to the wafer surface (e.g., the Si-terminated surface of a SiC wafer). For this reason, the edge grinding station 332 shown in Figure 41 is positioned upstream of the precision grinding station 334 so that any surface damage inflicted by the edge grinding station 332 can be removed by the precision grinding station 334. The precision grinding station 334 removes only a small thickness of the wafer, which may alter the rounded or chamfered edge profile created by the edge grinding station 332, but a sufficient amount of the rounded or chamfered edge profile remains to prevent wafer edge fracture.

[0161] A method for processing a crystalline material wafer having a first surface having surface damage, wherein the first surface is bounded by an edge, can be performed using the apparatus 320 shown in Figure 41. The method includes grinding the first surface using at least one first grinding apparatus to remove a first portion of the surface damage; edge grinding the edge after grinding the first surface using at least one first grinding apparatus to form a chamfered or rounded edge profile; and grinding the first surface using at least one second grinding apparatus after edge grinding to remove a second portion of the surface damage to the extent sufficient to make the first surface suitable for further processing by chemical mechanical planarization. In certain embodiments, the first grinding apparatus may be implemented in rough grinding stations 328A, 328B, edge grinding may be performed by an edge grinding station 332, and the second grinding apparatus may be implemented in a precision grinding station 312. In certain embodiments, the first surface using at least one first grinding apparatus A carrier removal step can be performed after grinding and before edge grinding to form a chamfered or rounded edge profile.

[0162] In certain embodiments, protective surface coatings may be employed to reduce the likelihood of inflicting additional surface damage during edge grinding and to prepare the crystalline material wafer for chemical planarization. Such surface coatings may include photoresists or any other suitable coating materials and can be applied before edge grinding and removed after edge grinding.

[0163] Figure 42 is a schematic diagram of a material processing apparatus 340 according to one embodiment, which is similar to the embodiment in Figure 40, but with a surface coating station 354 incorporated between the precision grinding station 352 and the edge grinding station 356, and a coating removal station 358 incorporated between the edge grinding station 356 and the CMP station 360. The material processing apparatus 340 further includes a laser processing station 342, a material crushing station 344, a first robotic handler 346, a plurality of parallel-arranged rough grinding stations 348A, 348B, and a second robotic handler 348 upstream of the precision grinding station 352. The coating station 354 may be configured to apply a protective coating (e.g., photoresist) by methods such as spin coating, dip coating, spray coating, or similar. The protective coating should be thick and robust enough to absorb any damage that may be inflicted by the edge grinding station 365. In the case of SiC wafers, the Si-terminated surface may be coated with a protective coating because the Si-terminated surface is typically the surface on which epitaxial growth takes place. The coating removal station 358 may be configured to remove the coating by chemical, thermal, and / or mechanical means.

[0164] A method for processing a crystalline material wafer having a first surface having surface damage, wherein the first surface is bounded by an edge, can be performed using the apparatus 340 shown in Figure 42. The method includes grinding the first surface using at least one first grinding apparatus (e.g., rough grinding stations 348A, 348B) to remove a first portion of the surface damage; then grinding the first surface using at least one second grinding apparatus (e.g., precision grinding station 352) to remove a second portion of the surface damage to the extent sufficient to make the first surface suitable for further processing by chemical mechanical planarization; then forming a protective coating on the first surface (e.g., using a surface coating station 354); then grinding the edges (e.g., using an edge grinding station 356) to form a chamfered or rounded edge profile; and then removing the protective coating from the first surface (e.g., using a coating removal station). The first surface can then be processed by chemical planarization (e.g., by a CMP station 360), which prepares the first surface (e.g., the Si-terminated surface of the wafer) for subsequent processing such as surface cleaning and epitaxial growth.

[0165] In certain embodiments, the gripping device may be configured to hold an ingot having an end face that is not perpendicular to the side wall in order to enable laser processing of the end face to form subsurface damage. In certain embodiments, the gripping effector may be configured to follow an inclined side wall having a circular cross-section when viewed from above. In certain embodiments, the gripping effector may include a joint to allow the gripping effector to follow the inclined side wall.

[0166] Figure 43A is a schematic lateral cross-sectional view of a first gripping device 362 for holding an ingot 364 having end faces 366, 368 that are not perpendicular to the side wall 370, according to one embodiment. The upper end face 366 is positioned horizontally to receive the laser beam 376. A carrier 372 is attached to the lower end face 368, and a chuck 374 (e.g., a vacuum chuck) may hold the carrier 372. A gripping effector 378 having a non-vertical surface is provided for gripping the side wall 370 of the ingot 364, and the gripping effector 378 is positioned at non-vertical angles A1, A2 with respect to a horizontal operating rod 380. When the ingot 364 is held using the gripping device 362 as shown (e.g., near its bottom), the upper end face 366 and the upper portion of the side wall 370 become available for processing using the method disclosed herein.

[0167] Figure 43B is a schematic lateral cross-sectional view of a second gripping device 362' for holding an ingot 364' having end faces 366' and 368' that are not perpendicular to the side wall 370', according to one embodiment. The upper end face 366' is positioned horizontally to receive a laser beam 376, while a carrier 372' is attached to the lower end face 368', and the carrier 372' may be held by a chuck 374'. A gripping effector 378' having a non-perpendicular surface is provided to grip the side wall 370' of the ingot 364', and the gripping effector 378' is positioned at angles A1 and A2 that are not perpendicular to a horizontal operating rod 380'. A pivot joint 382' is provided between the operating rod 380' and the gripping effector 378', thereby facilitating automatic alignment between the gripping effector 378' and the side wall 370' of the ingot 364'.

[0168] In one example, a 150 mm diameter single-crystal SiC substrate (ingot) with a thickness of over 10 mm is used as the starting material for producing a SiC wafer with a thickness of 355 microns. Laser radiation is applied through the C-terminated upper surface of the SiC substrate to form subsurface laser damage. A sapphire support is bonded to the upper surface of the SiC substrate using a thermoplastic adhesive material disclosed herein, and thermal-induced fracturing is performed to separate the upper (wafer) portion of SiC from the rest of the ingot. Both the Si-terminated surface of the separated wafer portion and the C-terminated surface of the rest of the ingot are roughly ground using a 2000-grit grinding wheel (e.g., a metal, glassy, ​​or resin-bonded grinding wheel) to remove all visible laser and fracturing damage. Subsequently, the Si-terminated surface of the separated wafer portion and the C-terminated surface of the remaining ingot are both precision-ground to 7,000 or more grits (e.g., up to 30,000 or more) (e.g., using a glassy grinding surface) to obtain a smoother surface, preferably with an average roughness (Ra) of less than 4 nm, more preferably in the range of 1 to 2 nm Ra. The remaining ingot requires a smooth surface to avoid any impact on subsequent laser processing. The wafer is prepared for CMP and smooth enough to minimize the amount of CMP removal required, as CMP is generally a more expensive process. Typical material removal during precision grinding to remove all residual subsurface damage from rough grinding and any remaining laser damage (both visible and invisible to the naked eye) may be in the thickness range of 5 to 10 microns. The remaining ingot is then returned to the laser for further processing, and the wafer is edge-ground and subjected to chemical mechanical planarization (CMP) to prepare it for epitaxial growth. To completely avoid the risk of scratching the precision-ground Si surface, edge grinding can be performed between rough and precision surface grinding. Material removal during CMP can be within a thickness range of approximately 2 microns. The total material consumed from the substrate (ingot) can be less than 475 microns. If the final wafer thickness is 355 microns, the kerf loss is less than 120 microns.

[0169] [Wafer-to-wafer thickness variation affected by laser power and crystal variations] As already noted in this specification, in order to form sufficient laser damage to break down the crystalline material so that the wafer is obtained at cross-sectional positions starting distal to the seed crystal and gradually approaching the seed crystal, it is necessary to gradually increase the laser power level. This is possible. When forming subsurface damage, using high laser power at each continuous depth position is thought to result in unwanted material loss. Furthermore, due to variability in both damage depth and resolution endpoints relative to the laser beam waist, the thickness difference from wafer to wafer is thought to become significantly larger. Such concepts can be understood by referring to Figures 44 and 45.

[0170] Figure 44 shows the incoming horizontal beam 400 in the propagation direction being focused by lens 404, with the minimum width W at the downstream position 406 corresponding to the focal length f of lens 404. f Figure 45 is a schematic lateral cross-sectional view of a conventional laser focusing device forming an output beam 402 having a beam waist pattern with a beam waist pattern. Downstream from this position 406, the beam widens to a wider region 408. Figure 45 is a schematic lateral cross-sectional view of a vertically oriented focused laser beam 402 that can be guided into a crystalline material, exhibiting a beam waist pattern (having its minimum width at position 406, corresponding to the focal length of the lens (not shown)), where the beam widens thereafter to a wider region 408. When the focused laser beam 402 is guided into a crystalline material (e.g., a substrate such as a SiC ingot), the crystalline material will be thermally decomposed at different threshold points (i.e., depths) determined by factors such as the laser power, the degree of absorption of radiation by the crystalline material (which can be affected by the presence or absence of dopants and / or crystal defects, which can vary depending on the depth (and width) location in the substrate), and the degree of focusing depending on the vertical position. Three different decomposition threshold points 410A to 410C are shown in Figure 45.

[0171] The methods and apparatus disclosed herein make it possible to address the above problem by imaging the top surface of a crystalline material substrate with subsurface laser damage to detect crack-free areas within the substrate, analyzing one or more images to identify conditions indicating the presence of crack-free areas within the substrate, and performing one or more actions in response to the analysis (e.g., when suitable conditions are achieved). Such actions may include performing an additional laser pass at the same depth position and / or changing the instruction set to create subsurface laser damage at the next depth position. Such methods and apparatus facilitate the production of substrate portions of uniform thickness without unwanted material loss.

[0172] Figures 46A to 46C provide plots of the laser output versus wafer sequential wafer identification (i.e., sequential wafer identification (ID) numbers from 1 to 55) obtained from three SiC ingots, respectively, where in each case, a higher wafer ID number corresponds to a greater proximity to the seed crystal (i.e., slice 1 is furthest from the seed crystal). Figure 46A shows the results for the first SiC ingot, where the first wafer group 411A was fractured after subsurface laser damage was formed at a laser power level of approximately 3.75 W, the second wafer group 412A was fractured after subsurface laser damage was formed at a laser power level of approximately 4 W, the third wafer group 413A was fractured after subsurface laser damage was formed at a laser power level of approximately 4.25 W, the fourth wafer group 414A was fractured after subsurface laser damage was formed at a laser power level of approximately 4.5 W, and the fifth wafer group 415A was fractured after subsurface laser damage was formed at a laser power level of approximately 4.6 W. Figure 46B shows the results for the second SiC ingot, where the first wafer group 411B was fractured after forming subsurface laser damage at a laser power level of less than 3.2W, the second wafer group 412B was fractured after forming subsurface laser damage at a laser power level of approximately 3.4W, and the third to fifth wafer groups 413B-415B were fractured after forming subsurface laser damage at higher laser power levels, each approximately 0.25W higher than the previous level. Figure 46C shows the laser power levels ranging from approximately 4W to approximately 5.5W between different wafer groups 411C-420C to successfully form 55 wafers that are sequentially cut from a single SiC ingot (by the formation of subsurface laser damage and subsequent fracture). This shows that 10 different laser power levels were required, according to wafer groups 411C–420C. Figures 46A–46C therefore show the large variation in laser power requirements for forming multiple wafers of substantially the same thickness by laser-assisted cutting, both per ingot and within each ingot.

[0173] Figure 47 is a plot of resistivity (Ohm-cm) versus slice number for 50 wafers per ingot, produced from approximately 50(50) SiC ingots. The superimposed polynomial fit shows that resistivity decreases with increasing slice number. In each case, a larger slice number represents a greater proximity to the seed crystal from which the ingot was epitaxially grown, with slice 1 representing the top of the ingot furthest from the seed crystal. While the resistivity range differs from ingot to ingot, each ingot exhibits consistently decreasing resistivity throughout as its proximity to the seed crystal increases. The resistivity value range on the y-axis in Figure 47 is consistent with N-type SiC. The decrease in resistivity corresponds to increased doping and increased laser absorption.

[0174] Figure 48 is a plot of laser power (watts) versus resistivity for wafers produced from approximately 50(50) SiC ingots, with polynomial fits superimposed, where the laser power represents the value required to achieve successful laser-assisted cutting (with fracturing after subsurface laser damage formation) by the method described herein. Figure 48 shows that while the laser power requirements vary considerably depending on the ingot, the laser power level required to achieve successful cutting decreases as the resistivity level of the ingot increases.

[0175] [Apparatus including a scattering light source and an imaging device] In certain embodiments, the material processing apparatus includes a laser processing station configured to process a crystalline material substrate to form subsurface laser damage therein, the laser processing station including an irradiation device and an imaging device configured to enable detection of conditions indicating the presence of crack-free regions within the crystalline material. The crack-free regions can be used as a visible indicator to determine when additional laser substrate damage is needed at a first mean depth position (to form a first thickness reduction portion of the substrate, such as a first wafer obtained from an ingot) and / or when additional laser power is needed to form laser damage at the next mean depth position (to form a next thickness reduction portion of the substrate, such as a next wafer obtained from an ingot), thereby providing a stable and reproducible laser cutting process with respect to the wafer thickness distribution while avoiding unwanted kerf loss. In this context, the term "average depth position" is used because slight variations in the laser focal depth position (e.g., typically less than 10 microns) may be used between subsurface laser damage formation processes to form the same thickness reduction portion of a substrate, and / or within a single laser damage formation process (e.g., to address the presence of highly doped areas such as doping rings).

[0176] Preferably, the irradiation device and imaging device are positioned to enable imaging of the substrate surface while the substrate is held by the laser processing chuck. This capability allows for inspection (e.g., imaging and analysis in an automated manner) of the substrate without the need to remove the substrate and reposition it in the laser processing chuck in order to quickly assess whether additional laser processing may be required before shredding. This in-situ inspection of the substrate while it is in the laser processing station avoids downtime and increases the utilization of the laser processing tools, thereby improving the throughput of the laser shredding process. In certain embodiments, the laser is moved away from the substrate held by the laser processing chuck so that imaging can be performed without the laser interfering with the irradiation or imaging of the substrate surface. obtain.

[0177] Figures 49A and 49B provide schematic lateral cross-sectional and top plan views, respectively, of the scattering light source 438 and the imaging device 442, which are positioned in close proximity to the crystalline material substrate 430 within the laser processing station 425. Referring to Figure 49A, the crystalline material substrate 430 includes a top surface 433 and subsurface laser damage 434 located inside the substrate 430 below the top surface 433. The subsurface laser damage 434 generally resembles an irregular sawtooth pattern in a direction parallel to the <11-20> direction, if the substrate has a hexagonal crystal structure. The substrate 430 includes a central axis 436. The scattering light source 438 is displaced laterally in a first direction 437A with respect to the central axis 436, and the imaging device 442 is displaced laterally in a second direction 437B opposite to the central axis 436. Both the scattering light source 438 and the imaging device 442 can be displaced upward relative to the top surface 433 of the substrate 430. Furthermore, the scattering light source 438 may be positioned on a first lateral side 431 of the substrate 430, and the imaging device 442 may be positioned on a second lateral side 432 opposite to the substrate 430. In certain embodiments, the definable angle between the light-emitting surface of the scattering light source 438 and the light-receiving surface of the imaging device 442 (which can optionally be expressed as the angle between the beam 440 emitted from the scattering light source 438 and the incident ray 444 received by the imaging device 442) may be in the range of about 100 degrees to about 170 degrees. In certain embodiments, the scattering light source 438 may include one or more suitable light-emitting devices (e.g., light-emitting diodes), and a diffuser plate is positioned between the light-emitting device and the ray 440 emitted from the scattering light source 438. In certain embodiments, the imaging device 442 may include one or more charge-coupled devices (CCDs) or complementary metal-oxide-semiconductor (CMOS) image sensors, optionally arranged in an array.

[0178] Figure 49B provides a top plan view of the same elements as those depicted in Figure 49A. The substrate 430 may include a primary flat 435 (shown in Figure 49B) substantially parallel to the <11-20> direction (shown in Figure 49A). The inventors have found that the orientation of the scattering light source 438 and imaging device 442 relative to the substrate 430 is important to assist in imaging subsurface laser damage. In certain embodiments, the light source can be positioned substantially perpendicular to the primary flat 435 of the substrate 430 and / or within ±5 degrees of the direction perpendicular to the <11-20> direction of the hexagonal structure of the substrate 430, thereby improving the visibility of crack-free areas through the top surface 433 of the substrate 430.

[0179] Figure 50A is an image of the top surface of a crystalline SiC substrate 450 with subsurface laser damage, taken using an apparatus similar to that shown in Figures 49A-49B. The substrate 450 in Figure 50A includes three regions 451-453 of different colors (in the original image) and an irregularly shaped dark region 456 corresponding to a crack-free region within the substrate 450. In the original image, the outermost roughly annular region 451 (including the primary flat 455) is predominantly green, the middle roughly annular region 452 is predominantly red, and the central roughly circular region 453 is predominantly gold. Within the central region 453, an irregularly shaped black region 456 is visible. The black region 456 corresponds to the presence of a crack-free region along the subsurface laser damage below the upper surface of the substrate 450. The different colors (green, red, and gold) of the different regions 451-453 are thought to correspond to the degree of damage caused by the cracks. In certain embodiments, the method disclosed herein involves analyzing at least one characteristic of an image to identify areas with varying degrees of damage due to cracks, and, for the purpose of addressing differences in substrate characteristics that may vary with vertical (depth) and / or horizontal positions within the substrate, forming auxiliary subsurface laser damage at the same (e.g., first) average depth position (to form a first thickness reduction portion of the substrate, e.g., a first wafer), and / or forming subsequent subsurface laser damage at different (e.g., second or next) average depth positions (to form second and subsequent thickness reduction portions of the substrate, e.g., second and next wafers), by adjusting one or more laser damage formation parameters (e.g., laser output, This includes adjusting the laser focusing depth, laser pulse duration, and / or the number of laser damage formation steps.

[0180] Figures 49 to 49B show the use of a scattering light source 438 and an imaging device 442 positioned along both sides of the substrate in the transverse direction, but in certain embodiments, other configurations and / or types of light sources and imaging devices may be used. In certain embodiments, at least one microscope, such as an optical microscope, a confocal microscope, a scanning electron microscope, and / or a transmission electron microscope, may be used to scan one or more portions (or all) of the top surface of the substrate by translating relative to the substrate and the microscope.

[0181] Figure 50B is a schematic diagram of the substrate representation 450A, showing the irregularly shaped dark region 456 of Figure 50A within the dotted-line regions 451A-453A (the outermost region 453A encompassing the primary flat 455A), which substantially correspond to the boundaries between the differently colored regions 451-453 on the top surface of the substrate 450 in Figure 50A. In certain embodiments, all regions except the dark region 456 may be removed from the captured image to facilitate the analysis of one or more area characteristics of the dark region 456.

[0182] Figure 50C is a magnified view of the irregularly shaped dark regions 456 shown in Figures 50A to 50B, where consecutive dark regions are individually numbered 456A to 456D, and rectangular boxes are added around each of the individual regions 456A to 456D. Each dark region 456A to 456D has a maximum length corresponding to L11 to L4 and a maximum width corresponding to W1 to W4, respectively. In certain embodiments, the analysis of the image obtained from the substrate includes identifying conditions indicating the presence of crack-free regions within the crystalline material (e.g., dark and / or black regions in certain embodiments) and quantifying the top area characteristics of one or more crack-free regions within (or at least one top area characteristic). In certain embodiments, the quantified top area characteristics include the total top area of ​​all crack-free regions. In certain embodiments, the quantified apex area characteristics include individually identifying any continuous crack-free region, quantifying the apex area of ​​each continuous crack-free region, and / or quantifying the maximum length and width dimensions of the continuous crack-free region, and / or specifying the length / width aspect ratio of the continuous crack-free region. In certain embodiments, the length and width may be established relative to the crystal orientation and / or the primary flat of the substrate (e.g., the length is perpendicular to the primary flat and the width is parallel to the primary flat). The inventors have found that the presence of a large continuous crack-free region with a given apex area makes it easier to prevent fracture than the presence of a number of discontinuous crack-free regions with the same total apex area. Furthermore, the inventors have found that the orientation and / or aspect ratio of the continuous crack-free region may affect the prevention of fracture. Small, localized black areas indicating crack-free regions generally do not hinder separation by fracturing; however, as the size of the black areas (particularly in the widthwise direction, substantially parallel to the primary flat and / or substantially perpendicular to the laser damage lines) increases, it may be recognized that in such regions, there is a need to add another laser damage formation stroke at the same mean depth position and / or increase the laser power when forming a laser damage region at the next mean depth position.Areas without long cracks (for example, perpendicular to the primary flat) may be less problematic in preventing fracture than areas without wide cracks.

[0183] Figure 51 is a schematic diagram of a material processing apparatus 458 according to one embodiment, which includes a laser processing station 459 comprising a laser 465, at least one translation stage 466 (e.g., preferably an xyz translation stage) configured to facilitate relative movement between the laser 465 and the substrate 460, a scattering light source 468 configured to illuminate the top surface 463 of the substrate 460, and an imaging device 472 configured to produce at least one image of the top surface 463 of the substrate 460. The substrate 460 may include a laser tool chuck. It is positioned on a support 464. Various items within the material processing apparatus 458 are electrically in contact with at least one computing device 474 having associated memory 476. The computing device 474 can control the operation of the scattering light source 468, the imaging device 472, the laser 465, and the translation stage 466. The memory 476 may further store a substrate-specific instruction set (e.g., fabrication recipes) that can be used and modified on individual substrate bases. In certain embodiments, the computing device 474 and memory 476 may be used in performing various steps of the method disclosed herein, including, but not limited to, analyzing a substrate image to identify conditions indicating the presence of crack-free areas within the substrate, quantifying one or more top area characteristics of the crack-free areas, and comparing the top area characteristics to one or more predetermined threshold area characteristics. In certain embodiments, the computing device 474 may be further used to analyze the image and detect the presence of areas of different colors (e.g., other than the presence of black or dark spots), and to adjust the operation of the laser 465 to compensate for contamination of the laser facets in response to such analysis. In certain embodiments, the computing device 474 may be further used to detect the presence of different doping conditions in different areas of the substrate and to change the laser power delivery accordingly. In certain embodiments, in response to the detection of conditions indicating non-uniform doping of the crystalline material, the laser power can be changed during the formation of a subsurface laser damage pattern to provide a first average power level of laser radiation when forming subsurface laser damage in a first doping area, and to provide a second average power level of laser radiation when forming subsurface laser damage in a second doping area, wherein the first and second average power levels are different from each other.

[0184] [Methods including imaging, comparison, and laser processing / power adjustment] Figure 52 is a flowchart 480 illustrating the steps of a first crystalline material processing method, which generally includes generating an image of the top surface of a substrate with subsurface laser damage; analyzing the image to identify the presence of conditions indicating one or more crack-free regions; comparing one or more properties of the crack-free regions to first and second thresholds; and, in response to the comparison, performing a procedure (i.e., (A) performing an additional laser pass at substantially the same depth position by optionally adjusting one or more laser parameters to form auxiliary laser damage, and / or (B) adjusting one or more laser parameters to form subsurface laser damage at second and subsequent depth positions). The method begins in block 482. Proceeding to block 484, the first step is to form subsurface laser damage on a crystalline material substrate along a first (or new) depth position below the top surface of the substrate, having at least one subsurface laser damage pattern (optionally including at least one group of substantially parallel lines as disclosed herein), wherein the at least one subsurface laser damage pattern is configured to facilitate the formation of at least one group of multiple cracks within the crystalline material, propagating outward from the at least one subsurface laser damage pattern. Proceeding to block 486, the second step is to generate at least one image of the top surface of the substrate. In certain embodiments, the image generation step is to irradiate the top surface with a scattering light source located on a first transverse side of the substrate (preferably substantially perpendicular to the primary flat of the substrate and / or within ±5 degrees from the perpendicular to the <11-20> direction of the hexagonal structure) and to acquire at least one image using an imaging device located on a second transverse side of the substrate opposite the first transverse side. In certain embodiments, one or more alternative or additional imaging methods disclosed herein may be used.

[0185] Moving on to block 488, the next step is to identify the conditions that indicate the presence of crack-free regions within the crystalline material (e.g., dark and / or black regions in a particular embodiment). This includes analyzing at least one image. Optionally, the top area characteristics of at least one of one or more crack-free regions within the crack-free region may be quantified, and the quantified top area characteristics may optionally include the total top area of ​​all crack-free regions. In certain embodiments, the quantified top area characteristics include individually identifying any consecutive crack-free regions, as well as quantifying the top area of ​​each consecutive crack-free region, and / or quantifying the maximum length and width dimensions of the consecutive crack-free regions, and / or determining the length / width aspect ratio of the consecutive crack-free regions. In certain embodiments, the length and width may be established relative to the crystal orientation and / or the primary flat of the substrate (e.g., the length is perpendicular to the primary flat and the width is parallel to the primary flat). The inventors have found that the presence of a large consecutive crack-free region with a given top area makes it easier to prevent fracture than the presence of a number of discontinuous crack-free regions with the same total top area. Furthermore, the inventors found that the orientation and / or aspect ratio of continuous crack-free regions can affect the prevention of fracturing. Small, localized black regions indicating crack-free areas generally do not hinder separation by fracturing, but as the size of the black regions (particularly in the width direction, substantially parallel to the primary flat and / or substantially perpendicular to the laser damage lines) increases, it may be recognized that in such regions, it is necessary to increase the laser power when adding another laser damage formation stroke at the same average depth position and / or forming a laser damage region at the next average depth position. Crack-free regions with a large length (e.g., perpendicular to the primary flat) may pose no problem in preventing fracturing than crack-free regions with a large width.

[0186] Proceeding to determination block 490, one or more properties of the crack-free region (optionally including at least one quantified top area property) are compared to at least one predetermined first threshold. The first threshold may be (but is not limited to) one or more of the following: a top area threshold for a continuous crack-free region, a threshold for the total top area of ​​a crack-free region, a threshold for the maximum width of a crack-free region, a threshold for the maximum length / width aspect ratio, or similar. If at least one property of the crack-free region does not exceed at least one predetermined first threshold, the method proceeds to block 498, in which the substrate is moved to a crushing station to produce a first thickness reduction portion of the substrate (e.g., a first wafer from an ingot) whose thickness roughly corresponds to a first mean depth position. Conversely, if at least one characteristic of a crack-free region exceeds at least one predetermined first threshold, the method proceeds to block 492, thereby modifying the instruction set associated with the substrate (e.g., fabrication recipe) by incrementally adjusting at least one laser parameter for forming subsurface laser damage when creating subsurface laser damage patterns at a second mean depth position and any next mean depth position of the substrate (e.g., for forming at least one additional thickness reduction portion of the substrate, e.g., a second and subsequent wafer from an ingot). Adjustable laser parameters include one or more of the following: laser power, laser focal depth, number of laser strokes, laser stroke interval, laser pulse width, etc. In certain embodiments involving a change in laser power, the instruction set is modified to increase the average laser power by a value within the range of 0.10 to 0.50 watts, or 0.15 to 0.35 watts, or 0.20 to 0.30 watts, or by a value of approximately 0.25 watts. After adjusting one or more laser parameters to form subsequent laser damage at the second and subsequent mean depth positions, the formation of additional damage at the previously established first mean depth position does not necessarily occur.In decision block 494, a determination is made as to whether additional laser damage at the first average depth position may be necessary to promote fragmentation.

[0187] The decision block 494 determines one or more characteristics of the crack-free region (optionally selected to be less The method includes the step of comparing a (including at least one quantified top area characteristic) with at least one second predetermined threshold. In certain embodiments, the second predetermined threshold is greater than the first predetermined threshold. The second threshold may be (but is not limited to) one or more of the following: a top area threshold for a continuous crack-free region, a threshold for the total crack-free top area, a threshold for the maximum crack-free width, a threshold for the maximum length / width aspect ratio, or similar. If at least one characteristic of the crack-free region does not exceed at least one second predetermined threshold, the method proceeds to block 498, according to which it is determined that no additional laser damage is needed to support the fracture of the substrate along the first mean depth position, and the substrate is moved to the fracture station. Conversely, if at least one characteristic of the crack-free region exceeds at least one second predetermined threshold, the method proceeds to block 496, according to which additional subsurface laser damage is formed along the first mean depth position. In certain embodiments, this involves forming auxiliary subsurface laser damage at or near a first mean depth position to assist at least one subsurface laser damage pattern, and performing relative movement between the laser and the substrate while supplying laser radiation focused into the interior of the substrate, at least in crack-free regions, but optionally across the entire substrate, to facilitate the formation of additional cracks inside the crystalline material propagating outward from the auxiliary at least one subsurface laser damage pattern. After the formation of this auxiliary subsurface laser damage, the method proceeds to block 498, according to which the substrate is moved to a crushing station.

[0188] Moving on to block 500, in a particular embodiment, a carrier can be bonded to a substrate in a crushing station to form a bonded assembly. Subsequently, according to block 502, the crystalline material is crushed along a first depth position to separate the bonded assembly (including the carrier and the removed portion of the substrate) from the rest of the substrate, such a step serves to expose a new top surface of the substrate. Subsequently, according to block 504, the substrate can be returned to the laser processing station (optionally after surface treatment such as grinding and / or polishing of the newly exposed substrate surface) to allow for another subsurface laser damage step according to block 484. If the instruction set associated with the substrate has been modified to increase the average laser power according to block 492, the modified instruction set will be used to perform the steps described in block 484 to form subsurface damage. This modified instruction set is preferably stored in memory and associated with a particular substrate (for example, in a record of a relational database including a substrate identifier and parameters for forming subsurface laser damage within the substrate). In this way, a substrate-specific recipe for forming subsurface laser damage is maintained, and this can be dynamically updated.

[0189] After the assembled parts joined in block 502 are broken off from the substrate, the assembled parts can be moved to one or more surface processing stations (according to block 506) to modify the substrate portion attached to the carrier. Examples of surface processing steps that can be performed include rough grinding, edge grinding, fine grinding, and cleaning, according to blocks 508, 510, 512, and 514, respectively. At this point, the processed substrate portion may be ready for epitaxial growth.

[0190] Figure 53 is a flowchart 520 illustrating the steps of a second crystalline material processing method, which generally involves generating an image of the top surface of a substrate with subsurface laser damage, analyzing the image to quantify the top area characteristics of one or more crack-free regions, comparing the top area characteristics to first and second threshold area characteristics, and taking action in response to the comparison (i.e., performing an additional laser pass at the same depth position and / or at the next depth position) to improve the reliability of producing a substrate portion (e.g., a wafer) from the substrate. The method includes adjusting the output for subsurface laser damage. The method begins in block 522. Proceeding to block 524, the first step includes forming subsurface laser damage on a crystalline material substrate along a (new) first depth position below the top surface of the substrate, having at least one subsurface laser damage pattern (optionally including at least one group of substantially parallel lines), wherein the at least one subsurface laser damage pattern is configured to facilitate the formation of at least one group of multiple cracks within the crystalline material, substantially propagating outward from the at least one subsurface laser damage pattern. Proceeding to block 526, the second step includes generating at least one image of the top surface of the substrate. In certain embodiments, the image generation step includes illuminating the top surface with a scattering light source located on a first transverse side of the substrate (preferably substantially perpendicular to the primary flat of the substrate and / or within ±5 degrees from the perpendicular to the <11-20> direction of the hexagonal structure) and acquiring at least one image using an imaging device located on a second transverse side of the substrate opposite the first transverse side. In certain embodiments, one or more alternative or additional imaging methods disclosed herein may be used.

[0191] Proceeding to block 528, further steps include analyzing at least one image to identify conditions indicating the presence of crack-free regions within the crystalline material (e.g., dark and / or black regions in a particular embodiment), and quantifying the top area characteristics (or at least one top area characteristics) of one or more crack-free regions within. In a particular embodiment, the quantified top area characteristics include the total top area of ​​all crack-free regions. In a particular embodiment, the quantified top area characteristics include individually identifying any consecutive crack-free regions, as well as quantifying the top area of ​​each consecutive crack-free region, and / or quantifying the maximum length and width dimensions of the consecutive crack-free regions, and / or determining the length / width aspect ratio of the consecutive crack-free regions. In a particular embodiment, the length and width may be established relative to the crystal orientation and / or the primary flat of the substrate (e.g., the length is perpendicular to the primary flat and the width is parallel to the primary flat).

[0192] Proceeding to decision block 530, at least one quantified top area characteristic is compared to at least one first predetermined area (or area characteristic) threshold. The first threshold may include one or more of the following: a top area threshold for a continuous crack-free region, a threshold for the total crack-free top area, a threshold for the maximum crack-free width, a threshold for the maximum length / width aspect ratio, or similar. If at least one quantified top area characteristic does not exceed at least one first predetermined threshold area characteristic, the method proceeds to block 538, in which case the substrate is moved to the crushing station. Conversely, if at least one quantified top area characteristic exceeds at least one first predetermined threshold area characteristic, the method proceeds to block 532, in which the instruction set associated with the substrate (e.g., fabrication recipe) is modified by progressively increasing the average laser power for forming subsurface laser damage when creating subsurface laser damage patterns at a second mean depth position and any next mean depth position of the substrate. (Adjustable laser parameters may include, in addition or alternatively, one or more of the following: laser focal depth, number of laser strokes, laser stroke interval, laser pulse width, etc.) In certain embodiments, the instruction set is modified to increase the average laser power by a value within the range of 0.10 to 0.50 watts, or 0.15 to 0.35 watts, or 0.20 to 0.30 watts, or by a value of approximately 0.25 watts. After gradually increasing the laser power to form the next laser damage at the second and subsequent average depth positions, the formation of additional damage at the previously established first average depth position does not necessarily occur. A determination is made in decision block 534 as to whether additional laser damage may be necessary to facilitate shattering.

[0193] Decision block 534 includes the step of comparing at least one quantified top area characteristic with at least one second predetermined threshold area characteristic. In certain embodiments, the second predetermined threshold area characteristic is greater than the first predetermined threshold area characteristic. The second threshold area characteristics may include one or more of the following: a top area threshold for a continuous crack-free region, a threshold for the total crack-free top area, a threshold for the maximum crack-free width, a threshold for the maximum length / width aspect ratio, or similar. If at least one quantified top area characteristic does not exceed at least one second predetermined threshold area characteristic, the method proceeds to block 538, according to which it is determined that no additional laser damage is required to support the fracture of the substrate along the first mean depth position, and the substrate is moved to the fracture station. Conversely, if at least one quantified top area characteristic exceeds at least one second predetermined threshold area characteristic, the method proceeds to block 536, according to which auxiliary subsurface laser damage is formed along the first mean depth position. In certain embodiments, this involves forming auxiliary subsurface laser damage to assist at least one subsurface laser damage pattern at or near a first mean depth position, and causing relative movement between the laser and the substrate while supplying focused laser radiation within the substrate to facilitate the formation of additional cracks within the crystalline material propagating outward from the auxiliary at least one subsurface laser damage pattern. After the formation of this auxiliary subsurface laser damage, the method proceeds to block 538, according to which the substrate is moved to a crushing station.

[0194] Moving on to block 540, in a particular embodiment, a carrier can be bonded to a substrate in a crushing station to form a bonded assembly. Subsequently, according to block 542, the crystalline material is crushed along a first depth position to separate the bonded assembly (including the removed portions of the carrier and substrate) from the remaining portion of the substrate, such a step serves to expose a new top surface of the substrate. Subsequently, according to block 544, the substrate can be returned to the laser processing station (optionally after surface treatment such as grinding and / or polishing of the newly exposed substrate surface) to allow for another subsurface laser damage step according to block 524. If the instruction set associated with the substrate has been modified to increase the average laser power according to block 532, the modified instruction set will be used to perform the steps described in block 524 to form subsurface damage. This modified instruction set is preferably stored in memory and associated with a particular substrate, for example, in a record of a relational database containing a substrate identifier and parameters for forming subsurface laser damage within the substrate.

[0195] After the assembled parts joined in block 542 are broken off from the substrate, the assembled parts can be moved to one or more surface processing stations (according to block 546) to modify the substrate portion attached to the carrier. Examples of surface processing steps that can be performed include rough grinding, edge grinding, fine grinding, and cleaning, according to blocks 548, 550, 552, and 554, respectively. At this point, the processed substrate portion may be ready for epitaxial growth.

[0196] [Typical computer systems usable with systems and methods] Figure 54 is a schematic diagram of a generalized representation of a computer system 600 (optionally embodied in a computing device) which may be included in any component of the systems or methods disclosed herein. In this regard, the computer system 600 is adapted to execute instructions from a computer-readable medium to perform these and / or any functions or processes described herein. In this regard, the computer system 600 in Figure 54 supports the communication services It may include a set of instructions that are executable to program and configure programmable digital signal processing circuits to support scaling. The computer system 600 may be connected to (e.g., networked) other machines in a LAN, intranet, extranet, or the internet. Although only a single device is shown, the term “device” may also be interpreted to include any set of devices that individually or in combination execute one or more sets of instructions for performing any one or more of the methodologies discussed herein. The computer system 600 may also be one or more circuits contained in an electrical circuit board card such as a printed circuit board (PCB), a server, a personal computer, a desktop computer, a laptop computer, a personal digital assistant (PDA), a computing pad, a mobile device, or any other device, and may represent, for example, a server or a user’s computer.

[0197] In this embodiment, the computer system 600 includes a processing device or processor 602, main memory 604 (e.g., read-only memory (ROM), flash memory, dynamic random-access memory (DRAM) such as synchronous DRAM (SDRAM), etc.), and static memory 606 (e.g., flash memory, static random-access memory (SRAM), etc.), which can communicate with each other via a data bus 608. Alternatively, the processing device 602 may be connected to the main memory 604 and / or static memory 606 directly or via some other means of connection. The processing device 602 may also be a control device, and the main memory 604 or static memory 606 may be any type of memory.

[0198] The processing device 602 represents one or more general-purpose processing devices, such as a microprocessor, a central processing unit, or the like. More specifically, the processing device 602 may be a composite instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a processor implementing another instruction set, or another processor implementing a combination of instruction sets. The processing device 602 is configured to execute processing logic in instructions for performing the operations and steps discussed herein.

[0199] The computer system 600 may further include a network interface device 610. The computer system 600 may also include, or may not include, an input unit 612 configured to receive inputs and selections to be communicated to the computer system 600 when an instruction is executed. The computer system 600 may also include, or may not include, an output unit 614, which may include, but is not limited to, a display, a video display unit (e.g., a liquid crystal display (LCD) or a cathode ray tube (CRT)), an alphanumeric input device (e.g., a keyboard), and / or a cursor control device (e.g., a mouse).

[0200] The computer system 600 may or may not include a data storage device containing instructions 616 stored in a computer-readable medium 618. Instructions 616 may also be present, in whole or in part, in main memory 604 and / or processing device 602 during their execution by the computer system 600, the main memory 604 and processing device 602 also constitute the computer-readable medium. Instructions 616 may be further transmitted or received via the network 620 and the network interface device 610.

[0201] In one embodiment, the computer-readable medium 618 is shown as a single medium. However, the term “computer-readable medium” should be interpreted to include a single or multiple mediums that store one or more sets of instructions (e.g., a centralized or distributed database, and / or associated caches and servers). The term “computer-readable medium” should also be interpreted to include any medium capable of storing, encoding, or transporting a set of instructions executed by a processing device, causing the processing device to execute one or more of the methodologies of the embodiments disclosed herein. The term “computer-readable medium” should therefore be interpreted to include, but not limited to, solid-state memory, optical media, and magnetic media.

[0202] The embodiments disclosed herein include a variety of steps. The steps of the embodiments disclosed herein can be implemented or executed by hardware components, or they can be embodied as machine-executable instructions, which can be used to cause a general-purpose or dedicated processor programmed with these instructions to perform the steps. Alternatively, these steps can be performed by a combination of hardware and software.

[0203] Embodiments disclosed herein may be provided as computer program products or software, which may include a machine-readable medium (or computer-readable medium) storing instructions that can be used to program a computer system (or other electronic device) to perform the steps of the embodiments disclosed herein. The machine-readable medium includes any mechanism for storing or transmitting information in a form readable by a machine (e.g., a computer). For example, the machine-readable medium includes machine-readable storage media (e.g., ROM, random-access memory ("RAM"), magnetic disk storage media, optical storage media, flash memory devices, etc.), and similar.

[0204] Unless explicitly stated otherwise, and as is evident from the above considerations, any discussion throughout the explanation using terms such as “analyze,” “process,” “calculate,” “determine,” and “display” is understood to refer to the operation and processing of a computer system or similar electronic computing device that manipulates and converts data and memory, expressed as physical (electronic) quantities in the registers of a computer system, into other data similarly expressed as physical volumes in the memory or registers of a computer system or other such information storage, transmission, or display devices.

[0205] The algorithms and displays presented herein do not inherently relate to any particular computer or other device. Various systems may be used with the program according to the teachings herein, or it may be convenient to construct a device specialized by performing the required method steps. The structures required for these various systems are disclosed in the above description. Furthermore, the embodiments described herein do not refer to any particular programming language in their description. It will be understood that various programming languages ​​may be used to implement the teachings of embodiments such as those described herein.

[0206] Those skilled in the art will further understand that various exemplary logic blocks, modules, circuits, and algorithms described in connection with the embodiments disclosed herein may be implemented as instructions stored in electronic hardware, memory, or another computer-readable medium and executed by a processor or other processing device, or a combination of both. The system components described herein may, for example, be used in any circuit, hardware component, integrated circuit (IC), or IC chip. The memory disclosed herein may be of any type and size and may be configured to store any desired type of information. To clearly demonstrate this compatibility, various examples are provided. Specific components, blocks, modules, circuits, and steps are described above in general terms of their functionality. How such functionality is implemented depends on the specific application, design choices, and / or design constraints imposed on the system as a whole. A person skilled in the art may implement the described functionality in various ways for each specific application, but such implementation decisions should not be construed as resulting in a departure from the scope of the embodiments specified herein.

[0207] The various exemplary logic blocks, modules, and circuits described in relation to the embodiments disclosed herein may be implemented or run using processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. Furthermore, the control device may be a processor. The processor may be a microprocessor, but alternatively, the processor may be any conventional processor, control device, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices (e.g., a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration).

[0208] Embodiments disclosed herein can be embodied in hardware and as instructions that may reside in hardware, for example, in RAM, flash memory, ROM, electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), registers, hard disks, removable disks, CD-ROMs, or any other form of computer-readable media known in the art. The storage medium is coupled to a processor to enable reading and writing of information to the storage medium. Alternatively, the storage medium may be integrated with the processor. The processor and storage medium may reside in an ASIC. The ASIC may reside in a remote station. Alternatively, the processor and storage medium may reside as separate components in a remote station, base station, or server.

[0209] It should also be noted that the operating steps described in any embodiment of this specification are provided for illustrative purposes and for consideration. The operations described may be performed in an order different from the numerous examples given. Furthermore, operations described as a single operating step may actually be performed as several different steps. Moreover, one or more operating steps considered in the embodiments may be combined. Those skilled in the art will also understand that information and signals can be represented using a variety of techniques and methods. For example, data, instructions, commands, information, signals, bits, symbols, and chips that may be mentioned throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields, particles, optical fields, or any combination thereof.

[0210] Unless explicitly stated otherwise, none of the methods described herein are intended to be construed as requiring the steps to be performed in a specified order. Therefore, unless the order in which the steps should be followed is not actually specified in the method claims, or unless it is otherwise explicitly stated in the claims or description that those steps are limited to a particular order, no particular order is intended to be inferred.

[0211] Unless expressly indicated otherwise in this specification, the information disclosed herein is It is intended that any or more features or characteristics of any one or more embodiments can be combined with any or more features or characteristics of other embodiments.

[0212] The technical benefits that may be obtained by one or more embodiments of the present disclosure include: improved reproducibility when manufacturing wafers of uniform thickness by laser processing and subsequent crushing from a crystalline material substrate (e.g., an ingot) while avoiding unwanted material loss; addressing variations in laser power requirements, both per substrate and at different depth locations within a single substrate, when performing laser-assisted cutting methods; improved detection of crack-free areas in crystalline material substrates with subsurface laser damage; reduced crystalline material kerf loss compared to wire sawing; reduced processing time and increased throughput of crystalline material wafers and resulting devices compared to wire sawing; reduced laser processing time compared to conventional laser-based methods; reduced force required to cause crushing along laser-damaged areas; reduced need for post-separation surface smoothing to remove laser damage after separation; and / or reduced bowing and breakage of crystalline material.

[0213] Those skilled in the art will recognize improvements and modifications to preferred embodiments of this disclosure. All such improvements and modifications are deemed to fall within the scope of the concepts disclosed herein and in the following claims.

Claims

1. The method involves supplying and executing laser radiation focused along a first average depth position within the crystalline material of a substrate to form subsurface laser damage having at least one subsurface laser damage pattern, wherein the at least one subsurface laser damage pattern is configured to facilitate the formation of at least one group of cracks within the substrate that propagate outward from the at least one subsurface laser damage pattern. After the formation of the at least one subsurface laser damage pattern, an image of the top surface of the substrate is generated. To identify the conditions indicating the presence of crack-free regions within the substrate, the analysis involves analyzing at least one image, wherein the analysis includes quantifying the top area characteristics of one or more crack-free regions within the substrate and comparing the top area characteristics with at least one predetermined threshold area characteristic. In response to the aforementioned analysis, the following steps (i) or (ii): (i) To form a first thickness reduction portion of the substrate, a step of causing relative movement between the laser and the substrate while supplying the radiation of the laser focused into the interior of the substrate in at least the crack-free region, in order to form an auxiliary subsurface laser damage to assist the at least one subsurface laser damage pattern and to facilitate the formation of additional cracks in the crack-free region along or adjacent to the first mean depth position, (ii) performing at least one of the steps of modifying the instruction set associated with the substrate for forming subsurface laser damage when creating subsurface laser damage patterns at a second mean depth position and any subsequent mean depth position in the substrate for the purpose of forming at least one additional thickness reduction portion of the substrate, A method for processing crystalline materials, including the processing method described above.

2. The at least one predetermined threshold area characteristic includes a first predetermined threshold area characteristic and a second predetermined threshold area characteristic, wherein the second predetermined threshold area characteristic is greater than the first predetermined threshold area characteristic. Step (i) is performed if the top area characteristic is at least the same size as the first predetermined threshold area characteristic, and then step (ii) is performed. The method for processing a crystal material according to claim 1, comprising performing step (i) when the top area characteristic is at least the same size as the second predetermined threshold area characteristic.

3. A method for processing a crystalline material according to claim 1, comprising performing both steps (i) and (ii) in response to the aforementioned analysis.

4. Step (ii) is a method for processing a crystalline material according to any one of claims 1 to 3, comprising adjusting at least one of (a) the average laser power, (b) the laser focusing depth relative to the exposed surface of the substrate, or (c) the number of laser damage formation steps when creating a subsurface laser damage pattern at the second average depth position and any subsequent average depth position in the substrate.

5. The method for processing a crystalline material according to claim 4, wherein modifying the instruction set in accordance with step (ii) includes increasing the average laser power by a value within the range of 0.15 to 0.35 watts.

6. A method for processing a crystalline material according to any one of claims 1 to 5, step (i) comprising adjusting at least one of (a) the average laser power or (b) the laser focusing depth relative to the exposed surface of the substrate when creating the auxiliary subsurface laser damage to assist the at least one subsurface laser damage pattern and to facilitate the formation of additional cracks in the crack-free region along or adjacent to the first average depth position.

7. A method for processing a crystalline material according to any one of claims 1 to 6, wherein the substrate has a substantially circular edge having a primary flat, and generating the at least one image comprises (a) illuminating the top surface with scattered light produced by a scattering light source located on a first lateral side of the substrate and substantially perpendicular to the primary flat, and (b) capturing the at least one image using an imaging device located on a second lateral side opposite to the substrate.

8. The aforementioned crystalline material has a hexagonal structure, Generating the at least one image includes (a) illuminating the top surface with scattered light generated by a scattering light source located on the first lateral side of the substrate and positioned within ±5 degrees from the direction perpendicular to the <11-20> direction of the hexagonal crystal structure, and (b) capturing the at least one image using an imaging device located on the second lateral side of the substrate opposite to the first lateral side. A method for processing crystalline material according to any one of claims 1 to 6.

9. The at least one subsurface laser damage pattern includes a first subsurface laser damage pattern and a second subsurface laser damage pattern formed after the first subsurface laser damage pattern. The first subsurface laser damage pattern includes a first plurality of substantially parallel lines, and the second subsurface laser damage pattern includes a second plurality of substantially parallel lines. The lines of the second plurality of substantially parallel lines are dispersed between the lines of the first plurality of substantially parallel lines, At least some of the second substantially parallel lines do not intersect any of the first substantially parallel lines. A method for processing a crystalline material according to any one of claims 1 to 8.

10. Each of the second substantially parallel lines is positioned between different pairs of adjacent lines among the first substantially parallel lines. The method for processing a crystalline material according to claim 9.

11. The aforementioned crystalline material has a hexagonal structure, Each of the first plurality of substantially parallel lines and each of the second plurality of substantially parallel lines are within ±5 degrees from the direction perpendicular to the <11-20> direction of the hexagonal structure and are substantially parallel to the surface of the substrate. The method for processing a crystalline material according to claim 9.

12. The at least one subsurface laser damage pattern includes a first subsurface laser damage pattern and a second subsurface laser damage pattern formed after the first subsurface laser damage pattern. The aforementioned at least one group of substantially parallel lines includes a first group of substantially parallel lines and a second group of substantially parallel lines. The first plurality of substantially parallel lines is non-parallel to the second plurality of substantially parallel lines. They are parallel, The angular direction of the second plurality of substantially parallel lines differs from the angular direction of the first plurality of substantially parallel lines by 10 degrees or less. At least some of the second substantially parallel lines do not intersect any of the first substantially parallel lines. A method for processing a crystalline material according to any one of claims 1 to 7.

13. The at least one subsurface laser damage pattern further includes a third subsurface laser damage pattern formed after the second subsurface laser damage pattern, The aforementioned group of substantially parallel lines further includes a third group of substantially parallel lines, The aforementioned group of at least one set of cracks includes a first, second, and third set of cracks. The first subsurface laser damage pattern forms a plurality of first cracks in the interior of the substrate, propagating transversely outward from a plurality of substantially parallel lines. The second subsurface laser damage pattern forms a second plurality of cracks in the interior of the substrate that propagate transversely outward from the second plurality of substantially parallel lines, and the second plurality of cracks do not connect with the first plurality of cracks. The third subsurface laser damage pattern forms a third plurality of cracks in the interior of the substrate that propagate transversely outward from the third plurality of substantially parallel lines, and at least some of the third plurality of cracks connect with at least some of the first plurality of cracks and at least some of the second plurality of cracks. The method for processing crystalline material according to claim 12.

14. The detection of conditions indicating non-uniform doping of the crystalline material over at least a portion of the surface of the substrate, wherein the non-uniform doping includes a first doping region and a second doping region. In response to the detection of the conditions indicating heterogeneous doping of the crystalline material, the following steps (A) or (B): (A) Modifying the laser output to provide a first power level of laser radiation when forming subsurface laser damage in the first doping region and a second power level of laser radiation when forming subsurface laser damage in the second doping region, during the formation of the at least one subsurface laser damage pattern, or (B) When forming subsurface laser damage in either the first doping region or the second doping region, perform at least one of the following steps: A method for processing a crystalline material according to any one of claims 1 to 13, further comprising:

15. A method for processing a crystalline material according to any one of claims 1 to 14, further comprising crushing the crystalline material substantially along the at least one subsurface laser damage pattern such that first and second crystalline material portions are produced, each having a reduced thickness compared to the substrate but substantially the same length and width as the substrate.

16. The method for processing a crystalline material according to any one of claims 1 to 15, wherein the substrate contains silicon carbide.

17. The method for processing a crystalline material according to any one of claims 1 to 16, wherein the substrate includes an ingot having a diameter of at least 150 mm.

18. A laser processing station configured to process a crystalline material substrate, A laser configured to form a subsurface laser damage region inside the substrate, At least one translation stage configured to perform relative movement between the laser and the substrate, A scattering light source configured to illuminate the top surface of the substrate, and positioned so as to be located on the first lateral side of the substrate, A laser processing station comprising: an imaging device configured to generate at least one image of the top surface of the substrate, and configured to be positioned on a second lateral side of the substrate opposite to the first lateral side; A material processing apparatus equipped with the following features.

19. The substrate has a substantially circular edge with a primary flat, The scattering light source is positioned substantially perpendicular to the primary flat on the first lateral side of the substrate. The material processing apparatus according to claim 18.

20. The aforementioned crystalline material has a hexagonal structure, The scattering light source is positioned on the first lateral side of the substrate, within ±5 degrees from a direction perpendicular to the <11-20> direction of the hexagonal crystal structure. The material processing apparatus according to claim 18 or 19.

21. The material processing apparatus according to claim 18 or 19, further comprising a computing device configured to analyze the at least one image to identify conditions indicating the presence of crack-free regions within the substrate.

22. The computing device, in response to the analysis performed by the computing device, performs the following steps (i) or (ii): (i) For the purpose of forming a first thickness reduction portion of the substrate, a step of performing relative movement between the laser and the substrate while supplying the radiation of the laser focused into the interior of the substrate in at least the crack-free region, in order to form an auxiliary subsurface laser damage to assist at least one subsurface laser damage pattern in the substrate and to facilitate the formation of additional cracks in the crack-free region along or adjacent to the first mean depth position, or (ii) The material processing apparatus according to claim 21, further configured to perform at least one of the steps of (ii) modifying the instruction set associated with the substrate for forming subsurface laser damage when creating subsurface laser damage patterns at a second mean depth position and a subsequent arbitrary mean depth position in the substrate for the purpose of forming at least one additional thickness reduction portion of the substrate.

23. The material processing apparatus according to claim 22, wherein the analysis performed by the computing device includes quantifying the top area characteristics of one or more crack-free regions within the substrate and comparing the top area characteristics with at least one predetermined threshold area characteristics.

24. The at least one predetermined threshold area characteristic includes a first predetermined threshold area characteristic and a second predetermined threshold area characteristic, wherein the second predetermined threshold area characteristic is the first predetermined threshold area characteristic Larger than sex, The computing device is configured to control the material processing apparatus to perform step (ii) if the top area characteristic is at least the same size as the first predetermined threshold area characteristic. The computing device is configured to control the material processing apparatus to perform step (i) when the top area characteristic is at least the same size as the second predetermined threshold area characteristic. The material processing apparatus according to claim 23.

25. The material processing apparatus according to any one of claims 22 to 24, further comprising a memory configured to store the instruction set associated with the substrate for forming subsurface laser damage on the substrate, wherein the memory is accessible by the computing device.

26. The material processing apparatus according to any one of claims 18 to 25, further comprising a crushing station configured to receive the substrate from the laser processing station.

Images