SEMI-CONDUCTIVITY SUBSTRATE MANUFACTURING SYSTEMS AND RELATED PROCESSES

Abstract

Method for separating a wafer from a rod comprising a semiconductor material, the method comprising: Creating a damage layer in a rod comprising semiconductor material, wherein the rod has a first end and a second end; Cooling the first end of the rod; and Warming the second end of the stick; where a thermal gradient between the cooled first end and the heated second end helps a silicon carbide wafer to detach from the rod at the damage layer, wherein the cooling of the first end of the rod and the heating of the second end of the rod occur simultaneously; and wherein the second end of the rod is heated by bringing a heated chuck into contact or by applying heat pulses using a heated chuck.

Description

BACKGROUND 1. Technical field

[0001] The aspects of this document relate generally to methods for forming semiconductor substrates from a rod. 2. Background

[0002] Semiconductor substrates can be formed by separating a larger section of a semiconductor material (a rod) created using various processing techniques. In some formation techniques, the rod is shaped to have a uniform crystal structure throughout (single crystal), ensuring that all semiconductor substrates formed from the rod have the same crystalline structure.

[0003] Documents EP 2 221 138 A1, DE 10 2015 006 971 A1, DE 10 2015 000 450 A1 and DE 10 2010 013 549 A1 are known from the prior art. SUMMARY

[0004] Implementations of a method for separating a wafer from a rod containing semiconductor material may include: generating a damage layer in a rod semiconductor material with a first end and a second end. The method may involve cooling the first end of the rod and heating the second end. A thermal gradient can be created between the cooled first end and the heated second end of the rod, which can help a silicon carbide wafer separate from the rod at the damage layer.

[0005] Implementations of a method for separating a wafer from a silicon carbide rod may include one, all, or any of the following: The damaged layer can be created by laser irradiation.

[0006] Cooling the first end of the rod and heating the second end can be done simultaneously. The second end of the rod can be heated by bringing it into contact with a heated chuck.

[0007] Heating the second end of the rod may involve applying heat pulses using a heated chuck.

[0008] Cooling the first end of the rod may also involve bringing the first end of the rod into contact with liquid nitrogen. Heating the second end of the rod may also involve applying a hot liquid to the second end of the rod.

[0009] Cooling the first end of the rod may further involve bringing the first end of the rod into contact with liquid nitrogen, and heating the second end of the rod may further involve rapidly immersing the second end of the rod in a hot liquid.

[0010] The procedure may also include placing the second side of the rod on a heated chuck and pulling, chiseling or twisting the first end of the rod with a handle while applying heat to the second side of the rod.

[0011] Implementations of a method for separating a wafer from a silicon carbide rod may include: creating a damage layer in a silicon carbide rod with a first end and a second end. The method may also include applying a material to the second end of the rod and cooling the first end. The method may include heating the material at the second end of the rod. A thermal gradient can be created by heating the second end of the rod and cooling the first end. The thermal gradient can help a silicon carbide wafer separate from the rod at the damage layer.

[0012] Implementations of a method for separating a wafer from a silicon carbide rod may include one, all, or any of the following: The damaged layer can be created by laser irradiation. The material can increase heat conduction to the second end of the rod.

[0013] Cooling the first end of the rod and heating the material at the second end of the rod can be done simultaneously, and the material at the second end of the rod can be heated by bringing it into contact with a hot chuck and heating it quickly.

[0014] Heating the second end of the rod may involve applying heat pulses using a heated chuck.

[0015] Cooling the first end of the rod may further include bringing the first end of the rod into contact with liquid nitrogen, and heating the second end of the rod may further include applying a hot liquid to the second end of the rod.

[0016] Cooling the first end of the rod may further involve bringing the first end of the rod into contact with liquid nitrogen, and heating the second end of the rod may further involve rapidly immersing the second end of the rod in a hot liquid.

[0017] The procedure may further include placing the second side of the rod on a heated chuck and pulling, chiseling or turning off the first end of the rod with a handle while applying heat to the second side of the rod.

[0018] Implementations of a method for separating a wafer from a silicon carbide rod may include: generating a damage layer in a silicon carbide rod. The rod may have a first end and a second end. The method may include cooling the first end of the rod by contacting it with liquid nitrogen and heating the second end of the rod. The method may include generating a thermal gradient between the first and second ends of the rod by simultaneously heating the first end and cooling the second end. The method may include mechanically separating a silicon carbide wafer from the rod at the damage layer.

[0019] Implementations of a method for separating a wafer from a silicon carbide rod may include one, all, or any of the following: The second end of the rod can be brought into contact with a hot-dip chuck and heated quickly. Heating the second end of the rod can also include applying heat pulses using a heated chuck. Heating the second end of the rod may also include applying a hot liquid to the second end of the rod. Heating the second end of the rod may also involve immersing the second end of the rod in a hot liquid. The foregoing and further aspects, features and advantages are evident to the person skilled in the art from the DESCRIPTION and the DRAWINGS as well as from the CLAIMS. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The following describes implementations in conjunction with the accompanying drawings, wherein identical reference numerals denote identical elements, and: Fig. 1 is a side view of an implementation of a method for forming a damage layer in a silicon carbide rod; Fig. 2 is a top view of an implementation of a method for forming a damage layer in a silicon carbide rod; Fig. 3 is a perspective side view of an implementation of a system for cooling a first end of a rod and for heating a second end of a rod; Fig. 4 is a side view of an implementation of a system for cooling a first end of a wafer and for heating a second end of a wafer; Fig. 5 is a perspective side view of an implementation of a material coupled to a second end of a rod; Fig. 6 is a perspective side view of an implementation of a system for pouring liquid nitrogen onto a first end of a rod; Fig. 7 a side perspective view of an implementation of a rod immersed in a liquid; and Fig. Figure 8 shows a perspective side view of an implementation of a handle for twisting off an upper surface of a silicon carbide rod containing a damage layer. DESCRIPTION

[0021] This disclosure, its aspects and implementations are not limited to the specific components, assembly procedures or process elements disclosed herein. Many other components, assembly procedures and / or process elements known in the prior art, compatible with the intended optimization of semiconductor substrate manufacturing systems, and related methods are disclosed in this disclosure for use with specific implementations.Accordingly, for example, although particular implementations are disclosed, such implementations and implementing components may include any shapes, sizes, designs, types, models, versions, dimensions, concentrations, materials, quantities, process elements, steps and / or the like as are known in the prior art for such optimization of semiconductor substrate manufacturing systems and processes, and implementing components and processes that are compatible with the intended operation and processes.

[0022] Single-crystal silicon carbide is a wide-bandgap semiconductor material with physical and chemical properties well-suited for high-power microwave, temperature-tolerant, and radiation-resistant applications. Silicon carbide exhibits a critical electric field breakdown strength approximately ten times greater than that of silicon. It also displays a high-field electron velocity similar to that of gallium nitride. Furthermore, silicon carbide exhibits a thermal conductivity similar to copper, which is higher than that of both silicon and gallium arsenide. These properties make silicon carbide an excellent material for certain semiconductor-based devices, such as monolithic microwave integrated circuits (MMICs), field-effect transistors (FETs), and other power devices.

[0023] Silicon carbide for use in semiconductor devices is often fabricated as a large single crystal. Silicon carbide is formed from equal parts silicon and carbon through covalent bonding. This process results in a highly ordered configuration that is extremely hard. The hardness of silicon carbide makes it expensive to cut and thin wafers / substrates from a silicon carbide rod or ingot, as the saws used for cutting silicon carbide wear out quickly. Cutting a silicon carbide rod using a laser damage process can increase the efficiency of silicon carbide wafer fabrication. The lasers are used to create damage layers beneath the surface of the rod to remove a silicon carbide wafer.An example of such a process that uses laser-induced damage is the process marketed under the trade name KABRA by DISCO Corporation, Tokyo, Japan.

[0024] Thermal gradients can be applied to a silicon carbon rod to create stress or to shock an embedded damage layer within the rod. The stress on the embedded damage layer can create or propagate a crack through the material of the damage layer, which can aid in removing a wafer from the rod. The upper portion of the rod, including the damage layer, can then be separated from the rest of the rod as a substrate. After heat treatment, the upper portion can be mechanically removed, or it can separate purely through thermal forces caused by the pressure of the temperature change within the rod.

[0025] If the rod has been thinned by removing substrates to the point where only a few remain, completing the separation process can be very difficult because it can be challenging to create a significant thermal gradient between the first and second ends of the rod. At the rod end, the stresses used to create and propagate a crack through the damage layer may be insufficient to separate a wafer from the rod if the rod is too thin. Therefore, the lower section or end of the rod remains unusable, even though several more substrates could be formed from the material at the rod end. Various methods for improving the thermal gradient between the first and second ends of the rod can help optimize / increase the number of wafers that can be singulated from a single rod.

[0026] A wide variety of semiconductor substrate types exist and are used in the process of fabricating semiconductor devices. Non-restrictive examples of semiconductor substrates that can be processed using the principles disclosed in this document include single-crystal silicon, silicon dioxide, glass, silicon on insulator, gallium arsenide, sapphire, ruby, silicon carbide, polycrystalline or amorphous forms of any of the foregoing, and any other substrate type useful for constructing semiconductor devices. Particular implementations disclosed herein may use silicon carbide semiconductor substrates of any polytype. In this document, the term "wafer" is also used together with "substrate" because a wafer is a common type of substrate, but not as an exclusive term used to refer to all semiconductor substrate types.The various semiconductor substrate types disclosed in this document can, as a non-limiting example, be round, rounded, square, rectangular, or any other closed shape in various implementations. Each of the system and / or method implementations disclosed in this document can be used with any of the semiconductor substrate types disclosed in this document.

[0027] Referring to Fig. Figure 1 illustrates a section of a semiconductor rod 2 containing a damage layer 4. As illustrated, a laser beam 6 irradiates the first surface 8 of the semiconductor rod. A focal point 10 of the laser beam 6 is adjusted such that it is located in the semiconductor substrate below the first surface 8. The wavelength of the laser light 6 used to irradiate the first surface 8 is one for which the material of the specified semiconductor substrate is at least partially optically transparent, whether translucent or transparent. The focal point 10 creates a damage region 4 in the rod at and around the focal point 10.The degree of damage is determined by many factors, including, as a non-limiting example, the power of the laser light, the duration of exposure of the material, the absorption of the substrate material, the crystallographic orientation of the substrate material relative to the direction of the laser light, the atomic structure of the substrate, and any other factor to regulate the extinction of light energy and / or the transfer of induced damage or heat into the substrate.

[0028] The in Fig. The illustrated section of rod 2 is a silicon carbide (SiC) substrate, and thus the laser wavelength that can be used can be any wavelength capable of transmission into the SiC substrate material. In certain implementations, the wavelength can be 1064 nm. In various implementations, the laser light source can be a pulsed Nd:YAG laser or a pulsed YVO4 laser. In an implementation using an Nd:YAG laser, a spot size of 10 micrometers and an average power of 3.2 W can be used, along with a repetition rate of 80 kHz, a pulse width of 4 ns, and a numerical aperture (NA) of the focusing lens of 0.45. In another implementation, an Nd:YAG laser with a repetition rate of 400 kHz, an average power of 16 W, a pulse width of 4 ns, a spot diameter of 10 micrometers and an NA of 0.45 can be used.In various implementations, the laser power can vary from approximately 2 W to approximately 4.5 W. However, in other implementations, the laser power can be less than 2 W or more than 4.5 W.

[0029] As illustrated, the focal point 10 of the laser light forms a point of rapid heating and can lead to complete or partial melting of the material at this focal point. The point of rapid heating and the resulting stress on the hexagonal single-crystal structure of the SiC substrate cause cracking of the substrate material along a c-plane. Depending on the type of SiC single crystal used to fabricate the rod, the c-plane can be oriented at an oblique angle to the second surface of approximately 1 degree to approximately 6 degrees. In various implementations, this angle is determined at the time the rod is fabricated. In certain implementations, the oblique angle can be approximately 4 degrees.

[0030] During operation, the laser is operated in pulsed mode to generate numerous overlapping patches of pulsed light as it scans the surface of the substrate. As a result, a continuous / semi-continuous layer / band of modified material is formed within the wafer. In other implementations, the laser can be operated in continuous-wave mode instead of pulsed mode to generate the band of modified material. As illustrated, the stress induced by focal point 12 causes cracking along the c-plane in the material of the SiC substrate 2 in one or both directions along the c-plane. These cracks are described as extending from the region of focal point 10 (where the modified layer / band is located) at the oblique angle shown in Fig. Figure 1 illustrates angled propagation. In different implementations, the cracks can be distributed on the material below, above, or directly from focal point 10, depending on the laser's properties and application method. In different implementations, the crack length into the substrate is a function of the applied laser power. As a non-restrictive example, the focal depth was set to 500 µm into the substrate; when the laser power was 3.2 W, the crack propagation from the modified layer / ribbon was approximately 250 µm; when the laser power was 2 W, the crack lengths were approximately 100 µm; when the laser power was set to 4.5 W, the crack lengths were approximately 350 µm.

[0031] As in Fig. As illustrated in Figure 2, the rod 2 can be moved stepwise beneath the laser beam 6 (or the laser beam 6 can be moved stepwise above the rod 2, or both can be moved stepwise together) to produce a plurality of spaced spots 12 where damage has been induced in the substrate. The width between the plurality of spaced spots can be a function of the crack lengths in the wafer material and / or the amount of damage layer that forms when a wafer is initially scanned. As a non-restrictive example, the width can be set between approximately the length of the cracks in the wafer and approximately twice the length of the cracks in the wafer. In situations where the damage layer initially forms on one side of the wafer during scanning, the width between the spaced spots can initially be reduced.As a non-restrictive example, the width can initially be set to 200 µm until the cracks begin to propagate from the modified layer, at which point the width (travel amount) can be set to 400 µm. The feed rate of the wafer under the laser (or the laser over the wafer) can, as a non-restrictive example, be 400 mm / second, although smaller or larger feed rates can be used in various implementations.

[0032] Although a multitude of spaced spots 12 are illustrated, in various implementations the laser beam 6 may not be applied in a stepwise manner, but may be applied to the substrate material in a continuous or essentially continuous manner to create continuous or essentially continuous zones / areas of damage in the substrate. These damage areas may include corresponding continuous or discontinuous cracking of the substrate material. The multitude of spaced spots 12 or the continuous lines / areas affected by the laser irradiation form a damage layer within the semiconductor substrate after the laser has completed the stepwise process / traverse of the semiconductor substrate material.

[0033] Referring to Fig. Figure 2 illustrates a diagram of a silicon carbide wafer 14 with an alternating single-pass laser irradiation path 16 (single-pass path). This figure is used to illustrate an alternative method for producing a wafer for singulation from a rod using a laser. The particular semiconductor substrate illustrated here has two wafer flats corresponding to a SiC wafer, although the principles disclosed herein could be applied to many different substrate types. As illustrated, a path 16 is shown that a laser follows when irradiating the substrate with laser light, the path 16 indicating locations where the light irradiation occurs and a focal point is formed within the substrate.In other implementations, however, path 16 can illustrate the path of the laser as it moves across the surface of the substrate and irradiates the wafer in continuous-wave rather than pulsed mode. The one in . Fig. Path 16, illustrated in Figure 2, is an alternating single-pass path where the laser first traverses the wafer in the y-direction and then in the x-direction, and then in the opposite y-direction in several steps. In different implementations, the spacing of the steps in the x-direction can be the same, as shown in Figure 2. Fig. Figure 2 illustrates this. However, in other implementations, the step spacing across the wafer can vary, either for an initial period or for the entire distance across the wafer in the x-direction, depending on how the damage layer forms. The step spacing can be any spacing disclosed in this document.

[0034] However, different implementations can utilize multi-pass paths and paths that do not follow straight lines (spirals, etc.). Similarly, more than one laser beam can be used in different implementations, scanning the substrate in the same or different directions.

[0035] Referring to Fig. Figure 3 illustrates an implementation of a system that employs a method for generating a thermal gradient in a silicon carbide rod 18. The method can include generating a damage layer 20 in the rod 18 by laser irradiation using a process described in Fig. 1 to Fig. Figure 2 illustrates this. The damage layer 20 can also be produced by any other method described in this document. The method then includes cooling 22 of a first end 24 of the rod 18 and heating 26 of a second end 28 of the rod 18. In various implementations, cooling the rod involves cooling it to a temperature of about -196 °C using boiling liquid nitrogen. In some implementations, heating the rod involves heating it to a temperature of about 100 °C. Heating and cooling the rod creates a temperature gradient in the rod 18 and can aid in the propagation of the damage layer across the width of the rod 18.In some implementations of a method for separating a wafer from a silicon carbide rod, generating a thermal gradient in the rod 18 can involve simultaneously cooling 22 the first end 24 of the rod 18 and heating 26 the second end 28 of the rod 18. As illustrated, in various system implementations, heating the second end 28 of the rod can be accomplished by placing the rod 18 on a heated chuck 30. In some implementations, heating 26 of the rod 18 can involve applying heat pulses (time-varying heat flux) to the rod using the heated chuck 30. In other implementations, heating the second end of the rod can involve applying a constant heat flux (essentially a time-invariant heat flow) to the second end 28 of the rod 18.In other implementations, heating the second end 28 of the rod 18 can include heating while the hot chuck is being heated rapidly. As a non-restrictive example, rapid heating can include heating at a rate of 40 to 100 °C / second. In various implementations, the rod 18 can be placed on the hot chuck 30 before cooling 22 of the first end 24 of the rod 18 begins.

[0036] Referring to Fig. Figure 4 illustrates another implementation of a system that employs a method for separating a wafer from a silicon carbide rod. As previously described, all methods described herein can begin with the formation of a damage layer 32 in the rod and / or wafer 34 before a temperature gradient is applied to the rod and / or wafer 34. The method can include cooling 36 of the silicon carbide wafer 34 before heat 38 is applied to the wafer. In various implementations, the heat 38 can then be applied by applying heat pulses through a heating chuck 40.

[0037] Referring to Fig. Figure 5 illustrates an additional implementation of a method for separating a wafer from a silicon carbide rod 42. The method may include the application of a material 44 to the second end 46 of the rod 42. The material 44 may include any material / substance capable of facilitating heat conduction to the second end of the rod, including, as a non-limiting example, rubber coatings, temporary or permanent gels, or thermally conductive pastes or gels. In some implementations, the rod may be cooled before the material is applied. In other implementations, the material may be applied to the rod before the first end of the rod is cooled.After the material has been applied, the heating and cooling of the rod may include any method described in this document, such as, as a non-limiting example, heating the second end of the rod by a hot chuck 48 using any of the approaches disclosed in this document.

[0038] Referring to Fig. Section 6 illustrates an additional method for generating a thermal gradient in a rod made of semiconductor material. This method includes cooling the first end 50 of the rod 52 by bringing the first end of the rod into contact with liquid nitrogen 54. In other implementations (such as others disclosed in this document), other cooling liquids / gases / solids may be used. The method may also include heating the second end 56 of the rod 52 by applying a hot liquid 58 to the second end 56 of the rod 52. As illustrated, applying a hot liquid 58 to the second end 56 of the rod 52 may involve placing the rod in a hot water bath or another bath containing any other heated liquid. In other implementations, the hot liquid may be poured onto the second end of the rod. Referring to Fig. 7. In various implementations, the procedure can include a rapid immersion of the rod in the hot liquid in the bath after cooling (quenching).

[0039] Referring to Fig. Figure 8 illustrates another implementation of a system that employs a method for separating a wafer from a silicon carbide rod. The method includes generating a damage layer 60 in a silicon carbide rod 62, cooling a first end of the rod, and heating the second end of the rod. Cooling the rod may include any process / system disclosed in this document, such as, as a non-limiting example, contacting the rod with liquid nitrogen or other cooling fluids, applying cold gas to the first end of the rod, or other suitable methods for cooling a semiconductor rod relative to an initial rod temperature. In various implementations, methods for cooling the rod may include rapid cooling of the rod at a rate of about 40 to 100 °C / second.Heating the rod may involve any process / system disclosed in this document, including, as a non-limiting example, applying heat by means of a heated chuck, immersion in hot liquid, application of hot liquid to the second end of the rod, or any other suitable method for heating a rod of semiconductor material relative to an initial rod temperature. As illustrated, in various system / process implementations, the wafer may be separated from the rod using mechanical methods. Various mechanical methods may include peeling a wafer from a rod, chipping a wafer from a rod, turning a wafer from a rod, or other methods for mechanically separating two parts of the same material. As shown in... Fig. As illustrated in Figure 8, the rod can be placed on a stable surface, such as, as a non-restrictive example, a heated chuck 64. The procedure can involve bringing the first end of the rod into contact with a handle 66 and twisting the handle, as illustrated by the curved arrows 68. In some implementations, the heated chuck can apply heat to the second end of the rod while the handle 66 twists the first end of the rod. In other implementations, a second handle can be attached to the second end of the rod. The first end of the rod can be twisted in a first direction while the second end of the rod can be twisted in a second direction by the second handle, the second direction being the opposite direction to the first.As a result of the mechanical action of the handle(s), substrates / wafers made of semiconductor material can then separate from a rod through the combination of thermal and mechanical forces at the damage layer 60.

[0040] During the implementation, which in Fig. As illustrated in Figure 8, the handle 66 is depicted as having a smaller diameter than the rod 62. However, in other implementations, the handle may have a larger diameter than the rod. In some implementations, the handle may also extend over a section of the rod's sides, making contact with the flats or other structures of the rod to aid in gripping the rod and / or applying a mechanical force to it.

[0041] In various process implementations, the damaged layer can be created by laser irradiation.

[0042] Various process implementations can include heating the second end of the rod, bringing the second end of the rod into contact with a heating chuck and rapid heating, applying heat pulses through a heating chuck, applying a hot liquid to the second end of the rod, or immersing the second end of the rod in a hot liquid.

[0043] It should be readily understood that where the foregoing description relates to particular implementations of the optimization of semiconductor substrate manufacturing systems and processes and implementing components, subcomponents, processes and subprocesses, a number of modifications may be made without departing from its spirit, and that these implementations, implementing components, subcomponents, processes and subprocesses may also be applied to another optimization of semiconductor substrate manufacturing systems and processes.

Claims

[1] Method for separating a wafer from a rod comprising a semiconductor material, the method comprising: Creating a damage layer in a rod comprising semiconductor material, wherein the rod has a first end and a second end; Cooling the first end of the rod; and Warming the second end of the stick; where a thermal gradient between the cooled first end and the heated second end helps a silicon carbide wafer to detach from the rod at the damage layer, wherein the cooling of the first end of the rod and the heating of the second end of the rod occur simultaneously; and wherein the second end of the rod is heated by bringing a heated chuck into contact or by applying heat pulses using a heated chuck. [2] The method of claim 1, wherein cooling the first end of the rod further comprises bringing the first end of the rod into contact with liquid nitrogen; and wherein heating the second end of the rod further comprises one of: Applying a hot liquid to the second end of the rod; or rapidly immersing the second end of the rod in a hot liquid. [3] Method according to claim 1, further comprising placing the second side of the rod on a heated chuck and pulling, chiseling or twisting the first end of the rod with a handle while applying heat to the second side of the rod. [4] Method for separating a wafer from a silicon carbide rod, the method comprising: Creating a damage layer in a silicon carbide rod, wherein the rod has a first end and a second end; Applying a material to the second end of the rod; and heating the material at the second end of the rod; a thermal gradient, created by heating the second end of the rod, helps a silicon carbide wafer to detach from the rod at the damaged layer. [5] Method according to claim 4, wherein the material increases heat conduction to the second end of the rod. [6] Method according to claim 4, further comprising cooling the first end of the rod and wherein the cooling of the first end of the rod and the heating of the material at the second end of the rod are carried out simultaneously; and wherein the material at the second end of the rod is heated by bringing a heating chuck into contact and rapidly heating or by applying heat pulses through a heating chuck. [7] The method of claim 6, wherein cooling the first end of the rod further comprises bringing the first end of the rod into contact with liquid nitrogen; and wherein heating the second end of the rod further comprises one of: Applying a hot liquid to the second end of the rod; or rapidly immersing the second end of the rod in a hot liquid. [8] Method according to claim 4, further comprising placing the second side of the rod on a hot chuck, pulling, chiseling or twisting the first end of the rod with a handle while applying heat to the second side of the rod. [9] Method for separating a wafer from a silicon carbide rod, the method comprising: Cooling one end of a rod and heating a second end of the rod; Creating a damage layer in a silicon carbide rod; Creating a temperature gradient between the first end and the second end of the rod; and mechanical separation of a silicon carbide wafer from the rod at the damaged layer, wherein the cooling of the first end of the rod and the heating of the second end of the rod occur simultaneously; and wherein the second end of the rod is heated by bringing a heated chuck into contact or by applying heat pulses using a heated chuck.

Images