Method of separating semiconductor wafers from a layer stack and method of producing semiconductor wafers

Abstract

A method for separating a semiconductor wafer (4') from a layer stack (1) comprising the following steps: A) providing a layer stack (1) comprising a seed substrate (2), a separation layer (3) on the seed substrate (2), and a semiconductor layer (4) formed on the separation layer (3), the semiconductor layer (4) being preferably prepared on the separation layer (3) by epitaxial growth; B) separating the semiconductor wafer (4') from the layer stack (1), further comprising a first step B1) in which the semiconductor layer (4) is pre-damaged in a region to reduce the fracture stress of the semiconductor layer (4) in this region, wherein step B1) is performed before and / or during step B); and a second step B2) in which a heat energy means from a source is applied to the semiconductor layer (4) along a separation line (11), wherein the separation line (11) starts from the pre-damaged region of the semiconductor layer (4), and wherein said heat energy means from a source locally heats and / or cools the semiconductor layer (4) to generate a thermal induced stress at least in the region of the semiconductor layer (4) which is heated and / or cooled, thereby initiating a crack (13) which extends along the separation line (11) and separates the semiconductor wafer (4') from the layer stack (1). Furthermore, the invention relates to a method for preparing a semiconductor wafer (4').

Description

[0001] manual

[0002] The present invention relates to a method for separating a semiconductor wafer from a stack of layers according to claim 1, and a method for fabricating a semiconductor wafer according to claim 19.

[0003] To produce large-area light-emitting elements or electronic components such as solar cells, low-cost semiconductor wafers are required. Semiconductor wafers are typically flat substrates made of semiconductor materials such as silicon (Si), gallium arsenide (GaAs), germanium (Ge), or silicon carbide (SiC).

[0004] The manufacturing or production process of semiconductor wafers typically begins with a large monolithic crystal of semiconductor material, which is then cleaved to separate individual semiconductor wafers. This is done using machine tools such as diamond wire saws. While high quality can be achieved, the manufacturing cost is high, primarily due to material loss.

[0005] Epitaxial chemical vapor deposition (CVD) processes can avoid these problems by forming layer stacks in which the desired semiconductor wafer is formed as a layer on a seed substrate, hereinafter referred to as the semiconductor layer. A porous separation layer is typically disposed between the seed substrate and the semiconductor layer, enabling the separation of the semiconductor layer from the seed substrate with almost no loss of seed wafer material. This allows for both high-quality semiconductor wafers and good reusability of the seed substrate.

[0006] To separate semiconductor wafers from seed substrates, laser beams are typically used to create separation cuts through laser ablation. When using this method, the laser beam usually penetrates completely through the semiconductor layer and reaches the separation layer and / or the interior of the seed substrate, thereby damaging or reducing the size of the seed substrate.

[0007] However, the seed substrate and the separation layer are also affected by these cuts. Therefore, the seed substrate is often not reusable for the production of other semiconductor layers, or it must be processed at a high cost, which limits the reusability (number of reuse cycles) of the seed substrate.

[0008] Laser ablation can also be used to trim the stacked layers from the side, but this operation also reduces the size of the seed substrate, thus limiting its reusability. In addition, the laser ablation process can severely affect the edge quality of the diced semiconductor layers in the form of laser-induced damage, thereby reducing the fracture strength of the resulting wafer and the material quality as measured by minority carrier lifetime (MCLT).

[0009] A further problem is that the semiconductor layer is formed with a non-planar geometry, especially at the edges, which is undesirable for manufacturing electronic components based on this layer. For this reason, it is also desirable to separate the individual regions of the semiconductor layer from the rest of the semiconductor layer.

[0010] To separate semiconductor wafers, laser radiation can be used to create separation kerfs, as described in, for example, DE 10 2015 118 042 A1. However, there are risks associated with using laser radiation: the laser radiation can penetrate the semiconductor layers and reach the separation layer and the seed substrate. This can damage the seed substrate, impairing its subsequent use for growing other semiconductor layers. On the other hand, if the laser power is set too low, it may not be sufficient to separate individual areas of the semiconductor wafer from the layer stack, or it may result in a low cutting speed, limiting wafer production capacity. Another drawback of conventional laser ablation is the generation of particles, especially on the wafer itself, which require careful handling during manufacturing.

[0011] US2014 / 0038392 A1 discloses a system and method for laser splitting of semiconductor wafers, wherein a focused laser beam scans the entire surface of a semiconductor layer to generate a laser beam array within the layer. The thermal stress and / or external mechanical stress following laser irradiation can enhance the separation of the wafer from the semiconductor layer. Scanning the entire surface of the wafer is time-consuming, costly, and requires careful operation to ensure that all areas are processed. Otherwise, the wafer may be damaged during the stripping process.

[0012] Therefore, one object of the present invention is to provide a reliable method for separating semiconductor wafers from layer stacks, which enables good reusability of seed substrates and high quality of semiconductor wafers at high process speeds.

[0013] This objective is achieved by the method for separating a semiconductor wafer from a layer stack according to claim 1, and the method for fabricating a semiconductor wafer according to claim 19. Preferred embodiments are subject to the dependent claims.

[0014] According to the present invention, a method for separating a semiconductor wafer from a stack of layers includes the following steps: A) Provide a layer stack comprising a seed substrate, a separation layer on the seed substrate, and a semiconductor layer on the separation layer, wherein the semiconductor layer is preferably epitaxially grown on the separation layer; B) Separate a semiconductor wafer, which is at least a part of the semiconductor layer (4), from the layer stack.

[0015] The key point is that, in the first step B1), the semiconductor layer is pre-damaged in a region to reduce the fracture stress of the semiconductor layer in that region. Step B1) is performed before and / or during step B). Furthermore, in the second step B2), a thermal energy device from a source is applied to the semiconductor layer along a separation line, wherein the separation line includes the pre-damage, or specifically originates from the pre-damage in the semiconductor layer. The thermal energy device locally heats and / or cools the semiconductor layer, particularly below the surface of the semiconductor layer, thereby generating thermal stress to initiate cracks that extend substantially along the separation line.

[0016] This invention is based on the fact that thermal stress can be generated within a semiconductor layer simply by heating and / or cooling with a thermal energy device, and this stress is released by forming cracks from pre-damaged areas—in particular, without altering the microstructure of the semiconductor layer. This invention is also based on the understanding that thermal energy devices can be used to create clean separation lines along which semiconductor wafers can be separated from the rest of the stack without affecting the quality of the separation layer and / or the seed substrate, or generating particulate matter such as that caused by laser ablation. It has now been discovered that only a single separation line is needed to separate a semiconductor wafer from the stack. This eliminates the need to process the entire surface area, enabling faster processing speeds at a lower cost.

[0017] Preferably, the thermal energy device is configured to directly heat and / or cool the semiconductor layer, and in particular, not to heat and / or cool any intermediate layers, especially any layer located above the semiconductor layer in a layer stack. More preferably, the thermal energy device is configured to heat and / or cool the semiconductor layer within the layer thickness range, without applying heat energy to the surface of the semiconductor layer.

[0018] By applying thermal stress to the semiconductor layer, one or more cracks are generated, and the semiconductor wafer separates from the rest of the layer stack along the cracks. By applying a peeling force to the semiconductor wafer, the wafer is peeled off from the separation layer.

[0019] Preferably, after the application of thermal stress, the semiconductor wafer remains at least partially adhered to the separation layer. This separation method is more efficient than known methods because it does not compromise the reusability of the seed structure and avoids unnecessary high material loss. Another advantage is that the formation of the separation line is not limited by a specific geometry. Within the scope of this invention, the separation line can, in principle, be configured with any orientation. This is particularly advantageous because the orientation of the separation line can be adapted to any contour of the layer stack, or to a semiconductor wafer to be cut that may have both straight and curved edges. In a preferred embodiment, the separation line is a completely closed line, particularly formed as a generally rectangular line.

[0020] This invention relates to separating layers from a stack to prepare a wafer, which can be further processed, for example, for the production of electronic components or solar panels. In particular, the semiconductor wafer is a single element having a flat surface. This invention does not particularly relate to processing the wafer to prepare localized regions.

[0021] Specifically, peeling a semiconductor layer from a stack means that the semiconductor layer is no longer connected to the stack; in other words, it is removed from the stack. Separating a semiconductor layer from a stack means that the semiconductor layer is separated from the stack, but may still adhere to the stack, and a peeling force is required to remove the semiconductor layer from the stack.

[0022] In a preferred embodiment, the thermal energy device is laser radiation from a laser source. Preferably, the laser radiation is continuous emission, and specifically, the laser source is a continuous-wave laser.

[0023] This invention is based on the understanding that relatively low-intensity laser radiation, particularly low-intensity laser radiation, can be used to form clean separation lines along which the semiconductor wafer can be separated from and peeled off from the rest of the stack. Specifically, it is unnecessary to set the laser power to a level that would cause laser ablation, which could negatively impact the quality of the separation layer and / or the seed substrate. Instead, thermal stress can be generated within the semiconductor layer simply by heating with laser radiation, and this stress is released through crack formation—without altering the microstructure of the semiconductor layer. Compared to pulsed lasers, continuous-wave lasers emit laser radiation uninterruptedly. Therefore, thermal stress can be generated uniformly, particularly along the separation line without spatial discontinuity. This is beneficial for crack initiation and propagation. Furthermore, pulsed lasers may have drawbacks: the energy density of each pulse may exceed a threshold that would cause changes in the crystal structure of the semiconductor layer, which is undesirable.

[0024] Therefore, in a preferred embodiment, the power of the laser radiation used as a thermal energy device in step B2) is selected such that the laser radiation is substantially absorbed within the semiconductor layer, and the selected laser radiation power allows the semiconductor layer to retain its original crystal structure. The pre-damage generated in step B1) is used to reduce the fracture stress of the semiconductor layer, so that even a low thermally induced stress is sufficient to induce the desired cracking within the semiconductor layer along the separation line. Furthermore, the pre-damage can serve as a guide line and / or initiation point and / or end point for thermally induced crack propagation.

[0025] The study also showed that the semiconductor layer typically needs to maintain a distance of 1-3 mm from the edge of the stack along the separation line (11) from the stack. However, by forming pre-damage, the crack path can be better controlled, thus achieving a shorter distance, which can improve the yield of semiconductor wafer production.

[0026] The crack path extends along a predetermined separation line. However, cracks that deviate from the intended separation line can also form. This deviation can occur "naturally" due to stress, particularly near the edge of the semiconductor wafer, or through separate pre-damage and / or laser radiation.

[0027] The separation line preferably represents the line along which the crack extends substantially. However, the separation line and the crack path do not necessarily coincide perfectly, and there may be some degree of deviation between them.

[0028] This invention offers the following advantages: extremely high cutting speeds, typically 150-500 mm / s or even several meters per second, can be achieved when cracks are formed. Furthermore, compared to machining, laser-based semiconductor layer separation processes have fewer vulnerable components. Additionally, wafer contamination or dirt caused by abrasion and / or particle generation due to saw blade or laser ablation is reduced. Moreover, laser-induced material damage in the heat-affected zone is reduced, or even completely avoided, by this invention. Furthermore, there is no need to form so-called reference marks or alignment marks, so both surfaces of the semiconductor layer can be used. Edge chipping is also easily avoided, resulting in good edge quality and high mechanical stability of the semiconductor layer after peeling. In addition, the entire process can be performed without the use of coolant, thus eliminating any drying steps for the semiconductor layer.

[0029] The semiconductor layer and semiconductor wafer are preferably made of silicon, specifically single-crystal silicon. However, the semiconductor layer and semiconductor wafer may also be made of germanium or silicon carbide, gallium arsenide, indium phosphide or indium arsenide or other III-V compound semiconductor materials, or gallium nitride, aluminum nitride, indium nitride or other III-V nitrides or their compounds, which are also within the scope of this invention.

[0030] The semiconductor wafer to be separated is an epitaxially grown layer with a higher density than the separation layer below it, which falls within the scope of this invention. Specifically, the porosity of the separation layer is higher than that of the semiconductor layer.

[0031] Preferably, the thickness of the semiconductor layer is from 5µm to 2000µm, more preferably from 10µm to 500µm, particularly preferably from 50µm to 200µm, and specifically from 60µm to 160µm. The thickness of the seed substrate can be relatively small, from 1µm to 800µm, preferably from 200µm to 600µm, but can also reach several millimeters in thickness.

[0032] According to the present invention, in step B2, the thermal stress is caused by a temperature gradient generated within the semiconductor layer due to localized heating of the semiconductor layer by a thermal energy device (particularly a laser). Specifically, the temperature gradient only needs to be generated by the temperature difference between the heated or cooled regions and the unheated or uncooled regions of the semiconductor layer. Therefore, it is sufficient that at least the unheated or uncooled regions of the semiconductor layer are at ambient temperature.

[0033] Preferably, step B2) is performed using laser direct cleavage (LDC), which processes only the semiconductor layer and does not affect the separation layer and / or the seed substrate.

[0034] In a preferred embodiment, the thermal energy device is a heating device, and a cooling medium is applied along the separation line before, during, and / or after the application of the heating device. Preferably, the heating device and the cooling medium are applied in a laterally offset manner. This increases the temperature difference and the resulting thermal stress within the semiconductor layer, better guiding the extension of cracks along the separation line. This facilitates the formation of cracks along the thickness direction within the semiconductor layer and makes it easier to separate the semiconductor wafer from the rest of the stack. The invention does not limit the type of cooling medium used, but the use of water-based aerosols has proven advantageous. Preferably, the cooling medium itself and / or the application process of the cooling medium can be adjusted by setting the temperature of the cooling medium before and / or during its application to the semiconductor layer, or by setting the pressure and / or flow rate of the cooling medium during its application to the semiconductor layer. When an additional cooling medium is used, method step B2) is preferably performed using thermal laser separation (TLS).

[0035] In another preferred embodiment, the thermal energy device is a cooling device, and a heating medium is applied along a separation line before, during, and / or after the application of the cooling device. Preferably, the heating medium and the cooling device are applied in a laterally offset manner.

[0036] In a preferred embodiment, the wavelength of the laser radiation is in the infrared range, preferably from 780 nm to 2000 nm, more preferably from 800 nm to 1500 nm, particularly preferably from 1000 nm to 1100 nm, and especially 1060 nm or 1064 nm. Studies have shown that the above wavelength range is particularly advantageous for achieving the beneficial effects of the present invention. However, apart from this, the present invention does not limit the wavelength range in principle, and the wavelength range can be determined according to the material of the semiconductor layer.

[0037] In another preferred embodiment, the laser radiation is applied to the semiconductor layer after being deflected by a reflector and / or a galvanometer scanner. With this embodiment, in principle, no relative movement between the semiconductor layer and the laser source is required to form the separation line. However, it is also possible to first induce spatial displacement between the semiconductor layer and the laser source, positioning the laser source appropriately relative to the layer stack, and then heat the semiconductor layer along the separation line by deflecting the laser radiation; this also falls within the scope of the present invention.

[0038] Preferably, the thermal energy device is applied to the semiconductor layer in a continuous or pulsed manner.

[0039] Preferably, the layer stack is moved relative to the source according to the shape of the separation line. The scope of this invention includes both a fixed layer stack and a moving source or a portion thereof relative to the layer stack, and a fixed source and a moving layer stack relative to the source. Alternatively, both the source or a portion thereof and the substrate may be moved. Specifically, guiding and moving devices may be provided to achieve relative movement between the source (specifically a portion of the source) and the layer stack. In a preferred embodiment, when the source is a laser source, a portion of the laser source may be a laser head or an optical element of the laser source.

[0040] In another preferred embodiment, the layer stack is stationary relative to the source, and the thermal energy device is guided to the semiconductor layer via a guide device. With this embodiment, it is essentially unnecessary to create relative movement between the semiconductor layer and the source to form the separation line. However, it is also possible to first induce spatial displacement between the semiconductor layer and the source, positioning the source of the thermal energy device appropriately relative to the layer stack, and then guide the thermal energy device along the separation line to heat and / or cool the semiconductor layer, which is also within the scope of this invention.

[0041] In a preferred embodiment, the pre-damage is formed by a laser source or another laser source, specifically by a pulsed laser source, preferably by a laser source with a pulse length in the nanosecond range. Alternatively or preferably additionally, the pre-damage is formed by scribing.

[0042] Studies have shown that pre-damage is used to establish the starting point for cracking and separation of semiconductor layers. In this regard, the beneficial effects of the present invention are sufficient as long as the pre-damage is substantially point-like, i.e., without a preferred extension direction on the semiconductor layer. Preferably, the shape of the pre-damage is substantially linear. Specifically, the separation line originates from the pre-damage and extends spatially beyond the pre-damage, which is within the scope of the present invention.

[0043] In a preferred embodiment, the pre-damage is an initial pre-damage that covers only a small area of ​​the semiconductor layer. Preferably, it can be formed as an initial scribing with a length of 1µm to 2000µm, and more preferably as an initial scribing with a length of 100µm to 1000µm. The initial pre-damage can also be visible on the surface of the semiconductor layer.

[0044] In another preferred embodiment of the invention, the pre-damage is a deep pre-damage, which extends particularly along the separation line and can reach a length up to the size of the semiconductor layer. Preferably, the deep pre-damage extends within the semiconductor layer, rather than on its surface. This deep pre-damage can also be referred to as a deep scribing. The deep scribing can provide separation lines, particularly closed separation lines, within the semiconductor layer. Therefore, the area covered by a deep, long pre-damage is larger than that of a short initial pre-damage. However, the deep pre-damage does not cause complete cracking of the semiconductor layer it surrounds. Cracking is achieved in step B2 by a thermal device, particularly by laser radiation.

[0045] Preferably, the deep pre-damage can be a modified region within the semiconductor layer, where the semiconductor material is modified compared to the surrounding semiconductor material, the modification being achieved, particularly through laser absorption. The modification can be a change in the crystal structure of the semiconductor material. This can be achieved, for example, by briefly melting or vaporizing the material within the semiconductor layer (particularly through laser absorption) and then recrystallizing or solidifying the molten material. For deep scribing, it is preferable to analyze the surface structure of the semiconductor layer, especially its thickness variation, to set the deep pre-damage at the desired depth. However, prior analysis of the semiconductor layer's surface structure is not mandatory before generating the deep pre-damage.

[0046] In another preferred embodiment, a plurality of pre-damages are generated on the semiconductor layer, the plurality of pre-damages being distributed on the semiconductor layer along the direction of the separation line.

[0047] Preferably, additional pre-damage is created at the ends and / or edges and / or bends of the separation line, and / or at the intersections of at least two separation lines. This additional pre-damage prevents chipping of the separation line (especially at its ends and / or edges and / or bends). Furthermore, this additional pre-damage can provide direction for the further extension of the separation line and / or crack path.

[0048] Within the scope of this invention, multiple pre-damage regions may be provided, and these multiple pre-damage regions are distributed on the semiconductor layer along the direction of the separation line. However, within the scope of this invention, pre-damage regions may also be continuously formed along the direction of the separation line.

[0049] In another preferred embodiment, as a first mode, the semiconductor layer is heated by a continuous-wave laser, which may have an additional mode in which laser radiation deviates from the first mode to generate pre-damage. Specifically, the laser radiation in the additional mode may be pulsed and / or may have a different focus and / or intensity than the first mode. Alternatively or preferably additionally, the wavelengths of the laser radiation in the first mode and the additional mode may be different. This allows both pre-damage and heating to be achieved using the same device.

[0050] The method of the present invention is advantageous regardless of how the semiconductor layers are disposed on the layer stack. In particular, the semiconductor layers can be separated by separation lines, such that a portion of the semiconductor layers constituting the semiconductor wafer is separated from another portion of the semiconductor layers remaining on the layer stack. In other words, the semiconductor layers may overlap with the sides of the seed structure and / or the separation layer, thus it may be necessary to separate planar or high-quality regions of the semiconductor layers from non-planar or defective regions.

[0051] Preferably, the semiconductor layer includes an overlap that protrudes from one side of the seed substrate and / or the separator layer, particularly at least partially covering one or more sides of the seed substrate and / or the separator layer. In a preferred embodiment, the overlap is formed of the same material as the semiconductor layer.

[0052] Preferably, the separation line can be formed such that the semiconductor wafer is separated from the overlap and / or layer stack in process step B).

[0053] Preferably, the semiconductor layer includes a core region, which is the area of ​​the semiconductor layer located inside the separation line and surrounded by the separation line. Specifically, the core region does not include an overlap. Preferably, the core region constitutes a semiconductor wafer.

[0054] As described above, the object of the present invention is also achieved by a method for manufacturing a semiconductor wafer, the method comprising the following steps:

[0055] V1) Provides a seed substrate; V2) Create a separation layer on the seed substrate; V3) Preferably, a semiconductor layer is formed on the separation layer by chemical vapor deposition epitaxy to provide a layer stack including a seed substrate, a separation layer, and a semiconductor layer; and V4) In particular, a semiconductor wafer, which is at least a part of the semiconductor layer, is peeled off from the layer stack by a peeling force.

[0056] The key point is that the semiconductor wafer separation method according to the present invention or its preferred embodiments is performed before or during method step V4).

[0057] Preferably, the thermal energy device directly heats and / or cools the semiconductor layer, particularly without heating and / or cooling any intermediate layers, especially any layer located above the semiconductor layer in the layer stack. More preferably, the thermal energy device heats and / or cools the semiconductor layer within the layer thickness range without applying heat energy to the surface of the semiconductor layer.

[0058] Regarding the production or manufacturing process according to the present invention, the advantages that can be achieved by the semiconductor layer separation method according to the present invention or its preferred embodiments can be referred to.

[0059] The advantages of the present invention will be explained by way of example in conjunction with embodiments and accompanying drawings.

[0060] Figure 1 The layer stacking is shown in side view a) and top view b). Figure 2 Perspective views a)-d) illustrate a first embodiment of separating a semiconductor wafer from a stacked layer using deep pre-damage and subsequent thermal laser separation; Figure 3 The first embodiment of separating a semiconductor wafer from a stacked layer using deep pre-damage and subsequent thermal laser separation is shown in side views a)-d). Figure 4 Perspective views a)-d) illustrate a second embodiment in which a semiconductor wafer is separated from a stacked layer by first forming pre-damage through laser ablation and then performing subsequent thermal laser separation. Figure 5 Perspective views a)-d) illustrate a second embodiment in which a semiconductor wafer is separated from a stacked layer by first forming pre-damage through laser ablation and then performing subsequent thermal laser separation. Figure 6 Perspective views a)-d) illustrate a third embodiment of separating semiconductor wafers from a stack of layers using direct laser cleaving; Figure 7 The third embodiment of separating a semiconductor wafer from a stacked layer using direct laser cleavage is shown in side views a)-d). Figure 1 View a) is a side view of layer stack 1, which comprises multiple layers deposited on top of each other. According to Figure 1 In the embodiment shown in view a), the crystalline seed substrate 2 is covered by a separation layer 3, which is porous silicon. A semiconductor layer 4 is epitaxially grown above the separation layer 3; in this embodiment, the semiconductor layer 4 is made of monocrystalline silicon.

[0061] The porous separation layer 3 helps to cleanly separate the semiconductor layer 4 from the seed substrate 2, so that the semiconductor wafer 4' constituting at least a portion of the semiconductor layer 4 can be used to manufacture electronic components, and the seed substrate 2 can be reused to grow more semiconductor layers 4.

[0062] Depend on Figure 1 As can be seen in view a), the semiconductor layer 4 includes a core region 5 disposed above the flat surfaces of the seed substrate 2 and the separator layer 3. Furthermore, the semiconductor layer 4 has an overlap portion 6 through which it extends beyond the side surfaces of the seed substrate 2 and the separator layer 3, and laterally surrounds the seed substrate 2 and the separator layer 3. This is typically because the epitaxially grown layer is formed, for example, by chemical vapor deposition, over the entire surface of the seed substrate 2 or the separator layer 3, and is therefore not limited to a specific spatial region.

[0063] Figure 1 View b) is a top view of the layer stack 1. As can be seen from this view, the layer stack 1 has a roughly rectangular lateral profile / shape, with the overlapping portion 6 surrounding the core area 5 circumferentially.

[0064] For semiconductor wafer 4' to be used in the manufacture of electronic components, a low total thickness variation on its surface is typically required. Specifically, the thickness of semiconductor layer 4 needs to be within a range of less than 25 µm in total thickness variation. Therefore, it is necessary to separate the core region 5 of semiconductor layer 4 from the layer stack 1, especially from the typically uneven overlap 6. This will be discussed in conjunction with... Figures 2 to 5 Describe the methods used for this purpose.

[0065] according to Figure 2 View a), providing and combining Figure 1 The structure described is the same layer stack 1. (As in combination) Figure 1 As explained, the layer stack 1 has an outer semiconductor layer 4, which includes a core region 5. The core region 5 needs to be separated from the layer stack 1, specifically from the overlap 6.

[0066] In order to separate the semiconductor wafer 4' from the stacked layer 1, as Figure 2 As shown in view b), a pre-damage 8 is formed using a pre-damage device 7 (e.g., a laser). This laser-type pre-damage device 7 forms the pre-damage 8 in the semiconductor layer 4, thereby locally reducing the fracture stress of the semiconductor layer 4. Specifically, the laser power is adjusted so that the pre-damage 8 is formed within the semiconductor layer 4, without extending to the layer stack 1 or the surface of the semiconductor layer 4. Alternatively, the laser power of the pre-damage device 7 can be adjusted to form the surface recess in the form of the trench by laser ablation. The trench can also be formed by scribing within the semiconductor layer 4.

[0067] In principle, the pre-damage 8 should only be limited to the selected area, especially small dots or thin lines, and should only cover a portion of the separation line 11; such as Figures 4 to 7 As shown in other embodiments, along the separation line 11, the semiconductor wafer 4', including region 5, will be separated from the layer stack 1. In this case, the length of the pre-damage 8 is approximately 100 µm to 1000 µm.

[0068] Figure 2 and Figure 3The pre-damage device 7 includes a pulsed laser, which is used to introduce the desired pre-damage 8 into the semiconductor layer 4, which is part of the layer stack 1. The pre-damage 8 can also be positioned at any starting point of the separation line 11 to be formed by moving the pre-damage device 7 relative to the layer stack 1. In this embodiment, the pre-damage device 7 is moved relative to the layer stack 1 to form a pre-damage 8 with a length equivalent to that of the layer stack 1. This also allows for the formation of deep pre-damage 8 within the semiconductor layer 4. In another embodiment, the layer stack 1 may remain stationary relative to the pre-damage device 7, and the laser radiation from the pre-damage device 7 may be deflected by a mirror. Through this mirror, the laser radiation 10 is guided along the desired separation line 11 and applied to the semiconductor layer 4.

[0069] Subsequently, as Figure 2 As shown in view c), a thermal laser separation (TLS) device 9 applies laser radiation 10 to the semiconductor layer 4 along the separation line 11; along the separation line 11, the semiconductor layer 4' consisting of region 5 will be separated from the layer stack 1 through crack 13.

[0070] The starting point of the operation of device 9 is the pre-damage 8 formed in the preceding steps. Laser radiation is applied directly to the interior of semiconductor layer 4. As a result, semiconductor layer 4 is locally heated, causing thermal stress to be generated inside semiconductor layer 4. This will induce cracks 13 in the pre-damage, which will extend along the separation line 11. This allows region 5 of semiconductor layer 4 to be separated from its overlap 6 and the rest of the stack 1, wherein the semiconductor wafer 4' formed by region 5 will be further processed into electronic devices. Crack 13 is formed along separation line 11. In this embodiment, the shape of crack 13 is substantially consistent with separation line 11. However, crack 13 may also be offset relative to separation line 11, for example, due to stress caused by defects or faults in the crystal structure of the inner edge region of semiconductor layer 4.

[0071] To promote the generation of thermal stress, the device 9 is designed to apply a cooling medium 12 (especially an aerosol) during the heating of the semiconductor layer 4 by laser radiation 10, wherein the cooling medium 12 is applied at a position laterally offset from the laser radiation 10. As described above, the aerosol, as the cooling medium 12, can enhance and guide the generation of thermal stress, thereby initiating cracks 13 and enabling the cracks 13 to extend in a controlled and directional manner within the semiconductor layer 4.

[0072] The laser radiation 10 in device 9 is generated by a continuous-wave laser. This allows thermal stress to be continuously applied to the semiconductor layer 4. The wavelength of the laser is 1060 nm or 1064 nm. In embodiments not shown in detail here, the layer stack 1 may also remain stationary relative to device 9, and the laser radiation 10 may be deflected by a mirror or other guiding device. Through such a mirror, the laser radiation 10 is guided along the desired separation line 11 and applied to the semiconductor layer 4.

[0073] The key is that the laser power of the laser radiation 10 must be selected such that, apart from the heating effect, the laser itself will not cause any other degradation of the semiconductor layer 4. Specifically, applying laser radiation 10 to the semiconductor layer 4 will not cause changes in the microstructure or crystal structure of the semiconductor layer 4, the separation layer 3, or the seed substrate 2.

[0074] Once crack 13 forms, the semiconductor wafer 4' formed by the core region 5 separates from the stacked layer 1 and can be peeled off from the stacked layer 1 and the overlapping portion 6, as shown below. Figure 2 As shown in view d).

[0075] Figure 3 It shows Figure 2 The process steps shown are as follows: Figure 3 The side views a)-d) correspond to Figure 2 (a)-d) 3D views

[0076] Figure 3 View a) with Figure 1 The same view a) shows a layer stack 1, which includes a seed substrate 2, a separation layer 3 and a semiconductor layer 4, the semiconductor layer 4 including a region 5 and an overlap 6.

[0077] Figure 3 View b) exemplarily illustrates the formation of the pre-damage 8 in semiconductor layer 4: the pre-damage device 7 forms a deep scribing as a deep pre-damage 8. It can be clearly seen from view b) that the depth of the deep scribing-type pre-damage 8 does not reach the separation layer 3, therefore the separation layer 3 and the underlying seed substrate 2 are unaffected. Furthermore, the deep pre-damage 8 does not reach the surface of semiconductor layer 4. In fact, the purpose of forming the pre-damage 8 is solely to reduce the fracture stress of semiconductor layer 4, thereby separating region 5 from the rest of the layer stack 1 and the overlapping portion 6. The pre-damage 8 is formed along the length direction of the layer stack 1.

[0078] Figure 3 View c) shows that device 9 applies laser radiation 10 starting from the pre-damage 8, followed immediately by cooling of semiconductor layer 4 with cooling medium 12, which is applied at a position laterally misaligned with the laser radiation 10. This generates thermal stress within semiconductor layer 4, thereby initiating a crack 13 that originates from the pre-damage 8 and extends along separation line 11. This crack 13 extending along separation line 11 causes the core region 5 constituting the semiconductor wafer 4' to be further processed to separate from the layer stack 1, particularly from the overlap 6. Figure 3 As shown in view d), in order to peel the semiconductor wafer 4' from the stack 1, a peeling force is applied to the semiconductor wafer 4'.

[0079] Figure 4 and Figure 5This illustrates another method for separating the semiconductor wafer 4' from the layer stack 1. (Compared to...) Figure 2 and Figure 3 Unlike other embodiments, the pre-damage 8 in this example refers to both the initial pre-damage 8 and the end pre-damage 8' located at the end of the cutting line; and... Figure 2 and Figure 3 Unlike the deep scribing of deep pre-damage 8, the initial pre-damage 8 and the end pre-damage 8' are formed in a small area of ​​semiconductor layer 4 and do not extend along the length direction of the stacked layers 1.

[0080] like Figure 4 b) and Figure 5 As shown in b), in this embodiment, the initial scribing 8, which serves as the initial pre-damage 8, and the end pre-damage 8' are formed as grooves approximately 200 μm in length. These grooves are also formed by laser processing. However, they can also be formed by other methods such as scribing. Except... Figure 4 outside the first pre-damage 8 on the right side in b), Figure 4 Additional pre-damage 8' is also formed on the left side of b) to prevent crack 13 from penetrating at the end of separation line 11, especially at the edge of semiconductor layer 4. Figure 2 and Figure 3 The deep pre-damage 8 is different from the separation line 11 extending from it. Figure 4 and Figure 5 The separation line 11 is only the movement path of device 9.

[0081] Apart from the aforementioned differences between pre-damage 8 and 8', Figure 4 and Figure 5 The process shown is the same as Figure 2 and Figure 3 The processes shown are no different in principle. Figure 4 and Figure 5 a) View, providing layer stack 1, and according to Figure 4 and Figure 5 View b) shows initial pre-damage 8, 8' formed within semiconductor layer 4, which is particularly useful for reducing fracture stress as described above. Furthermore, according to... Figure 4 and Figure 5 View c) shows laser radiation 10 irradiating semiconductor layer 4. As described above, the laser radiation 10 generates thermal stress in semiconductor layer 4, causing cracks 13 to form along separation line 11. Semiconductor wafer 4' separates from layer stack 1 along separation line 11 and can be peeled off from layer stack 1 by applying peeling force.

[0082] Figure 6 and Figure 7 This illustrates another method for separating semiconductor wafer 4 from layer stack 1. (Compared to...) Figures 2 to 5Unlike other processes, this process does not use an additional cooling medium 12 to generate thermal stress. Instead, the thermal stress required for the propagation of crack 13 within semiconductor layer 4 is generated solely by the temperature difference between the heated area of ​​semiconductor layer 4 and the ambient temperature of approximately 300K.

[0083] besides, Figure 6 and Figure 7 The process shown is the same as Figure 2 and Figure 3 The processes shown are no different in principle. Figure 6 and Figure 7 a) View, providing layer stack 1, and according to Figure 6 and Figure 7 View b) shows a deep pre-damage 8 formed in the form of deep scribing within the semiconductor layer 4, which is particularly used to reduce fracture stress. This deep pre-damage 8 forms a separation line 11 within the semiconductor layer 4. In this embodiment, the layer stack 1 remains stationary relative to the pre-damage device 7, and the laser radiation 10 emitted by the pre-damage device 7 is deflected by a mirror. To deflect the laser radiation 10, the pre-damage device 7 includes one or more galvanometer scanners. Through the aforementioned mirrors, especially the galvanometer scanners, the laser radiation 10 is guided and irradiated onto the semiconductor layer 4 along the desired separation line 11. As described above, according to... Figure 6 and Figure 7 In view c), laser radiation 10 is irradiated onto semiconductor layer 4, which also generates the thermal stress described above, causing cracks 13 to form along separation line 11. Semiconductor wafer 4' separates from the stack 1 along separation line 11 and is then peeled off from the stack 1 by applying peeling force.

[0084] List of reference numerals

[0085] 1-layer stack

[0086] 2 Seed substrate

[0087] 3 Separation layer

[0088] 4 Semiconductor layer

[0089] 4' Semiconductor wafer

[0090] 5 areas

[0091] 6. Overlapping parts

[0092] 7. Pre-damage equipment

[0093] 8, 8' Pre-damage

[0094] 9 Equipment

[0095] 10. Laser radiation

[0096] 11. Separation line

[0097] 12 Cooling medium

[0098] 13 Cracks

Claims

1. A method for separating a semiconductor wafer (4') from a layer stack (1), comprising the following steps: A) Provide a layer stack (1), the layer stack (1) including a seed substrate (2), a separation layer (3) located on the seed substrate (2), and a semiconductor layer (4) formed on the separation layer (3), the semiconductor layer (4) preferably being prepared on the separation layer (3) by epitaxial growth; B) Separating the semiconductor wafer (4') at least partially from the layer stack (1), the semiconductor wafer (4') being at least a portion of the semiconductor layer (4), Its features are: Step B1), wherein pre-damage (8) is performed on the semiconductor layer (4) in a region to reduce the fracture stress of the semiconductor layer (4) in that region, wherein step B1) is performed before and / or during step B); and The second step (B2) involves applying thermal energy from a source to the semiconductor layer (4) along the separation line (11). The separation line (11) includes the pre-damage (8) or specifically originates from the pre-damage (8) in the semiconductor layer (4). Furthermore, the thermal energy device locally heats and / or cools the semiconductor layer (4) to generate thermal stress, thereby inducing cracks (13) extending along the separation line (11).

2. The method according to claim 1, characterized in that: The thermal energy device is a heating device, and a cooling medium (12) is applied along the separation line (11) before, during, and / or after the application of the heating device. The cooling medium is preferably an aerosol. In particular, the cooling medium (12) is applied in a laterally offset manner from the heating device, or The thermal energy device is a cooling device, and a heating medium is applied along the separation line (11) before, during and / or after the application of the cooling device, in particular, the heating medium is applied in a laterally offset manner from the cooling device.

3. The method according to any one of the preceding claims, characterized in that: The thermal energy device is laser radiation (10) from a laser source, preferably, the laser radiation (10) is continuously emitted and, in particular, the laser source is a continuous wave laser.

4. The method according to claim 3, characterized in that: The power of the laser radiation (10) is selected such that it is substantially absorbed within the semiconductor layer (4), and the power of the laser radiation (10) is selected such that the semiconductor layer (4) retains its original crystal structure.

5. The method according to any one of claims 3 to 4, characterized in that: The wavelength of the laser radiation (10) is in the infrared range, preferably 780 nm to 2000 nm, and the wavelength is preferably 800 nm to 1500 nm, particularly preferably 1000 nm to 1100 nm, especially 1060 nm to 1064 nm, and / or The laser radiation (10) is deflected by a mirror and / or a galvanometer scanner and applied to the semiconductor layer (4).

6. The method according to any one of the preceding claims, characterized in that: The thermal energy device applies heat in a continuous or pulsed manner, and / or The layer stack (1) is moved relative to the source according to the shape of the separation line (11).

7. The method according to any one of the preceding claims, characterized in that: The stacked layers (1) are stationary relative to the source, and the thermal energy device is guided to the semiconductor layer (4) by a guiding device.

8. The method according to any one of the preceding claims, characterized in that: The pre-damage (8) is formed by the laser source or another laser source, particularly by a pulsed laser source, preferably a laser source with a pulse length in the nanosecond range, and / or by scribing.

9. The method according to any one of the preceding claims, characterized in that: The pre-damage (8) is an initial pre-damage, wherein the initial pre-damage only covers a small area of ​​the semiconductor layer (4), and the length is preferably at least about 1 µm to 2000 µm, more preferably at least about 100 µm to 1000 µm.

10. The method according to any one of the preceding claims, characterized in that: The pre-damage (8) is a deep pre-damage, which extends in particular along the separation line (11) and has a length that can reach the size of the semiconductor layer (4), wherein the deep and long pre-damage preferably extends inside the semiconductor layer (4) rather than on its surface.

11. The method according to any one of the preceding claims, characterized in that: Multiple pre-damages (8) are created on the semiconductor layer (4), wherein the multiple pre-damages (8) are distributed on the semiconductor layer (4) according to the orientation of the separation line (11).

12. The method according to claim 11, characterized in that: Additional pre-damage (8) is created at the ends and / or edges and / or bends of the separation line (11), and / or at the intersection of at least two separation lines (11).

13. The method according to any one of the preceding claims, characterized in that: The semiconductor layer (4) includes an overlap (6) that protrudes from one side surface of the seed substrate (2) and / or the separation layer (3) and, in particular, at least partially covers one or more side surfaces of the seed substrate (2) and / or the separation layer (3).

14. The method according to claim 13, characterized in that: The overlapping portion (6) is formed of the same material as the semiconductor layer (4).

15. The method according to claim 13 or 14, characterized in that: The separation line (11) is formed such that in process step B), the overlap (6) is separated from the semiconductor wafer (4') and / or the layer stack (1).

16. The method according to any one of the preceding claims, characterized in that: The separation line (11) is completely closed, and in particular, the separation line (11) is formed such that the semiconductor wafer (4') obtained by stripping is generally rectangular.

17. The method according to any one of the preceding claims, characterized in that: The semiconductor layer (4) is made of silicon, particularly monocrystalline silicon, and / or the separation layer (3) is porous silicon.

18. The method according to any one of the preceding claims, characterized in that: The thickness of the semiconductor layer (4) is from 5 µm to 2000 µm, preferably from 10 µm to 500 µm, and particularly preferably from 50 µm to 200 µm.

19. A method for preparing a semiconductor wafer (4'), comprising the following steps: V1) Provide seed crystal substrate (2); V2) Create a separation layer (3) on the seed crystal substrate (2); V3) A semiconductor layer (4) is preferably formed on the separation layer (3) by chemical vapor deposition epitaxial growth to provide a layer stack (1) including the seed substrate (2), the separation layer (3) and the semiconductor layer (4); as well as V4) Specifically, the semiconductor wafer (4') is obtained by peeling it from the layer stack (1) using a peeling force, the semiconductor wafer (4') being at least a portion of the semiconductor layer (4). Its features are: The method according to any one of claims 1 to 18 is performed before or during method step V4).

Images