Method for manufacturing a pre-treated composite substrate, and pre-treated composite substrate

Abstract

A preprocessed composite substrate (18) is provided that is used as a basis for further processing into electronic semiconductor components. A method for producing a pre-processed composite substrate (18) includes doping a donor substrate (12) with a first layer (21) of SiC by ion implantation using an energy filter (20), creating a predetermined damage site (26) in the donor substrate (12), and creating a bonded connection between the donor substrate (12) and an acceptor substrate (28). The first layer (21) is disposed in a region between the acceptor substrate (28) and a remaining portion (22) of the donor substrate (12). Finally, the donor substrate (12) is divided at the region of the predetermined damage site (26) to produce a pre-processed composite substrate (18), which includes an acceptor substrate (28) and a doped layer (32) connected to the acceptor substrate and constituting at least a portion of the first layer (21) of the donor substrate (12).

Description

[Technical Field]

[0001] The present invention relates to a method for manufacturing a pre-treated composite substrate, and to a pre-treated composite substrate that serves as a basis for further processing into electronic semiconductor components.

[0002] Individual high-blocking power semiconductor components with nominal blocking voltages exceeding 600V are typically constructed vertically in both silicon and SiC. For diodes, specifically MPS (Merge PIN Schottky) diodes, Schottky diodes, or pn diodes, this means the cathode is on the front side of the substrate and the anode is on the back side. A similar configuration applies to vertical power MOS (metal oxide semiconductor) components, where the gate and source electrodes are on the front side of the substrate and the drain electrode is on the back side. The actual transistor elements or channel regions of a power MOSFET can be arranged parallel to the surface (D-MOS) or perpendicular to the surface (trench MOS). For SiC MOSFETs, such as trench transistors, specific structures have been established.

[0003] The width of the drift zone (active zone, voltage absorption layer) is adjusted according to the required reverse blocking voltage. For example, the drift zone width of a 600V MOSFET component in silicon is approximately 50 μm.

[0004] In the case of so-called superjunction components, the width of the voltage absorption layer may be slightly narrower compared to a "simple" vertical MOSFET. A special feature of this type of vertical component is that the drift zone is characterized by the alternating arrangement of vertical p-doped and n-doped columns. The p-doping, additionally introduced for blocking, compensates for the increased charge in the n-doped region that determines the resistance between the source and drain electrodes in the on state. Therefore, with the same blocking capability, the on-resistance can be reduced to approximately one-tenth or even tenth compared to a conventional vertical MOS transistor. The channel region of the actual transistor element, i.e., the superjunction MOSFET architecture, can be arranged parallel to the surface (D-MOS) or perpendicular to the surface (trench MOS).

[0005] The specific material properties of SiC for vertical power semiconductor components require the provision of specific manufacturing methods and the use of specific architectures for the channel and transistor regions.

[0006] Typically, the active zone of many vertical power diodes, or all power transistors (MOSFETs), is formed in a single-crystal epitaxial layer. These epitaxial layers are formed or deposited on a crystalline carrier wafer. This means that the doping and vertical spread (thickness) of the active epitaxial zone can be matched to their respective blocking voltages, and highly doped carrier wafers can be optimized in terms of doping to minimize their contribution to on-resistance.

[0007] In particular, with SiC substrates, the deposition of epitaxial layers and the preparation of single-crystal carrier wafers incur enormous costs, making the manufacturing of the aforementioned layered structures complex and expensive.

[0008] DE 10 2019 112 985 A1 proposes, as an alternative, the fabrication of semiconductor components without epitaxial deposition by separating the substrate from the SiC wafer and then implanting ions into the drift zone using an energy filter.

[0009] The present invention aims to identify a method for manufacturing a pre-treated composite substrate, and, based thereon, a pre-treated composite substrate that can reduce complexity and enable the industrial manufacture of high-performance, high-quality semiconductor components at a lower cost.

[0010] This objective is achieved by the features of independent claims 1 and 43. The advantageous configuration is the subject matter of the dependent claims.

[0011] According to the present invention, a method for manufacturing a pre-treated composite substrate that serves as a base for further processing into an electronic semiconductor component, wherein the pre-treated composite substrate comprises an acceptor substrate and a doped layer bonded thereto, a) A step of preparing a donor substrate containing single crystal SiC, b) A step of doping a first layer of a donor substrate by ion implantation using an energy filter, wherein the energy filter is a microstructured film having a predetermined structural profile for adapting the dopant depth profile and / or defect depth profile resulting from the implantation into the first layer of the donor substrate, the doping creating a predetermined dopant depth profile and / or defect depth profile in the first layer of the donor substrate, the first layer extending from a first surface of the donor substrate facing the ion beam to a predetermined dopant depth, and subsequently the remaining portion of the donor substrate extending thereafter. c) A step of creating an intended damaged area within the donor substrate, d) A step of preparing an acceptor substrate and creating a bond between a donor substrate and an acceptor substrate, wherein a first layer is placed in the region between the acceptor substrate and the rest of the donor substrate, e) A step of dividing a donor substrate in the region of the intended damaged area to create a pre-treated composite substrate, wherein the pre-treated composite substrate includes an acceptor substrate and a doped layer bonded thereto, the doped layer including at least a section of a first layer of the donor substrate, It holds.

[0012] The first layer is always composed of single-crystalline SiC. The donor substrate is preferably made entirely of single-crystalline SiC.

[0013] The thickness of the first layer is preferably 3 to 15 μm. Ion implantation can be carried out over a thickness of this magnitude.

[0014] In a preferred embodiment, the donor substrate is a crystal made of a high-purity high-quality semi-insulating SiC material (HQSSiC). In particular, this is understood to mean a material in which the concentrations of elemental impurities, especially N, B, and P, are mainly less than 5E15 cm -3 It is understood that what "mainly" means in this context is that the criterion is applicable substantially throughout the depth profile, but there may be deviations in certain regions, such as the surface.

[0015] In a preferred embodiment, the donor substrate is composed of 4H, 6H, or 3C polytype SiC. These polytypes have been found to be advantageous for the manufacture of semiconductor components.

[0016] The surface of the donor substrate facing the ion beam has a deviation from the perpendicular to the c direction of less than 6°, more preferably less than 4°, even more preferably less than 3°, and even more preferably 0°. The 4° orientation is currently used in most component architectures. The special advantage in the case of 0° is that the donor wafer can be cut parallel to the surface, so more individual wafers can be obtained from one cylinder.

[0017] The donor substrate preferably has a thickness exceeding 100 μm, preferably exceeding 200 μm, more preferably exceeding 300 μm, up to 15 cm, preferably up to 10 cm.

[0018] Alternatively, the donor substrate can preferably have a carrier wafer made of SiC and an epitaxial layer, and the epitaxial layer is either not doped or has a doping of less than 1E15 cm -3 less than, preferably less than 1E14 cm-3 Having a doping below, the first layer is part of an epitaxial layer.

[0019] In this case, the epitaxial layer preferably has a thickness exceeding 10 μm, preferably exceeding 50 μm, more preferably exceeding 80 μm. The maximum thickness of such a layer is usually 120 μm.

[0020] Here too, it is preferable that the surface of the epitaxial layer facing the ion beam has a deviation of less than 6°, more preferably less than 4°, more preferably less than 3°, and even more preferably 0° from the perpendicular to the c direction.

[0021] The epitaxial layer here is preferably composed of 4H, 6H, or 3C polytype SiC.

[0022] Generally, the doping of the first layer gives p or n doping with a doping concentration or defect concentration in the first layer of 1E15 cm -3 to 5E17 cm -3 This doping concentration or defect concentration is very suitable for the drift zones (active layers, power absorption layers) of many high-performance components. The doping may be constant throughout the thickness of the first layer or may exhibit different doping profiles.

[0023] The first layer is preferably doped with ions of one of the elements N, P, B, or Al.

[0024] The primary energy range of the ion beam in the doping of the first layer is preferably between 1 MeV and 50 MeV.

[0025] In a preferred embodiment, doping of the first layer gives a constant dopant depth profile and / or defect depth profile, or a substantially constant dopant depth profile and / or defect depth profile. This is understood to mean a profile in which the deviation from a perfectly flat dopant depth profile and / or defect depth profile is less than 20%, preferably less than 10%. In practice, the plateau is adjacent to a descending side; that is, the decline in the profile is neither vertical nor abrupt in the region of doping depth.

[0026] In the alternative configuration, doping of the first layer results in a stepped dopant depth profile and / or defect depth profile, where the steps are formed in the region near the surface of the first layer facing the ion beam, up to a maximum of 40%, preferably up to 30%, of the total depth of the first layer.

[0027] Preferably, the difference in concentration between the highest and lowest stages is at least 10 times, preferably at least 100 times, more preferably at least 500 times, and particularly preferably at least 1000 times.

[0028] Here, the depth-direction range of the lateral region of the stairs is more dominant than the depth-direction range of the stepped plateau.

[0029] In the alternative configuration, doping the first layer results in a continuously decreasing dopant depth profile and / or defect depth profile.

[0030] Here, it is preferable that the continuously decreasing dopant depth profile and / or defect depth profile follows the following formula. TIFF0007886034000001.tif23162 Here, D max This is the maximum doping concentration, α is a value between 10 and 10,000. z is the distance from the surface, b is the thickness of the layer, f is a tolerance coefficient between 0.95 and 1.05, D0 is the background doping, Here, TIFF0007886034000002.tif13162 Here, E max is the maximum field, ε r is the relative permittivity of the semiconductor, ε0 is the permittivity in vacuum, e0 is the elementary charge of an electron, V br is the breakdown voltage, And here, TIFF0007886034000003.tif13162

[0031] Generally, it is preferable to perform a further step of creating a contact layer on the surface region of the first layer where the bonding between the donor substrate and the acceptor substrate is established, or applying a contact layer on the surface of the first layer. Through the contact layer, an order of the acceptor substrate, the contact layer, the remaining part of the first layer or the first layer, and the remaining part of the donor substrate occurs. Thereby, a particularly good low-resistance connection between the donor substrate and the acceptor substrate can be achieved.

[0032] The contact layer is preferably created by ion implantation.

[0033] Preferably, the dopant concentration in the contact layer is at least 100 times, preferably at least 1000 times, more preferably at least 10000 times, even more preferably at least 100000 times higher than the average dopant concentration in the remaining part of the first layer or in the first layer. Thereby, a very low-resistance connection is realized and punch-through of the electric field to the interface of the semiconductor component is prevented.

[0034] In a preferred configuration, the dopant concentration in the contact layer exceeds 1E17 cm -3 and more preferably exceeds 1E19 cm-3 It exceeds.

[0035] The intended fracture site is preferably located within the region of the first layer, more preferably within the edge region of the first layer close to a predetermined doping depth, and the edge region is more preferably 1 μm or less. In this way, a minimum amount of doped material remains on the donor substrate after splitting.

[0036] In an alternative configuration, the intended fracture site is located in the remaining portion of the donor substrate, and furthermore, after step e), the additional step of ion implantation into the composite substrate is performed from the side away from the acceptor substrate. This has the advantage of allowing the formation of an active zone with a larger total thickness. In this way, overlap between the two different implantations is possible, making it possible to create different preferred doping profiles.

[0037] In connection with this alternative configuration, it is preferable to provide a dopant depth profile and / or defect depth profile to an auxiliary doping layer extending at least to the doped layer by ion implantation into the composite substrate.

[0038] Ion implantation into the composite substrate is preferably carried out such that the combination of two dopant depth profiles and / or defect depth profiles of the doped layer and the auxiliary doped layer results in a constant profile, a profile that gradually increases toward the acceptor substrate, or a profile that continuously increases toward the acceptor substrate.

[0039] The obliquely descending sides in the transition region of the two dopant depth profiles and / or defect depth profiles of the doped layer and the auxiliary doped layer may overlap each other.

[0040] It is preferable to create the intended fracture site by ion implantation of segmented trigger ions.

[0041] It is preferable that the splitting trigger ions are introduced across the entire width of the donor substrate to form a very uniform separation surface.

[0042] Alternatively, the segmented trigger ions may be introduced only over a portion of the width of the donor substrate. This reduces the complexity of ion implantation.

[0043] It is preferable to introduce split trigger ions only to the edge region of the donor substrate.

[0044] In a preferred embodiment, the splitting trigger ion is selected from H, H2, He, and B.

[0045] In principle, it is advantageous when the splitting trigger ion is a high-energy ion having an energy between 0.5 and 10 MeV, preferably between 0.5 and 5 MeV, and more preferably between 0.5 and 2 MeV.

[0046] The particle dose of the split trigger ion is 1E15cm in each case. -2 and 5E17cm -2 It is preferable that the dose be within this range. This dose ensures reliable fractionation.

[0047] The energy spread (ΔE / E) of the ion beam of the split trigger ion is preferably 10 -2 Less than 10 -4 It is less than this. In this way, it is ensured that the intended fracture site has the minimum thickness and that the ion energy loss peak at the intended fracture site is very sharp.

[0048] The splitting of the donor substrate is preferably caused by heat treatment of the composite substrate at a temperature between 600°C and 1300°C, preferably between 750°C and 1200°C, and more preferably between 850°C and 1050°C. Alternatively, a mechanical method may also be considered.

[0049] In a preferred embodiment, the bond is established by heat-treating the composite substrate at a temperature between 800°C and 1600°C, preferably between 900°C and 1300°C.

[0050] This method is considered simplified because both the establishment of the bond and the separation of the donor substrate are performed by heat treatment, and both steps are carried out simultaneously.

[0051] Preferably, prior to the step of establishing the bond, at least one, preferably both, of the surfaces to be bonded is pretreated, particularly by wet chemical treatment, plasma treatment, or ion beam treatment.

[0052] The acceptor substrate is preferably thermally stable up to at least 1500°C and has a coefficient of thermal expansion that is 20% or less, preferably 10% or less, of the SiC coefficient of thermal expansion. This effectively prevents bending of the composite substrate.

[0053] In a particularly preferred configuration, the acceptor substrate is formed from polycrystalline SiC or graphite.

[0054] Preferably, the splitting step is followed by post-treatment of the composite substrate surface in the area of ​​the intended damage, particularly polishing and / or removal of defects (near the surface).

[0055] A preferred extension of this method involves tempering the injection defects of the pre-treated composite substrate at temperatures between 1500°C and 1750°C. This can be achieved either during the manufacturing of the pre-treated composite substrate or solely through subsequent processing of the electrical components.

[0056] The pre-treated composite substrate according to the present invention, which serves as a base for further processing into electronic semiconductor components, comprises an acceptor substrate and a doped layer of single-crystal SiC bonded thereto.

[0057] The doped layer preferably has injection defects.

[0058] The thickness of the doped layer is preferably 3 μm to 25 μm, and more preferably 3 μm to 15 μm.

[0059] In a preferred configuration, the doped layer is composed of 4H, 6H, or 3C polytype SiC.

[0060] Preferably, the surface of the doped layer has a deviation of less than 6°, preferably less than 3°, and more preferably 0° from the perpendicular to the c direction.

[0061] The doping layer has a doping concentration of 1E15cm³. -3 From 5E17cm -3 It is preferable to have p or n doping.

[0062] The doped layer is preferably doped with an ion of one of the elements N, P, B, or Al.

[0063] The doped layer preferably has a substantially constant dopant depth profile and / or defect depth profile.

[0064] Preferably, the doped layer has a dopant depth profile and / or defect depth profile that increases stepwise toward the acceptor substrate, and the steps are formed in the region of the doped layer facing the acceptor substrate up to a maximum of 40% of the total depth of the doped layer, preferably up to a maximum of 30% of the total depth of the doped layer.

[0065] It is also preferable that the difference in concentration between the highest and lowest stages is at least 10 times, preferably at least 100 times, more preferably at least 500 times, and particularly preferably at least 1000 times.

[0066] It is even more preferable that the depth-direction range of the lateral region of the stairs is dominant over the depth-direction range of the stepped plateau.

[0067] Preferably, the doped layer provides a dopant depth profile and / or defect depth profile that increases continuously in the direction toward the acceptor substrate.

[0068] The continuously increasing dopant depth profile and / or defect depth profile is preferably a profile that follows the following formula: TIFF0007886034000004.tif22162 Here, D max This is the maximum doping concentration, α is a value between 10 and 10000. z is the distance from the surface, b is the thickness of the layer, f is a tolerance coefficient between 0.95 and 1.05. D0 is background doping, Here, TIFF0007886034000005.tif13162 Here, E max This is the maximum electric field, ε r This is the relative permittivity of a semiconductor, ε0 is the permittivity in a vacuum, e0 is the elementary charge of an electron, V br This is the breakthrough voltage, And here, TIFF0007886034000006.tif13162

[0069] Two aspects are considered in the aforementioned stepped profile, or continuously rising profile. First, this dopant profile provides an optimal compromise between on-resistance and a given voltage stability. Second, the very high concentration of the doping profile near the acceptor substrate eliminates electric field punch-through to the interface.

[0070] In one embodiment, in addition to the doped layer, an auxiliary doping layer of single-crystal SiC is provided, and the overlapping region of the respective dopant depth profile and / or defect depth profile exists in the transition area between the doped layer and the auxiliary doping layer. The two dopant depth profiles and / or defect depth profiles have overlapping, diagonally descending sides. The combination of the two dopant depth profiles and / or defect depth profiles of the doped layer and the auxiliary doping layer may be a constant profile, a profile that gradually increases toward the acceptor substrate, or a profile that continuously increases toward the acceptor substrate.

[0071] Preferably, the doped layer and the auxiliary doped layer are doped with the same type of ions. The combined thickness of the doped layer and the auxiliary doped layer is a maximum of 40 μm.

[0072] The acceptor substrate is preferably thermally stable up to at least 1500°C and has a coefficient of thermal expansion that deviates from the coefficient of thermal expansion of SiC by 20% or less, preferably 10% or less.

[0073] In a particularly preferred configuration, the acceptor substrate is formed from polycrystalline SiC or graphite. [Brief explanation of the drawing]

[0074] Figure 1 is a schematic cross-sectional view of the first configuration of a donor substrate that can be used in the method of the present invention.

[0075] Figure 2 is a schematic cross-sectional view of a second configuration of a donor substrate that can be used in the method of the present invention.

[0076] Figure 3 is a schematic diagram of an irradiation setup equipped with an energy filter for irradiating the donor substrate.

[0077] Figure 4 is a schematic diagram of the operating modes of the energy filter that can be used in the method of the present invention.

[0078] Figure 5 is a schematic diagram of different doping profiles that can be generated by energy filters of different structures.

[0079] Figure 6 shows the doping profile of the first layer of the donor substrate and a schematic diagram of the resulting doping profile of the donor substrate.

[0080] Figure 7 shows various options for the doping profile of the first layer of the donor substrate.

[0081] Figure 8 shows a schematic diagram of the creation or coating of the contact layer on the donor substrate.

[0082] Figure 9 shows a schematic diagram of the first modification that creates the intended fracture site within the donor substrate.

[0083] Figure 10 shows a schematic diagram of a second modification that creates an intended fracture site within the donor substrate.

[0084] Figure 11 shows a schematic diagram of the formation of a bond between the donor substrate and the acceptor substrate.

[0085] Figure 12 shows a schematic diagram of separating the remaining portion of the donor substrate from the composite substrate.

[0086] Figure 13 shows a schematic diagram of the post-treatment of the composite substrate surface in the divided region.

[0087] Figure 14 shows a schematic cross-sectional view of one embodiment of a pre-treated composite substrate according to the present invention.

[0088] Figure 15 shows a cross-sectional view of a further embodiment of a pre-treated composite substrate according to the present invention and the corresponding doping profile.

[0089] Figure 16 is a schematic diagram of the division of a wafer bar that functions as a donor substrate when used to create multiple composite substrates from a donor substrate.

[0090] Figure 17 shows a schematic diagram of the doping profile of the first layer of the donor substrate using partial masking of the donor substrate, and the resulting alternative doping profile of the donor substrate.

[0091] The present invention's method for manufacturing a pre-treated composite substrate begins with preparing a donor substrate 12 containing or composed entirely of single-crystal silicon carbide (SiC). See Figures 1 and 2.

[0092] The embodiment of the donor substrate 12 shown in Figure 1 is a wafer made of high-purity, high-quality semi-insulating SiC material (HQSSiC). In particular, it has a concentration of elemental impurities, such as N, B, and P, of 5E15cm². -3 It is understood to mean materials that are less than [a certain value]. In this context, "primarily" means that the standard is substantially applicable to the entire depth profile, but differences may exist in certain areas, such as the surface.

[0093] The donor substrate 12 shown in Figure 1 preferably has a thickness of more than 100 μm, more preferably more than 200 μm, more preferably more than 300 μm, and up to 15 cm, preferably up to 10 cm. In particular, it may take the form of an undoped or lightly n-doped wafer bar. See Figure 16.

[0094] In a preferred embodiment, the donor substrate is composed of 4H, 6H, or 3C polytype SiC. These polytypes have been shown to be advantageous for the properties of the semiconductor components manufactured using them.

[0095] In the illustrated embodiment, the upper surface of the donor substrate 12 has a deviation of 0° from the line perpendicular to the c direction. Alternatively, a deviation of up to 3° or up to 6° from the line perpendicular to the c direction is possible.

[0096] The embodiment of the donor substrate 12 shown in Figure 2 is preferably a wafer having a carrier wafer 14 made of SiC and an epitaxial layer 16 made of SiC, wherein the epitaxial layer 16 is undoped or 1E15cm -3 Less than 1E14cm -3 It has less than 1% doping. The epitaxial layer here is preferably composed of 4H, 6H, or 3C polytype SiC.

[0097] In this case, the thickness of the epitaxial layer 16 is preferably greater than 10 μm, more preferably greater than 50 μm, and more preferably greater than 80 μm. The maximum thickness of such an epitaxial layer 16 is generally 120 μm.

[0098] Here, the deviation of the upper surface of the epitaxial layer 16 from the line perpendicular to the c direction is preferably less than 6°, more preferably less than 3°, and even more preferably 0°.

[0099] After the donor substrate 12 is prepared, the first layer 21 of the donor substrate 12 is doped (see Figure 6) and subsequently assumes, or partially assumes, the function of a drift zone (also called an active zone or voltage absorption zone) in the finished component. This doping of the first layer 21 of the donor substrate 12 is performed by ion implantation using an energy filter 20. The corresponding basic structure is shown in Figure 3.

[0100] Figure 3 shows an irradiation chamber 8, typically in a high vacuum. The irradiation chamber 8 houses a donor substrate 12 to be doped in a substrate holder 30.

[0101] The ion beam 10 is generated by a particle accelerator (not shown) and guided into the irradiation chamber 8. The energy of the ion beam 10 is diffused by the energy filter 20, where it collides with the donor substrate 12 and irradiates the irradiation chamber 8. Alternatively, the energy filter 20 may be located in the irradiation chamber 8 or in a separate vacuum chamber that can be closed by a valve immediately adjacent to the irradiation chamber 8.

[0102] The substrate holder 30 does not need to be stationary and may optionally be equipped with a device for moving the donor substrate 12 in the xy direction (in a plane perpendicular to the sheet surface). Another useful substrate holder 30 is a wafer wheel on which the donor substrate 12 to be injected is fixed and rotated during injection. It is also possible to move the substrate holder 30 in the beam direction (z direction). Furthermore, the substrate holder 30 may be equipped with a heater or cooler as needed.

[0103] The basic principle of the energy filter 20 is shown in Figure 4. The energy of the single-energy ion beam 10 changes depending on the entry site as it passes through the energy filter 20, which is configured as a microstructured film. The resulting ion energy distribution of the ion beam 10 results in a modification of the depth profile of the implanted material in the matrix of the donor substrate 12. E1 represents the energy of the first ion, E2 represents the energy of the second ion, c represents the doping concentration, and d represents the depth of the donor substrate 12. This figure shows a typical Gaussian distribution with reference numeral A on the right, which occurs when the energy filter 20 is not used. In contrast, reference numeral B shows, as an example, a rectangular distribution that can be achieved when the energy filter 20 is used.

[0104] The layout or three-dimensional structure of the energy filter 20 shown in Figure 5 illustrates basic options for creating multiple dopant depth profiles or defect depth profiles using the energy filter 20. c again indicates the doping concentration, and d again indicates the depth within the donor substrate 12. In principle, the filter structure profiles can be combined with each other to obtain new filter structure profiles, and consequently, new dopant depth profiles or defect depth profiles.

[0105] Such energy filters 20 are generally manufactured from silicon. They have a thickness between 3 μm and 200 μm, preferably between 5 μm and 50 μm, and more preferably between 7 μm and 20 μm. They may be held within a filter frame (not shown). The filter frame may be removably housed in a filter holder (not shown).

[0106] For the preferred formation of the n-doped first layer 21, implantation of nitrogen or phosphorus ions is particularly suitable, while for the p-doped layer, implantation of boron or aluminum ions is particularly suitable.

[0107] In the embodiment of the method step for doping the first layer 21 shown in Figure 6, ions are implanted into the donor substrate 12 from the front surface of the donor substrate 12. Short solid black arrows indicate ions with the lowest energy passing through the energy filter 20, and long solid black arrows indicate ions with the highest energy passing through the energy filter 20. The resulting doping profile in section A-A' is shown on the right side of the coordinate system. c represents the doping concentration. The doping profile is based on the configuration of the donor substrate 12 according to Figure 1 and is substantially uniform throughout the first layer 21. The first layer 21 extends from the surface of the donor substrate 12 facing the ion beam 10 to a predetermined doping depth T, and subsequently becomes the remaining portion 22 of the donor substrate 12 that is not affected by ion implantation by the energy filter.

[0108] The thickness of the first layer 21 preferably substantially corresponds to the pre-determined thickness of the active layer in the subsequent components, or substantially corresponds to the combination of the active layer and the field-stopping layer, or the combination of the active layer, the field-stopping layer and the surface functional zone. Therefore, the total thickness of the first layer 21 is determined by the properties of the semiconductor component being manufactured, particularly the voltage class. The higher the voltage class, the thicker the first layer 21 becomes. For particularly high voltage classes, refer to Figure 15 and its accompanying description.

[0109] The thickness of the first layer 21 is preferably between 3 and 15 μm. This corresponds to the currently feasible doping depth T for the preferred ion types in SiC described above.

[0110] Figures 7a to 7c show possible preferred doping profiles in the first layer 21 of the donor substrate 12.

[0111] As a general rule, doping in the first layer 21 is defined as a doping concentration or defect concentration within the first layer 21 of 1E15cm². -3 From 5E17cm -3 This results in p or n doping.

[0112] Figure 8 shows the results of any step of creating a contact layer 24 in the surface region of the first layer 21, or applying the contact layer 24 to the surface of the first layer 21.

[0113] It is preferable to create the contact layer 24 by ion implantation into the first layer 24. The contact layer 24 has a thickness of only 10 nm to 1 μm. It is preferable to use P, N, or Al ions for implantation (without using an energy filter).

[0114] The dopant concentration in the contact layer 24 is preferably at least 100 times, more preferably at least 1000 times, more preferably at least 10000 times, and even more preferably at least 100000 times higher than the average dopant concentration in the rest of the first layer 21 or within the first layer 21.

[0115] The dopant concentration in contact layer 24 is 1E17cm². -3 It is preferable that it exceeds 1E19cm -3 It is preferable to exceed this.

[0116] For example, a thin contact layer 24 with a thickness of several nanometers can be applied to the first layer 21. This can be achieved, for example, by sputtering deposition, vapor deposition, or CVD deposition. The contact layer 24 does not need to completely cover the layer and may be composed of nanoparticles.

[0117] Further surface treatment, such as physical etching, may be performed simultaneously with or after the application of the contact layer 24.

[0118] In the next step, as shown in Figure 9, the intended failure site 26 is created within the donor substrate 24. In the example of Figure 9, the intended failure site 26 is preferably located within the region of the first layer 21, in the edge region of the first layer 21 close to a predetermined doping depth T, and is away from the doping depth T, and therefore preferably 1 μm or less, more preferably 500 nm or less, and more preferably 100 nm or less from the edge of the first layer 21. In particular, in the case of a rectangular profile with a descending side surface, the intended failure site 26 should still be within the plateau region.

[0119] The intended fracture site 26 is preferably created by ion implantation of splitting trigger ions, schematically shown as black dots in Figure 9. No energy filter is used here. According to Figure 9, the splitting trigger ions are introduced across the entire width of the donor substrate 12. The splitting trigger ions are preferably selected from H, H2, He, and B. The splitting trigger ions are high-energy ions having energies between 0.5 and 10 MeV, preferably between 0.5 and 5 MeV, and more preferably between 0.5 and 2 MeV. For hydrogen, when the ion energy is 0.6 MeV, the intended fracture site 26 is formed at a depth of approximately 5 μm; when the ion energy is 1.0 MeV, it is formed at a depth of approximately 10 μm; and when the ion energy is 1.5 MeV, it is formed at a depth of approximately 20 μm.

[0120] The particle dose of the split trigger ion is 1E15cm in each case. -2 and 5E17cm -2 It is preferable that it be between [values]. The energy spread (ΔE / E) of the ion beam of the split trigger ion is preferably 10 -2 Less than 10 -4 It is less than . In implanting split trigger ions, it is advantageous to maintain the temperature inside the donor substrate 12 at less than 300°C, preferably less than 200°C. For this purpose, the chuck on which the donor substrate 12 is placed is optionally cooled.

[0121] Using these parameters, a doping profile with sharp peaks is created (see the Gaussian distribution shown in Figure 4A). In this way, high doping can be reliably distributed over an extremely thin thickness at the intended fracture site 26. The change in ion penetration in the donor substrate 12 (longitudinal dispersion σ) is only between 100 nm and 500 nm, preferably between 200 nm and 400 nm, depending on the primary energy of the ion beam.

[0122] Alternatively, as shown by the arrows and black horizontal bars in Figure 10, the splitting trigger ions may be introduced only over a portion of the width of the donor substrate 12, preferably only to one or both end regions of the donor substrate 12. In this way, the intended fracture sites 26 are predefined section by section.

[0123] As an alternative to ion implantation, the intended damage site 26 can also be formed by electron irradiation or laser irradiation.

[0124] Next, as shown in Figure 11, the donor substrate 12 is bonded to the acceptor substrate 28 with adhesive, with the side of the first layer 21 facing outwards. In this way, the first layer 21 is positioned in the region between the acceptor substrate 28 and the remaining portion 22 of the donor substrate 12. As indicated by the curved arrow in Figure 11, which also shows the donor substrate 12 being turned over, it is not important whether the donor substrate 12 is moved toward the acceptor substrate 28 or the acceptor substrate 28 is moved toward the donor substrate 12 in order to establish the bond.

[0125] An intermediate result of the coupling process is shown in the lower left of Figure 11. Similarly, if, for example, the acceptor substrate 28 moves toward the donor substrate 12, it is also possible to reverse the layer order.

[0126] The acceptor substrate 28 can be made from a range of materials. The acceptor substrate 28 is preferably thermally stable up to at least 1500°C and has a coefficient of thermal expansion that deviates by 20% or less, ideally 10% or less, from that of SiC. Suitable examples of materials for the acceptor substrate 28 are polycrystalline SiC or graphite.

[0127] Although Figures 10 and 11 do not show the contact layer 24 in each case, it is preferable that the contact layer 24 be present. In that case, the bond between the donor substrate 12 and the acceptor substrate 28 is established via the contact layer 24, resulting in the order of acceptor substrate 28, contact layer 24, the rest of the first layer 21, or the first layer 21, the rest of the donor substrate 12 22.

[0128] Low-resistance coupling is preferably established by heat-treating the substrate obtained as an intermediate result at a temperature between 800°C and 1600°C, more preferably between 900°C and 1300°C.

[0129] Prior to the step of establishing the bond, at least one, preferably both, of the surfaces to be bonded may be pretreated, particularly by wet chemical treatment, plasma treatment, or ion beam treatment. The treated surface may be the contact layer 24. Alternatively, a thin layer several nanometers thick may be applied to create a subsequent low-resistance bond between the acceptor substrate 28 and the donor substrate 12. In principle, extremely low-resistance contact and high-temperature resistant bonding between the acceptor substrate 28 and the donor substrate 12 are important.

[0130] Figure 12 shows a schematic diagram of the steps for separating the donor substrate 12 in the region of the intended damage area 26, creating a pre-treated composite substrate 18 including an acceptor substrate 28 and a doped layer 32 bonded thereto. Here, the doped layer 32 constitutes at least a portion of the first layer 21 of the donor substrate 12. The portion 34 of the donor substrate 12 separated from the acceptor substrate 28 is removed.

[0131] The splitting of the donor substrate 12 is preferably caused by heat treatment of the composite substrate 18 at a temperature between 600°C and 1300°C, preferably between 750°C and 1200°C, and more preferably between 850°C and 1050°C. In one embodiment (see Figures 9 and 10), implanted ions form bubbles, which then coalesce to cause splitting.

[0132] Alternatively, an external force may be applied to the composite substrate 18 so that the donor substrate 12 is damaged at the intended damage site 26. A combination of heat treatment and external force may also be necessary or useful. In particular, the application of an external force is unavoidable when ions are introduced only partially into the donor substrate 12.

[0133] If both the establishment of the bond and the separation of the donor substrate 12 are performed by heat treatment, the two steps can be carried out simultaneously in some cases.

[0134] As schematically shown by the arrows in Figure 13, after the splitting step, the surface of the composite substrate 18 in the area of ​​the intended damaged portion 26 may be post-treated, in particular by polishing and / or removal of defects.

[0135] The injection defects 42 schematically shown in Figure 14 can be ultimately tempered in the doped layer 32 of the pre-treated composite substrate 18, preferably at a temperature between 1500°C and 1750°C. This is preferably performed during subsequent component processing in a heat treatment step for tempering low-energy injections, such as source-drain contact injections, channel injections, and p-JFET injections.

[0136] Furthermore, if the corresponding high temperature is used during the splitting of portion 34 of the donor substrate 12 and / or during engagement formation between the donor substrate 12 and the acceptor substrate 28, and thus radiation defects can be tempered, then it can also be considered that the step of tempering injection defects 42 has already been performed.

[0137] Figures 8 to 13 illustrate and explain the steps of the conventional method using the donor substrate 12 shown in Figure 1, but these steps can be similarly performed using the donor substrate 12 shown in Figure 2. In that case, it is important that the epitaxial layer 16 of the donor substrate 12 is bonded to the acceptor substrate 28.

[0138] Contrary to the above description, the step of creating a bond between the donor substrate 12 and the acceptor substrate 28 can also be carried out in two steps. For example, first, the bonding process is performed at a low temperature and low bonding energy, and then solidification is performed in a subsequent second step to create a bond with higher bonding strength or bonding energy at a higher temperature and lower contact resistance. Solidification may be performed, for example, during or after splitting, during or after surface treatment of the composite substrate, or during or after tempering of injection defects.

[0139] The pre-treated composite substrate 18 produced in this manner serves as a base for further processing into electronic semiconductor components, as shown again in Figure 14. This includes an acceptor substrate 28 and a doped layer 32 of single-crystal SiC bonded thereto, where the doped layer 32 preferably contains injection defects 42 (radiation defects). A contact layer 24 may also be present between the acceptor substrate 28 and the doped layer 32.

[0140] The doped layer 32 preferably has a thickness of 3 μm to 30 μm, more preferably 3 μm to 15 μm. It is preferably composed of 4H, 6H, or 3C polytype SiC. The surface of the doped layer 32 preferably has a deviation of less than 6°, preferably 0°, from the line perpendicular to the c direction. The doped layer 32 is 1E15cm -3 From 5E17cm -3 It is preferable to have p or n doping with a doping concentration or defect concentration. The doped layer 32 is preferably doped with an ion of one of the elements N, P, B, or Al as a dopant.

[0141] The dopant depth profile and / or defect depth profile of the doped layer 32 is preferably obtained essentially from the inversion of the dopant depth profile and / or defect depth profile of the first layer 21 of the donor substrate 12.

[0142] Therefore, the doped layer 32 may have, for example, a substantially constant dopant depth profile and / or defect depth profile.

[0143] Similarly, the doped layer 32 may have a dopant depth profile and / or defect depth profile that rises in a stepwise manner toward the acceptor substrate 28, where the step is formed in the region of the doped layer 32 facing the acceptor substrate 28 up to 40%, preferably up to 30%, of the total depth of the doped layer 32.

[0144] The doped layer 32 can also provide a dopant depth profile and / or defect depth profile that increases continuously in the direction toward the acceptor substrate 28.

[0145] The injection defect profile essentially follows the depth profile of the injected external atom concentration.

[0146] The acceptor substrate 28 is thermally stable up to at least 1500°C and has a coefficient of thermal expansion that deviates from the coefficient of thermal expansion of SiC by 20% or less, preferably 10% or less. The acceptor substrate 28 is more preferably formed from polycrystalline SiC or graphite.

[0147] Figure 15 shows a cross-sectional view of another configuration of the pre-treated composite substrate 18 of the present invention, with below it showing the dopant concentration profile along the cross-section of the composite substrate 18 corresponding to arrow F. This is particularly suitable for the manufacture of very high-level blocking components, for example, greater than 1200V.

[0148] In this case, the pre-treated composite substrate 18 has an auxiliary doping layer 38 of single-crystal SiC in addition to the doped layer 32. Preferably, there is an overlapping region 40 of the respective dopant depth profiles and / or defect depth profiles in the transition area between the doped layer 32 and the auxiliary doped layer 38.

[0149] In the embodiment shown in Figure 14, the active layer (drift zone, voltage absorption layer) required in the later semiconductor component is formed solely by the doped layer 32, and therefore simultaneously by the first layer 21 or a (preferably larger) portion of the first layer 21 in the donor substrate 12.

[0150] In contrast, the active layer in the embodiment shown in Figure 15 is formed by a combination of a doped layer 32 and an auxiliary doped layer 38. In Figure 15, a substantially constant cumulative doping profile can be obtained from the superposition of the two component profiles, while other doping profiles can be formed by juxtaposing and partially overlapping the doping profiles of the doped layer 32 and the auxiliary doped layer 38. Thus, the composite overall doping profile, consisting of a combination of the two dopant depth profiles and / or defect depth profiles of the doped layer 32 and the auxiliary doped layer 38, can also be a profile that increases stepwise toward the acceptor substrate 28, or a profile that increases continuously toward the acceptor substrate 28.

[0151] Such a combination profile is obtained by creating the intended fracture site 26 within the donor substrate 12 not within the first layer 21, but within the remaining portion 22 of the donor substrate 12 that is not doped by ion implantation into the donor substrate 12.

[0152] As shown in Figure 12, after splitting at the intended fracture site 26, doping of the auxiliary doping layer 38 can be performed from the side furthest from the acceptor substrate 28 by further ion implantation using an energy filter. The description of ion implantation using an energy filter, as described above with respect to Figures 3 to 7, can also be applied to ion implantation of the auxiliary doping layer 38. The thickness of the auxiliary doping layer 38 is generally between 3 and 15 μm. In this way, a total thickness of the ion-implanted active zone of up to 30 μm can be obtained.

[0153] In principle, if the epitaxial layer 16 of the donor substrate 12 from Figure 1 or the donor substrate 12 from Figure 2 is at least twice the thickness of the dope layer 32 required for the composite substrate 18, then by the method of the present invention, it is possible to manufacture two or more composite substrates 18, and even a large number of composite substrates 18, from the donor substrate 12. This effect is particularly high when the donor substrate 12 is a thick wafer bar. In this way, it is possible to significantly reduce manufacturing costs. This is schematically shown in Figure 16.

[0154] As shown in Figure 17, in ion implantation into the first layer 21 of the donor substrate 12 (and / or the auxiliary doping layer 38 of the composite substrate 18) by the energy filter 20, a mask 46 can be used to create one or more undoped regions 44 within the first layer 21 of the donor substrate 12 (and / or the auxiliary doping layer 38 of the composite substrate 18).

[0155] The composite substrate 18 can also be characterized by further intermediate steps leading to the finished semiconductor component, such as the injection of further active regions, the fabrication of oxides, and the deposition of gate electrodes, contacts, wires, or vias.

[0156] In the context of this invention, “bonded” is understood to mean directly bonding or indirectly bonding, that is, bonding with the interposition of further elements. The “bonding” between two elements may be direct or indirect.

Claims

1. A method for manufacturing a pre-treated composite substrate (18) that serves as a base for further processing into electronic semiconductor components, wherein the pre-treated composite substrate (18) comprises an acceptor substrate (28) and a doped layer (32) bonded thereto. a) A step of preparing a donor substrate (12) equipped with single crystal SiC, b) Doping a first layer (21) of the donor substrate (12) by ion implantation using an energy filter (20), wherein the energy filter (20) is a microstructured film having a predetermined structural profile for adapting the dopant depth profile and / or defect depth profile produced by implantation into the first layer (21) of the donor substrate (12), the doping creates a predetermined dopant depth profile and / or a predetermined defect depth profile within the first layer (21) of the donor substrate (12), the first layer (21) extends from a first surface of the donor substrate (12) facing the ion beam (10) to a predetermined dopant depth (T), and subsequently the remaining portion (22) of the donor substrate (12) extends, the doping of the first layer is doping of N, P, or Al ions, and the doping of the first layer is between 1E15 cm⁻³ and 5E17 cm⁻³ in doping concentration within the first layer The steps include administering p or n doping up to a certain point, c) The step of creating an intended damaged area (26) within the donor substrate (12), d) A step of preparing an acceptor substrate (28) and manufacturing a bond between the donor substrate (12) and the acceptor substrate (28), wherein the first layer (21) is placed in the region between the acceptor substrate (28) and the remaining portion (22) of the donor substrate (12), e) A step of dividing the donor substrate (12) in the region of the intended damaged area (26) to create the pre-treated composite substrate (18), wherein the pre-treated composite substrate (18) includes the acceptor substrate (28) and a doped layer (32) bonded thereto, the doped layer (32) includes at least a section of the first layer (21) of the donor substrate (12), A method characterized by having the following:

2. The method according to claim 1, wherein the first layer (21) has a thickness of 3 to 15 μm.

3. The donor substrate (12) is a crystal made of high-purity, high-quality semi-insulating SiC material (HQSSiC), The method according to claim 1 or 2, wherein the donor substrate (12) is made of 4H, 6H, or 3C polytype SiC.

4. The method according to claim 3, wherein the donor substrate (12) has a thickness of more than 100 μm and up to 15 cm.

5. The donor substrate (12) comprises a carrier wafer (14) and an epitaxial layer (16), wherein the epitaxial layer (16) is undoped or 1E15cm －３ Having less than doping, the first layer (21) is part of the epitaxial layer (16), The method according to claim 1, wherein the epitaxial layer (16) has a thickness of more than 10 μm, and the epitaxial layer (16) is composed of 4H, 6H, or 3C polytype SiC.

6. Doping the first layer (21) yields a continuously decreasing dopant depth profile and / or defect depth profile, which follows the following formula: [Math 1] D ｍａｘ This is the maximum doping concentration, α is a value between 10 and 10,000. z is the distance from the surface, b is the thickness of the layer, f is a tolerance coefficient between 0.95 and 1.

05. D ０ The method according to claim 1, wherein background doping is used.

7. The process further includes creating a contact layer (24) on the surface region of the first layer (21), or coating the contact layer (24) onto the surface of the first layer (21), wherein the bond between the donor substrate (12) and the acceptor substrate (28) is established via the contact layer (24), resulting in the order of acceptor substrate (28), contact layer (24), the remainder of the first layer or the first layer (21), the remainder of the donor substrate (12) (22), and the dopant concentration in the contact layer (24) is 1E17cm². －３ The method according to claim 1, which exceeds the claim.

8. The method according to claim 1, wherein the intended damage area (26) is located within the edge region of the first layer (21) near a predetermined doping depth (T), and the thickness of the edge region is 1 μm or less.

9. The method according to claim 1, wherein the intended damaged area (26) is located within the area of ​​the remaining portion (22) of the donor substrate (12), and further, after step e), a further step of performing ion implantation into the composite substrate (18) using an energy filter (20) is performed from the side away from the acceptor substrate (28).

10. The ion implantation into the composite substrate (18) extends at least to the doped layer (32), The method according to claim 9, wherein the ion implantation into the composite substrate (18) is performed such that the combination of two dopant depth profiles and / or defect depth profiles of the dope layer (32) and the auxiliary dope layer (38) is a constant profile, a profile that increases stepwise toward the acceptor substrate (28), or a profile that increases continuously toward the acceptor substrate (28).

11. The intended fracture site (26) is created by ion implantation of fragmented trigger ions. The method according to claim 1, wherein the divided trigger ions are introduced over the entire width of the donor substrate (12).

12. The intended fracture site (26) is created by ion implantation of fragmented trigger ions. The method according to claim 1, wherein the divided trigger ions are introduced only over a portion of the width of the donor substrate (12).

13. The aforementioned split trigger ions are H, H ２ Selected from He, B, The aforementioned split trigger ion is a high-energy ion having an energy between 0.5 and 10 MeV. The particle dose of the split trigger ion is 1E15cm in each case. －２ and 5E 17cm －２ The method according to claim 11, wherein the method is between the two.

14. The aforementioned split trigger ions are H, H ２ Selected from He, B, The aforementioned split trigger ion is a high-energy ion having an energy between 0.5 and 10 MeV. The particle dose of the divided trigger ions is 1E15 cm in each case －２ and 5E17 cm －２ The method according to claim 12, wherein the particle dose is between

15. The method according to claim 1, wherein the acceptor substrate (28) is thermally stable up to at least 1500°C and has a linear expansion coefficient that is biased to 20% or less than that of SiC.

16. The method according to claim 15, wherein the acceptor substrate (28) is formed from polycrystalline SiC or graphite.

Images