Semiconductor layer stack and method of manufacturing the same

Abstract

The invention relates to a semiconductor layer stack, a component and a component module composed thereof, and a manufacturing method, wherein the semiconductor layer stack is characterized by at least two layers (A, B) which as individual layers each have an energy position of a Fermi level (103) in a semiconductor band gap (104, 105) which for the layer (A) applies formula (I) and for the layer (B) applies formula (II), wherein E F is the energy position of the Fermi level (103), E V is the energy position of the valence band (102), E L is the energy position of the conduction band (101), E L -E V is the energy difference of the semiconductor band gap E G (104, 105), wherein the thicknesses (106, 107) of the layers (A, B) are chosen such that there is a contiguous space-charge region (110) on the layers (A, B).

Description

Technical Field

[0001] This invention relates to a semiconductor layer stack and a method for manufacturing the same. Background Technology

[0002] The insulating semiconductor layer is irreplaceable for electrical insulation and low high-frequency attenuation in semiconductor component structures. Here, so-called deep impurities (Störstellen) are mostly used, meaning only a low degree of impurity at operating temperatures (i.e., 50% ionized impurities. If the semiconductor has intrinsic or ionized electronic conduction, deep doping is typically used, either with an Akzeptor or with a deep donor; conversely, if holes are intrinsically conductive, deep donor doping is used. This is because, in the first case, electron injection into the isolation layer is effectively prevented, and in the second case, hole injection is effectively prevented. Insulation properties are improved by simultaneously doping InP with both Akzeptors and deep donors (e.g., Akzeptor Fe and deep donor Ti) [T. Wolf, T. Zinke, A. Krost, H. Scheffler, H. Ullrich, P. Harde, and D. Bimberg, J. Appl. Phys. 75, 3870 (1994)]. However, usually only one dopant is used, as this is sufficient to achieve the insulation effect in most cases and is also easier to control in terms of process technology. Ideally, deep impurities exist near the center of the semiconductor bandgap at a concentration such that they can trap all free charge carriers, and thus the Fermi level is located or fixed at the energy position of the impurities, even with charge carrier injection. Ideally, the free charge carrier concentrations of electrons n and holes p correspond to the intrinsic charge carrier concentration nintrinsic. i That is, n = p = n i .

[0003] In many semiconductors, including those with large band gaps such as group III nitrides, although many dopants capable of constructing deep impurities and high-resistivity materials exist in principle, their locations are typically not near the center of the semiconductor band gap. Therefore, even with compensated deep impurities of sufficient N concentration (N>n,p), i.e., concentrations higher than the electron or hole concentrations n or p without a compensator, residual conductivity exists despite the Fermi level being fixed, and n or p is much higher than the intrinsic charge carrier concentration n. i .

[0004] Even if the charge carrier concentration is below 10 10 cm -3This is considered high even in some semiconductors, because, for example, very high breakdown field strengths exceeding 3 MV / cm can occur in semiconductors such as GaN, and low leakage current is significantly detrimental at high applied voltages or field strengths. Furthermore, this is because even performance achieved locally at high voltages with low leakage currents can lead to heating and ultimately thermal breakdown. Therefore, the residual conductivity or charge carrier concentration should be as low as possible, as close as possible to the intrinsic charge carrier concentration. In GaN, for example, carbon acceptors are about 0.9 eV higher than the valence band, while iron acceptors, which can be used alternatively, are about 0.6 eV lower than the conduction band. However, at room temperature, the band gap of GaN is about 3.4 eV, and the ideal location for deep impurities is about 1.7 eV. Therefore, for GaN:C, the free hole concentration is about 1 × 10⁻⁶. 6 cm -3 For GaN:Fe, the free electron concentration is approximately 3 × 10⁻⁶. 6 cm -3 Much higher than about 10 -9 cm -3 The intrinsic charge carrier concentration is low. As a result, the layer resistance using these dopants is more than 10 orders of magnitude lower than the theoretically achievable value with ideal dopants (which have energy positions near the center of the band gap). Because such dopants, for example in GaN, are currently unknown or not available in suitable forms for layer fabrication, the performance capabilities of these layers are limited, thus restricting the application areas of components made from these layers. Summary of the Invention

[0005] The current objective is to achieve improved layer insulation. This task is accomplished by means of semiconductor layer stacking according to the invention, component modules according to the invention, and methods according to the invention.

[0006] A semiconductor layer stack is proposed, characterized in that at least two layers (A, B) are used as monolayers, each layer having the following energy position of the Fermi level (103) in a semiconductor bandgap (104, 105): For layer (A) applicable Applicable to layer (B) E F It is the energy position of the Fermi level (103), E V It is the energy position of the valence band (102), E L It is the energy position of the conduction band (101), E L -E V It is the semiconductor band gap E GThe energy difference (104, 105) is such that the thicknesses (106, 107) of layers (A, B) are chosen such that a coherent space charge region (110) exists on layers (A, B). These layers can be separated directly from one another or through other layers.

[0007] According to the present invention, the problem of positioning the Fermi level near the center of the bandgap in a semiconductor is solved.

[0008] A coherent space charge region is achieved when consecutive segments with decreasing electric fields or different potential change processes are interconnected throughout the region. The width of this space charge region between two layers depends primarily on their doping levels (i.e., the number of ionizable acceptors and donors they contain) and the embedded potential difference. (This corresponds to the difference in Fermi levels within a single layer) and the dielectric constant. The value of . This width can be calculated. For N with shallow acceptor A And shallow benefactor N D The pn junction, in the p region W p Neutralizing region W n The estimation result of the space charge region width in is as follows:

[0009]

[0010] Where q is the unit charge (Elementarladung).

[0011] The calculations are slightly more complex for deep impurities and multiple dopants or other defects that result in residual conductivity. However, in principle, the width of the space charge region can be well estimated from the difference in Fermi level positions and the concentrations of ionized acceptors and donors in the space charge region (which can be estimated from their energy positions). The aim here is to achieve a value much larger than the thickness of a single layer so that the Fermi level positions between the initial dopants can be achieved on an average basis. Depending on the conductivity of the initial material, the dopants can be either acceptors or donors. Alternatively, in the case of an undoped n-conductive material, it is also possible to dope deep donors at a concentration higher than the electron concentration in the first layer, and dope acceptors and deep donors at an even higher concentration in the second layer.

[0012] Individually, each layer depletes charge carriers; however, the Fermi level is located close to the location of the dominant or donor level (which, in rare cases, is near the center of the band gap), thus retaining a relatively high residual conductivity for a single layer. However, by growing these layers alternately, a space charge region is created, which allows the Fermi level to be positioned between the various impurity levels, and thus enables a lower concentration of charge carriers throughout the layer.

[0013] In principle, the insertion of impurities and thus the position of the Fermi level can be controlled within a constrained framework by adjusting the manufacturing conditions. For directional doping, it is ideal to use dopants provided during growth from the initial material (intrinsic) or through a special source (extrinsic).

[0014] A particularly advantageous implementation involves inducing intrinsic or extrinsic doping at a concentration in at least two impurity energy levels, each having an energy position suitable for use in a single layer (A). And applicable in another layer (B) This allows the Fermi level to occupy the same energy position as the impurity level in such an infinitely thick monolayer, with a tolerance of ±50 meV, wherein the thickness of the layers in the semiconductor stack is chosen such that a coherent space charge region (110) is constructed over the entire stack.

[0015] Advantageously, not only are two layers grown sequentially, but they are also staggered, where other layers with different dopants or conductivities can be integrated into the stack as long as the space charge region extends across the entire stack. Typically, a space charge region also exists when only one deep impurity is used as a compensator within the building structure. However, the subsequent pinnen of the Fermi level energy position can be disadvantageous. Through the alternating layers and the resulting variation in Fermi level positions, potential fluctuations are also generated between the layers when observing a single layer, thus creating space charge regions and electric fields. This leads to a further reduction in the average charge carrier concentration, reaching close to n. i This value implies a higher layer resistance. This, in turn, allows for a reduction in the concentration of dopants used, which positively impacts the switching characteristics of the building blocks on top.

[0016] Therefore, effective Fermi energy positions can be achieved by doping. These effective Fermi energy positions, in addition to the Fermi energy positions in a single layer, also depend on the concentration of the dopant and the thickness of each layer. The obtained positions can be optimally set by the thickness parameter.

[0017] Preferably, the energy positions of the deep impurity energy levels generated by doping in the semiconductor layer stack are as follows: applicable in layer (A) Applicable in layer (B) This refers to the region where the average Fermi level is below or above the midpoint of the bandgap energy, where this midpoint totals 40% of the bandgap energy. In many semiconductors, the value of available impurities as sole dopants exceeds 50%, leading to unsatisfactory compensation, as is the case in GaN:Fe. If the resulting energy level is close to the center of the bandgap energy, i.e., within the aforementioned 40% range, satisfactory results in terms of layer insulation can generally be achieved even with this single dopant.

[0018] However, the method according to the invention can further improve this, and thus can be used meaningfully in individual cases as well. Background doping of the layer caused by inherent defects or unintentional impurities is also considered a dopant, and in this case, an intrinsic dopant. This allows for the alternation of intentionally doped layers in a single layer. However, due to the generally slightly fluctuating background doping of semiconductor layers, the use of intentionally doped layers is always preferred. Intrinsic dopant can be, for example, carbon or oxygen, and depending on the manufacturing method, they can be derived from the original material or the carrier gas. By appropriately selecting the growth parameters, carbon intercalation, particularly from alkyl groups of the metal-organic material used, can be reproducibly achieved by means of metal-organic vapor phase epitaxy (MOVPE) using these dopants.

[0019] Since defects (especially deviations in breakdown strength and layer resistance, in GaN as an example) also use lower limiting values ​​as limiting factors, it is also meaningful for the Fermi level caused by deep impurities to have a sufficient center position at a certain distance from the theoretically most favorable position.

[0020] In semiconductor layer stacking, the energy range in the space charge region (110) is sought. The average Fermi energy position E in F (108), that is, the average Fermi energy position is within the following ranges, respectively above or below the average bandgap energy: said ranges are up to 20% of the bandgap energy. For simplicity, the center of the bandgap is defined as n = p = n i The ideal Fermi level location for a semi-insulating semiconductor. Average location means that the Fermi level lies, on average, within this range over the thickness of the buffer according to the invention, i.e., also within ranges above or below this range, as may occur in alternating doping according to the invention. The Fermi level is ideally always located within this range by selecting sufficiently thin layers, stacked monolayers doped with dopant, to optimize the Fermi level. So thin that the energy should decrease to less than 1 / 4 of the possible valence band or conduction band energy variation in the layer pair due to the space charge region, preferably less than 1 / 4 for an infinitely extended layer pair.

[0021] In principle, simultaneously doping two dopants in a single layer offers advantages according to the invention; however, highly precise control of the corresponding concentrations is required to ideally locate the Fermi level. However, if a single, alternately doped layer is involved, the following effect is utilized: a space charge region with a slightly curved band is constructed between the two doped regions, and this slightly curved band depends very insensitively on the corresponding dopant concentration, provided the Fermi level in the monolayer is fixed at the impurity level induced by the dopant. In this case, the band-change process can also be calculated and reproducibly fabricated.

[0022] Due to the low carrier concentration during excess carrier compensation, the space charge region typically extends over a few micrometers (or even smaller if the different doped layers are chosen to be thin enough), resulting in an almost flat bandgap process with slight modulation of the corresponding band energy in each region, such as... Figure 1 As shown in d.

[0023] Here, the ideal layer thickness depends on the intrinsic background doping, the type of impurities (i.e., acceptors or donors), and the energy position. In individual cases, the optimal doping must be found, either simulated or experimentally. In principle, in a semiconductor with n-type background doping, most of the charge carriers have already been captured by acceptors located in the upper half of the bandgap. The charge carriers remaining in the conduction band are then captured by a second acceptor in the lower half of the bandgap. Thus, the resulting space charge region is so wide that the Fermi level is located near the lower acceptor almost throughout the entire layer stack, and therefore these layers must be very thin so that the Fermi level does not drop too sharply, and on average, the location of the Fermi level is close to the center of the bandgap, i.e., without inducing significant hole conduction. However, even without choosing the optimal thickness, as shown in Figure 6, improved layer insulation can be achieved because strong depletion of charge carriers can be locally achieved. Furthermore, even without a voltage applied to the layer stack, the position of the Fermi level already depends on how the adjacent layers are doped.

[0024] For example, if one chooses a lower level through a shallow donor (i.e., around 10) 16 -10 17 cm -3Within the range of the band gap, but with targeted n-doping, and this is compensated for by deep doping in the lower half of the band gap, it is advantageous to have a high concentration (far higher than the electron concentration) of deep doping, but this is done only in thin layers. Unlike the case where only a layer with deep doping is used, this results in the complete depletion of electrons in the semiconductor, which has only a low hole concentration passing through the thin layer with deep doping. These highly doped deep-doped thin layers must be placed so densely that the resulting space charge regions overlap. In this case, it is also necessary to dope deep doping at the beginning and end of the intentional insulating layer stack separately in order to obtain complete insulation and no conductive edge regions. Such a layer is advantageous for the switching characteristics of the component because, under varying applied voltages, reemission of charge carriers from deep impurities is minimized by the reduced number of deep impurities and the simultaneously decreasing residual conductivity.

[0025] The doping according to the invention in a semiconductor layer stack can be doped with acceptor-type or donor-type dopants, or with both acceptor-type and donor-type dopants. While it is possible to achieve doping with only acceptor-type or only donor-type dopants, and sometimes is meaningful, theoretically, a combination of donor and acceptor dopants is ideal because they better compensate for electron and hole injection, as are known as Fe and Ti compensators in InP, where they are very close to the center of the band gap in energy, and therefore do not require tuning of the Fermi level according to the invention by combining two dopants. If acceptors and donors are used, it is not easy to dop them in alternating layers due to the background doping of electrons or holes that is usually present, since deep donors do not trap or compensate for electrons, and deep donors do not trap or compensate for holes. In the case of, for example, n-type background doping, this can be solved by co-doping donors with shallow p-type dopants at a concentration higher than the electron concentration.

[0026] Alternatively, one of the deep donors in the upper half of the bandgap is used as a donor, and the other in the lower half as an acceptor. In GaN, the latter, for example, enables the deep donor to be a deep donor in the first layer (C) and without p-co-doping in the upper half of the bandgap in the second layer. If a space charge region is constructed in this region, slight p-type conduction induced by the acceptor is compensated by the donor in the second layer, despite having two energy positions far outside the central region of the energy bandgap. Ideally, a structure with a deep donor in layer A and a deep donor in layer B is desired. In this combination, it is preferable to use the acceptor in the lower half of the bandgap and the donor in the upper half.

[0027] The structure or semiconductor layer stack according to the invention consists of at least two layers, i.e., a sequence of at least two layer groups (Schichtpaket), which includes at least two layers (A, B) having the following Fermi levels: for said Fermi level, it is applicable as a single layer in layer (A). As a single layer, it is applicable in layer (B). Advantageously, these layers are grown in multiple alternations, i.e., ABABAB, ABABA or BABAB, and layer sequences such as ABCBA or ACBADB are also possible, where C and D are arbitrary intermediate layers, but their thickness and doping do not prevent the construction of a coherent space charge region (110).

[0028] According to the present invention, a component module is provided, which includes at least one component comprising a semiconductor layer stack according to the present invention.

[0029] In principle, the structure according to the invention may also have only shallow dopants, or a combination of shallow and deep donors and / or acceptors. That is, deep dopants and shallow donors, or vice versa, such as... Figure 5 The diagram shows only shallow donors and shallow acceptors. Here, in this case, multiple layers are required for insulation, since a pn junction, i.e., a diode structure, is generated in only one pair of layers A and B. In principle, a pn structure is also generated at the deep Fermi level according to the invention; however, these layers are then already high-resistivity, resulting in a significant reduction in the current flowing through them, and the diode characteristic curve is only very weakly significant.

[0030] If shallow dopants are used, very thin layers are usually required to achieve a coherent space charge region because the charge carrier concentration is typically higher than 10. 16 cm -3 For the same reason, multiple layers are needed to achieve a sufficiently high insulation effect on a sufficient layer thickness, since the breakdown field strength is limited by the material.

[0031] A similar approach with shallow dopant has been implemented in semiconductor laser structures. There, a laser diode with an inverted diode structure (i.e., an inverted layer sequence of p and n conductive layers compared to a laser diode structure) is overgrown by etching, creating an exposed pn junction. This results in a cutoff diode around the laser diode operating in the conduction direction during laser operation, restricting current flow to regions within the laser structure. However, this cannot be compared to embodiments according to the invention used for the layers mentioned herein, because, on the one hand, the 3D constructed sample is overgrown, and on the other hand, the layer thickness of the reverse-grown or later-operated pn structure is typically greater than the space charge region.

[0032] In this example, the target is not the center of the Fermi level, but the performance of the diode operating in the cutoff direction. By limiting the thickness of these diode layers that perform the cutoff, the voltage achievable up to the breakdown voltage is low, but in a laser structure, this is, in any case, limited by the current of the laser diode in the conduction direction through the active region.

[0033] An advantageous extension of the semiconductor layer stacking in a group III nitride material system, which is illustrated below by way of example, is that, in the group III nitride semiconductor, one of the following dopants is alternately doped in the first layer (A) and a second dopant is alternately doped in the second layer (B):

[0034] Iron and carbon, or

[0035] Carbon and donors, or

[0036] Iron and magnesium, or

[0037] Iron and zinc.

[0038] Here, combinations of these dopants are also possible in more than two layers or partially in one of multiple layers. In the case of doping with shallow donors (e.g., Si or Ge in GaN) or acceptors Mg in GaN, it is also advantageous in individual cases to continuously dope them at low concentrations and only dope the compensating impurities (i.e., deep acceptors or donors) in each layer segment, which is also covered by the configuration scheme according to the invention, since the variation of the position of the Fermi level in the hypothetical, infinitely extended monolayer is decisive for successful implementation, as is given in this case.

[0039] In group III nitrides, donors can be deep, such as those achieved with C, or shallow, such as those with Si, Ge, or O. However, when using shallow dopants, due to the small space charge region width, very thin layers or very low dopant concentrations are necessary to avoid creating layers with high residual conductivity; that is, to obtain regions that completely deplete charge carriers, such as... Figure 5 As shown, a thin monolayer of 10 nm is used. Achieving such a thin layer is more difficult due to flow effects and the delay in the addition of certain dopants (e.g., Mg) in modern growth processes (e.g., MOVPE).

[0040] For implementations of the invention, or rather, for methods of manufacturing semiconductor layer stacks, at least the following steps are advantageous:

[0041] Provide a substrate in an apparatus used for depositing semiconductors;

[0042] A sequence of at least two layers (A, B) is applied, each layer as a monolayer having the following energy positions of the Fermi level (103) in the semiconductor band gap (104, 105): For layer (A) applicable For layer (B) ;

[0043] Among them, layer D is selected in this way. A and D B The thicknesses (106, 107) of (A, B) result in a coherent space charge region (110) on layers (A, B). And applicable and Among them, W A and W B It is the space charge region, N A and N B It refers to the dopant concentration in layers A and B. q is the dielectric constant, and q is the unit charge. It is the embedded potential difference that is the same as the energy difference of the Fermi level.

[0044] The maximum thickness of the layer is chosen by numerical estimation of the width of the constructed space charge region. It is advantageous to keep it significantly (i.e., at least twice, preferably five times, ideally more than ten times) lower than the calculated layer thickness, because that sets a low band modulation, i.e. a very uniform band change process.

[0045] The following description of some embodiments and figures is shown in the gallium nitride examples that have been introduced. Detailed Implementation

[0046] Gallium nitride (GaN) is an important semiconductor in a variety of applications today, such as LEDs for general lighting and power electronics. Commercially, it is often used to fabricate components as a thin-film material via MOVPE.

[0047] Due to inherent defects and impurities, GaN typically exhibits slightly n-type conduction and is rarely high impedance, primarily due to process-related carbon impurities. Early operations used acceptor zinc or magnesium, which have activation energies greater than 150 meV to achieve high resistance, yet with moderate cutoff performance, thanks to the higher resistances (above 10) produced by these dopants. 10 cm -3The relatively high hole concentration. Iron is a commonly used compensator in III-V semiconductors and is now partially used in GaN. However, energy positions below the conduction band of 0.6 eV result in relatively high residual electron conductivity, which is unfavorable in most electronic components. This is especially true since unipolar components in GaN systems are typically electronically conductive, and therefore this layer only moderately and well blocks electron injection into the insulating layer. Carbon is used instead. Intrinsic doping results in a relatively high hole concentration in the lower half of the GaN bandgap (approximately E). v The deep host in the band gap (+0.9 eV) and the donor in the upper half of the band gap. Under standard growth conditions, in the case of a doped precursor (e.g., propane) or other precursors containing hydrocarbons or carbon (e.g., CBr4), carbon mainly acts as the deep host intercalation.

[0048] It is not surprising that no favorable properties have been reported regarding embedded carbon acting as both a deep acceptor and a deep donor, as it is currently impossible to set an appropriate acceptor / donor ratio to ensure the Fermi level is as close as possible to the center of the band gap due to a lack of knowledge. However, in principle, such a single-doped (Einzeldotand) configuration producing these two energy levels is suitable if the ratio of the two levels can be set through growth conditions. (See Figure 2 and...) Figure 3 , 4 Figure 5 shows the energy position of the band relative to the Fermi level, which is defined here as having a value of 0 eV and is indicated by a dashed line.

[0049] In the corresponding graph, the upper continuous curve represents the conduction band, and the lower one represents the price band.

[0050] Figure 2a (a) shows the band structure when Fe is doped, and (b) shows the band structure when C is doped. In both cases, the Fermi level is relatively close to the conduction band or valence band. Therefore, the remaining charge carrier concentration remains relatively high.

[0051] To obtain a high-resistivity monolayer, it is possible, for example, to co-dopate with Fe and C, where the dopant ratio must take into account the energy positions. For Fe, the E... L -0.6eV and E VAt an energy position of +0.9 eV, a concentration ratio of approximately 5000 is obtained. This must be satisfied as precisely as possible, but this depends on the exact, but often only inaccurately known, energy positions within the bandgap. Therefore, it is much easier to dope in alternating layers, as this allows for more pronounced variations in dopant concentration and exact energy positions, or they do not need to be precisely known. Thus, in n-type semiconductors, the Fe-doped layer is thicker than the C-doped layer because the acceptors only trap electrons, and the Fermi level is approximately located at the energy position of C in the multilayer stack due to the excessively thick GaN:C layer. Depending on the application, simulations show that the GaN:Fe layer thickness is 5 to 20 times that of the GaN:C layer. This is in Figure 3 This is illustrated exemplarily. In this example, the Fermi level is further away from the valence band by 100-200 meV compared to using a single dopant, which further reduces the hole concentration and thus increases the resistance.

[0052] In another embodiment—a combination of shallow donor Si and deeply rooted donor C—the Si concentration, present either continuously or in a single layer, is, for example, 1 × 10⁻⁶. 17 cm -3 In the case of, for example, a C concentration of 2 × 10⁻⁶ 18 cm -3 The doping is only done in a thin layer, which is thicker than is needed for pure calculations, in order to trap all the free electrons generated through Si, i.e., occupying more than 1 / 10 of the volume in this example. This is in Figure 4 The diagram shows a short stacking sequence. If the C-doped layer is too thick, the Fermi level approaches the C impurity, and slight p-conductivity appears. If similarly implemented using Mg and Fe, a sufficient n-type background must be present, whether intrinsic or, for example, through successive doping, to successfully combine them, similar to the combination of C and Fe. Only excessively high Mg concentrations or excessively thick GaN:Mg layers lead to significant p-type conductivity.

[0053] In this case, a combination or alternation of deep donors with Mg in the upper half of the bandgap is more suitable. Ti or deep C donors can also be used, provided they can be selectively embedded. In principle, slight n-doping and p-doping can also be achieved by alternating shallow impurities, allowing complete depletion according to the invention to be achieved through these layers. Figure 5 The image shows thin, alternating layers, where the Fermi level is located almost perfectly at the center of the bandgap.

[0054] Because Si and Mg are typically present in, for example, group III nitride layer deposition equipment, and therefore at most one deep dopant must be retained instead of two, combinations of deep and shallow impurities, or only shallow impurities, are of interest. However, processes are more readily handled with dopants whose energy levels are so deep that they, in principle, deplete the material, even when used with a carrier concentration of less than 10. 14 cm -3 Therefore, when the carrier concentration is less than 10... 14 cm -3 At this point, a space charge region width greater than 1 μm is given. If dopants such as C or Fe are used at concentrations higher than the electron concentration in the undoped case, the carrier concentration in GaN is typically below 10⁻⁶. 10 cm -3 This will result in a space charge region greater than 100 μm. Therefore, band bending between the alternating doped layers is only slightly noticeable and mainly bends toward (if present) the stronger conductive layers above and below them.

[0055] exist Figure 1 Figure a schematically illustrates an energy position diagram, showing how to begin with an alternating layer structure of layers A and B with different doping (which are not yet electrically connected), aligned along the electrochemical potential, or Fermi level. Here, 101 is the conduction band E. C 102 is the price band E L 103 is the Fermi level E Dopant,Fermi ,exist Figure 1 The Fermi level in layer a should be at the same energy position as the dopant. However, if the dopant concentration is so high that the Fermi level is fixed at the dopant's energy position, this is not necessarily required for the embodiments according to the invention. That is, compensation for the partial presence of background charge carriers and the position of the Fermi level (which is not at the same energy position as the dopant) can lead to the desired result, but is difficult to achieve reproducibly. The thicknesses of layers A and B are denoted by 10⁶ and 10⁷, and the corresponding band gap energies by 10⁴ and 10⁵. They are homoepitaxially identical, but in principle, they can also be different. If these layers A and B are electrically contacted, a potential difference is generated, which is compensated by charge shift and the resulting band bending.

[0056] exist Figure 1Figure b illustrates the case of a very thick monolayer, so thick that the resulting space charge region (109) is thinner than half the thickness of the layer pair. In this case, separate space charge regions are created between the different doped regions, denoted as 109 for the first three exemplary cases as shaded areas. However, the space charge regions do not extend into the adjacent transition regions, but rather beyond the space charge regions, where bands 101 and 102 are flat, and the Fermi level 108 is fixed at the impurity level. Only within the space charge regions are their energies between the two impurity levels. If the layer thickness is reduced, a region is created that is completely traversed by the space charge region or a coherent space charge region (110), i.e., an electric field modulated through the individual layers exists throughout the structure, such as Figure 1 c and Figure 1 As shown in d, for thinner layers, the band change process remains flat, where the Fermi level 108 is located at an energy position between the impurity levels. Here, for clarity, the reduction in layer thickness is not graphically represented, or in relation to... Figure 1 b. Different scales show the horizontal axis of the spacing. Because the charges that move by the potential difference or the charges that are fixed at the position of ionized impurities are still present at a low concentration, the band bend extends over a large area and results in an average almost flat band change process. In actual structures, additional band bends are obtained from the surface or doped layer at the edge of the alternating layer structure, which are not taken into account here; the band bends mostly cover the entire structure.

[0057] Figure 6 exemplarily illustrates a comparison of the current-voltage characteristic curves (a) and the calculated resistance (b) between the metal junction and the conductive substrate of a GaN-based field-effect transistor structure on silicon (111) with different doping levels. Also, from the position of the dopant (Fe≈E) L -0.6eV; C≈E CAs expected (+0.9 eV), the sample with the Fe-doped GaN buffer layer exhibits the highest current, followed by the C-doped buffer layer. If this doping is performed now, and a sample with a buffer layer of about 3 μm thickness (consisting of alternating 200 nm thick Fe and C-doped GaN layers) is fabricated, the current is lower, although either the Fe-doped or C-doped layer alone has a higher current, i.e., the resistivity of the monolayer is low. However, in the combination of alternating Fe and C-doped monolayers according to the invention, it has the lowest current and the highest resistance. In this structure on silicon, an aluminum-containing intermediate layer is introduced into the GaN to avoid thermally confined cracking. Here, it is advantageous to dope the GaN with Fe after the Al-containing intermediate layer and with C before it due to the charge generated at the boundary surface, since, from the growth direction, there is a potential accumulation of holes above the intermediate layer and an accumulation of electrons below, which can be ideally compensated for.

[0058] Applications according to the invention can be examined most simply, for example by means of mass spectrometry methods such as secondary ion mass spectrometry (SIMS), or based on defect luminescence in high-resolution methods, such as cathodoluminescence in scanning electron microscopy or scanning transmission electron microscopy. Sometimes, high position-resolution Raman spectroscopy can also be used to identify dopants and their embedding sites.

[0059] This invention relates to all semiconductors and dopants. Particularly for some semiconductors with smaller band gaps, the charge carrier concentration values ​​sought through compensators are higher than those described for GaN (E0). G (≈3.4 eV), because these semiconductors have high charge carrier concentrations and intrinsic conductivity. Like other values, it must also be matched to the bandgap energy and density of states, and the resulting intrinsic charge carrier concentration, in order to achieve the effects according to the invention. In principle, more than two dopants and dopants in more than two alternating layers are also possible and are part of embodiments according to the invention. Instead of dopants in the layers, intrinsically conductive n or p-type materials typically present in the layers can be used. ni. A doped layer can also overlap with another layer; that is, doping does not necessarily have to terminate abruptly at the nominal end of the layer, and overlap is possible, which can also be advantageous depending on the type and energy position of the dopant. The order of layers A and B mentioned in the configuration according to the invention is interchangeable, and there may be other layers in between, provided that the condition of a coherent space charge region is satisfied. The description of the location of the Fermi level always involves the case where no external voltage is applied to the layer.

Claims

1. A semiconductor layer stack, characterized in that, A sequence of at least two layer groups, said layer group comprising at least one layer A and at least one layer B, said at least two layers having, as monolayers, Fermi levels (103) in the semiconductor band gaps (104, 105) at the following energy positions: applicable to said layer A And applicable to layer B , of which E F It is the energy position of the Fermi level (103), E V It is the energy position of the valence band (102), E L It is the energy position of the conduction band (101), E L -E V The semiconductor bandgap E G The energy difference (104, 105) is such that the thicknesses (106, 107) of the two layers A and B are chosen such that a coherent space charge region (110) is generated on the two layers A and B, wherein, in the case of undoped n-type conductive material, layer A has deep donors, the first concentration of which is higher than the electron concentration, and layer B has deep donors, wherein the concentration of donors in layer B is greater than the concentration of acceptors in layer A.

2. The semiconductor layer stack according to claim 1, characterized in that, Intrinsic or intrinsic doping is achieved in at least two impurity levels at such concentrations that the Fermi level at an infinitely thick monolayer occupies the same energy position as the impurity level, with a tolerance of ±50 meV, wherein the thickness of the layers in the semiconductor stack is chosen such that a coherent space charge region (110) is constructed over the entire stack, wherein the at least two impurity levels each have the following energy positions: for which the energy positions are applicable in layer A And layer A is an infinitely thick single layer, while layer B is applicable Furthermore, layer B is an infinitely thick single layer.

3. The semiconductor layer stack according to claim 2, characterized in that... The energy positions of the deep impurity energy levels generated by the doping are as follows: For these energy positions, the following applies in layer A. The energy location applies in layer B. .

4. The semiconductor layer stack according to claim 1 or 2, characterized in that... The average Fermi energy position E in the following energy range within the space charge region (110) F (108): .

5. The semiconductor layer stack according to claim 1 or 2, characterized in that, Doping with acceptor-type or donor-type dopants.

6. The semiconductor layer stack according to claim 1 or 2, characterized in that, In a group III nitride semiconductor, one of the following dopants is alternately doped in layer A and a second dopant is alternately doped in layer B: Iron and carbon, or Carbon and donors, or Iron and magnesium, or Iron and zinc.

7. The semiconductor layer stack according to claim 2, characterized in that, The energy position of the deep impurity level generated by the doping is the average Fermi level in the range below or above the average bandgap energy, and this applies to the deep impurity level in layer A. And the deep impurity energy level is applicable in layer B. The average range is 40% of the bandgap energy or the range is up to 20% of the bandgap energy.

8. The semiconductor layer stack according to claim 1 or 2, characterized in that, Layer A and layer B are implemented using a group III nitride material system.

9. The semiconductor layer stack according to claim 1 or 2, characterized in that, Layer A and layer B are implemented using GaN.

10. A component module comprising at least one component, the component comprising a semiconductor layer stack according to any one of claims 1 to 9.

11. A method for manufacturing a semiconductor layer stack according to any one of claims 1 to 9, the method comprising at least the following steps: Provide a substrate in an apparatus used for depositing semiconductors; A sequence of at least two layer groups is applied, the layer groups comprising at least one layer A and at least one layer B, the at least one layer A and the at least one layer B having, as monolayers, Fermi levels (103) in the semiconductor band gap (104, 105) at the following energy positions: applicable to layer A And applicable to layer B ,in, Thus, the thicknesses of layers A and B are chosen to be (10⁶, 10⁷)D. A and D B This results in a coherent space charge region (110) on layers A and B. And applicable and Among them, W A and W B It is the space charge region, N A and N B It refers to the dopant concentration in layers A and B. q is the dielectric constant, and q is the unit charge. It is the embedded potential difference that is the same as the energy difference of the Fermi level, wherein, in the case of an undoped n-type conductive material, layer A has deep donors, the first concentration of which is higher than the electron concentration, and layer B has deep donors, wherein the concentration of donors in layer B is greater than the concentration of acceptors in layer A.

Images