Semiconductor device having an internal-field-guarded active region

Abstract

A semiconductor device comprises a layer sequence formed by a plurality of polar single crystalline semiconductor material layers that each has a crystal axis pointing in a direction of crystalline polarity and a stacking direction of the layer sequence. A core layer sequence is formed by an active region made of an active layer stack or a plurality of repetitions of the active layer stack. The active layer stack has an active layer having a first material composition associated with a first band gap energy, and carrier-confinement layers embedding the active layer on at least two opposite sides thereof, having a second material composition associated with a second band gap energy larger than the first band gap energy. A pair of polarization guard layers is arranged adjacent to the active region and embedding the active region on opposite sides thereof.

Claims

1. A semiconductor device, comprising a layer sequence formed by a plurality of polar single crystalline semiconductor material layers that each have a crystal axis pointing in a direction that coincides with a direction of crystalline polarity and with a stacking direction of the layer sequence; wherein the layer sequence is formed by a core layer sequence and shell layer sequences on opposite sides of the core layer sequence in the stacking direction; and wherein the core layer sequence is formed by an active region made of an active layer stack or a plurality of repetitions of the active layer stack, the active layer stack being formed by an active layer having a first material composition that is associated with a first band gap energy, and by carrier-confinement layers embedding the active layer on at least two opposite sides thereof and having a second material composition that is associated with a second band gap energy larger than the first band gap energy, wherein the active layer and the carrier-confinement layers are configured to effect a quantum-confinement of charge carriers in the active layer in one, two or three spatial dimensions; and a pair of polarization guard layers adjacent to the active region and embedding the active region on opposite sides thereof, wherein at least one of the polarization guard layers is formed by a semiconductor material layer having a third material composition that differs from the first and second material compositions and that is associated with a third band gap energy larger than the first band gap energy, but smaller than the second band gap energy.

2. The semiconductor device of claim 1, wherein both polarization guard layers have the third material composition.

3. The semiconductor device of claim 1, wherein only one of the polarization guard layers, herein called the first polarization guard layer, has the third material composition, and wherein the other of the polarization guard layers, herein called the second polarization guard layer, has the first material composition.

4. The semiconductor device of claim 3, wherein the active layer and the first polarization or the active layer and the second polarization guard layer have an identical thickness.

5. The semiconductor device of claim 1, wherein the first, second and third material compositions are selected so as to provide a band-gap energy of the at least one polarization guard layer that is larger than a transition energy associated with optical transitions of the quantum confined charge carriers in the active layer in operation of the semiconductor device.

6. The semiconductor device of claim 1, wherein the second material composition and a thickness of the carrier-confinement layers are selected to allow a tunneling transport of the charge carriers between the polarization guard layers and the carrier-confinement layers under application of the operating voltage.

7. The semiconductor device of claim 1, wherein the active layer has a thickness of less than 25 nanometer.

8. The semiconductor device of claim 1, wherein a thickness of the polarization guard layers is at least one monolayer.

9. The semiconductor device of claim 1, wherein a thickness of the polarization guard layer on at least one of the opposite sides of the active region is either smaller or at most equal to a thickness of the active layer.

10. The semiconductor device of claim 1, wherein the shell layer sequences each comprise a respective outer layer, each outer layer forming one of two opposite end faces of the layer sequence, which end faces interface with a dielectric or a contact material of metallic electrical conductivity, and wherein each of the outer layers has the first material composition.

11. The semiconductor device of claim 8, wherein the shell layer sequence comprises a first outer layer forming a top cover layer or second outer layer forming a bottom carrier layer of the layer sequence, the first or second outer layer having a thickness of at least 20 nanometer and having the first material composition.

12. The semiconductor device of claim 1, wherein the layer sequence forms a sequence of epitaxial layers deposited on a carrier layer.

13. The semiconductor device of claim 1, wherein the first material composition is GaN, the second material composition is AlN, and the third material composition is Al.sub.xGa.sub.1-xN, 1>x>0.

14. The semiconductor device of claim 13, wherein the third material composition is Al.sub.xGa.sub.1-xN, with x smaller than or equal to 0.5.

15. The semiconductor device of claim 1, wherein the layer sequence forms a diode and the active region comprises a layer stack configured to emit light under application of an operating voltage to the layer sequence, which operating voltage is suitable for allowing an electric current across the diode.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Further embodiments will be described in the following with respect to the attached figures. In the figures:

(2) FIG. 1 is a schematic cross-sectional view of an embodiment of a quantum-dot semiconductor device forming a light emitter;

(3) FIG. 2 is an enlarged view of a section of the quantum-dot semiconductor device of FIG. 1;

(4) FIG. 3 is a schematic cross-sectional view of an embodiment of a quantum-well semiconductor device;

(5) FIG. 4 is an enlarged view of a section of the quantum-well semiconductor device of FIG. 3;

(6) FIG. 5 is a schematic cross-sectional view of another embodiment of a quantum-dot semiconductor device;

(7) FIG. 6 is a schematic cross-sectional view of a section of another embodiment of a semiconductor device;

(8) FIG. 7 is a schematic top view of the semiconductor device of FIG. 6;

(9) FIG. 8 is a schematic cross-sectional view of another embodiment of a quantum dot semiconductor device forming a photonic crystal;

(10) FIG. 9 is a schematic top view of the embodiment of FIG. 8;

(11) FIG. 10 is a band-structure profile along a stacking direction of a layer sequence of a semiconductor device according to the prior art, as determined by 8-band kp-based band structure calculation;

(12) FIG. 11 is a band-structure profile along a stacking direction of a layer sequence of an embodiment of a semiconductor device according to an embodiment of the present invention, as determined by 8-band kp-based band structure calculation;

(13) FIG. 12 is a schematic three-dimensional view of a further embodiment of a semiconductor device according to the present invention;

(14) FIG. 13 is a schematic cross-sectional view of a layer sequence of the semiconductor device of FIG. 12; and

(15) FIG. 14 shows a plot of emission energies from different semiconductor devices having a structure corresponding to that of the semiconductor device underlying the band-structure calculation of FIG. 11.

DETAILED DESCRIPTION

(16) FIG. 1 is a schematic cross-sectional view of an embodiment of a quantum-dot semiconductor device 100 forming a light emitter. Reference is made in parallel to FIG. 2, which shows an enlarged view of a section S outlined by a dashed line in FIG. 1.

(17) The quantum dot semiconductor device 100 comprises a layer sequence 102 formed by a number of polar semiconductor material layers, which will be described in the following. The semiconductor material layers are group-III nitride semiconductors and each have a hexagonal crystal structure. The layer sequence 102 consists of semiconductor material layers that each has a crystal axis pointing in a direction c indicated by a vertical arrow in FIG. 1. The c direction also forms the stacking direction of the layer sequence 102. Typically, the layer sequence 102 is fabricated by epitaxial growth, the growth direction of the layer sequence 102 thus being the c direction.

(18) Specifically, the layer sequence 102 consists of the following layers: an active region 104, which is formed of an active layer stack that consists of an active layer 104.1 and carrier-confinement layers 104.2 embedding the active layer 104.1. In the present case, the active layer 104.1 is made of quantum dots made of gallium nitride GaN, which thus forms the first semiconductor material. In the present embodiment, the quantum dots 104.1 are fully embedded by the carrier-confinement layers 104.2, which are made of aluminum nitride AlN. In a variant of the present embodiment, which is not shown in the figures, the active layer additionally comprises a wetting layer of one or two monolayers of gallium nitride. The quantum dots of the active layer can be fabricated in a self-organized manner in an epitaxial growth process using a Stranski-Krastanov growth mode on an AlN growth surface. Due to a small extension of the quantum dots 104.1 in the three spatial dimensions a three-dimensional confinement of charge carriers is created, as is per se well-known in the art. The active layer stack 104 is embedded by a pair of polarization guard layers 106.1 and 106.2. At least one of the polarization guard layers 106.1 and 106.2 is made of a third material composition that differs from the first material of the quantum dots of the active layer 104.1 and from the second material of the carrier-confinement layers 104.2. The third material composition is associated with a third band gap energy that is larger than the first band gap energy of the quantum dot material, i.e., GaN, but smaller than the second band gap energy of the material of the carrier-confinement layers 104.2, i.e. AlN. The polarization guard layers 106.1 and 106.2 are made of aluminum gallium nitride in the present example. The active region 104 and the polarization guard layers 106.1 and 106.2 together form a core region of the semiconductor device 100.

(19) A shell region 108 of the semiconductor device 100 is made of shell layer sequences 108.1 and 108.2. The shell layer sequences may each comprise one or more layers. In the present embodiment, a lower shell layer sequence is formed by a substrate layer 108.1 and an upper shell layer sequence is formed by a cover layer 108.2. The substrate layer 108.1 and the cover layer 108.2 are made of the same material. In the present example, these two layers forming the outer region of that layer sequence 102 are made of aluminum gallium nitride, which is also the material of the first and second polarization guard layers 106.1 and 106.2. The substrate layer 108.1 and the cover layer 108.2 form outer layers and hence constitute the opposite end faces 110 and 112 of the layer sequence. The end faces form an interface of the layer sequence with a contact material of an electrical contact structure, which is not shown in FIG. 1.

(20) In a variant of this embodiment, only one of the polarization guard layers is made of aluminum gallium nitride, and the other polarization guard layer has the first material composition of the quantum dots in the active layer. It is particularly suitable to choose the third material composition, that is, in the present case, AlGaN for that polarization guard layer, which forms a part of the optical pathway of photons between the active region 104 and an end face of the semiconductor device 100, from which photons are to be emitted from the semiconductor device in light-emitter embodiments, or at which photons are to enter the semiconductor device in light-detector embodiments.

(21) In another variant, wherein the material composition of both polarization guard layers 106.1 and 106.2 differs from that of the active layer 104, the polarization guard layers 106.2 and 106.2 have mutually different material compositions. For the purpose of definition, both of these material compositions form a third material composition in that they differ from the first and second material compositions, but are associated with a band-gap energy that is in the range of energies larger than a transition energy associated with optical transitions of the quantum confined charge carriers in the active layer and larger than the band-gap energy of the carrier-confinement layers.

(22) A suitable third material composition of the polarization guard layers 106.1 and/or 106.2 is Al.sub.xGa.sub.1-xN, with x, the aluminum fraction being smaller than or equal to 0.5. In one variant, the aluminum fraction x is smaller than 0.01. In another variant, the aluminum fraction x is between 0.01 and 0.03, inclusively. In another variant, the aluminum fraction x is between 0.03 and 0.06, inclusively. In another variant, the aluminum fraction is between 0.06 and 0.09, inclusively. In another variant, the aluminum fraction x is between 0.09 and 0.12, inclusively. In another variant, the aluminum fraction is between 0.12 and 0.15, inclusively. In another variant, the aluminum fraction is between 0.15 and 0.20, inclusively. In another variant, the aluminum fraction is between 0.20 and 0.30, inclusively. In another variant, the aluminum fraction x is between 0.30 and 0.50, inclusively.

(23) The polarization guard layers 106.1 and 106.2 have different aluminum fractions in some variants.

(24) In another variant of the present embodiment (not shown), there is an intermediate layer sequence between the bottom layer 108.1 and the polarization guard layer 106.1. The intermediate layer sequence is formed by a buffer layer sequence comprising AlGaN layers of alternating material compositions, such as a superlattice or as a layer sequence of stepwise or gradient composition changes (not shown in FIG. 1). Such intermediate layer sequences serve for achieving a particularly low defect density in the active region 104. In this variant, the intermediate layer sequence is embedded between two shell polarization guard layers made of GaN (the material of the active layer) on each side, wherein the GaN substrate layer 108.1 thus forms one of the two shell polarization guard layers, and the other of the two shell polarization guard layers is not shown in FIG. 1. The shell polarization guard layers are arranged adjacent to the intermediate layer sequence on both sides of it. This way, the active region 104 is shielded from polarization fields due to the polar hetero-interfaces in the intermediate layer sequence, which thus cannot influence the band profile in the active region 104 along the stacking direction.

(25) In operation of the semiconductor device 100, an operation voltage is applied across the layer sequence, which creates a current across the layer sequence in vertical direction, which is the direction c or a direction pointing opposite to that direction c. Due to the presence of the polarization guard layers 106.1 and 106.2, a tunneling transport of charge carriers is achieved, and carriers are localized in the quantum dots 104.1, where light is generated through an optical interband or intraband transition occurring after capture of charge carriers injection into the active layer 104.1 via a tunneling process across a tunnel barrier formed by the carrier-confinement layer 104.2. Further details of the effects of the particular layer sequence will be described further below with reference to FIGS. 10 and 11.

(26) An optical interband transition involves in some embodiments an emission of one photon corresponding at least approximately to the band gap energy. In other embodiments, a two-photon emission process is used, involving an emission of two photons of lower energy to bridge the band gap energy. In an optical intraband transition, a transition between different states within the same energy band involves the emission of a photon of infrared energy. Intraband light emitters, for instance from an active region made of nitride semiconductor materials, can be used in communication technology because the emission energies are in a range suitable for optical transmission media, such as glass fibers.

(27) The structure of the semiconductor device of FIGS. 1 and 2 overcomes challenges faced by prior-art solutions and linked via the Quantum-Confined Stark Effect (QCSE). Specifically, in layer sequences without polarization guard layers, the QCSE spatially separates an electron wave function from a hole wave function in the active layer, due to strong built-in electric fields originating from the (spontaneous) pyro- and piezoelectricity of the polar semiconductor materials. The residual electron-hole overlap does not only limit the luminescence intensity, but also introduces large excitonic dipole moments in the QDs' growth direction. Hence, these excitonic dipole moments cause a large coupling to external, e.g. defect-induced, electric fields evoking spectral diffusion that originates highly undesirable emission line broadening in such prior-art devices. In addition, large effective hole masses in GaN in combination with the specific shape of GaN QDs resembling an upright, truncated, hexagonal pyramid, originate the existence of hybrid-biexciton states. While a standard, c-plane GaN QD of common size (e.g., height in c direction 2 nm, width 10 nm) provides a strong three-dimensional confinement for the electron situated at the top of the QD, the hole is most frequently only weakly confined in lateral directions. In this sense, common GaN QDs only achieve a full 3D-confinement of an exciton through the fully confined electron, which localizes the accompanying hole via the Coulomb interaction. As a direct consequence, a subtle balance of interactions in e.g. a biexcitonic complex is heavily altered to such an extent that a lateral separation of the hole wave functions occurs and enables parallel hole spin orientations. Hence, the total spin projection t (orbital angular momentum+spin) of a biexciton does not amount to zero anymore, but can reach up to t=3. Consequently, a biexcitonic decay in most GaN QDs evolves via the long-lived dark exciton states (t=2), enabling a large variety of non-radiative decay processes before the final decay of the bright excitons (t=1) occurs subsequent to a phonon-mediated spin-flip process.

(28) In contrast, the inclusion of polarization guard layers 106.1 and 106.2 enhances the overall luminescence output, as a partial shielding of the polarization fields can be obtained precisely at the position of the active layer 104.1. The key idea of this concept comprises the depicted guard layers 106.1 and 106.2 that consist of exactly the same material as the active layer 104.1. As a result of this particular layer sequence, the band structure reaches a highly desirable flat-band condition. The use of a third material composition for at least one of the polarization guard layers 106.2 and/or 106.2 avoids the potential draw-back of absorption of a certain fraction of the overall emission in growth direction by the polarization guard layers. In-plane light propagation, as required for light-guiding purposes and resonators, is of course not directly limited by the polarization guard layers.

(29) Next, reference will be made to FIGS. 3 and 4 and parallel. FIG. 3 is a schematic cross sectional view of an embodiment of a quantum wells semiconductor device. FIG. 4 is an enlarged view of a section of the quantum well semiconductor device of FIG. 3. The semiconductor device 300 of FIG. 3 resembles the semiconductor device 100 of FIG. 1 in many aspects. For this reason, reference labels are used for the structural elements of the semiconductor device 300, which differ from those of corresponding structural elements the embodiment of FIGS. 1 and 2 only in their first digit, which is 3 instead of 1. The following description will first turn to those features of the semiconductor device 300 distinguishing it from the semiconductor device 100.

(30) In particular, as can also be seen in the enlarged view of FIG. 4, the semiconductor device 300 comprises an active layer stack 304, which comprises a quantum well formed by an active layer 304.1 and carrier confinement layers 304.2 embedding the active layer 304.1 in the stacking direction c. Otherwise, the layer sequence is identical in comparison with that of the semiconductor device 100. In the active region 304 of the semiconductor device 300, the quantum well formed by the active layer 304.1 and the carrier confinement layers 304.2 and 304.3 achieves a confinement of charge carriers or excitons in one dimension, and thus allows motion of the charge carriers or excitons in the remaining two spatial dimensions.

(31) As for the embodiment of FIGS. 1 and 2, the active layer stack 304 is embedded by a pair of polarization guard layers 306.1 and 306.2. The polarization guard layers are made of a material differing from that of the active layer 304.1, which is the quantum well layer. In the present example, they are made of gallium nitride AlGaN with an aluminum fraction among the group-III elements of at most 0.5 (50%), while the active layer 304.1 is made of GaN, and the carrier confinement layers 304.2 are made of AlN. A shell region 308 of the semiconductor device 300 is made of a substrate layer 308.1 and of a cover layer 308.2. The substrate layer 308.1 and the cover layer 308.2 are made of the same material as the active layer 304.1. In the present example, these two layers forming the outer region of that layer sequence 302 are thus made of GaN. The two outer layers 308.1 and 308.2 form opposite end faces 310 and 312 of the layer sequence. The end faces form an interface of the layer sequence with a contact material of an electrical contact structure, which is not shown in FIG. 3.

(32) Both semiconductor devices 100 and 300 can be used as edge light emitters using emission from the edges 112, 114 and 312, 314, respectively. However, in view of the third material composition of at least one of the polarization guard layers surface emission can be used without having to accept a strong loss of intensity of light emission. This can be further improved by choosing a suitable material composition and thickness of the layers of the shell layer sequence 308.1, or 308.2.

(33) Next, reference will be made to FIGS. 5 to 7 in parallel. FIG. 5 is a schematic cross sectional view of another embodiment of a quantum dot semiconductor device. FIG. 6 is an enlarged view of a section of the quantum dot semiconductor device of FIG. 5. FIG. 7 is a schematic view of a section of a lowermost active layer in the active region of the semiconductor device of FIG. 5, as seen from above when neglecting any layer grown on top of the subject active layer.

(34) The semiconductor device 500 of FIGS. 5 and 6 resembles the semiconductor devices 100 of FIGS. 1 and 2 and 300 of FIGS. 3 and 4 in many aspects. For this reason, reference labels are used for the structural elements of the semiconductor device 500, which differ from those of corresponding structural element of the embodiment of FIGS. 1 to 4 only in the first digit, which is 5 instead of 1 or 3. The following description will first turn to those features of the semiconductor device 500, which distinguish it from the semiconductor devices 100 and 300.

(35) In the present example, the quantum dots 504.1 extend from an initial wetting layer of at most one mono layer thickness, which is typical for a Stranski-Krastanov growth mechanism. In order to facilitate the positioning of single quantum dots 504.1 by a self-assembled quantum dot growth based on the Stranski-Krastanov growth mechanism on the thin AlN carrier-confinement layer 504.2a, a positioning technique via a strain aperture is employed. A strain aperture is formed by an AlN micro-ring 514 fabricated in a surface region of the adjacent AlGaN polarization guard layer 506.1 by standard etching techniques. The GaN material inside the AlN ring 514 is laterally strained and for that reason forms an energetically favorable strain aperture for quantum-dot growth on the carrier-confinement layer. This micro-ring technique is scalable, allowing the positioning of more than one quantum dot 504.1 within one and the same micro-ring 514.

(36) Furthermore, the active region 504 contains a plurality of active layer stacks, each formed by two carrier confinement layers 504.2a, 504.2b embedding a quantum dot layer forming the active layer 504.1. In the sequence of active layer stacks within the active region 504, an upper carrier-confinement layer 504.2b of a given active layer stack at the same time forms a lower carrier-confinement layer 504.2a of a respective next active layer stack in the stacking direction c. The number of repetitions of the active layer stack within the active region 504 can be selected according to the requirements of a given application case. In the example of FIG. 5, five active layer stacks with active layers labeled D, E, F, G, and H are shown in the active region, wherein the region between the active layers F and G includes additional repetitions of the active layer stack, which are not shown, but indicated by three dots in vertical sequence. As shown in FIG. 7, the micro rings can be fabricated and positioned in a regular array.

(37) Next, reference will be made to FIGS. 8 and 9 in parallel. FIG. 8 is a schematic cross-sectional view of another embodiment of a quantum dot semiconductor device forming a photonic crystal. FIG. 9 is a schematic top view of the embodiment of FIG. 8. The embodiment of the semiconductor device 800 shown in FIG. 8 is based on the layer sequence of the semiconductor device 500 of FIGS. 5 to 7. In addition, the semiconductor device 800 has a regular array of holes 818 fabricated to extend through in the layer sequence between its end faces 810 and 812, i.e., from the bottom of the carrier layer 808.1 to the top of the cover layer 808.2. This way a photonic crystal is realized. The photonic crystal can be formed by that layer sequence as a thin freestanding membrane. Next, reference will be made in parallel to FIGS. 10 and 11. FIG. 10 is a band-structure profile along a stacking direction of a layer sequence of a quantum well semiconductor device according to the prior art without polarization guard layers, as determined by an 8-band kp-based band structure calculation. FIG. 11 is a band-structure profile along a stacking direction of a layer sequence of an embodiment of a semiconductor device according to an embodiment of the present invention that in its quantum well layer structure corresponds to the embodiment of FIG. 3. The band-structure profile shown in FIG. 11 was also determined by 8-band kp-based band structure calculation.

(38) On the abscissa, a position along the c direction in units of nanometers is given. The origin of the abscissa corresponds to a position of an interface between an active layer 304.1 and the adjacent carrier-confinement layer 304.2 shown in FIG. 3. On the ordinate, an energy level is indicated in units of electron volts. The band structure profiles of FIGS. 10 and 11 show the energy levels of a lower conduction band edge LCE and of an upper valance band edge UVE as a function of the position along the direction c in the different layers of the respective layer sequence. Reference labels of corresponding layers in the embodiment of FIG. 3 are indicated in the upper part of the diagram of FIG. 11, and interfaces between the different layers are indicated by dashed vertical lines. As in the embodiment of FIG. 3, the semiconductor device shown in FIG. 11 comprises a single quantum well structure.

(39) As described earlier, the layer sequence shown in FIG. 3 comprises polarization guard layers 306.1 and 306.2 made of aluminum gallium nitride, carrier confinement layers 304.2 and 304.3 made of aluminum nitride and embedding the active layer 304.1 made of gallium nitride. With respect to the group-III-component, the AlGaN polarization guard layers 306.1 and 306.2 have an aluminum fraction of 0.2 and a gallium fraction of 0.8. In order to assess the effect of the polarization guard layers 306.1 and 306.2 on the band profile along the c direction, a comparative calculation was made for a layer sequence, which omits the polarization guard layers 306.1 and 306.2 and only comprises an active layer Y of gallium nitride embedded between carrier confinement layers X of aluminum nitride. The total geometrical extension of the layer sequences in the c direction was assumed to be the same for both layer sequences of FIGS. 10 and 11.

(40) A comparison of the band profiles shown in FIGS. 10 and 11 shows that the introduction of the polarization guard layers and the thickness of their carrier confinement layers in the layer structure of FIG. 11 achieves a near-flat band situation in the active layer 304.1. In contrast, the reference structure shown in FIG. 10 exhibits a strong gradient of the band edge energies along the c direction in the active layer, which is due to strong built-in electric fields originating from the pyro- and piezoelectricity of the polar semiconductor materials. The residual electron-hole overlap does not only limit the luminescence intensity, but also introduces large excitonic dipole moments in the growth direction of the active layer Y. Occupation probability density calculations for electrons (DSE) and holes (DSH) reveal spatially separated Occupation probability densities in the reference structure underlying that band structure calculations of FIG. 10, while a strong spatial overlap of the occupation probability densities is achieved in the embodiment of the present invention according to FIG. 11. In addition, due to the larger band gap of the polarization guard layers 306.1 and 306.2 in comparison with the band-gap energy of the active layer, which approximately equals an emission or resonant absorption energy of photons in the active layer 304.1, the polarization guard layers are transparent for the purpose of emission or absorption of photons in the relevant energy range. Both effects are particularly advantageous for applications in the fields of optics and optoelectronics.

(41) For the purpose of the present band structure calculations along the c direction a quantum well-based layer sequence was assumed. Similar results are achieved when the active layer is formed by a quantum wire or a quantum dot.

(42) Next, reference is made to FIGS. 12 and 13 in parallel. FIG. 12 is a schematic three-dimensional view of a further embodiment of a semiconductor device 1200 according to the present invention. FIG. 13 is a schematic cross-sectional view of a layer sequence 1202 of the semiconductor device 1200 of FIG. 12.

(43) FIG. 12 illustrates a layer sequence 1202 that forms a nitride nano-membrane structure that is deposited on a silicon (111) carrier 1220. The nitride nano-membrane structure consisting of the layer sequence 1202 bridges a gap region 1222 between bridge posts 1224 and 1226 on the carrier, the carrier layer of the layer sequence being exposed to an ambient atmosphere in the gap region 1222 between the bridge posts 1224 and 1226. The particular layer sequence inside of the nitride nano-membrane is detailed in FIG. 13 and comprises AlGaN guard layers 1206.1 and 1206.2 adjacent to AlN carrier confinement layers 1204.2. This layer sequence enhances the overall luminescence output from an active layer 1204.1 made of GaN because, as a result of this particular layer sequence, the commonly strongly inclined band structure of nitride devices that are affected by the QCSE reaches a highly desirable near-flat-band condition as shown in FIG. 11 above.

(44) It is noted that the graphical illustrations of FIGS. 1 through 9, 12 and 13 are schematic, in particular with respect to layer thicknesses shown, and for instance also with respect to a number of quantum dots that form the active layer in the embodiment of FIGS. 1 and 2 and others. No indication of actual extensions of structural elements is intended to be given by way of the graphical illustrations.

(45) FIG. 14 shows a plot of emission energies from different semiconductor devices having a structure corresponding to that of the semiconductor device underlying the band-structure calculation of FIG. 11. The semiconductor devices comprise a GaN/AlN quantum well and polarization guard layers made of AlGaN. The emission energies are plotted as a function of the Al faction, which is given in percent of the total group-III component. The sum of the Ga fraction and Al fraction amounts to 100% in each case. Seven plots are shown, wherein each plot is associated with a given number of monolayers of the GaN well layer of the GaN/AlN quantum well. The number of monolayers varies from 6 monolayers (uppermost curve), over 7, 8, 9, 10, and 12 to 14 monolayers (lowest curve).

(46) An observable decrease of emission energy with increasing amount of GaN monolayers in the quantum well is due to a decrease of quantum confinement (in the c direction) of the charge carriers localized in the well layer. The quantum confinement decreases with increasing extension of the well layer in c direction.

(47) In addition, as can be seen from FIG. 14, the emission energy from a quantum well with a given number of GaN monolayers decreases with increasing Al content. This decrease as a function of the Al content of the polarization guard layers is due to the Quantum Confined Stark Effect. The curves show that the Quantum Confined Stark Effect gets stronger with increasing Al content in the polarization guard layers. It is thus desirable to keep the Al content as low as possible while keeping the Al content high enough to achieve a band gap energy of the polarization guard layers that avoids absorption of the photons emitted from the quantum well.

(48) Also shown in FIG. 14 is an almost straight line showing a dependency of an energy band gap of AlGaN as a function of Al content. The line exhibits crossing points with each of the emission energy curves for different well-layer thicknesses. Suitable Al content values for the polarization guard layers are above the respective Al content values at the crossing points. To achieve a low Quantum Confined Stark Effect, it is desirable to use an Al content close to that of the respective crossing point, for a given well-layer thickness. As a side note, the calculations underlying the plots of FIG. 14 were performed without taking into account Coulomb interactions between electrons and holes. Taking Coulomb interaction also into account, it is to be expected that actual photon energies will be a little lower than those to be considered in the context of FIG. 14, or, in other words, the Al content exactly at the crossing points shown in FIG. 14 is expected to also be suitable for use in the polarization guard layers.

(49) In summary, a semiconductor device comprises a layer sequence formed by a plurality of polar single crystalline semiconductor material layers that each has a crystal axis pointing in a direction of crystalline polarity and a stacking direction of the layer sequence. A core layer sequence is formed by an active region made of an active layer stack or a plurality of repetitions of the active layer stack. The active layer stack has an active layer having a first material composition associated with a first band gap energy, and carrier-confinement layers embedding the active layer on at least two opposite sides thereof, having a second material composition associated with a second band gap energy larger than the first band gap energy. A pair of polarization guard layers is arranged adjacent to the active region and embedding the active region on opposite sides thereof. Both polarization guard layers have the first material composition. The pair of polarization guard layers shields the active layer(s) from internal polarization fields.