Embedded Contact Layer for UV Emission Device

Abstract

In some embodiments, a light-emitting structure comprises a layered semiconductor stack including a first set of doped layers, a second layer, a light-emitting layer disposed between the first and second sets of doped layers, and electrical contacts to the first set of doped layers. The first set of doped layers may include a first sublayer, a second sublayer, and a third sublayer, with the third sublayer adjacent to the light-emitting layer. Electrical contacts to the first set of doped layers may be made to the second sublayer. The first, second, and third sublayers may be n-type doped, and the electrical conductivity of the second sublayer may be higher than the electrical conductivity of the first and third sublayers. In some cases, the second sublayer may absorb more light emitted from the light-emitting layer than the first or third sublayers.

Description

Technical Field

[0001] Related Applications This application claims priority to U.S. Non-Provisional Patent Application No. 16 / 864,838, titled "Embedded Contact Layer for UV Emitting Devices," filed on May 1, 2020, which is hereby incorporated by reference in its entirety for all purposes.

Background Art

[0002] Deep-ultraviolet (deep-UV) light-emitting diodes (LEDs) have great potential for sterilization, water treatment, scientific analysis, and other applications. However, the performance of conventional ultraviolet C-band (UVC) LEDs has been plagued by high turn-on voltages and driving voltages. High-bandgap semiconductor materials are used to emit high-energy UVC light in conventional LEDs. However, high-bandgap materials are also difficult to dope efficiently into n-type or p-type, and thus, contacts made of high-bandgap materials typically suffer from low conductivity. The low conductivity of the high-bandgap materials used in UVC LEDs is one of the causes contributing to the high driving voltages required for devices that emit light through wide-bandgap n-type semiconductors used particularly for n-contacts.

Summary of the Invention

Problems to be Solved by the Invention

[0003] For example, conventional UVC LEDs that utilize wurtzite-type semiconductors typically use AlGaN with a high Al content in the active layer that emits UVC light. Some such devices use a quantum well structure having wells with a low Al content and barriers with a high Al content in the active layer. In some cases, conventional LEDs emit light through the edge of the device or through the p-side of the structure. In these cases, the emitted light need not pass through the n-type contact layer, and a low bandgap material having high conductivity can be used for the n-contact layer of the structure. In conventional devices where light is emitted through the n-side of the structure, an optically transparent material having a higher bandgap and lower conductivity is used for the n-contact layer. However, the low conductivity of these high bandgap materials degrades the performance of the device, for example, by increasing the turn-on voltage and drive voltage required to operate the device.

Means for Solving the Problems

[0004] In some embodiments, the light-emitting structure comprises a layered stack comprising a first set of doped layers, a second layer, a light-emitting layer disposed between the first set of doped layers and the second layer, and a layer providing an electrical contact to the first set of doped layers, wherein the first set of doped layers, the second layer, and the light-emitting layer comprise a semiconductor material. In some cases, the first set of doped layers comprises a first sublayer, a second sublayer, and a third sublayer, and the third sublayer is adjacent to the light-emitting layer. The first sublayer, the second sublayer, and the third sublayer may each comprise a first superlattice, a second superlattice, and a third superlattice, respectively. The well layer of the second superlattice may be thicker than the well layers of the first and third superlattices. The barrier layer of the second superlattice may be thinner than the barrier layers of the first and third superlattices. The electrical contact to the first set of doped layers can be made to the second sublayer. The first, second, and third sublayers can be doped n-type, and the electrical conductivity of the second sublayer can be higher than the electrical conductivities of the first and third sublayers.

[0005] In some embodiments, the light emitting structure comprises a layered stack comprising a first set of doped layers, a second layer, a light emitting layer disposed between the first set of doped layers and the second layer, and an electrical contact to the first set of doped layers, wherein the first set of doped layers, the second layer, and the light emitting layer comprise a semiconductor material. The first set of doped layers may comprise a first sublayer, a second sublayer, and a third sublayer, and the third sublayer is adjacent to the light emitting layer. The electrical contact to the first set of doped layers can be made to the second sublayer. The first, second, and third sublayers can be doped n-type, and the electrical conductivity of the second sublayer can be higher than the electrical conductivity of the first and third sublayers. The first, second, and third sublayers can each comprise a first, second, and third superlattice, respectively. The light emitting layer can comprise a fourth superlattice, and the second layer can comprise a fifth superlattice. The first, second, third, fourth, and fifth superlattices can each comprise a set of GaN well layers and AlN barrier layers.

[0006] In some embodiments, the light emitting structure comprises a layered stack comprising a first set of doped layers, a second layer, a light emitting layer disposed between the first set of doped layers and the second layer, and an electrical contact to the first set of doped layers, wherein the first set of doped layers, the second layer, and the light emitting layer comprise a semiconductor material. The first set of doped layers may comprise a first sublayer, a second sublayer, and a third sublayer, and the third sublayer is adjacent to the light emitting layer. The electrical contact to the first set of doped layers can be made to the second sublayer. The first, second, and third sublayers can be doped n-type, and the electrical conductivity of the second sublayer can be higher than the electrical conductivity of the first and third sublayers. Light having a wavelength shorter than 300 nm can be emitted from the light emitting layer, can pass through the first set of doped layers before being emitted from the structure, and the second sublayer can absorb more of the light emitted from the light emitting layer than the first or third sublayer.

Brief Description of the Drawings

[0007]

Figure 1

Figure 2

Figure 3

Figure 4A

Figure 4B

Figure 5A

Figure 5B

Figure 6A

Figure 6B

Figure 6C

Figure 6D

Figure 6E

DETAILED DESCRIPTION OF THE INVENTION

[0008] An embedded layer inserted as an n-contact layer into a semiconductor epitaxial structure is described herein. The semiconductor epitaxial structure described herein can be used for a light-emitting diode (LED) that emits light at short wavelengths (e.g., in the UVC band or at wavelengths less than 300 nm). This embedded layer has a high n-type conductivity, which is beneficial for current injection from the n-contact into the active layer of the structure. The high-conductivity material of the embedded layer also has a high absorption coefficient in the target wavelength range (e.g., at wavelengths below 300 nm). Thus, the thickness of the embedded layer is adjusted to have an acceptable amount of light absorption in the embedded layer and an acceptable electrical resistance (i.e., between the n-contact and the active layer of the structure). In other words, the thickness of the embedded layer is adjusted to improve the device performance (e.g., turn-on voltage, drive voltage, and output power efficiency) compared to conventional structures by trading off an increase in absorption and a decrease in layer resistance. In some embodiments, precise epitaxial growth thickness control is used to form an embedded n-contact layer of an accurate thickness, and the n-contact layer is exposed using a precise etching process so that n-metal can be deposited thereon.

[0009] In some embodiments, the semiconductor epitaxial structure includes a first set of doped layers, a second layer, and a light-emitting layer disposed between the first set of doped layers and the second layer. The first set of doped layers is n-type doped, and an n-type electrical contact (i.e., using n-metal) is made to the first set of doped layers. The first set of doped layers provides a highly conductive embedded n-contact layer having an acceptable light transmittance (for the wavelength of the emitted light) that enables obtaining a low drive voltage without significantly reducing the output power. In some cases, the embedded n-contact layer can increase the output power of the structure compared to a structure having a transparent n-contact layer with lower conductivity.

[0010] In some embodiments, the first set of doped layers comprises a first sublayer, a second sublayer, and a third sublayer, and the third sublayer is adjacent to the light-emitting layer (i.e., such that the second sublayer is “embedded” under the third sublayer, or the third sublayer is between the active layer and the second sublayer). The second sublayer is an n-contact layer, i.e., the second sublayer is a layer within the structure that is connected to an n-metal contact. The second sublayer may have a higher doping density and / or higher electrical conductivity than the first and third sublayers. In some embodiments, the second sublayer has a higher absorption coefficient for light emitted from the light-emitting layer than the first or third sublayer. In some embodiments, the second sublayer can be disposed between the first sublayer and the third sublayer, or the first sublayer can be disposed between the second sublayer and the third sublayer.

[0011] In some embodiments, the embedded n-contact layer has the following characteristics. First, the embedded n-contact layer can be deposited using the same epitaxial growth process as the rest of the layers within the structure, thereby enabling it to be easily incorporated into the structure fabrication. Second, the embedded n-contact layer has a higher conductivity than the surrounding layers (e.g., the first and third sublayers described above), thereby improving current injection compared to similar devices without a highly conductive n-contact. Third, the embedded n-contact layer can be made thin enough to absorb an acceptable amount of light emitted from the active layer of the structure (e.g., the embedded n-contact layer absorbs less than 60%, less than 50%, less than 40%, or less than 30% of the emitted light, or absorbs between 10% and 50%, or between 10% and 60% of the emitted light), and the layer is not so thin that it is inaccessible through an etching process to expose the embedded n-contact layer (e.g., the embedded n-contact layer is thicker than about 10 nm, or thicker than about 20 nm).

[0012] In some embodiments, the semiconductor epitaxial structures described herein are grown using molecular beam epitaxy (MBE), metalorganic chemical vapor deposition (MOCVD), or hydride vapor phase epitaxy (HVPE). In some embodiments, the buried n-contact layer is doped during the formation of the epitaxial structure (e.g., using one or two dopant sources in an MBE system). Following the growth of the semiconductor epitaxial structure, standard semiconductor manufacturing methods can be used to process the semiconductor epitaxial structure, including etching (e.g., using dry etching) to form a mesa structure and metal deposition (e.g., using evaporation or sputtering) to deposit n and p metal contacts. In some embodiments, the n-contact layer can be processed after epitaxial growth to increase the conductivity and / or doping density of the layer. For example, a laser or heat treatment can be used to improve dopant activation within the layer. In some embodiments, such a laser or heat treatment can be performed during the epitaxial growth process, e.g., during a growth pause immediately after the formation of the buried n-contact layer. In some embodiments, ion implantation can also be used to increase the doping density of the layer.

[0013] In some embodiments, the first, second, and third sub-layers of the doped layer are each a single layer of semiconductor material. In some embodiments, the first and third sub-layers include a semiconductor having a higher bandgap than the semiconductor of the second sub-layer. Since materials with lower bandgaps are typically doped more efficiently (e.g., with extrinsic dopants), the lower bandgap of the second sub-layer allows that layer to be doped higher than the first or third sub-layers and have a higher electrical conductivity. However, the lower bandgap can also increase the optical absorption (at the emission wavelength) of the second sub-layer relative to the optical absorption of the first and third sub-layers, and in some cases, it will reduce the amount of light emitted from the structure. The higher electrical conductivity of the second sub-layer will reduce the required drive voltage (for a given amount of current injection) by reducing the resistance between the n-metal contact of the device and the active layer (i.e., it will be easier to inject carriers from the n-contact into the active layer). In this embodiment, the thickness of the second sub-layer is adjusted such that, in particular, the performance degradation due to increased absorption in the second sub-layer can be offset by improved carrier injection, and unexpectedly, the output power of the entire device can be approximately the same or even improved compared to a structure in which the second sub-layer has the same bandgap as the first and third sub-layers.

[0014] For example, the first and third sub-layers can be Al x Ga 1-x N (i.e., AlGaN) with 0 ≦ x ≦ 1, and the second sub-layer can be AlGaN with a lower Al content than the first and third sub-layers (i.e., having a lower value of x in the Al x Ga 1―x N material). In other words, the first and third sub-layers can be single layers of Al x Ga 1-x N, and the second sub-layer can be Al y Ga 1-yIt can be a single layer of N, and x can be made larger than y. In some embodiments, the first, second, and third sub-layers are single layers having different compositions from each other, and the second sub-layer has a lower bandgap than the first and third sub-layers. For example, the first sub-layer can be Al x Ga 1-x N, the second sub-layer can be a single layer of Al y Ga 1-y N, the third sub-layer can be Al z Ga 1-z N where 0≦z≦1, x can be made larger than y, x can be made larger than z, and y and z can be made different from each other. In some embodiments, AlGaN with a lower Al content reduces the bandgap of that layer and can be doped more efficiently (e.g., with an extrinsic dopant such as Si, Ge, or Se), and the lower Al content of the second sub-layer will allow that layer to be doped higher and have a higher electrical conductivity than the first or third sub-layer. However, the reduced bandgap of the second sub-layer can also increase the light absorption (at the emission wavelength) of the second sub-layer compared to the first and third sub-layers. The composition of such sub-layers can be different depending on the wavelength of the light emitted from the structure. For example, when the emission wavelength is 290 nm, the first and third sub-layers can be Al x Ga 1-x N (x is approximately equal to 0.45), and the second sub-layer can be Al y Ga 1-y N (y is approximately equal to 0.1). In such a structure, the first and third sub-layers can minimally absorb light at 290 nm, and the low bandgap of the second sub-layer can improve the conductivity of that sub-layer. However, the low Al content of the second sub-layer will cause that layer to absorb some of the 290 nm emitted light. In another example, when the emission wavelength is 250 nm, the first and third sub-layers can be Al x Ga 1-x N (x is approximately equal to 0.6), and the second sub-layer can be Al yGa 1-y It can be N (y is approximately equal to 0.1). In this structure, the higher Al content in the first and third sub-layers may enable them to minimally absorb light with a shorter wavelength of 250 nm. In this example, the low bandgap of the second sub-layer can improve the conductivity of that sub-layer, but the low Al content of the second sub-layer will also cause that layer to absorb a part of the emitted light at 250 nm. In both of the above examples, the y value of 0.1 for the second sub-layer is approximate, and higher or lower y values may provide different degrees of conductivity and absorption, which can be adjusted for different structures (for example, by changing the composition and / or thickness of the second sub-layer). Furthermore, the composition at the wavelengths in these examples is approximate, partly due to the variations in the bandgap and absorption coefficient of AlGaN compounds manufactured using different growth techniques. In these examples, the high conductivity of the second sub-layer can improve the operating voltage of the structure compared to a structure without the second sub-layer. Furthermore, in these examples, by adjusting the thickness of the second sub-layer, the performance degradation due to increased absorption in the second sub-layer can be offset by improved carrier injection, and unexpectedly, the output power of the entire device can be approximately the same or even improved compared to a structure in which the second sub-layer has the same Al content as the first and third sub-layers.

[0015] In some embodiments, the first, second, and third sub-layers of the doped layer each include a superlattice having a well layer and a barrier layer composed of a semiconductor material, and the well layer has a lower bandgap than the barrier layer. The semiconductors and / or thicknesses of the wells and / or barriers constituting the superlattice of the second sub-layer may be different from those constituting the superlattices of the first and / or third sub-layers, which may enable the second sub-layer to have a higher conductivity than the first and / or third sub-layers. In some embodiments, the second sub-layer (i.e., the n-contact layer) is a single layer, while the first and third sub-layers are superlattices. In some embodiments, the second sub-layer (i.e., the n-contact layer) is a superlattice, while the first and third sub-layers are each single layers. In some embodiments, one of the first, second, and third sub-layers is a superlattice and the other two are single layers. In some embodiments, two of the first, second, and third sub-layers are superlattices and the other layer is a single layer. For example, the third sub-layer may be a single layer (e.g., of AlN or AlGaN), while the first sub-layer is a superlattice and the second sub-layer is either a single layer or a superlattice. Further, in some embodiments, the third sub-layer includes a superlattice having a thick barrier layer adjacent to the active layer. For example, the third sub-layer may include a superlattice having repeating unit cells including wells (e.g., composed of GaN or AlGaN) and barriers (e.g., composed of AlN or AlGaN), and the barrier thickness is not constant throughout the superlattice and the last barrier may be very thick (e.g., greater than 1 nm or greater than 10 nm).

[0016] In some embodiments, the semiconductor forming the wells and / or barriers of the superlattice of the second sublayer has a lower bandgap than that forming the superlattice of the first and / or third sublayers. In some embodiments, the well layer of the superlattice of the second sublayer is thicker than the well layer of the superlattice of the first and / or third sublayers, and / or the barrier layer of the superlattice of the second sublayer is thinner than the barrier layer of the superlattice of the first and / or third sublayers. In some embodiments, the semiconductor forming the wells and / or barriers has a lower bandgap, the wells are thicker, and / or the barriers are thinner in the superlattice of the second sublayer compared to the superlattice of the first and / or third sublayers.

[0017] In all of the above cases, the effective bandgap of the superlattice of the second sublayer can be lower than the effective bandgaps of the superlattices of the first and third sublayers. The wells and barriers of the superlattice can create minibands, and the effective bandgap can reflect the optical and electronic properties of the entire superlattice, including the effects of the minibands. The effective bandgap of the second sublayer is lower than that of the first and / or third sublayers. This is because the superlattice of the second sublayer has a lower bandgap material and / or a higher proportion of lower bandgap material than the superlattice of the first or third sublayer. Further, when the superlattice of the second sublayer has thicker wells and / or thinner barriers, carriers are less likely to be confined within the wells. In other words, the superlattice minibands in the second sublayer occur at lower energies than those in the superlattices of the first and third sublayers, thereby enabling transitions between minibands to occur at lower energies compared to the superlattices of the first and third sublayers and efficiently reducing the bandgap (and potentially increasing light absorption at the target wavelength).

[0018] Lower bandgap materials are usually doped more efficiently (e.g., with extrinsic dopants), so the lower effective bandgap of the second sublayer will allow that layer to be doped higher than the first or third sublayers and have higher electrical conductivity. However, the lower effective bandgap can also make the light absorption (at the emission wavelength) of the second sublayer higher than that of the first and third sublayers. For the same reasons as above, the higher electrical conductivity of the second sublayer will reduce the turn-on voltage and drive voltage of the device by reducing the resistance between the n-metal contact of the device and the active layer (i.e., it will be easier to inject carriers from the n-contact into the active layer). In this case, if the thickness of the second sublayer is properly adjusted, the performance degradation due to the increased absorption in the second sublayer can be offset by the improved carrier injection, and unexpectedly, the output power of the entire device can be approximately the same or even improved compared to a structure where the second sublayer has the same effective bandgap as the first and third sublayers.

[0019] FIG. 1 is a schematic diagram of a conventional semiconductor light-emitting structure 100. The structure 100 includes a substrate 110, a buffer layer 120, an n-contact layer 130, an active layer 140, an optional layer 150, and a p-contact layer 160. The optional layer 150 can be configured as an electron blocking layer (EBL) in some cases. An n-metal 170 contacts the n-contact layer 130, and a p-metal 180 contacts the p-contact layer 160. An external bias can be applied between the metal contacts 170 and 180, whereby the active layer 140 can emit light, for example, in the UVC wavelength band. To expose the n-contact layer 130 to which the n-metal 170 will contact, an etching process is used to partially etch the n-contact layer 130 to generate a mesa structure. After the n-contact layer 130 is exposed by etching, the n-metal 170 can be deposited on the n-contact layer 130. In the structure 100, the n-metal and the p-metal cover most of the surface of the structure 100 on the side opposite to the substrate 110. Thus, the structure 100 can emit light from the edge of the device or through the n-contact layer 130, the buffer layer 120, and the substrate 110. As described above, in a conventional structure that emits light through the n-side of the structure, the n-contact layer 130 is typically a wide-bandgap material with low light absorption of the emitted light and also typically has a relatively high electrical resistance due to the wide bandgap. As a result, such a conventional UVC LED suffers from low performance including a high turn-on voltage and a driving voltage.

[0020] Figure 2 is a schematic diagram of a semiconductor light-emitting structure 200 incorporating an embedded n-contact layer (i.e., the second sub-layer) 230b according to some embodiments. The structure 200 includes a substrate 210, a buffer layer 220, a set of doped layers 230 (including a first sub-layer 230a, a second sub-layer 230b, and a third sub-layer 230c), an active layer 240, an EBL 250, and a p-contact layer 260. In some embodiments, the buffer layer 220, the set of doped layers 230, the active layer 240, the EBL 250, and the p-contact layer 260 all comprise semiconductor materials, which may vary between different layers. In some embodiments, the set of doped layers 230, the first sub-layer 230a, the second sub-layer 230b, and the third sub-layer 230c are all doped n-type. In some embodiments, the second sub-layer 230b is doped n-type (e.g., using an extrinsic dopant such as Si), and the first sub-layer 230a and / or the third sub-layer 230c are intentionally undoped (i.e., no extrinsic dopant is intentionally added, but impurities that may act as dopants may be unintentionally added in some cases). The structure is etched to form a mesa structure where the second sub-layer 230b is exposed and n-metal 270 can contact the second sub-layer 230b. Thus, the sub-layer 230b can be referred to as the n-contact layer. p-metal 280 contacts the p-contact layer 260. The structure 200 can emit light, for example, in the UVC wavelength band or at a wavelength less than 300 nm, when an external bias is applied between the n-metal 270 contact and the p-metal 280 contact. In this case, the light is emitted from the active layer 240 and passes through the set of doped layers 230, the buffer layer 220, and the substrate 210 before being emitted from the structure 200. As described above, in some embodiments, the sub-layer 230b has a higher electrical conductivity than the sub-layers 230a and 230c (e.g., due to a higher doping density). In some embodiments, the sheet resistance of the sub-layer 230b can be less than 10,000 ohms / square, or less than 1000 ohms / square, or between 10 ohms / square and 1000 ohms / square, or between 10 ohms / square and 10,000 ohms / square.In some embodiments, the sheet resistance of sub-layer 230b can be about 10 times, about 100 times, about 1000 times, or about 10,000 times higher than the sheet resistance of sub-layers 230a and / or 230c. The conductivity of the embedded n-contact layer can vary depending on the materials used (e.g., higher or lower than the values listed above). For example, it can be beneficial to add an embedded n-contact layer with moderate conductivity (e.g., sheet resistance exceeding about 1000 ohms per square or exceeding about 10,000 ohms per square) to a conventional structure having a very resistive (e.g., sheet resistance exceeding about 10,000 ohms per square or exceeding about 100,000 ohms per square) n-contact layer material. Further, in some cases, sub-layer 230b has higher light absorption at the wavelength emitted by active layer 240 as compared to sub-layers 230a and 230c.

[0021] In some embodiments, sub-layers 230a and 230b can be repositioned such that sub-layer 230b (the n-contact layer) is adjacent to buffer layer 220 and sub-layer 230a is between sub-layers 230b and 230c. In some cases, layer 230a can be omitted from structure 200. In such a case, the structure would include buffer layer 220, followed by n-contact layer 230b, sub-layer 230c, and active layer 240. In some cases, layer 230c can be omitted from structure 200. In such a case, the structure would include buffer layer 220, followed by sub-layer 230a, n-contact layer 230b, and active layer 240. In some embodiments, the first and third sub-layers 230a and 230c have the same composition, thickness, and / or structure (in the case of multi-layer or superlattice sub-layers), while in some embodiments, the first and third sub-layers 230a and 230c have different compositions, thicknesses, and / or structures (in the case of multi-layer or superlattice sub-layers).

[0022] The n-metal 270 can contact the upper surface of the n-contact layer 230b (i.e., the surface of layer 230b on the side opposite the substrate), or the sub-layer 230b can be partially etched away so that the n-metal can contact the interior of the sub-layer 230b. In embodiments where the sub-layer 230b includes a superlattice, the n-metal 270 can contact the wells or barriers of the superlattice of the sub-layer 230b. In some embodiments, the n-metal 270 contacts a layer within the superlattice of the sub-layer 230b and contacts a highly doped well layer within the superlattice of the sub-layer 230b to create a Schottky n-barrier that is minimized as much as possible. In some embodiments, the sub-layer 230b includes a superlattice, the wells of the superlattice are very thin (e.g., about 1 monolayer (ML) thick, or less than 1 ML thick), and the etching process used to expose the sub-layer 230b etches the sub-layer 230b to slightly different depths, making the exposed surface of the sub-layer 230b slightly rough, bumpy, or non-uniform. In such cases, the n-metal can potentially contact both the wells and the barriers at different lateral (i.e., parallel to the substrate surface) positions within the superlattice of the sub-layer 230b.

[0023] Throughout this disclosure, layers having a fractional monolayer, i.e., a thickness less than 1 ML, or a thickness equal to a fractional ML (e.g., 1.5 ML) are described. A fractional ML of a semiconductor having a thickness less than 1 ML can include three-dimensional islands of the semiconductor, and thus the layer can be discontinuous. For example, a GaN well layer having a thickness less than 1 ML surrounded on both sides by AlN barriers can include three-dimensional islands of GaN surrounded by AlN. A fractional ML having a thickness greater than 1 ML can be described by the connection of a first layer including an integer number of MLs and a second fractional ML layer having a thickness less than 1 ML, and the second fractional ML layer having a thickness less than 1 ML has the above-described characteristics.

[0024] In some cases, the superlattice (e.g., a short-period superlattice (SPSL)) includes alternating layers of AlN and GaN and does not include ternary AlGaN layers, and the AlN and / or GaN layers each include AlN and / or GaN less than 1 ML. Thus, in some regions (or nano-regions) of such layers, there is a mixed compound Alx Ga 1-x N may still be present. Similarly, in some cases, the superlattice may have only GaN and AlN layers, and the thicknesses of the AlN and / or GaN layers may be non-integer MLs. In such cases, in some regions (or nano-regions) of the layer, the mixed compound Al x Ga 1-x N may still be present within the superlattice.

[0025] The substrate 210 can be many different materials such as sapphire, SiC, AlN, GaN, silicon, or diamond. In some embodiments, the substrate 210 has a low absorption coefficient for the light emitted from the active layer 240 and / or has a lattice constant similar to that of the materials forming the other layers of the epitaxial structure 200. In some embodiments, the substrate significantly absorbs light at the wavelength of interest, and the substrate is thinned or removed during device processing. In some cases, the substrate is locally thinned to form a window for the light emitted from the active layer 240 to escape from the structure.

[0026] In some embodiments, the buffer layer 220 has a thickness from 50 nm to 1000 nm, or from 50 nm to 5000 nm, and is composed of a semiconductor material having a low absorption coefficient for the light emitted from the active layer 240 and a lattice constant similar to that of the materials forming the other layers of the epitaxial structure 200. Some examples of materials that can be used for the buffer layer 220 are AlN, AlGaN, and InAlGaN. The buffer layer 220 can be a single layer, multiple layers, or a superlattice in various embodiments.

[0027] In some embodiments, the set 230 of doped layers includes a semiconductor epitaxial layer that is doped n-type. Some examples of materials for the set 230 of doped layers are GaN, AlN, AlGaN, and InAlGaN. Each of the sublayers 230a, 230b, and 230c in the set 230 of doped layers can have a thickness of less than about 10 nm to 3000 nm, or less than about 10 nm to 1000 nm, or 100 nm to 1000 nm, or 10 nm to 300 nm, or 10 nm to 100 nm, or about 50 nm, or less than 50 nm, depending on the layer and the overall structure. In some embodiments, the sublayer 230a has a thickness greater than 100 nm in order to improve the material quality of the active layer 240. For example, a thick (e.g., greater than 100 nm, or 100 nm to 500 nm, or 100 nm to 1000 nm) sublayer 230a can remove threading dislocations before they can reach the active layer 240 and degrade device performance. In some embodiments, the sublayer 230c has a thickness and a natural lattice constant (i.e., a relaxed or strain-free lattice constant) suitable for improving strain matching and better confining electrons in the active layer 240. In some embodiments, the bandgap or effective bandgap of the sublayers 230a and 230c is wider than the bandgap or effective bandgap of the sublayer 230b, such that the sublayer 230b absorbs more of the light emitted from the active layer 240. In some embodiments, the doping density of the sublayers 230a and 230c is lower than the doping density of the sublayer 230b, such that the sublayer 230b has a higher electrical conductivity. The higher electrical conductivity of the sublayer 230b can also result from reasons unrelated to the doping density from extrinsic dopants. For example, the electrical conductivity of the sublayer 230b can be higher than the electrical conductivity of 230a and 230c due to polarization doping of the sublayer, free carrier mobility, or other material properties, or due to the structure and / or dimensions of the sublayer. Different configurations of the set 230 of doped layers are discussed further below.

[0028] In some embodiments, the active layer 240 includes a semiconductor material configured to emit light. In some embodiments, the active layer 240 may include one or more narrower bandgap wells surrounded by wider bandgap barriers (e.g., in a quantum well structure, superlattice, or short period superlattice (SPSL)), where the bandgaps and thicknesses of the wells and barriers are selected to emit light at wavelengths less than 300 nm (e.g., in the UVC band). Some examples of materials for the wells and / or barriers of the active layer 240 are GaN, AlN, AlGaN, and InAlGaN. The active layer may have a thickness, for example, of less than about 10 nm to 1000 nm, or from 10 nm to 100 nm, or about 50 nm. In some embodiments, using a superlattice (or SPSL) in the active layer 240 may be beneficial in terms of the efficiency of light emission and / or light extraction from the structure.

[0029] In some embodiments, the EBL 250 includes a wide bandgap semiconductor material configured to block electrons from exiting the active layer 240 and entering the p-contact layer 260. Optionally, the EBL layer 250 can be an externally doped p-type. Optionally, the EBL 250 can be undoped or doped via a mechanism such as polarization doping. The EBL 250 can include a single layer having a conduction band offset configured to confine electrons within the active layer 240, or can include multiple layers. In some embodiments, the EBL is a chirped SPSL having wells and barriers, and the thickness of the wells and / or barriers varies across the EBL 250. For example, the thickness of the wells can vary linearly from a thinner thickness (e.g., less than 1 ML, about 0.5 ML, or about 0.25 ML) adjacent to the active layer 240 to a higher thickness (e.g., greater than 5 ML, or about 10 ML) adjacent to the p-contact layer 260. The thickness of the barriers can be constant across the EBL layer 250 or can vary. The EBL 250 can be a p-type chirped superlattice having a varying well layer thickness, a varying barrier layer thickness, or a varying well layer and barrier layer thickness through the layers. The thickness of the EBL layer can be, for example, from 5 nm to 50 nm, or about 20 nm.

[0030] The EBL including the chirped EBL for long wavelength LEDs is fully described by International Patent Application Publication No. WO2019 / 193487, which is hereby incorporated by reference in its entirety.

[0031] In some embodiments, the p-contact layer 260 is a material having a high conductivity that enables a low contact resistance between the p-metal and the active layer 240. The p-contact layer can be a narrow bandgap material (e.g., to provide high electrical conductivity) or a wide bandgap material (e.g., to reduce the secondary absorption of light emitted from the active layer 240). In some embodiments, the p-contact layer material has a bandgap that also provides a low resistance contact and a low absorption coefficient for the wavelength of light emitted from the structure. Some examples of materials for the p-contact layer 260 are GaN, AlN, AlGaN, and InAlGaN. The p-contact layer 260 is doped with a p-type dopant such as Mg. The thickness of the p-contact layer can be, for example, from 10 nm to 100 nm, or about 40 nm. In some cases, the p-contact layer 260 can be a superlattice, e.g., an SPSL having GaN wells and AlN barriers or AlGaN wells and barriers. In some cases, the p-contact layer 260 can have, for example, a graded composition from a first composition of Al x Ga 1-x N to a second composition.

[0032] The n-metal 270 and the p-metal 280 can include any combination of metals that form ohmic contacts to the n-contact layer 230b and the p-contact layer 260, respectively. Some examples of materials that can be used for the n-metal and / or the p-metal are Ti, Al, Ta, and Ni. For example, the n-metal and the p-metal can include a layer of Ti adjacent to the n-contact layer 230b or the p-contact layer 260, followed by a layer of Al. In some cases, the n-metal and p-metal contact layers each include 1 nm to 10 nm (or about 2 nm) of Ti deposited on the n-contact layer 230b or the p-contact layer 260, followed by 20 nm to 400 nm of Al. The total thickness of the n-metal 270 and the p-metal 280 can be from about 20 nm to about 400 nm.

[0033] In some embodiments, the set of doped layers 230 (including sub-layers 230a, 230b, and 230c), the active layer 240, the EBL 250, and the optional p-contact layer 260 are all composed of an SPSL consisting of alternating pairs of GaN wells and AlN barriers. In such cases, the well and / or barrier thicknesses can be adjusted to adjust the effective bandgap of each of the layers 230, 240, 250, and 260.

[0034] In some embodiments, the set of doped layers 230 (including sub-layers 230a, 230b, and 230c), the active layer 240, the EBL 250, and the p-contact layer 260 are all composed of an SPSL, and the average alloy content of each of the plurality of unit cells of the SPSL (i.e., the repeating unit of the SPSL, e.g., GaN / AlN) is either constant or not constant along the growth direction. In some embodiments, the set of doped layers 230 (including sub-layers 230a, 230b, and 230c) and the active layer 240 are all composed of an SPSL, and the average alloy content of each of the plurality of unit cells of the SPSL (i.e., the repeating unit of the SPSL, e.g., GaN / AlN) is either constant or not constant along the growth direction. Thickness t GaN of the GaN layer and thickness t AlN of the AlN layer, the average alloy content of a simple unit cell containing two compositions such as these is x ave =t AlN / (t AlN +t GaN ) where x ave represents the effective Al fraction in the pair in the unit cell. In alternative embodiments, the unit cells of the SPSL can include three or more Al x Ga 1-x N compositions, and in such embodiments, the effective alloy content can be determined similarly. The average alloy content of other layer compositions including binary, ternary, and quaternary materials can be defined according to one or more elemental components. For example, layers 230, 240, 250, and 260 can be AlN / Al x Ga 1-x N / GaN or AlN / Al x Ga 1-x N / Al yIn z Ga 1-y-z The SPSL may include a three-layer unit cell having three layers of GaN, and the average alloy content (e.g., Al ratio) in these unit cells can also be determined. By maintaining a certain average alloy content, lattice matching of the effective in-plane lattice constants of the unit cells of different superlattices becomes possible (e.g., in layers 230, 240, 250, and / or 260). In some embodiments, the unit cells adjacent to each other throughout the semiconductor structure have substantially the same average alloy content. In some embodiments, the average alloy content of each of the plurality of unit cells is constant in most of the semiconductor structure 200.

[0035] The set 230 of doped layers can have different configurations. In some embodiments, the sub-layers 230a, 230b, and 230c are all single layers (e.g., single layers of semiconductor material), and the bandgap of the second sub-layer 230b is lower than the bandgap of the first sub-layer 230a and / or the third sub-layer 230c. In other cases, the sub-layers 230a, 230b, and 230c are all superlattices (e.g., superlattices including wells of a first semiconductor material and barriers of a second semiconductor material), and the effective bandgap of the superlattice of the second sub-layer is lower than the effective bandgap of the superlattice of the first sub-layer 230a and / or the third sub-layer 230c superlattice. In still other cases, one or two of the sub-layers 230a, 230b, and 230c are single layers, one or two of the sub-layers 230a, 230b, and 230c include superlattices, and the bandgap or effective bandgap of the second sub-layer is lower than the bandgap or effective bandgap of the first sub-layer 230a and / or the third sub-layer 230c. Each of the sub-layers 230a, 230b, and 230c can have a bandgap, or in the case of a superlattice, an effective bandgap, which can be constant or vary across the sub-layer. Each of the sub-layers 230a, 230b, and 230c can have a composition, or in the case of a superlattice, an average composition of each unit cell (e.g., well / barrier pair), which can be constant or vary across the sub-layer. For example, the thickness of the wells and / or barriers within each superlattice of the sub-layers 230a, 230b, and / or 230c can be constant or can vary through the thickness of the sub-layer.

[0036] The thicknesses of sub-layers 230a, 230b, and 230c (i.e., in the growth direction perpendicular to the surface of substrate 210) may all be the same or different from each other. Sub-layer 230a may have a thickness of from 10 nm to 3000 nm, or from 10 nm to 1000 nm, or from 100 nm to 500 nm, or about 400 nm. Sub-layer 230b (i.e., the n-contact layer 230b) may have a thickness of from less than 10 nm to 100 nm, or about 50 nm, depending on the material system. Sub-layer 230c is adjacent to the active layer 240 and may have a thickness of from less than 10 nm to 100 nm, or about 50 nm. For example, sub-layer 230a may be thicker than sub-layers 230b and 230c and sub-layer 230a may have a thickness of about 400 nm. Sub-layer 230c can help improve strain matching in the structure and confine electrons to the active region (for example, because the bandgap or effective bandgap is typically larger than that of sub-layer 230b), and thus may have a thickness of less than 50 nm. The thickness of sub-layer 230b is important as it governs the resistance and light absorption of the layer. In some embodiments, sub-layer 230b may be thinner than sub-layers 230a and / or 230c and may have a thickness of from about 10 nm to about 100 nm, or from about 20 nm to about 100 nm. The thickness of sub-layer 230b can be adjusted to trade off the turn-on voltage and / or drive voltage with the output power emitted from the structure, as described above.

[0037] FIG. 3 shows a schematic diagram of an example of a set 230 of doped layers including sub - layers 230a, 230b, and 230c according to some embodiments. In this example, sub - layers 230a, 230b, and 230c are all superlattices. The superlattice of sub - layer 230a includes a repeating pair of wells 310 and barriers 315, the superlattice of sub - layer 230b includes a repeating pair of wells 320 and barriers 325, and the superlattice of sub - layer 230c includes a repeating pair of wells 330 and barriers 335. The number of repeating pairs of wells and barriers (i.e., repeating units) within sub - layers 230a, 230b, and / or 230c can range from small values (e.g., about 10 repeating units for thin layers) to large values (e.g., about 1000 repeating units or more for thicker layers). For example, sub - layers 230a, 230b, and / or 230c can include from about 10 to about 1000, or from about 10 to about 200, or from about 10 to about 100 repeating units. In some embodiments, the wells 320 and barriers 325 of the superlattice of sub - layer 230b are of a different composition and / or thickness than the wells 310 and 330 and barriers 315 and 335 of the superlattices of sub - layers 230a and 230c.

[0038] FIG. 3 shows that in each of the superlattices of the sub - layers, there can be many repeating pairs of wells 310, 320, and 330 and barriers 315, 325, and 335. Further, each of the superlattices of the sub - layers can start and end with either a well or a barrier. However, in some embodiments, there cannot be adjacent wells or barriers between adjacent superlattices. In other words, in some embodiments, adjacent wells 310 and 320 of adjacent superlattices have a barrier (either 315 or 325) between them, and adjacent wells 320 and 330 of adjacent superlattices have a barrier (either 325 or 335) between them. Similarly, in some embodiments, adjacent barriers 315 and 325 of adjacent superlattices have a well (either 310 or 320) between them, and adjacent barriers 325 and 335 of adjacent superlattices have a well (either 320 or 330) between them.

[0039] The thicknesses of wells 310, 320, and 330 and barriers 315, 325, and 335 can be varied according to the materials used for these layers (and other layers of structure 200). In some embodiments, wells 310, 320, and 330 can be from 0.1 ML to 10 ML, or from 0.1 ML to 4 ML, and the barriers can be from 2 ML to 20 ML.

[0040] In some embodiments, the set of doped layers 230 can include Al x Ga 1-x N (where x can range from 0 to 1) to form wells 310, 320, and 330 and / or barriers 315, 325, and 335. For example, well 320 can be Al x Ga 1-x N having a lower Al content than wells 310 and 330. In this case, sublayer 230b will have a lower effective bandgap than sublayers 230a and 230c because of the lower Al ratio in well 320. The lower Al content of well 320 within sublayer 230b allows that layer to be doped higher (e.g., with an extrinsic dopant such as Si) and have a higher electrical conductivity than the first sublayer 230a and the third sublayer 230c. However, the reduced effective bandgap of sublayer 230b caused by the lower Al content can also increase the light absorption of sublayer 230b compared to sublayer 230a and sublayer 230c. In this example, by adjusting the composition of well 320 in sublayer 230b and the total thickness of sublayer 230b, the performance degradation due to the increased absorption in sublayer 230b can be offset by the improved carrier injection, and the overall output power of the light emitted from the structure can be approximately the same or even improved compared to a structure where well 320 of sublayer 230b has the same Al content as wells 310 and 330 of the first sublayer 230a and the third sublayer 230c.

[0041] In some embodiments, the set of doped layers 230 includes different thicknesses (i.e., in the growth direction perpendicular to the surface of the substrate 210) of the wells 310, 320, and 330 and / or the barriers 315, 325, and 335. For example, the well 320 may be thicker than the wells 310 and 330. In another example, the well 320 may be thicker than the wells 310 and 330, and the barrier 325 may be thinner than the barriers 315 and 335. In another example, the well 320 may be the same thickness as the wells 310 and 330, and the barrier 325 may be thinner than the barriers 315 and 335. In all of these cases, the sub-layer 230b has a lower effective bandgap than the sub-layers 230a and 230c due to the thicker well 320 and / or the thinner barrier 325 as described above, which reduces the carrier confinement within the well of the sub-layer 230b. The lower effective bandgap of the sub-layer 230b allows that layer to be doped more highly (e.g., with an extrinsic dopant such as Si) and have a higher electrical conductivity than the first sub-layer 230a and the third sub-layer 230c. However, the reduced effective bandgap of the sub-layer 230b may also increase the light absorption (at the wavelengths emitted by the active layer 240) of the sub-layer 230b compared to the sub-layers 230a and 230c. In this example, by adjusting the thickness of the well 320 and / or the barrier 325 of the sub-layer 230b and the total thickness of the sub-layer 230b, the performance degradation due to the increased absorption in the sub-layer 230b can be offset by improved carrier injection, and the overall output power of the light emitted from the structure can be approximately the same or even improved compared to a structure in which the well 320 and / or the barrier 325 of the sub-layer 230b has the same thickness as the wells 310 and 330 and / or the barriers 315 and 335 of the first sub-layer 230a and the third sub-layer 230c.

[0042] In some embodiments, both the composition and thickness of the well 320 and / or the barrier 325 of the sublayer 230b are different from those of the wells 310 and 330 and the barriers 315 and 335 of the sublayers 230a and 230c. In this case, the effective bandgap of the sublayer 230b can be made lower than the effective bandgaps of the sublayers 230a and 230c by a combination of a lower bandgap composition of the well 320 and / or the barrier 325 and a thicker well 320 and / or a thinner barrier 325 compared to the wells 310 and 330 and / or the barriers 315 and 335 of the sublayers 230a and 230c. In some cases, the composition of the material forming the well 320 of the sublayer 230b can be changed such that the bulk bandgap of the material forming the well 320 increases, while at the same time, the thickness of the well 320 can be increased such that the effective bandgap of the entire superlattice of the sublayer 230b can decrease. In another case, the composition of the material forming the well 320 of the sublayer 230b can be changed such that the bulk bandgap of the material forming the well 320 decreases, while at the same time, the thickness of the well 320 can be decreased such that the effective bandgap of the superlattice of the sublayer 230b can decrease. Similar to the above cases, by adjusting the composition and thickness of the well 320 and / or the barrier 325 of the sublayer 230b and the total thickness of the sublayer 230b, the performance degradation due to increased absorption in the sublayer 230b can be offset by improved carrier injection, and the overall output power of the light emitted from the structure can be approximately the same or even improved compared to a structure in which the well 320 and / or the barrier 325 of the sublayer 230b has the same composition and thickness as the wells 310 and 330 and / or the barriers 315 and 335 of the first sublayer 230a and the third sublayer 230c.

[0043] In some embodiments, the LED structures described herein (e.g., having a structure similar to structure 200 of FIG. 2) emit incoherent light having a wavelength of less than 300 nm, or from 200 nm to 300 nm. In some embodiments, the LED structures described herein have a total wall plug efficiency (i.e., the ratio of optical output power to electrical input power consumed) of from 0.1% to 90%. In some embodiments, the LED structures described herein have a total output power of from 0.1 mW to 1 W. In some embodiments, the LED structures described herein have a turn-on voltage and / or an operating voltage of from 3 V to 10 V, or from 3 V to 20 V. In some embodiments, the LED structures described herein have an operating voltage that is less than 60%, or less than 50%, or less than 40%, or less than 35%, or less than 30% of the operating voltage of a similar LED structure that does not incorporate an embedded n-contact layer. In some embodiments, the LED structures described herein have a turn-on voltage that is less than 60%, or less than 50%, or less than 40%, or less than 35%, or less than 30% of the turn-on voltage of a similar LED structure that does not incorporate an embedded n-contact layer. In some embodiments, the LED structures described herein have an output power that is greater than 100%, or greater than 80%, or greater than 60%, or greater than 40%, or greater than 20%, or greater than 10% improved compared to the output power of a similar LED structure that does not incorporate an embedded n-contact layer. In some embodiments, the LED structures described herein have a wall plug efficiency (WPE) that is greater than 1000%, greater than 750%, or greater than 500%, or greater than 400%, or greater than 300%, or greater than 200%, or greater than 100%, or greater than 50% improved compared to the WPE of a similar LED structure that does not incorporate an embedded n-contact layer. In some embodiments, the output power of the LED structures described herein is approximately the same as or less than the output power of a similar LED structure that does not incorporate an embedded n-contact layer, but the WPE of the structure is improved by the improvement in the operating voltage.

Example

[0044] Figures 4A and 4B show simplified diagrams of two UVC LED structures 401 and 402, respectively, having an epitaxial layer manufactured using MBE, a mesa etched using dry etching, and metal contacts deposited by evaporation. Both Figures 4A and 4B included a substrate 410, a buffer layer 420, an active layer 440, an EBL layer 450, a p-contact layer 460, an n-metal 470, and a p-metal 480. However, Figure 4A shows a baseline structure (i.e., a control structure) incorporating a single n-contact layer 435, and Figure 4B shows an improved structure incorporating a set of doped layers 430. The layer 430 included a first sub-layer 430a, a second sub-layer 430b (functioning as an embedded n-contact layer 430b), and a third sub-layer 430c. Both structures 401 and 402 emit UV light of approximately 233 nm through the substrate rather than through the p-side of the structure. This is advantageous because the emitted light is hardly absorbed or blocked by the metal contacts 470 and 480.

[0045] In both structures 401 and 402, the substrate 410 was sapphire and the buffer layer 420 was a 400 nm thick AlN buffer layer. Both structures 401 and 402 also incorporated the same active layer 440, which was a 50 nm thick layer of undoped (i.e., no dopants were intentionally added during layer growth) short period superlattice (SPSL). The SPSL of the active layer 440 included repeating pairs of GaN wells 444 and AlN barriers 442, where the GaN wells 444 and AlN barriers 442 were of a thickness such that the active layer 440 in both structures 401 and 402 emitted UV light with a wavelength of approximately 233 nm. The EBL layer 450 in both structures 401 and 402 was a 20 nm thick chirped SPSL EBL, which included GaN wells and AlN barriers, and the thickness of the GaN wells varied linearly across the entire layer 450. The p - contact layer 460 in both structures 401 and 402 was a 40 nm thick GaN layer p - doped with Mg to provide sufficient holes. The n - and p - metal contact layers 470 and 480 in both structures 401 and 402 were Ti / Al layers, which were deposited after the LED mesa structure was created by etching the upper layers of the structure (i.e., any part of 460, 450, 440, and 435 or 430c).

[0046] The n - contact layer 435 of structure 401 was a 400 nm thick SPSL with alternating GaN wells 434 and AlN barriers 432. In this example, the GaN wells 434 and AlN barriers 432 were the same thickness as the GaN wells 444 and AlN barriers 442, respectively, in the active layer 440. The n - contact layer 435 was doped with Si to make the layer conductive. The sheet resistance of the n - contact layer 435 was from about 5x10 4 ohms / square to about 2x10 5It was ohm / square. As will be described later, the relatively high sheet resistance through the n-contact layer 435 resulted in a relatively high turn-on voltage and drive voltage necessary to achieve a current injection of 100 mA. It should be noted that the n-contact layer 435 has the same composition and thickness of SPSL wells and barriers as the active layer. However, the n-contact layer 435 does not absorb a significant amount of light from the active layer. This is because the emission energy is higher than the energy at the absorption edge of such SPSL.

[0047] The set 430 of doped layers within the structure 402 includes three SPSL sub-layers: a 400-nm-thick SPSL sub-layer 430a, a 50-nm-thick sub-layer 430b (i.e., the n-contact layer 430b), and a 30-nm-thick SPSL sub-layer 430c, each of which includes GaN wells and AlN barriers. The set 430 of doped layers is also n-type doped with all Si. In the improved structure 402, the n-contact layer 430b was made more conductive by varying the thicknesses of the GaN wells and AlN barriers compared to the well and barrier thicknesses of sub-layers 430a and 430c. In layers 430a and 430c, the thickness of the GaN well 434 and the thickness of the AlN barrier 432 were the same as those of the n-contact layer 435 in FIG. 4A. However, in layer 430b, the thickness of the GaN well 438 was approximately four times thicker than the thickness of the GaN well 434 in layers 430a and 430c (and 435 in FIG. 4A) in FIG. 4B. The thickness of the AlN barrier 436 was approximately half the thickness of the AlN barrier 432 in layers 430a and 430c (435 in FIG. 4A) in FIG. 4B. Since the thickness of the GaN well 438 in sub-layer 430b is wider and the thickness of the AlN barrier 436 is narrower compared to the thicknesses of sub-layers 430a and 430c, sub-layer 430b is made more conductive than sub-layers 430a and 430c. The sheet resistance of sub-layer 430b was from about 100 ohm / square to about 300 ohm / square, or had a conductivity about two to three orders of magnitude higher than that of sub-layers 430a and 430c.

[0048] Compared with the design of the n-contact layer 435 of the structure 401, the improved design of the n-contact layer 430b of the structure 402 significantly reduced the turn-on voltage and drive voltage required to achieve a current injection of 100 mA, as described below. However, the wider well 438 and narrower barrier 436 of the n-contact layer 430b of the structure 402 compared to those of the n-contact layer 435 of the structure 401 caused the n-contact layer 430b to absorb a significant amount of the light emitted from the active layer 440, as shown in the following data. Surprisingly, as shown in the following data, the output power from the structure 402 was also improved compared to the output power of the structure 401. Without being limited to the theoretical aspect, the thickness of the n-contact layer 430b, as well as the thicknesses of the GaN well 438 and AlN barrier 436 within the layer, were adjusted such that the reduced resistance of the n-contact layer 430b improved more than the absorption within the layer decreased the output power, as further discussed below, and thus the output power may have been improved. In other words, even though the sub-layer 430b absorbed light from the active layer, the improved structure of the n-contact layer 430b improved the turn-on voltage, the drive voltage during operation, and the power output from the structure 402 compared to those of 401.

[0049] Figure 5A shows the measured absorption spectra 510 from SPSLs of different total thicknesses, each having the same structure as the n-contact layer 430b of FIG. 4B (i.e., having GaN wells 438 and AlN barriers 436). A 405 nm thick SPSL was grown on a sapphire substrate with a thick AlN buffer layer. The thick SPSL was sequentially measured, thinned by etching, and measured again to generate seven absorption spectra 510. The y-axis of the plot in FIG. 5A is the absorption rate, and the x-axis is the wavelength in nm. The thicknesses of the SPSL samples from which the absorption spectra 510 were generated were 71 nm, 97 nm, 151 nm, 253 nm, 301 nm, 352 nm, and 405 nm. As indicated by arrow 520, the absorption of the thicker SPSL layer at 233 nm was higher. FIG. 5A also shows the absorption spectra 530 of three different thicknesses of AlN layers, which shows that the AlN layers (having the same thickness as those used in the SPSL measured to generate the absorption spectra 510) absorb about 10% of the light at 233 nm. FIG. 5A also shows the absorption data 540 of the sapphire substrate, which shows that the substrate absorbs a minimum amount of light at 233 nm.

[0050] Figure 5B shows the absorption data 550 of the SPSL measured in FIG. 5A at a wavelength of 233 nm and the polynomial fit 555 of the absorption data as a function of thickness. The y-axis of the plot in FIG. 5B is the absorption rate of the SPSL, and the x-axis is the thickness of the SPSL in nm. To obtain the absorption data 550 that is only the absorption of the SPSL, the absorption from the thick AlN layer (i.e., about 10% at 233 nm) and the substrate was removed from the experimental absorption data at 233 nm shown in FIG. 5A. FIG. 5B shows that the estimated absorption of such SPSL can be reduced from about 70% absorption at a 300 nm SPSL thickness to about 30% absorption associated with a 50 nm SPSL thickness.

[0051] Figures 6A - 6E show experimental data from structures 401 and 402, demonstrating the effect of the embedded n - contact layer 430b in reducing the driving voltage and improving the output power. The lateral dimensions (i.e., parallel to the surface of the substrate) of the tested structures were 930 microns × 930 microns.

[0052] Figure 6A shows the emission spectra from structures 401 and 402, including three spectra 610 from three structures 401 (without the embedded n - contact layer) and three spectra 620 from three structures 402 (with the embedded n - contact layer 430b). The y - axis is the spectral flux in units of W / nm, and the x - axis is the wavelength in nm. All of the spectra 610 and 620 have similar peak emission wavelengths. Surprisingly, the spectra 620 from structure 402 (with the embedded n - contact layer 430b) show a higher output intensity compared to the spectra 610 from structure 401 (without the embedded n - contact layer).

[0053] Figure 6B shows the output power 630 from three structures 401 (without the embedded n - contact layer) and the output power 635 from three structures 402 (with the embedded n - contact layer 430b). The y - axis is the output power in mW, and the x - axis is the radius in mm of the position on the wafer where the structure was tested. This plot shows that the incorporation of the embedded contact layer with higher conductivity as well as higher light absorption into the UVC LED structure improves the output power from about 0.8 mW to about 1.2 mW or about 1.3 mW, which is an improvement of about 60%.

[0054] FIG. 6C shows the peak emission wavelengths of 640 from three structures 401 (without an embedded n-contact layer) and 645 from three structures 402 (with an embedded n-contact layer 430b). The y-axis is the peak emission wavelength in nm, and the x-axis is the radius in mm of the position on the wafer where the structures were tested. All peak emission wavelengths are between 233 nm and 235 nm. The variation of the peak emission 640 compared to 645 (i.e., less than 2 nm) is consistent with the typical variation observed between different epitaxial growths of semiconductor structures having the same design, indicating that the embedded contact layer does not have a significant impact on the generated wavelength. Also, in this example, it should be noted that the device having the embedded n-contact layer 430b has a brighter emission (as shown, for example, in FIG. 6B), even though the peak emission wavelength 645 is shorter than the peak emission wavelength 640, and that the higher output power is due to the different n-contact designs and not an artifact due to the wavelength difference.

[0055] FIG. 6D shows the current-voltage (IV) responses 650 from three structures 401 (without an embedded n-contact layer) and the IV responses 655 from three structures 402 (with an embedded n-contact layer 430b). The y-axis is current in amperes and the x-axis is voltage in volts. Note that the current in the IV responses 655 from the three structures 402 saturated at the 200 mA current compliance limit of the measurement hardware beyond a voltage of about 7V. FIG. 6E shows the drive voltage 660 required for three structures 401 (without an embedded n-contact layer) and the drive voltage 665 required for three structures 402 (with an embedded n-contact layer 430b) to reach an injection current of 100 mA. The y-axis is the drive voltage (at 100 mA) in volts and the x-axis is the radius in mm of the position on the wafer where the structures were tested. FIGS. 6D through 6E show that the insertion of the embedded layer decreased the drive voltage at 100 mA (i.e., the voltage required to drive the device at 100 mA) from about 20V to about 6.5V. Due to the improved output power and lower operating voltage, the WPE increased from about 0.04% in structure 401 (FIG. 4A) to about 0.18% in structure 402 (FIG. 4B) with an embedded n-contact layer 430b, which was an increase of about 400%.

[0056] While reference has been made in detail to embodiments of the disclosed invention, one or more examples thereof are shown in the accompanying drawings. Each example is provided to illustrate the technology and not to limit it. In fact, although this specification has described in detail specific embodiments of the invention, it is understood that those skilled in the art, having obtained the foregoing understanding, will readily conceive of variations, modifications, and equivalents to these embodiments. For example, features illustrated or described as part of one embodiment may be used in another embodiment to yield yet another embodiment. Accordingly, it is intended that the subject matter cover such variations and modifications that fall within the scope of the appended claims and their equivalents. These and other modifications and variations to the invention may be practiced by those skilled in the art without departing from the scope of the invention as specifically set forth by the appended claims. Further, those skilled in the art will understand that the foregoing description is for purposes of illustration only and is not intended to limit the invention.

Claims

1. A light-emitting structure comprising: a first set of doped layers, a second layer, a light-emitting layer disposed between the first set of doped layers and the second layer, and a layer having an electrical contact to the first set of doped layers, wherein the first set of doped layers, the second layer, and the light-emitting layer comprise a semiconductor material; the first set of doped layers comprises a first sub-layer, a second sub-layer, and a third sub-layer, and the third sub-layer is adjacent to the light-emitting layer; the first, second, and third sub-layers each comprise a first, second, and third superlattice, respectively; the well layer of the second superlattice is thicker than the well layers of the first and third superlattices; the barrier layer of the second superlattice is thinner than the barrier layers of the first and third superlattices; the electrical contact to the first set of doped layers is made to the second sub-layer; the first, second, and third sub-layers are doped n-type; the electrical conductivity of the second sub-layer is higher than the electrical conductivities of the first and third sub-layers, the layer stack; the light-emitting layer comprises a fourth superlattice; the second layer comprises a fifth superlattice; the fifth superlattice is a chirped superlattice having a varying well layer thickness, a varying barrier layer thickness, or a varying well layer and barrier layer thickness through the fifth superlattice, the light-emitting structure.

2. Light having a wavelength shorter than 300 nm emitted from the light-emitting layer passes through the first set of doped layers before being emitted from the light-emitting structure, and the second sub-layer absorbs 10% to 60% of the light emitted from the light-emitting layer that reaches the second sub-layer, the light-emitting structure according to claim 1.

3. the well layer of the second superlattice comprises a material having a lower bandgap than the well layers of the first and third superlattices, the light-emitting structure according to claim 1.

4. the first, second, and third sub-layers each have a constant effective bandgap through the sub-layer, the light-emitting structure according to claim 1.

5. the first, second, and third sub-layers each have a varying effective bandgap through the sub-layer, the light-emitting structure according to claim 1.

6. the first, second, third, fourth, and fifth superlattices each comprise a set of GaN well layers and AlN barrier layers, the light-emitting structure according to claim 1.

7. At least one of the first, second, third, fourth, and fifth superlattices includes Al ｘ Ga １－ｘ N, and x is 0 or more and 1 or less. The light-emitting structure according to claim 1.

8. The first, second, third, fourth, and fifth superlattices each include Al ｘ Ga １－ｘ N, where x is 0 or more and 1 or less. The light-emitting structure according to claim 1.

9. The light-emitting structure according to claim 1, wherein at least one of the first, second, third, fourth, and fifth superlattices comprises InAlGaN.

10. The light-emitting structure according to claim 1, wherein each of the first, second, third, fourth, and fifth superlattices comprises InAlGaN.

11. The light-emitting structure according to claim 1, wherein at least one of the first, second, third, fourth, and fifth superlattices comprises a three-layer unit cell.

12. The three-layer unit cell includes AlN, Al ｘ Ga １－ｘ N, and GaN, and the light-emitting structure according to claim 11.

13. The three-layer unit cell is AlN, Al ｘ Ga １－ｘ N, and Al y In z Ga 1-y-z N, and the light-emitting structure according to claim 11.

14. The light-emitting structure according to claim 1, wherein the second layer has a thickness of 5 nm to 50 nm.

15. The light-emitting structure according to claim 1, wherein the chirped superlattice is a p-type chirped superlattice.

16. Comprising a substrate coupled to the first set of doped layers, The light-emitting structure according to claim 1, wherein the substrate comprises sapphire, SiC, AlN, GaN, silicon, or diamond.

17. Further comprising a second electrical contact coupled to the second layer, The light-emitting structure according to claim 1, wherein the second electrical contact comprises Ti, Al, Ta, and / or Ni.

18. The light-emitting structure according to claim 1, wherein the electrical contact comprises Ti, Al, Ta, and / or Ni.

19. The well layer of the second superlattice is thicker than the well layers of the first and third superlattices, The light-emitting structure according to claim 1, wherein the barrier layer of the second superlattice is thinner than the barrier layers of the first and third superlattices.

Images