Optoelectronic device and method of processing optoelectronic device

Abstract

The invention relates to an optoelectronic device comprising a mesa etch layer stack comprising a first layer doped with a first doping type, a second layer doped with a second doping type, and an active layer arranged between the first doped layer and the second doped layer, where the active layer comprises at least a quantum well. The mesa etch sidewalls of the active layer include a plurality of surface states that facilitate non-radiative recombination of charge carriers introduced into the active layer. An electrode is disposed on at least a sidewall of the active layer, the electrode comprising a dielectric layer covering the surface state, a conductive material on the dielectric layer, where the electrode is configured to change a density of the surface state when a voltage is applied to the conductive material to facilitate non-radiative recombination.

Description

[0001] This application claims priority to German patent application DE 10 2023 113 574.0, filed on May 24, 2023, the disclosure of which is incorporated herein by reference in its entirety. The present invention relates to optoelectronic devices and methods for operating such optoelectronic devices. Background Technology

[0002] Optical data communication is becoming an increasingly interesting field of interest, not only for transmitting information over long distances but also for transmitting information between hardware modules that are relatively closely spaced, ranging from tens of centimeters to about 10 meters apart. Beyond these new applications, optical data communication itself remains a growing market with new and evolving requirements regarding switching times to improve data transmission speeds.

[0003] Therefore, very high modulation speeds, up to the GHz range, are required for the transmitter's optical output to meet these demands. More lenient requirements are found in display applications using pulse-width modulation (PWM), although modulation speeds still increase to the tens to hundreds of kHz range. Due to bandwidth requirements, lasers are most often the choice of transmitter. VCSELs and edge-emitting lasers offer high switching speeds and relatively narrow bandwidths at reasonable power consumption.

[0004] However, light emitters tend to become smaller in size, with lengths in the range of less than 100 µm. As mentioned above, this has proven to be a design challenge for lasers. Furthermore, at the same illumination output, the current consumption in a laser tends to be higher than that of a conventional LED, where heat loss and heat generation become issues. Therefore, LEDs have been proposed as a replacement for lasers due to their smaller footprint and reduced power consumption requiring less cooling than lasers.

[0005] Furthermore, a major drawback of LEDs is related to their switching speed. Switching speed is primarily limited by the on-time and off-time (also known as rise and fall times). Lasers achieve on-times and off-times in the range of tens of ps, corresponding to hundreds of MHz to GHz, while conventional lighting diodes are much slower in terms of switching speed (within hundreds of kHz or µs). This slow switching speed is due to the carrier lifetime in the active layer, as charge carriers continue to fill the active layer even when the power is turned off.

[0006] Therefore, the purpose of this application is to propose measures to improve the switching speed of light-emitting diodes. Summary of the Invention

[0007] This and other objectives are addressed through the subject matter of the independent claims. Features and additional aspects of the proposed principles are outlined in the dependent claims.

[0008] Several methods have been employed to improve switching speed. Several methods primarily aim to reduce carrier lifetime, for example, by intentionally introducing impurities into the active layer. These impurities act as nonradiative recombination centers. Since the overall carrier lifetime is a combination of radiative and nonradiative recombination, one recombination path (in this case, nonradiative) can be increased at the expense of another. On the one hand, this effectively shortens the "afterglow" of light after the illumination diode is turned off, thus reducing the turn-off switching time. A shorter carrier lifetime may also result in faster turn-on switching due to the increasing rate of carrier density accumulation in the active layer. However, shortening the carrier lifetime by increasing the nonradiative recombination rate is accompanied by a decrease in internal quantum efficiency (IQE), and therefore a loss in overall efficiency.

[0009] An alternative approach is to design a refined drive scheme that includes a current spike or reverse bias at the start of the pulse to accelerate optical attenuation. This typically requires a complex driver design that demands a larger footprint and higher drive cost.

[0010] The inventors now propose a third method: dynamically controlling the mesa etching of nonradiative recombination at the sidewalls of an optoelectronic device by applying an electric potential. The mesa etching sidewalls of an optoelectronic device comprise multiple surface states, which are typically defect states leading to nonradiative recombination of charge carriers, thus reducing the overall internal quantum efficiency of the device. In conventional devices, the amount of such surface states is typically reduced and minimized by appropriate post-processing after mesa etching (e.g., passivation with a dielectric layer). The inventors now propose covering the sidewalls with an electrode that can be used to bend the semiconductor band at the mesa sidewalls via the applied potential. This applied potential can have two distinct effects.

[0011] In materials exhibiting rapid nonradiative surface recombination, such as in certain optoelectronic devices based on phosphide materials, the bending of the energy bands by an applied potential leads to the depletion of charge carrier types (depending on the polarity of the applied potential), thus reducing or virtually eliminating surface recombination. Therefore, tuning such a potential allows for increasing or decreasing (or even effectively “turning off”) nonradiative recombination at the surface. The resulting changes in carrier lifetime can be used to increase or decrease rise and fall times when switching optoelectronic devices or modulating their light emission.

[0012] This not only allows for a reduction in overall switching time, but also allows for individual and separate control of the on-time and off-time of the pulse-modulated optoelectronic devices at their rising and falling edges. Furthermore, this method avoids efficiency losses by “turning off” non-radiative surface recombination during the diode's on-state.

[0013] In a second concept, applicable to both phosphide-based optoelectronic devices and other material systems, bending the bandgap by applying a corresponding potential through electrodes allows for filling the interface at the sidewalls with a type of charge carrier, thus creating a leakage path capable of carrying significant current. The bandgap bending at the interface due to the potential acts as a quasi-ohmic path, leading to a significantly faster reduction of the active layer compared to normal radiative recombination, resulting in a significantly faster switching time.

[0014] By selectively activating additional current conduction mechanisms, such as those proposed in this application, the aforementioned concept allows for faster turn-on and turn-off times of optoelectronic devices, and thus reduces the rise and fall times of the emitted light pulses. This will provide a compact, low-cost, and efficient light source capable of rapid switching operations. The implementation using mesa etching of sidewalls is similar for both concepts and can be applied using various material systems such as phosphides, arsenides, nitrides, and others.

[0015] In some aspects, the inventors have proposed optoelectronic devices comprising a mesa-etched layer stack. The mesa-etched layer stack includes a first layer doped with a first doping type, such as a p-doped layer. The stack also includes a second layer doped with a second doping type, such as an n-doped layer. The mesa-etched layer stack according to the proposed principle further includes an active layer disposed between the first and second doped layers.

[0016] Two differently doped layers may comprise one or more sublayers with different doping concentrations or distributions. Various sublayers can provide different functions, including but not limited to current distribution and current transport. Different doping distributions and concentrations can be used to achieve rapid charge carrier transport and injection into the active layer. In this respect, the active layer comprises at least one quantum well, but may also comprise a multi-quantum-well structure having multiple alternating barrier layers and quantum well layers, respectively. Alternatively, the active region may comprise a bulk layer.

[0017] The mesa-etched sidewalls of the active layer now comprise multiple surface states that facilitate nonradiative recombination of charge carriers introduced into the active layer. These surface states are located at the mesa-etched sidewalls of the active layer, directly adjacent to or near the sidewalls. However, according to the proposed principle, the nonradiative recombination rate at the surface can be altered and tuned by adjusting the electron and hole densities. This, in turn, will change the effective carrier lifetime. Therefore, variations in the surface recombination rate are used to influence the rising and falling edges of optical pulses, and thus rapidly turn on or off, or modulate, the light emission of the optoelectronic device.

[0018] For this purpose, the optoelectronic device also includes electrodes disposed at least on the sidewalls of the active layer. The electrodes comprise a dielectric layer (or a stack of different layers) covering the sidewalls of the active layer (and more specifically, a significant portion of the etched surface). A conductive material is disposed at at least some portions of the dielectric layer stack. The electrodes are configured to alter the nonradiative recombination rate when a voltage is applied to the conductive material.

[0019] Depending on the design choices, at least one of the concepts mentioned above can be implemented. This flexibility allows for optimization of the rise and fall edges of the light pulses emitted by the optoelectronic device. In addition to simple switching, as in pulse width modulation applications, modulation speed can be increased, as in amplitude modulation, based on the proposed principles.

[0020] The proposed idea of ​​an additional current path at the surface of a dynamically modulated mesa can be applied to both large-signal and small-signal modulation, meaning not only completely turning optoelectronic devices on or off, but also simply modulating their light emission or brightness. This idea can also be applied to various chip designs, including thin-film or µ-LEDs.

[0021] In this regard, the description of mesa-etched optoelectronic devices includes stacked optoelectronic devices, the stack including sidewalls in the form of a cube, cylinder, truncated cone, or truncated pyramid. The sidewalls may be perpendicular to the base laser and included relative to the normal axis. The tilt angle may depend on the etching method and its parameters, as well as on the crystal parameters.

[0022] For example, an epitaxial growth process is used to deposit a stack of optoelectronic devices comprising different doped layers onto a growth surface, with multiple quantum well structures between the different doped layers. After the layer stack is deposited, a corresponding structured hard mask layer is provided, and one or more etching steps are performed to etch pixel facets and sidewalls out of the layer stack. The etched pixel facets can then be passivated or otherwise processed, and in particular covered with a dielectric material that forms part of the electrodes.

[0023] In this respect, the dielectric material of the electrodes can be deposited directly onto the exposed sidewalls after the mesa etching process. Alternatively, a semiconductor layer can be regrown on the sidewalls, and the dielectric layer then covers the regrown layer. The sidewalls can also be treated after the mesa etching process to influence the density of surface states. This can be achieved, for example, by several curing and healing steps to minimize the amount of surface states, and then sequentially adding the defined amount back to the sidewalls. Such a process is advantageous because it results in a defined density of surface states and improved predictability. Such a process can also provide greater flexibility in certain material systems where the mesa etching process does not lead to a high density of surface states that promote nonradiative recombination.

[0024] The dielectric layer is fabricated to be as thin as possible, having a diameter of less than 100 nm, and particularly in the range of 20 nm to 80 nm. A highly conductive layer, such as a metal, can then be arranged on the dielectric layer. In some cases, the metal is gold, silver, or another highly reflective material. Depending on the geometry, the reflective layer improves light emission. A potential can be applied to this highly conductive material.

[0025] In some aspects, the active layer includes at least one cladding layer, particularly an undoped cladding layer. In some aspects, the undoped cladding layer is disposed between a quantum well or multiple quantum well structure and each of the doped layers in a stacked configuration. The undoped cladding layer typically prevents dopant from diffusing from the first and second doped layers into the active layer of the multiple quantum well structure.

[0026] In another embodiment, the mesa etched sidewalls of the active layer are covered by a regenerated layer, on which electrodes are disposed. In some aspects, the band gap of the regenerated layer is larger than the band gap of at least one quantum well of the active layer. Furthermore, the band gap of the regenerated layer may also be smaller than the band gap of at least one cladding layer.

[0027] In this regard, it may be useful to electrically couple the conductive material of the electrode to at least one of the first and second layers. This can be achieved using a switch or similar device. In some aspects, the electrode can be switched between a potential applied to the first layer and a potential applied to the second layer. In other words, the electrode can be switched to the anode potential during device operation. With proper design of the interface between the active layer and the electrode, the applied potential can facilitate band bending, thus reducing or increasing nonradiative recombination.

[0028] To further improve switching speed, the capacitance of the electrodes must be considered. One factor determining this capacitance is the area of ​​the conductive material forming the electrodes applied to the sidewalls. To minimize this parasitic capacitance, the total area of ​​the electrodes covering the sidewalls can be reduced. For example, in some cases, the electrodes are located only above the etched surface of the active layer, and optionally on its adjacent portion. In some other aspects, the conductive material of the electrodes is located only above the etched surface of the active layer, and optionally on its adjacent portion. Thus, the conductive material of the electrodes can extend only above the active layer. Alternatively, the conductive material of the electrodes can extend over an adjacent portion of the active layer and one of the first and second layers, wherein the portion of the first and second layers covered by the conductive material of the electrodes is less than 15% of the respective width of the first and / or second layers. Thus, the electrodes do not cover the entire sidewall of the mesa etched layer stack, but only a portion adjacent to the active layer.

[0029] In some aspects, the dielectric layer of the electrode comprises one of silicon nitride (SiNx) or silicon oxide (SiO2) with a thickness of less than 100 nm. The conductive material of the electrode may include a metal, and particularly gold. Alternatively, another highly reflective metal such as silver may be used.

[0030] The proposed concept is not limited to specific types of optoelectronic devices or basic material systems, but can be flexibly integrated into various designs and material systems. In some aspects, the material of the active layer, and particularly the material of at least one quantum well, can be based not only on phosphide material systems such as InP, GaP, AlP, GaInP, GaAlP, InAlP, and InGaAlP, but also on arsenide systems such as GaAs and AlGaAs. Furthermore, nitride-based material systems such as GaN, InGaN, InAlGaN, AlN, InAlN, and AlGaN can be used.

[0031] Some other aspects involve methods for processing optoelectronic devices based on the proposed principles.

[0032] For this purpose, a layer stack is formed on a growth substrate, and multiple additional layers, both doped and undoped, are deposited thereon. Depending on the base material, different types of growth substrates, including some buffer layers, can be used. The layer stack ultimately comprises a first layer doped with a first doping type, a second layer doped with a second doping type, and an active layer disposed between the first and second doped layers. The doped layers may include multiple sublayers, each with different doping concentrations and / or different doping distributions. The active layer includes at least one quantum well and may, for example, contain a quantum well or a multi-quantum well structure. To initiate an optoelectronic device containing a mesa, a structured hard mask is applied to the layer stack, and one or more mesa etching steps are performed to define mesa facets at least along the active layers of the layer stack. The mesa facets of the active layer include multiple surface states that facilitate nonradiative recombination of charge carriers introduced into the active layer.

[0033] Then, electrodes are deposited at least on the sidewalls of the active layer, wherein the electrode layer comprises a dielectric layer covering the surface states and a conductive material disposed on the dielectric layer. When a voltage or potential is applied to the conductive material, the resulting structure achieves band bending. Band bending alters the carrier density at the mesa sidewalls, thereby affecting the radiative-to-nonradiative recombination ratio. Therefore, by altering nonradiative recombination, the carrier density in the active region within the active layer can be injected or depleted significantly faster, thereby reducing the falling or rising edge of the light pulse emitted by the device.

[0034] Mesa etching steps typically provide multiple surface states, although the surface state density is random and somewhat unpredictable. Therefore, in some respects, it is appropriate to redefined multiple surface states to provide a defined surface state density, and in particular, a surface state density within a specific range.

[0035] For this purpose, the mesa facets can be cured after the etching process, thereby minimizing the amount of surface states. This step can be performed using a wet chemical process within the gas phase stage, employing an etchant such as KOH, or by any other suitable means. Multiple new surface states are then reintroduced onto the mesa asset, at least in the active layer, to provide a specific density of such surface states. Various methods can be used for this purpose, including, for example, ion etching and other methods.

[0036] Additional steps in one or more mesa etching steps may include a step of regrowing a semiconductor layer. The regrown semiconductor layer comprises a bandgap larger than the bandgap of the quantum well within the active layer, but optionally smaller than the bandgap of the cladding layer.

[0037] In some other aspects, electrode deposition involves depositing a conductive material on the dielectric layer at the location of the mesa facets covering the active layer, but not necessarily extending beyond the mesa facets of the active layer onto the p-doped or n-doped layer. This ensures that the actual electrode only covers a portion of the active layer on the sidewalls, thereby reducing parasitic capacitance. Attached Figure Description

[0038] Further aspects and implementations based on the proposed principles will become apparent relative to the various embodiments and examples described in detail with reference to the accompanying drawings, in which...

[0039] Figure 1 A first embodiment of an optoelectronic device based on some aspects of the proposed principles is shown;

[0040] Figure 2 A diagram illustrating some aspects of the proposed principle is shown;

[0041] Figure 3 Equivalent circuits of optoelectronic devices, including several current paths corresponding to different composite mechanisms, are shown to illustrate some additional aspects of the proposed principle.

[0042] Figure 4 The current-time plots show the switching behavior simulated based on some aspects of the proposed principles and the results.

[0043] Figure 5 Another current-time plot is shown, illustrating the switching behavior and its results as indicated by some aspects of the proposed principles.

[0044] Figure 6 Another current-time plot is shown, illustrating the switching behavior and its results as indicated by some aspects of the proposed principles.

[0045] Figure 7 A second embodiment of an optoelectronic device based on some aspects of the proposed principles is shown;

[0046] Figure 8 A third embodiment of an optoelectronic device based on some aspects of the proposed principles is shown;

[0047] Figure 9 A model of bandgap bending based on some other aspects of the proposed principle is shown;

[0048] Figure 10 Instructions are shown on some aspects of the proposed principles. Figure 9 The current-time curves of the simulated behavior of the model;

[0049] Figure 11 A method for operating an optoelectronic device based on the proposed principle is shown. Detailed Implementation

[0050] The following embodiments and examples disclose various aspects and combinations thereof based on the proposed principles. The embodiments and examples are not always to scale. Similarly, different elements may be shown enlarged or reduced in size to emphasize aspects. It goes without saying that the various aspects, embodiments, and examples shown in the figures can be combined with each other without contradiction to the principles of the invention. Some aspects illustrate regular structures or forms. It should be noted that in practice, minor differences and deviations from the ideal form may occur; however, these will not contradict the concept of the invention.

[0051] Furthermore, the various figures and aspects are not necessarily shown at the correct dimensions, and the proportions between the elements are not necessarily substantially correct. Some aspects are highlighted by showing them enlarged. However, terms such as "above," "over," "below," "under," "larger," and "smaller" are correctly used to represent the elements in the figures. Therefore, such relationships between elements can be inferred from the figures.

[0052] Figure 1An embodiment of an optoelectronic device 1 based on the proposed principles is shown. The optoelectronic device 1 includes a semiconductor layer stack 10 having a mesa-etched sidewall structure for forming a vertical µ-LED. The layer stack 10 includes a first n-side electrode made of conductive metal and a second p-side electrode of a reflective conductive layer to reflect light to the main emitting surface on the other side. The layer stack includes a plurality of n-doped layers 13, 13′, and 13′′. During processing of the optoelectronic device, the n-doped layers are deposited on corresponding growth substrates (not shown herein). During processing, a buffer layer of n-doped material is provided, and the various n-doped layers 13, 13′, and 13′′ are deposited on the buffer layer. Due to the growth parameters generated for forming defect-free surface regions, the thickness of the n-doped layers is typically greater than the thickness of the corresponding p-doped layers 11, 11′, and 11′′.

[0053] The dopant concentrations in the various n-doped layers 13, 13′, and 13′′ vary and depend on the desired function. For example, in the current case, a conductive n-electrode is applied to the highly n-doped layer 13′ for current distribution and charge carrier injection into the subsequently deposited transport layer 13. The top surface of the transport layer 13 also forms the bottom of the mesa etched surface of the device and is covered on its top surface by the material of the dielectric layer 41 forming part of the electrode. Another n-doped layer 13′′ is disposed on the current transport layer 13 with a similar but different doping concentration, which optionally includes a doping distribution for injecting charge carriers into the active layer 12.

[0054] The active layer 12 includes a very thin undoped cladding layer 120, which also includes a different bandgap compared to its adjacent layers. The thin cladding layer 120 serves to prevent the diffusion of dopant from layer 13′′ into the active layer 12. The active layer 12 also includes a multi-quantum-well structure with multiple alternating quantum well layers 123 and quantum barrier layers 124. The quantum well layers 123 typically include a smaller bandgap compared to the corresponding barrier layers. In some instances, the quantum well layers and barrier layers comprise a semiconductor material containing Al. The varying Al content then provides the necessary bandgap difference between the barrier layers and the quantum well layers.

[0055] Both layers are typically undoped, but may include small amounts of silicon or other dopants to improve carrier injection, transport, or radiative recombination. The active layer 12 includes another undoped cladding layer 121 adjacent to the p-doped layer stacks 11′, 11′, and 11′′. Similar to the n-doped layer structure, the p-doped layer structure includes various sublayers with different doping concentrations and / or different dopant distributions. These sublayers include carrier transport layers 11 and 11′′ and a current distribution layer 11′. A conductive layer 20, such as a highly doped semiconductor, is deposited adjacent to layer 11′. A p-electrode made of a transparent conductive material such as TCO is disposed thereon. Figure 1The layer stack shown is for illustrative purposes only and can be adjusted based on the desired emission wavelength, operating parameters and other characteristics.

[0056] like Figure 1 The optoelectronic device shown comprises a mesa-etched structure with sloping sidewalls extending from the p-electrode down to the top surface of the n-doped layer 13. During the mesa etching process, multiple distinct surface states are generated along the sidewalls, and particularly along and partially within the active layer 12. These surface states typically originate from various types of defects, dangling bonds, crystal mismatches, or small steps, as well as other factors leading to energy states within the bandgap of the multi-quantum-well structure, particularly provided by the quantum well layer 123. The surface states within the bandgap act as traps for charge carriers, primarily generating nonradiative recombination during device operation. The amount of surface states per surface region corresponds to the surface state density. Density is typically a design or processing parameter and depends on the etching steps and their parameters; however, the surface state density is adjustable and should be set within a defined range to improve the predictability of device operation.

[0057] The inventors now propose controlling nonradiative recombination caused by these surface states by actively bending the conduction and valence bands of the active layer near the surface using an electrode on which a dielectric layer 14 and a conductive metal layer 41 are deposited. Figure 1 In this embodiment, the sidewalls of the optoelectronic device are covered by a dielectric layer 40 from the top surface of layer 13 to the p-electrode. The dielectric layer material includes, for example, SiO2 or SiNx. The dielectric material 40 is generally transparent to light generated within the active layer. To improve emission toward the main emitting surface (where a top emitter is required), a reflective and conductive material 41 is deposited on the dielectric material.

[0058] The dielectric material 40 covers the surface states of the active layer, maintaining the fundamental density of such surface states. The thickness of the dielectric material is, for example, in the range of 20 nm to 80 nm. Thickness is one of the possible parameters that will determine the amount of band bending at the surface for a given voltage applied to the conductive material.

[0059] in this regard, Figure 2 A model of the bandgap of a quantum well layer close to a dielectric material is shown, where a subsequent gold layer is attached to the dielectric material as a conductive material. More specifically, Figure 2 A horizontal cut of the band distribution at a given interface of an InAlGaP-based optoelectronic device is presented, on which silicon nitride (SiNx) has been deposited as the dielectric material and gold as the conductive material layer. The conductive gold material is fixed to the potential of the n-contact, such as the ground potential. This results in the band bending shown, indicated by the slight curvature around the dashed region at the interface between the quantum well and the dielectric material.

[0060] Although the bending appears small in the model, it is adjustable by applying a correspondingly larger or smaller potential to the conductive material deposited on the dielectric material. Figure 2 Examples are provided where more possible choices can be made regarding the dielectric material, its thickness, and other parameters to tune the band bending at the interface between the quantum well and the dielectric material. This, along with the surface recombination rate at the mesa etching, affects the voltage required to achieve the desired target in a given electrode.

[0061] Figure 3 Equivalent circuit diagrams of the devices are shown to provide a deeper understanding of the different recombination processes. Figure 3 This includes simple schemes in which various composite mechanisms are represented by ideal voltage-dependent current sources, the corresponding current levels of which indicate the respective parts of the composite mechanism.

[0062] The first mechanism, SRH_volume, is a nonradiative recombination mechanism in the host material of the active layer, known as Shockley-Read-Hall recombination. This part dominates in low current density and large optoelectronic devices, but its contribution decreases as device size shrinks, favoring the second recombination mechanism. This mechanism is a nonradiative recombination caused by surface states, known as surface recombination (surface_rec). This recombination mechanism is essentially caused by defects at the mesa facets of the optoelectronic device and leads to a reduction in the effective lifetime of charge carriers during operation. The third mechanism is radiative recombination (rad_rec), and the last mechanism is Auger recombination (auger_rec), which is also nonradiative.

[0063] On the one hand, the total carrier lifetime is given by the combined effect of all these recombination mechanisms. In smaller devices, surface recombination may become dominant and effectively determine the carrier lifetime; the key factor is the ratio of the exposed surface area of ​​the active layer to its volume. On the other hand, the amount of emitted light depends on the ratio between radiative recombination processes and all non-radiative recombination processes. Therefore, in small devices, very high surface recombination ratios can be used to significantly reduce carrier lifetime and decrease emitted light intensity.

[0064] For the purpose of illustrating the various effects and the proposed principles, [the following has been used]. Figure 3 Spice simulations were performed using an equivalent circuit in AlGaInP material systems. Typical recombination parameters for AlGaInP material systems were used to define voltage-dependent current sources. In the simulations, surface recombination can be dynamically switched at least between on and off states, and more generally modulated between these two extremes. In response to the modulation of surface recombination, such modulation or switching alters the ratio between radiative and nonradiative recombination.

[0065] Figure 4 Simulation results are shown in the time-current plot for a given current pulse that switches from its off state to its on state at time 0. Without non-radiative surface recombination, approximately 16.6 ns (not shown here) is required before radiative recombination reaches a sufficiently high level (90% of the final level). This is the time required to establish a sufficient carrier density in the active region.

[0066] exist Figure 4 In the curve, surface recombination is initiated at t=0, leading to a significant amount of nonradiative recombination, a shortened carrier lifetime, and therefore a faster build-up of carrier density. Charge carriers supplied by the current pulse at t=0 recombine in the surface states without any significant amount of light emission. Therefore, during the dead time period from approximately 0 ns to about 15 ns, surface recombination absorbs most of the charge carriers, preventing them from radiative recombination.

[0067] At approximately 15 ns, surface recombination is deactivated by applying a corresponding third potential. Therefore, radiative recombination now begins to increase significantly and quite rapidly. This rapid rise is supported by the already filled active layer. Compared to conventional solutions where no surface recombination occurs, the speed improvement is in the range of approximately 17 ns to 12 ns. The 12 ns rise time is approximately 30% faster than the rise time of the conventional solution.

[0068] Therefore, compared to conventional solutions, the increase in radiative recombination during the rising edge is significantly faster, thus reducing the turn-on time. Furthermore, reducing the surface state density by applying an appropriate potential improves the internal quantum efficiency, thereby increasing the total illumination output while maintaining the total power consumed. Nevertheless, as... Figure 4 As shown, due to activated surface recombination, the potential decreases rapidly at the rising edge of the optical pulse, which is the initial dead time. However, this time is known and can be compensated for by appropriately designing the turn-on and turn-off signals.

[0069] in this regard, Figure 5 Another simulation example is shown, in which the current through the optoelectronic device is always activated, and the modulation of the light output is performed only by applying a corresponding third potential. Figure 5 As shown, the current pulse is continuously activated with approximately 5 µA, and the surface recombination is periodically activated and deactivated, generating pulses approximately every 50 ns. The example shown corresponds to large-signal modulation, but the same concept can be applied to obtain small-signal modulation.

[0070] Therefore, the rise time of each pulse is approximately 10 ns to 12 ns, while the fall time is even faster, ranging from 1 ns to 2 ns. By applying a third potential at the electrodes to change the surface recombination ratio, the optoelectronic device is continuously supplied with a DC current and directly switched or modulated, providing fast rise and fall times or fast amplitude modulation of the corresponding pulses, respectively. Continuous current supply to the optoelectronic device allows for simpler driver design options but also results in higher overall power consumption. In an alternative implementation, the current through the device can be turned on slightly earlier than the deactivation of surface recombination to ensure a fast rise time. When the optoelectronic device is turned off by surface recombination activation (e.g., approximately 50 ns or 150 ns in the graph), the total DC current through the device can also be turned off shortly after surface recombination has been activated. This reduces current consumption to an acceptable level without compromising switching time or increasing heat generation in the device.

[0071] Activating surface recombination to reduce the fall time of the optical pulse also behaves as... Figure 6 As shown in the diagram, surface recombination activated at the end of the pulse significantly increases nonradiative recombination, thereby very rapidly clearing the active layer and resulting in a greatly reduced optical decay time. Simultaneously, the DC current pulse is also set to 0, so the activation of surface recombination can be turned off shortly thereafter.

[0072] The proposed concept can be modified and altered based on the underlying material system and its physical properties, including parameters such as total carrier lifetime, diffusion length, and others. For the proposed concept, a potential for very fast turn-off and therefore very short fall-off appears attractive. The reduction in turn-on time, as shown above, is useful if a small additional DC current consumption or compensation for the additional dead time in the device modulation is acceptable.

[0073] Another aspect related to switching speed is the capacitance of the electrodes disposed on the sidewalls of the optoelectronic device. However, applying a potential to the electrodes requires adding or subtracting certain charge carriers from the corresponding electrodes. Therefore, the electrodes, together with the dielectric material, create a parasitic capacitance that needs to be charged or discharged to switch nonradiative recombination. Therefore, this parasitic capacitance should be minimized to avoid delays in switching speed. Since most surface recombination occurs on the mesa facets of the active layer, the inventors propose restricting the location of the corresponding electrodes to the sidewalls of the active layer.

[0074] Figure 7 An exemplary embodiment of an optoelectronic device is shown, in which electrodes are substantially adjacent to the active layer 12. Identical elements in optoelectronic devices generally have the same reference numerals.

[0075] In this embodiment, the optoelectronic device includes electrodes having a dielectric material 40 and a conductive material 41. While the dielectric material 40 is substantially disposed over the entire sidewall of the various surface states 30 covering the active layer and doped layers, the conductive material 41 is supplied only over the regions corresponding to the sidewalls of the active layer 12 and the cladding layers 121 and 120, respectively. In some instances, the conductive material 41 of the third electrode may also extend slightly over the leading layer of the active layer 12 into the p-doped and n-doped layers without completely covering the sidewalls of these layers. The reduced size of the electrode material 41 lowers the parasitic capacitance associated with the electrode, thereby allowing for faster capacitance changes and thus allowing for faster switching or modulation of radiative recombination.

[0076] Figure 8 Another implementation of the proposed principle involving a concept known as a channel leakage concept is shown. After structuring and etching, layers are stacked to form sloping sidewalls. As... Figure 8 As shown, a thin semiconductor layer 42 can be regrown on the corresponding sidewall of the regenerated length.

[0077] The regrown semiconductor layer 42 has a thickness ranging from tens to hundreds of nanometers and a band gap larger than that of the active layer and the multi-quantum-well structure therein, particularly larger than the band gap of the corresponding multi-quantum-well. However, the band gap of the regrown semiconductor layer 42 can be lower than that of the corresponding cladding layers 120 and 121, respectively. This will reduce the turn-on voltage of the leakage channel compared to the turn-on voltage of the active region, thus requiring a lower potential in the third electrode to open the leakage channel. The dielectric material 40 of the electrode and the conductive layer material 41 in the form of a conductive metal are then deposited on the sidewalls of the mesa etched layer stack to form the electrode.

[0078] In some implementations, the bandgap potential bend is large enough to form a channel along the multi-quantum-well structure, thereby creating a leakage path capable of carrying a significant current of one type of charge carrier. The leakage channel created by the band bend acts as a quasi-ohmic path and significantly reduces the fall-edge time of the light pulse emitted by the optoelectronic device by sweeping one type of charge carrier out of the active layer.

[0079] Figure 9 The diagram illustrates the pathway created at the interface between the dielectric material of the electrode and the multi-quantum-well structure. As shown, due to band bending caused by the applied voltage, an electronic (or more generally, charge carrier) pathway is created at the interface, similar to the channel of a field-effect transistor passing through the n-doped and p-doped portions of the multi-quantum-well structure of the device. This causes charge carriers to overflow into the doped region without emitting light or recombinating in the active layer.

[0080] exist Figure 10The simulations show the results of such leakage current along the interface between the dielectric portion of the electrode and the active layer. During operation, the DC current continues to flow through the device. A leakage path is created to open the current channel at approximately 50 ns. Therefore, radiative recombination decreases significantly within an extremely short timeframe of a few nanoseconds. The additional current path in parallel with the active layer corresponds to a parallel quasi-ohmic path in which radiative recombination does not occur. If the resistance of the generated leakage path is low enough, the voltage within the active layer drops rapidly when the channel opens, and optical decay is accelerated due to the rapid depletion of charge carriers capable of radiative recombination in the active layer.

[0081] The suggested electrodes on the sidewalls can now be connected to a separate third potential, or via a switch to one of the n-electrodes or p-electrodes. Through appropriate dielectric layer design and selection of conductive materials, flat band conditions can be obtained at the surface of the active layer, corresponding to the maximum value of the nonradiative recombination mechanism when connected to one of the aforementioned electrodes. Therefore, when switching to the other electrode, the bandgap at the surface of the active layer exhibits increased band bending and decreased nonradiative recombination. This allows for tuning the on / off ratio of nonradiative recombination.

[0082] Figure 11 A method for processing optoelectronic devices is illustrated. In step S1, a growth substrate with a buffer layer deposited thereon is provided to provide a smooth and substantially defect-free surface. The buffer layer and the growth substrate are adjusted to meet the growth requirements of the optoelectronic device and the material system of the stacked layers deposited thereon.

[0083] Then, multiple layers with different doping types are deposited, wherein the doping concentration and doping distribution are adjusted to provide specific functions. Specifically, one or more first layers with different concentrations of a first doping type are deposited. Then, a small and thin cladding layer is deposited on the uppermost layer with the first doping type as part of the active layer. The cladding layer is used to prevent dopant from diffusing into the active region.

[0084] Then, multiple alternating quantum well and quantum barrier layers are deposited. The band gap of the quantum barrier layer is slightly larger than that of the quantum well layer, which can be achieved, for example, by slightly altering the composition of the underlying material system. The number of quantum well and quantum barrier layers can be varied, ranging from 2 to approximately 30 different layers. Another undoped coating layer is deposited on top of the last quantum well or quantum barrier layer.

[0085] In step S2 of the exemplary embodiment, a structured mask is applied to the layer stack. The structured mask is a hard mask layer that is substantially insensitive to the next step S3, in which one or more mesa etching steps are performed. The mesa etching process in step S3 removes the material of the layer stack exposed by the structured mask on the layer stack, thereby forming mesa facets along the sidewalls of the layer stack. More specifically, the mesa facets of the active layer are exposed by the mesa etching steps.

[0086] Mesa etching typically yields multiple distinct surface states that facilitate nonradiative recombination of charge carriers when introduced into the active layer during device operation. To confine the surface state density within a specific range, the surface of the mesa assets on the active layer is cured and annealed in step S4 to minimize the total number of surface states. After curing, multiple surface states are reintroduced to confine the surface state density within a defined and desired range.

[0087] In the final step S5, an electrode is deposited on the sidewall of the active layer, wherein the electrode comprises a dielectric layer covering the surface states and a conductive material on the dielectric layer. During device operation, the conductive material can be electrically connected to either the first or second layer, thereby providing a potential equal to that applied to one of the contacts in the layer stack. This avoids the need for a separate potential source for the third contact.

[0088] In an alternative implementation, the mezzanine asset may be covered with a material of the regrowth layer prior to the dielectric layer on which the electrodes are applied.

[0089] Reference List

[0090] 1 Optoelectronic devices

[0091] 10-layer stack

[0092] 11, 11′, 11′′ First doped layer

[0093] 12 active layers

[0094] 13, 13′, 13′′ Second doped layer

[0095] 123 quantum well layers

[0096] 124 quantum barrier layers

[0097] 20 current distribution layers

[0098] 30 surface states

[0099] 40 dielectric layers

[0100] 41 conductive layer

Claims

1. An optoelectronic device, comprising: - a mesa etch layer stack; the mesa etch layer stack comprising: a first layer doped with a first doping type; a second layer doped with a second doping type; and an active layer arranged between the first doped layer and the second doped layer, wherein the active layer comprises at least a quantum well; - wherein the mesa etch sidewalls, through or in the immediate vicinity of the active layer, comprise a plurality of surface states that contribute to non-radiative recombination of charge carriers injected into the active layer; - an electrode arranged at least on the sidewalls of the active layer, the electrode comprising: a dielectric layer covering the surface states; a conductive material on the dielectric layer; - wherein the electrode is configured to change the semiconductor band bending when a voltage is applied to the conductive material, thereby changing the non-radiative recombination rate.

2. The optoelectronic device of claim 1, wherein, The active layer comprises at least one cladding layer, in particular an un-doped cladding layer.

3. The optoelectronic device of any of the preceding claims, wherein, The mesa edge sidewalls of the active layer are covered by a regrowth layer, the electrode being arranged on the regrowth layer.

4. The optoelectronic device of claim 3, wherein, The bandgap of the regrowth layer is larger than the bandgap of at least one of the quantum wells of the active layer and, optionally, smaller than the bandgap of at least one of the cladding layers.

5. The optoelectronic device of any of the preceding claims, wherein, The conductive material of the electrode is electrically connected to at least one of the first and second layers.

6. The optoelectronic device of any of the preceding claims, wherein, The conductive material of the electrode extends only over the active layer and / or over less than 15% of the width of one of the first and second layers adjacent to the active layer.

7. The optoelectronic device of any of the preceding claims, wherein, The dielectric layer of the electrode comprises one of SiNx and SiO2 and / or the conductive material of the electrode comprises a metal, in particular Au.

8. The optoelectronic device of any of the preceding claims, wherein, The dielectric layer of the electrode is less than 100 nm, in particular less than 60 nm.

9. The optoelectronic device of any of the preceding claims, wherein, The material of the active layer, in particular the material of the at least one quantum well, is based on InP, GaP, AlP, AlGaP, InGaP, AlGaInP, GaAs, AlGaAs.

10. A method of processing an optoelectronic device, comprising: - providing a layer stack, the layer stack comprising: a first layer doped with a first doping type; a second layer doped with a second doping type; and an active layer arranged between the first doped layer and the second doped layer, wherein the active layer comprises at least a quantum well; - applying a structured mask on the layer stack; - performing one or more mesa etching steps to define a mesa facet at least along the active layer of the layer stack, wherein the mesa facet of the active layer comprises a plurality of surface states that contribute to non-radiative recombination of charge carriers injected into the active layer; - depositing an electrode at least on the sidewalls of the active layer, the electrode layer comprising: a dielectric layer covering the surface states; a conductive material on the dielectric layer.

11. The method of claim 10, wherein, The step of performing one or more mesa etching steps comprises: - redefining the plurality of surface states to provide a density of surface states in a specific range.

12. The method according to claim 11, comprising the steps of: - curing the mesa facet, thereby minimizing the amount of surface states; - reintroducing surface states on the mesa facet of at least the active layer.

13. The method of any one of claims 10-12, wherein, The step of performing one or more mesa etching steps comprises: - regrowing a semiconductor layer, the semiconductor layer comprising a bandgap that is larger than the bandgap of a quantum well within the active layer.

14. The method of any one of claims 10 to 12, wherein, The step of depositing an electrode comprises: - depositing a conductive material that covers essentially only the mesa facet of the active layer.

Images