Search

This tutorial paper focuses on the physical origin of thermal droop, i.e., the decrease in the luminescence of light-emitting diodes (LEDs) induced by increasing temperature. III-nitride-based LEDs are becoming a pervasive technology, covering several fields from lighting to displays, from automotive to portable electronics, and from horticulture to sensing. In all these environments, high efficiency is a fundamental requirement, for reducing power consumption and system cost. Over the last decade, a great deal of effort has been put in the analysis of the efficiency droop, the decrease in LED internal quantum efficiency (IQE) induced by high current density. On the other hand, an IQE decrease is observed also for increasing temperature, a phenomenon usually referred to as thermal droop. For commercial LEDs, the IQE decrease related to thermal droop can be comparable to that of efficiency droop: for this reason, understanding thermal droop is a fundamental step for making LEDs capable of operating at high temperature levels. In several fields (including street lighting, automotive, photochemical treatments, projection, entertainment lighting, etc.), compact and high-flux light sources are required: typically, to reduce the size, weight, and cost of the systems, LEDs are mounted in compact arrays, and heat sinks are reduced to a minimum. As a consequence, LEDs can easily reach junction temperatures above 85–100 °C and are rated for junction temperatures up to 150–175 °C (figures from commercially available LED datasheets: Cree XHP70, Osram LUW HWQP, Nichia NVSL219CT, Samsung LH351B, and LedEngin LZP-00CW0R) and this motivates a careful analysis of thermal droop. This paper discusses the possible physical causes of thermal droop. After an introduction on the loss mechanisms in junctions, we will individually focus on the following processes: (i) Shockley–Read–Hall (SRH) recombination and properties of the related defects; (ii) Auger recombination and its temperature dependence, including the discussion of trap-assisted Auger recombination; (iii) impact of carrier transport on the thermal droop, including a discussion on carrier delocalization, escape, and freeze out; (iv) non-SRH defect-related droop mechanisms. In addition, (v) we discuss the processes that contribute to light emission at extremely low current levels and (vi) the thermal droop in deep ultraviolet LEDs, also with reference to the main parasitic emission bands. The results presented within this paper give a tutorial perspective on thermal droop; in addition, they suggest a pathway for the mitigation of this process and for the development of LEDs with stable optical output over a broad temperature range. [ABSTRACT FROM AUTHOR]

We present a detailed simulation study on plasmonic-organic hybrid electro-optic modulators based on coupled symmetric or asymmetric plasmonic slots. An electro-optic polymer is exploited as an active material, and the device is compatible with a silicon photonics platform. The proposed device operates at 1550 nm wavelength, typical of data center or long-haul telecommunication systems. The device performance in terms of area, plasmonic losses, optical bandwidth, intrinsic modulation bandwidth and energy dissipation are comparable to already proposed Mach-Zehnder solutions, but with potentially better extinction ratio, coupling losses due to photonic-plasmonic transitions, and flexibility in exploiting, without any performance penalty, asymmetric slots to shift the ON and OFF states bias. Finally, the bias dependence of the modulation chirp is investigated, comparing through and cross-coupling configurations. [ABSTRACT FROM AUTHOR]

Herein, an overview on the physics‐based and system‐oriented modeling of organic electro‐optic Mach–Zehnder modulators for optical communication systems is presented. State‐of‐the‐art solutions in electro‐optic organic materials and modulator designs are reviewed, with particular stress on silicon organic hybrid (SOH) and plasmonic organic hybrid (POH) solutions. Then, the physics‐based simulation of concentrated and traveling‐wave modulators is discussed both through 3D optical simulations and a combination of 2D and 3D models, where the modulator is partitioned into convenient subdomains. Neural network behavioral models are finally discussed and case studies are proposed on the physics‐based and system‐level simulation of SOH and POH Mach–Zehnder modulators. [ABSTRACT FROM AUTHOR]

Defects can significantly modify the electro-optical characteristics of InGaN light-emitting diodes (LEDs); however, modeling the impact of defects on the electrical characteristics of LEDs is not straightforward. In this paper, we present an extensive investigation and modeling of the impact of defects on the electrical characteristics of InGaN-based LEDs, as a function of the thickness of the quantum well (QW). First, we demonstrate that the density of defects in the active region of III-N LEDs scales with increasing thickness of the InGaN QW. Since device layers with high indium content tend to incorporate more defects, we ascribed this experimental evidence to the increased volume of defects-rich InGaN associated to thicker InGaN layers. Second, we demonstrate that the current–voltage characteristics of the devices are significantly influenced by the presence of defects, especially in the sub turn-on region. Specifically, we show that the electrical characteristics can be effectively modeled in a wide current range (from pA to mA), by considering the existence of trap-assisted tunneling processes. A good correspondence is obtained between the experimental and simulated electrical characteristics (I–V), by using—in the simulation—the actual defect concentrations/activation energies extracted from steady-state photocapacitance, instead of generic fitting parameters. [ABSTRACT FROM AUTHOR]

The ICT scene is dominated by short-range intra-datacenter interconnects and networking, requiring high speed and stable operations at high temperatures. GaAs/AlGaAs vertical-cavity surface-emitting lasers (VCSELs) emitting at 850–980 nm have arisen as the main actors in this framework. Starting from our in-house 3D fully comprehensive VCSEL solver VENUS, in this work we present the possibility of downscaling the dimensionality of the simulation, ending up with a multiphysics 1D solver (D1ANA), which is shown to be capable of reproducing the experimental data very well. D1ANA is then extensively applied to optimize high-temperature operation, by modifying cavity detuning and distributed Bragg's reflector lengths. [ABSTRACT FROM AUTHOR]

In state-of-the-art, large format, HgCdTe-based infrared focal plane arrays the typical pixel size (3–10 μ m ) is of the order of the operating wavelength and much smaller than the carrier diffusion length. This makes inter-pixel crosstalk a limiting factor, especially in planar structures. Employing three-dimensional electromagnetic and electrical simulations we show that, besides reducing the dark current through Auger suppression, majority carrier depletion of the detector absorber is also effective in curtailing the inter-pixel crosstalk due to carrier diffusion. In the case of a 5 μ m -pitch pixel, a proper design of the absorber composition and doping profile allows to reduce inter-pixel crosstalk by more than a factor of two when increasing the reverse bias from - 0.1 to - 0.5 V , keeping the contribution to crosstalk coming from carrier diffusion between 2 and 12% in the mid- and long-wavelength infrared spectrum. [ABSTRACT FROM AUTHOR]

Probing the thermal effects that limit the performance of vertical-cavity surface-emitting lasers (VCSELs) reveals an intricate interplay between carrier transport, recombination mechanisms, thermal conduction, and optical features. An understanding of this interplay requires an accurate yet computationally-efficient multiphysical approach at the microscopic level. We address this task by characterizing oxide-confined multimode 850 nm VCSELs over a wide temperature range, and by simulating them with VENUS, our in-house multiphysics simulator. The agreement with the experimental results in the whole explored temperature range provides physical insight into the mechanisms that determine VCSEL rollover and turn off, and demonstrates the predictive capabilities of the model. A complete set of model parameters determined through this combined experiment-simulation approach is presented. [ABSTRACT FROM AUTHOR]

The properties of vertical-cavity surface-emitting lasers (VCSELs) are investigated by means of a multiphysical Vcsel Electro-opto-thermal NUmerical Simulator (VENUS). VENUS includes a three-dimensional vectorial electromagnetic code, a description of the quantum well optical response, a heat equation solver, and a quantum-corrected drift-diffusion simulator. The proposed suite includes coupling mechanisms often overlooked in VCSEL simulation and allows to reassess the impact of parameters, which can be critically dependent on implementation details inaccessible in commercial codes. The agreement with experimental results holds the promise of the application of this framework to the computer-aided design of innovative VCSEL concepts. [ABSTRACT FROM AUTHOR]

Highlights • Simulation of HgCdTe infrared detectors with realistic alloy composition profiles essential for their optimization. • Realistic full-wave electromagnetic with compositionally-graded layers is difficult with commercial simulators. • We give practical hints for discretizing the grading into sublayers. • We test several different discretizations. • We assess when the grading may be neglected. Abstract The ability to simulate HgCdTe infrared detectors with realistic alloy composition profiles is essential for their optimization. We give practical guidelines for the realization of combined three-dimensional, realistic full-wave electromagnetic and electrical modeling for the description of detectors with compositionally-graded layers, suitable to most available numerical solvers. Following the known procedure to sample the compositionally-graded layers into a number N of sublayers, we assess the effects of different choices for N , both on calculation accuracy and computational cost. Quantum efficiency spectra calculated with the proposed approach are compared with those determined through two shortcuts: ray tracing, based on classical optics, and full-wave simulation where graded layers are replaced by constant, spatially-averaged optical properties. It is shown that the former is not generally a valid alternative, since it does not address interference effects due to back-reflections, while the latter can lead to incorrect estimates of the detector cutoff wavelength. [ABSTRACT FROM AUTHOR]

The potential applications of germanium and its alloys in infrared silicon-based photonics have led to a renewed interest in their optical properties. In this letter, we report on the numerical determination of Auger coefficients at T=300K for relaxed and biaxially strained germanium. We use a Green's function based model that takes into account all relevant direct and phononassisted processes and perform calculations up to a strain level corresponding to the transition from indirect to direct energy gap. We have considered excess carrier concentrations ranging from 1016 cm-3 to 5×1019 cm-3. For use in device level simulations, we also provide fitting formulas for the calculated electron and hole Auger coefficients as functions of carrier density. [ABSTRACT FROM AUTHOR]

A conformal-mapping approach to model the skin-effect conductor losses of thick multiconductor coplanar waveguides and multiconductor coplanar strips is proposed. The model allows for arbitrary strip or slot number and widths and is accurate for conductor thicknesses up to around 40% of the minimum strip and slot width. Examples are presented to demonstrate the accuracy of the approach when compared to the results from a finite-element method numerical code. [ABSTRACT FROM PUBLISHER]

A self-energy-based model for the calculation of the electron capture time through phonon emission in presence of carrier population is applied to nitride-based quantum wells (QWs). The role of plasma dynamical screening on the Frölich interaction is investigated, highlighting its central role in the dependence of capture time on carrier density. The proposed approach yields a closed-form expression for capture time as a function of the carrier densities in QW and barrier states. Its application to simple two-population rate equations allows to reproduce available experimental data with excellent agreement, offering an accurate yet practical alternative to the usual approximation of a constant capture time in modeling light-emitting diodes and lasers. [ABSTRACT FROM AUTHOR]

Recent experiments of electron emission spectroscopy (EES) on III-nitride light-emitting diodes (LEDs) have shown a correlation between droop onset and hot electron emission at the cesiated surface of the LED p-cap. The observed hot electrons have been interpreted as a direct signature of Auger recombination in the LED active region, as highly energetic Auger-excited electrons would be collected in long-lived satellite valleys of the conduction band so that they would not decay on their journey to the surface across the highly doped p-contact layer. We discuss this interpretation by using a full-band Monte Carlo model based on first-principles electronic structure and lattice dynamics calculations. The results of our analysis suggest that Auger-excited electrons cannot be unambiguously detected in the LED structures used in the EES experiments. Additional experimental and simulative work are necessary to unravel the complex physics of GaN cesiated surfaces. [ABSTRACT FROM AUTHOR]

What has turned into highly complex and somewhat misunderstood efficiency loss mechanisms occurring in light-emitting diodes (LEDs) based on the InN-GaN material system at high injection levels are discussed. Suggestions are made as to the dominant mechanism (s) in an open forum format as well as pointing out some of the shortcomings of the methodologies used and premises forwarded. It is unequivocally known that increased junction temperature would cause a reduction in radiative power due to mainly the reduction in radiative recombination efficiency. Another obvious mechanism is the asymmetry in doping in wide bandgap semiconductors, such as GaN, wherein the hole concentration lags well behind that of electrons in the active region. Because an electron and a hole are required for radiative recombination, the radiative efficiency cannot keep up with increasing carrier injection due to progressively lagging hole population. This results in either electron escape without radiative recombination or electron accumulation, which in turn changes the internal bias of the device, manifested as reduced internal forward bias, which reduces the rate of increase in light intensity. Some of the reports ascribe the efficiency loss at high injection levels to Auger recombination (mainly through indirect and recently reportedly direct deductions) as the main and or the only source of efficiency loss by in many cases simply relying on the temperature and injection independent (not well taken) A, B, C coefficients to fit a third order polynomial to the efficiency vs. injection current. As for the direct deduction, the spectroscopic analysis of Auger kicked hot electrons as they traverse through the G and L bands before being emitted into the vacuum by means of cesiated surface challenges the existing theories and some experiments regarding carrier scattering and G-L separation. Despite just a few reports to the contrary, the bulk of the resonant optical emission experiments do not support the Auger argument as being the main cause. In parallel, there exists a body of theoretical and experimental reports for electron overflow of ballistic/quasi-ballistic electrons traversing the active region to p-GaN, escaping recombination altogether in the active region. In fact and from the get go, the LED industry ubiquitously employed, and continues to do so, an (Al,In)GaN electron-blocking layer (EBL) to prevent electron escape for improved light output that in and of itself would more than suggest that the electron escape (overflow) does indeed occur. The only adverse effect of EBL is that it impedes hole injection due to the valence band offset between the p-type (Al,In)GaN EBL and p-GaN and also generates piezoelectric (if not lattice matched) and differential spontaneous polarization induced fields that pull down the conduction band edge at the interface reducing EBL's effectiveness. To at least reduce the aforementioned aggravating factors to some extent, the electron overflow and the associated efficiency loss can be reduced substantially (particulars of which depend on the active layer design) by inserting a stair-case electron injector (SEI) with a step-like indium composition to act as an "electron cooler" or by linearly graded cooler in some form or another prior to the active region. Use of multiple layered heterostructures for the active layers also plays the role of electron cooler, albeit not necessarily in the most optimum fashion. As if oblivious to the raging issues alluded to above, the LED industry has been moving along very successfully with a 2-prong approach. In one, dubbed the "high voltage LED", a set of LEDs (most likely configured in the form of a full-wave bridge rectifier) operating at low currents, where the efficiency is at its maximum, is used, which also has the added benefit of much reduced power supply complexity and weight. The other is the continual improvement of layer quality and optimum active layer design, taking technological parameters into consideration, which at the time of writing sported 63% wall-plug efficiency at an injection current of 350 mA for ~1.1 mm x 1.1 mm LEDs, which translates to approximately 75% efficiency once the voltage (about 3 V) and phosphor conversion efficiencies are taken into account. Assuming the same extraction and external quantum efficiencies, one obtains about 86% for each. It would not be an exaggeration to conclude that mid 90% internal quantum efficiencies are very likely in play here, which means that the purported inherent problems dominating the discussion are practically reduced to an academic exercise. Despite aggravating factors involving holes and hetero-barriers, commercial LEDs almost invariably use multilayer heterostructure designs for the active layers, loosely termed as multiple quantum wells, presumably to circumvent technological challenges and the ramifications of the polarization induced field even though double heterojunction varieties could be preferable from the point of view of hole transport. Eventually, the real limiting factors are the extent of hole supply (can be mitigated by increased hole concentration, which is well known and continually explored) and the proverbial thermal wall (can be mitigated by increased efficiency and efficient heat removal) of course. [ABSTRACT FROM AUTHOR]