GaN micro- (or nano-) column structures have been studied in an attempt to enhance the LED performance by improving the light extraction and reducing the strain due to the large lattice mismatch between GaN and InGaN. Nowadays columnar LEDs are drawing attention as a candidate for a multicolor emission source as illustrated in Fig. 1. Though numerous studies have focused on selective area growth, the top-down approach has often been the subject of study since it can realize a columnar structure more easily than when using the bottom-up method. However, predominant issue in the top-down approach is dry-etch damage resulting from the inductively coupled plasma (ICP) and reactive ion etching (RIE) systems. According to the previous reports the damage usually results in the increase in the sheet-resistance of GaN, along with the decrease in the reverse breakdown voltage and the reductions in the Schottky barrier height in the diodes formed on GaN. The roughened sidewall after etching also exhibited higher series resistance in the device and deteriorates the electrical characteristics. KOH treatment has been found to improve the electrical characteristics through the removal of the damaged region and providing the vertical profile in the sidewall with a smooth surface. Moreover, the size control of GaN micro- (or nano-) columns also becomes possible by adopting the KOH treatment.
We introduce ITO on graphene as a current-spreading layer for separated InGaN/GaN nanorod LEDs for the purpose of passivation-free and high light-extraction efficiency. Transferred graphene on InGaN/GaN nanorods effectively blocks the diffusion of ITO atoms to nanorods, facilitating the production of transparent ITO/graphene contact on parallel-nanorod LEDs, without filling the air gaps, like a bridge structure. The ITO/graphene layer sufficiently spreads current in a lateral direction, resulting in uniform and reliable light emission observed from the whole area of the top surface. Using KOH treatment, we reduce series resistance and reverse leakage current in nanorod LEDs by recovering the plasma-damaged region. We also control the size of the nanorods by varying the KOH treatment time and observe strain relaxation via blueshift in electroluminescence. As a result, bridge-structured LEDs with 8 min of KOH treatment show 15 times higher light-emitting efficiency than with 2 min of KOH treatment.
In this study, we produce InGaN/GaN microcolumn LED (MC-LED) arrays having nonpolar metal sidewall contacts using a top-down method, where the metal contacts only with the sidewall of the columnar LEDs with an open top for transparency. The trapezoidal profile of the as-etched columns was altered to a rectangular profile through KOH treatment, exposing the nonpolar sidewalls. While the MC-LED with no treatment emitted no light because of the etch-damaged region, the MC-LEDs with KOH treatment exhibited much improved the electrical properties with the much higher shunt resistance due to the removal of the etch-damaged region. The optical output power was strongest for the MC-LED with a 5-min treatment indicating an almost complete removal of the damaged region.
This paper addresses the training issues associated with neural network-based automatic speech recognition (ASR) under noise conditions. In particular, conventional joint training approaches for a pipeline comprising speech enhancement (SE) and end-to-end ASR model surfer from a conflicting problem and a frame mismatched alignment problem because of different goals and different frame structures for ASR and SE. To mitigate such problems, a knowledge distillation (KD)-based training approach is proposed by interpreting the ASR and SE models in the pipeline as teacher and student models, respectively. In the proposed KD-based training approach, the ASR model is first trained using a training dataset, and then, acoustic tokens are generated via K-means clustering using the latent vectors of the ASR encoder. Thereafter, KD-based training of the SE model is performed using the generated acoustic tokens. The performance of the SE and ASR models is evaluated on two different databases, noisy LibriSpeech and CHiME-4, which correspond to simulated and real-world noise conditions, respectively. The experimental results show that the proposed KD-based training approach yields a lower character error rate (CER) and word error rate (WER) on the two datasets than conventional joint training approaches, including multi-condition training. The results also show that the speech quality scores of the SE model trained using the proposed training approach are higher than those of SE models trained using conventional training approaches. Moreover, the noise reduction scores of the proposed training approach are higher than those of conventional joint training approaches but slightly lower than those of the standalone-SE training approach. Finally, an ablation study is conducted to examine the contribution of different combinations of loss functions in the proposed training approach to SE and ASR performance. The results show that the combination of all loss functions yields the lowest CER and WER and that tokenizer loss contributes more to SE and ASR performance improvement than ASR encoder loss.
In the realm of optical technologies, the integration of nature's designs and modern engineering has paved the way for groundbreaking innovations. Bio-inspired tunable optics and photonics, drawing from the intricate mechanisms found in biological systems, offer a new frontier in adaptive and efficient light management. Here, this review presents a comprehensive examination of the principles, advancements, and applications of natural light-manipulation and adaptation mechanisms, highlighting their translation into artificial tunable optics and photonic structures. Emphasizing the remarkable potential of bio-inspired systems, particularly those emulating the tunable optical functionalities of biological eyes and skins, it explores the current state of bio-inspired tunable optics and photonic devices. Our review categorizes these tunable bio-inspired systems into two foundational mechanisms: light-manipulation and light-adaptation, illustrating their wide-ranging implications from consumer electronics to next-generation technologies. This review also highlights the challenges and prospects of bio-inspired tunable optics and photonics. It emphasizes their role in promoting tunable optical properties for multifunctional devices, providing revolutionary opportunities across various sectors, including the military and everyday life, thus surpassing current cutting-edge optical technologies.
We demonstrate a cost-effective top-down approach for fabricating InGaN/GaN nanorod arrays using a wet treatment process in a KOH solution. The average diameter of the as-etched nanorods was effectively reduced from 420 nm to 180 nm. The spatial strain distribution was then investigated by measuring the high-resolution cathodoluminescence directly on top of the nanorods. The smaller nanorods showed a higher internal quantum efficiency and lower potential fluctuation, which can subsequently be exploited for high-efficiency photonic devices.
Blue to green tunable GaN-based LEDs having a dual-junction structure have been fabricated. We observed that the color of the dual-junction LEDs can be tuned by controlling both current path and density through three terminals.
Ag nanoparticles are embedded in intentionally etched micro-circle p-GaN holes by means of a thermal agglomeration process to enhance the light absorption efficiency in InGaN/GaN multi-quantum-well (MQW) solar cells. The Ag nanoparticles are theoretically and experimentally verified to generate the plasmon light scattering and the localized field enhancement near the MQW absorption layer. The external quantum efficiency enhancement at a target wavelength region is demonstrated by matching the plasmon resonance of Ag nanoparticles, resulting in a Jsc improvement of 9.1%. Furthermore, the Ag-nanoparticle-embedded InGaN solar cell is effectively fabricated considering the carrier extraction that more than 70% of F.F. and 2.2 V of high Voc are simultaneously attained.
Dance learning through online videos has gained popularity, but it presents challenges in providing comprehensive information and personalized feedback. This paper introduces DanceSculpt, a system that utilizes 3D human reconstruction and tracking technology to enhance the dance learning experience. DanceSculpt consists of a dancer viewer that reconstructs dancers in video into 3D avatars and a dance feedback tool that analyzes and compares the user's performance with that of the reference dancer. We conducted a comparative study to investigate the effectiveness of DanceSculpt against conventional video-based learning. Participants' dance performances were evaluated using a motion comparison algorithm that measured the temporal and spatial deviation between the users' and reference dancers' movements in terms of pose, trajectory, formation, and timing accuracy. Additionally, user experience was assessed through questionnaires and interviews, focusing on aspects such as effectiveness, usefulness, and satisfaction with the system. The results showed that participants using DanceSculpt achieved significant improvements in dance performance compared to those using conventional methods. Furthermore, the participants rated DanceSculpt highly in terms of effectiveness (avg. 4.27) and usefulness (avg. 4.17) for learning dance. The DanceSculpt system demonstrates the potential of leveraging 3D human reconstruction and tracking technology to provide a more informative and interactive dance learning experience. By offering detailed visual information, multiple viewpoints, and quantitative performance feedback, DanceSculpt addresses the limitations of traditional video-based learning and supports learners in effectively analyzing and improving their dance skills.
In recent years, research into implementing microdisplays for use in virtual reality and augmented reality has been actively conducted worldwide. Specifically, inorganic light-emitting diodes (LED) have many advantages in microdisplays, so much effort has been made to implement them in various ways. However, it is still challenging to realize a display with high resolution using only inorganic LEDs without color conversion layers because a typical LED chip is designed to emit only one color on a single wafer. In this study, we integrated high-efficiency red, green, and blue (RGB) LED material systems on the same sapphire substrate. Since the blue and green LED structures comprised the same GaN semiconductor, a metalorganic chemical vapor deposition method was used to integrate them. Meanwhile, the red LEDs made from another semiconductor were incorporated into the blue/green LEDs using a wafer-bonding technique. The fabricated hybrid RGB LEDs were able to cover a wide color space. In addition, the RGB LED material systems consisting of a high-quality single crystal were stably combined on the sapphire substrate without any structural defects. We show the possibility of their use in displays by integrating the RGB LEDs on one chip and suggest that their utilization could range from large-area LED displays to ultrahigh-resolution small displays.
'%3e%3cg id='b'%3e%3cpath id='a' stroke='%23000' stroke-width='2' d='M-6-25H6m-12 3H6m-12 3H6'/%3e%3cuse xlink:href='%23a' y='44'/%3e%3c/g%3e%3cpath stroke='%23fff' d='M0 17v10'/%3e%3ccircle r='12' fill='%23cd2e3a'/%3e%3cpath fill='%230047a0' d='M0-12A6 6 0 0 0 0 0a6 6 0 0 1 0 12 12 12 0 0 1 0-24z'/%3e%3c/g%3e%3cg transform='rotate(-123.69)'%3e%3cuse xlink:href='%23b'/%3e%3cpath stroke='%23fff' d='M0-23.5v3M0 17v3.5m0 3v3'/%3e%3c/g%3e%3c/svg%3e)