Recent years have witnessed a tremendous development of vertical gallium nitride (GaN) power devices, a new class of device technology that could be the key enabler for next-generation high ...performance power electronics. In this comprehensive review, we discuss the recent progress made on vertical GaN power devices, highlighting their important device design principles and fabrication processes. Part I of the two-part review series introduces the basic design principles of vertical GaN devices using Schottky barrier diodes (SBDs) and p-n diodes as examples. We provide a comprehensive review and in-depth analysis on their basic structures, fabrication processes, and device physics. Materials engineering, including buffer layer and drift layer design, and device engineering, including various edge termination methods, are elucidated. Key device topics, including avalanche breakdown and leakage mechanisms of vertical GaN power devices, are also discussed. We also examine various approaches and provide detailed descriptions of the key knowledge obtained. This timely review provides valuable information for the power electronics community and can inspire researchers for future interdisciplinary collaborations in this emerging and exciting research field.
Background
The most common neurodegenerative disease, Alzheimer’s disease (AD) constitutes about two thirds of cases of dementia overall 1. Mild cognitive impairment (MCI) is considered to be a ...transitional stage between aging and AD 2. This general AD staging information and ample MCI data motivate a further investigation on the subcortical shape analysis incorporating MCI data in AD study. We therefore designed a novel logistic regression model to explore more morphometric patterns associated with the disease progression of AD.
Method
In this study, we used T1‐weighted MRI data of 225 controls, 398 MCI patients and 195 AD patients from the ADNI‐1 dataset. We designed an ordinal logistic regression model 3 regularized by total variation (TV) and L1 penalization term 4. Unlike the binary or conventional multi‐class classifiers, our TV‐L1 ordinal classifier is capable of sparse‐spatially utilizing ordinal morphometric features to learn potential ordinal patterns and predict normal control, MCI and AD all at once in a single model. We extracted vertex‐wise thickness features in 7 subcortical regions 5 and trained our model with all 14 regions of interest (ROI) across the whole brain. To evaluate the performance and find the optimal hyper‐parameters in our classification task, we computed the precision, recall score and f‐score with a 4‐fold cross‐validation (CV) grid search. For comparison we also trained a TV‐L1 logistic regression model binary classifiers for controls and AD.
Result
Table 1 shows the confusion matrix, precision, recall score and f‐score. Based on the 4‐fold CV, the average precision and recall score for control, MCI and AD are 0.457 and 0.443 respectively. The weights map of the trained TV‐L1 logistic regression model is shown in Figure 1, and that of the TV‐L1 ordinal classifier in Figure 2.
Conclusion
By designing an TV‐L1 ordinal logistic regression machine learning model, we performed a shape feature‐based classification for controls, MCI and AD with f1 score substantially above chance across all three AD stages. The trained weights map reveals subtle morphometric patterns associated with the disease progression from a cognitively normal state to MCI and eventually AD stage. This could potentially provide more insights for clinical studies of AD.
We report on the thermal performance of the electroluminescence of 12-nm-thick single-quantum-well (SQW) InGaN blue light-emitting diodes (LEDs) grown on the semipolar ($20\bar{2}\bar{1}$) plane. At ...a current density 100 A/cm 2 , the external quantum efficiency (EQE) decreased by 9.7% when the temperature was increased from 20 to 100 °C. Hot/cold factors were more than 0.9 at current densities greater than 20 A/cm 2 . A high characteristic temperature of 900 K and low junction temperature of 68 °C were also measured using bare LED chips.
This paper gives an experimental demonstration of non-line-of-sight (NLOS) visible light communication (VLC) using a single 80 μm gallium nitride (GaN) based micro-light-emitting diode (micro-LED). ...This device shows a 3-dB electrical-to-optical modulation bandwidth of 92.7 MHz. IEEE 802.11ac modulation scheme with 80 MHz bandwidth, as an entry level of the fifth generation of Wi-Fi, was employed to use the micro-LED bandwidth efficiently. These practical techniques were successfully utilized to achieve a demonstration of line-of-sight (LOS) VLC at a speed of 433 Mbps and a bit error rate (BER) of 10−5 with a free space transmit distance 3.6 m. Besides this, we demonstrated directed NLOS VLC links based on mirror reflections with a data rate of 433 Mbps and a BER of 10−4. For non-directed NLOS VLC using a print paper as the reflection material, 16 QAM, 195 Mbps data rate, and a BER of 10−5 were achieved.
Vertical gallium nitride (GaN) power devices are enabling next-generation power electronic devices and systems with higher energy efficiency, higher power density, faster switching, and smaller form ...factor. In Part I of this review, we have reviewed the basic design principles and physics of building blocks of vertical GaN power devices, i.e., Schottky barrier diodes and p-n diodes. Key topics such as materials engineering, device engineering, avalanche breakdown, and leakage mechanisms are discussed. In Part II of this review, several more advanced power rectifiers are discussed, including junction barrier Schottky (JBS) rectifiers, merged p-n/Schottky (MPS) rectifiers, and trench metal-insulator-semiconductor barrier Schottky (TMBS) rectifiers. Normally- OFF GaN power transistors have been realized in various advanced device structures, including current aperture vertical electron transistors (CAVETs), junction field-effect transistors (JFETs), metal-oxide-semiconductor field-effect transistors (MOSFETs), and fin field-effect transistors (FinFETs). A detailed analysis on their performance metrics is provided, with special emphasis on the impacts of key fabrication processes such as etching, ion implantation, and surface treatment. Lastly, exciting progress has been made on selective area doping and regrowth, a critical process for the fabrication of vertical GaN power devices. Various materials characterization techniques and surface treatments have proven to be beneficial in aiding this rapid development. This timely and comprehensive review summarizes the current progress, understanding, and challenges in vertical GaN power devices, which can serve as not only a gateway for those interested in the field but also a critical reference for researchers in the wide bandgap semiconductor and power electronics community.
Atomic-scale nanofacets were observed in semipolar and InGaN quantum wells (QWs)/GaN quantum barriers interfaces. Transmission electron microscopy studies showed that these nanofacets were mainly ...composed of , and planes, which led to significant fluctuations in QW thickness. Atom probe tomography studies were carried out to visualize the nanofacet structure. The In composition in the InxGa1−xN alloys followed a binominal distribution despite the formation of the nanofacet structure. One-dimensional (1D) Schrödinger-Poisson drift-diffusion simulation showed that these nanofacets and associated QW thickness fluctuations will lead to a large wavelength shift and a broadened spectral linewidth for semipolar QWs.
This paper reports a comprehensive study on the anisotropic electrical properties of vertical (<inline-formula> <tex-math notation="LaTeX">\overline {\textsf {2}}01 </tex-math></inline-formula>) and ...(010) <inline-formula> <tex-math notation="LaTeX">\beta </tex-math></inline-formula>-Ga 2 O 3 Schottky barrier diodes (SBDs). The devices were fabricated on single-crystal substrates grown by an edge-defined film-fed growth method. The temperature-dependent current-voltage (I-V) and capacitance-voltage (C-V) characteristics were systematically measured, analyzed, and compared. The (<inline-formula> <tex-math notation="LaTeX">\overline {\textsf {2}}01 </tex-math></inline-formula>) and (010) SBDs exhibited on-resistances (<inline-formula> <tex-math notation="LaTeX">{R}_{{ \mathrm{\scriptscriptstyle ON}}} </tex-math></inline-formula>) of 0.56 and <inline-formula> <tex-math notation="LaTeX">0.77~\textsf {m}\Omega \cdot \textsf {cm}^{{\textsf {2}}} </tex-math></inline-formula>, turn- ON voltages (<inline-formula> <tex-math notation="LaTeX">{V}_{{ \mathrm{\scriptscriptstyle ON}}} </tex-math></inline-formula>) of 1.0 and 1.3 V, Schottky barrier heights (SBHs) of 1.05 and 1.20 eV, electron mobilities of 125 and 65 cm 2 /(<inline-formula> <tex-math notation="LaTeX">\textsf {V}\cdot ~\textsf {s} </tex-math></inline-formula>), respectively, with an on-current of ~1.3 kA/cm 2 and on/off ratio of ~10 9 . The (010) SBD had a larger <inline-formula> <tex-math notation="LaTeX">{V}_{{ \mathrm{\scriptscriptstyle ON}}} </tex-math></inline-formula> and SBH due to anisotropic surface properties (i.e., surface Fermi level pinning and band bending), as supported by X-ray photoelectron spectroscopy measurements. Temperature-dependent I-V also revealed the inhomogeneous nature of the SBH in both devices, where the (<inline-formula> <tex-math notation="LaTeX">\overline {\textsf {2}}01 </tex-math></inline-formula>) SBD showed a more uniform SBH distribution. The homogeneous SBH was also extracted: 1.33 eV for the (<inline-formula> <tex-math notation="LaTeX">\overline {\textsf {2}}01 </tex-math></inline-formula>) SBD and 1.53 eV for the (010) SBD. The reverse leakage current of the devices was well described by the two-step trap-assisted tunnelingmodel and the 1-D variable range hopping conduction model. The (<inline-formula> <tex-math notation="LaTeX">\overline {\textsf {2}}01 </tex-math></inline-formula>) SBD showed a larger leakage current due to its lower SBH and/or smaller activation energy, and thus a smaller breakdown voltage. These results indicate that the crystalline anisotropy of <inline-formula> <tex-math notation="LaTeX">\beta </tex-math></inline-formula>-Ga 2 O 3 can affect the electrical properties of vertical SBDs and should be taken into consideration when designing <inline-formula> <tex-math notation="LaTeX">\beta </tex-math></inline-formula>-Ga 2 O 3 electronics.
III-nitrides material systems have attracting growing interests in photovoltaic (PV) applications after huge success in optoelectronics. In this work, a semi-analytical model is used to analyze the ...PV performance of single junction InGaN solar cells. Through clarifying four basic types of loss mechanisms, including transmission loss, thermalization loss, spatial relaxation loss and recombination loss, we discover that transmission loss accounts for the primary part of efficiency loss due to the large bandgaps of III-nitride materials. As for all recombination-related losses, Shockley-Reed-Hall (SRH) recombination loss is dominant over others. By incorporating non-step-like absorptance and emittance with below-bandgap absorption, we discover that reducing SRH recombination current by improving the material quality of InGaN layers proves an efficient approach to optimize the cell performance. Furthermore, the energy conversion efficiency increases with higher material quality and larger solar concentration. Our calculations show that energy conversion efficiency of 7.35% can be achieved under one sun and maximum efficiency of 8.43% under 1000 suns. This theoretical study offers detailed guidance for the future design of high-performance thin film InGaN solar cells.
Abstract
Background
Disease progression models (DPM) of Alzheimer’s disease (AD) based on non‐invasive biomarkers have received significant attention in recent years. Crucially for drug development, ...DPM promises to identify pre‐clinical stages of AD. However, traditional DPMs consider only a single canonical sequence of neurodegeneration, which may ignore important population clusters and individual variation. Addressing this, we propose a method to use individual dMRI‐based connectomes as an informative prior in AD DPM.
Method
We extracted regional cortical thickness measures and tractography‐based connectomes from T1 and diffusion MRI of 32 AD patients and 52 controls. We build a DPM using the event‐based (EBM) approach. EBM models AD staging discreetly. Every stage is identified with a neurodegenerative event, e.g. a regional thickness falling below some threshold. Disease stage is then defined by the last abnormal biomarker based on a canonical event order. The goal of the model is to find an event order specific to AD in a Baeysian setting. We replace the uninformative prior on the order with one based on the subject connectomes. Every regional biomarker pair is assigned a probability that one region degenerates following the other based on the shortest connectome path length between the regions. The idea follows closely the amyloid network diffusion hypothesis. We test the original EBM and our models on cross‐sectional data and test them on 12‐ and 24‐month follow‐up scans. Predicted patient stage and visit order are ordinally correlated to assess model performance.
Result
Our model outperforms the original EBM (Kendall tau=0.34 vs 0.49). Our model systematically assigns higher progression scores for AD subjects (Fig.1). We also compare DPM performance using predicted stage in disease classification. Our model again outperforms traditional EBM, achieving ROC AUC=0.88 (std=0.046) vs 0.816 (std=0.008).
Conclusion
We have developed an individualized neuroimaging disease progression model of AD based on the widely used EBM. Individual connectome variation allows the computational model to more finely pin‐point an individual’s specific stage of disease progression based on the observed pattern of neurodegeneration. The connectome‐based EBM shows significant improvement over standard EBM.
PDF
