Search for: All records

Abstract In this paper we present a study of distribution polarization doped Al_xGa_1−xN layers and their use in quasi-vertical configuration pn-diodes which exhibited a high breakdown field of ∼8.5 MV cm⁻¹and a large forward current density (∼23 kA cm⁻²). We also establish their potential use in UVC light emitters by studying the optical emission from a quantum well inserted at the distribution polarization doped pn-junction interface. 

An extreme bandgap Al0.64Ga0.36N quantum channel HEMT with Al0.87Ga0.13N top and back barriers, grown by MOCVD on a bulk AlN substrate, demonstrated a critical breakdown field of 11.37 MV/cm—higher than the 9.8 MV/cm expected for the channel’s Al0.64Ga0.36N material. We show that the fraction of this increase is due to the quantization of the 2D electron gas. The polarization field maintains electron quantization in the quantum channel even at low sheet densities, in contrast to conventional HEMT designs. An additional increase in the breakdown field is due to quantum-enabled real space transfer of energetic electrons into high-Al barrier layers in high electric fields. These results show the advantages of the quantum channel design for achieving record-high breakdown voltages and allowing for superior power HEMT devices. 

Abstract High voltage (∼2 kV) Al_0.64Ga_0.36N-channel high electron mobility transistors were fabricated with an on-resistance of ∼75 Ω. mm (∼21 mΩ. cm²). Two field plates of variable dimensions were utilized to optimize the breakdown voltage. The breakdown voltage reached >3 kV (tool limit) before passivation however it reduced to ∼2 kV after Si₃N₄surface passivation and field plate deposition. The breakdown voltage and on-resistance demonstrated a strong linear correlation in a scattered plot of ∼50 measured transistors. The fabricated transistors were electrically characterized and benchmarked against the state-of-the-art high-voltage (> 1 kV) Al-rich (>40%) AlGaN-channel transistors in breakdown voltage and on-resistance, indicating significant progress. 

State-of-the-art semiconducting aluminum nitride (AlN) films were characterized by cathodoluminescence (CL) spectroscopy in the range of 200–500 nm in an attempt to identify the energy levels within the bandgap and their associated defects. Near-band edge emission (around 206 nm) and high-intensity peaks centered in the near UV range (around 325 nm) are observed for both n- and p-type AlN films. The near UV peaks are potentially associated with oxygen contamination in the films. The p-type AlN films contain at least two unidentified peaks above 400 nm. Assuming that the dopant concentration is independent of compensation (i.e., in the perfect doping limit), three effective donor states are found from Fermi–Dirac statistics for Si-doped AlN, at ∼0.035, ∼0.05, and ∼0.11 eV. Similarly, a single effective acceptor energy of ∼0.03–0.05 eV (depending on the degeneracy factory considered) was found for Be doped AlN. CL investigation of doped AlN films supports claims that AlN may be a promising optoelectronic material, but also points to contaminant mitigation and defect theory as major areas for future study. 

Abstract Molecular Dynamics (MD) simulation of biomolecules provides important insights into conformational changes and dynamic behavior, revealing critical information about folding and interactions with other molecules. The collection of simulations stored in computers across the world holds immense potential to serve as training data for future Machine Learning models that will transform the prediction of structure, dynamics, drug interactions, and more. Ideally, there should exist an open access repository that enables scientists to submit and store their MD simulations of proteins and protein-drug interactions, and to find, retrieve, analyze, and visualize simulations produced by others. However, despite the ubiquity of MD simulation in structural biology, no such repository exists; as a result, simulations are instead stored in scattered locations without uniform metadata or access protocols. Here, we introduce MDRepo, a robust infrastructure that provides a relatively simple process for standardized community contribution of simulations, activates common downstream analyses on stored data, and enables search, retrieval, and visualization of contributed data. MDRepo is built on top of the open-source CyVerse research cyber-infrastructure, and is capable of storing petabytes of simulations, while providing high bandwidth upload and download capabilities and laying a foundation for cloud-based access to its stored data. 

We present a fully integrated AI-driven framework for rapid endurance prediction in NVDRAM ferroelectric capacitors. Endurance testing is one of the most time- and resource-intensive steps in memory characterization, often requiring up to 10¹² cycles per device. To overcome the scarcity of endurance training data, we propose an experimentally calibrated synthetic data generation pipeline using kinetic Monte Carlo (kMC) simulations in Ginestra™, seeded with experimentally extracted defect parameters. We train a transformer-based AI surrogate using this high-fidelity dataset, achieving an R² of 0.992 and enabling ~105x speedup in defect evolution prediction. The surrogate generates large-scale synthetic datasets by sampling initial defect profiles, which are then used to train a hybrid multi-layer perceptron (MLP)- attention model that maps early-life defect characteristics to Weibull endurance distributions. This final endurance prediction model achieves strong agreement with ground truth Weibull parameters, with R² values of >0.98 for η and ~0.9 for β, demonstrating its reliability in capturing endurance distribution characteristics. Wafer-scale prediction of breakdown distributions is demonstrated in one-shot, reducing characterization time by over 10 orders of magnitude. This framework enables scalable, high-throughput reliability screening for ferroelectric memory technologies. 

While negative capacitance (NC) has been demonstrated in ferroelectric-dielectric (FE-DE) heterostructures in the form of capacitance enhancement, all experimental evidence, to date, suggests the existence of domains therein. Here, we address the question: what are the conditions to achieve ideal, domain-free NC in FE-DE heterostructures? Our main claim is that for given thicknesses of the ferroelectric and the dielectric layers, there is a critical value of domain wall energy parameter— above which the system would be stabilized in an ideal and robust domain-free NC state and would be robust against domain formation. Our analyses suggest that to achieve ideal NC, efforts should lie in understanding the means to control the domain wall energy on all fronts, both theory and experiments via high throughput design, discovery, and engineering of ferroelectrics.