Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “GANs”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 181 records · Page 10
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Using GANs to predict milling stability from limited data

Milling is a key manufacturing process that requires the selection of operating parameters that provide efficient performance. However, the presence of chatter, a self-excited vibration causing poor surface finish and potential damage to the machine and cutting tool, makes it challenging to select the appropriate parameters. To predict chatter, stability maps are commonly used, but their generation requires expensive data, making it difficult to employ these maps in industry. Therefore, there is a pressing need for an approach that can accurately predict stability maps using limited experimental data. This study introduces the new Encoder GAN (EGAN) approach based on Generative Adversarial Networks (GANs) that predicts stability maps using limited experimental data. The approach consists of the encoder, generator, and discriminator subnetworks and uses the trained encoder and generator to predict the target stability map. This versatile method can be applied to various tool setups and can accurately predict stability maps with limited experimental data (five to 10 cutting tests) even when there is little information available for unknown parameters. In conclusion, the study evaluates the proposed approach using both numerical data and experiments and demonstrates its superior performance compared to state-of-the-art benchmarks.

42 ENGINEERING↗
DOE OSTI2020

Inverse design of two-dimensional graphene/h-BN hybrids by a regressional and conditional GAN

Design of materials with desired properties is currently laborious and heavily relies on intuition of researchers through a trial-and-error process. To tackle this challenge, in this work we propose a novel regressional and conditional generative adversarial network (RCGAN) for inverse design of representative two-dimensional materials, the graphene and boron-nitride (BN) hybrids. RCGAN incorporates a supervised regressor network, thus overcoming the common technical barrier in the traditional unsupervised GANs, which cannot generate data when fed with continuous and quantitative labels. RCGAN can autonomously generate graphene/BN hybrids given any target bandgap values. These structures are distinguished from the ones used for training and exhibit high diversity for a given bandgap. Moreover, they exhibit high fidelity, yielding bandgaps within ~10% MAE F of the desired bandgaps as validated by density functional theory (DFT) calculations. Analysis by the principle component analysis (PCA) and modified locally linear embedding (MLLE) reveals that the generator has successfully generated structures following the statistical distribution of the real structures. It implies the possibility of the RCGAN in recognizing physical rules hidden in the high-dimensional data. The novel strategy for designing regressional GAN architecture together with the successful application to inverse design of materials would inspire further exploration in research fields beyond materials.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Multi-agent voltage control in distribution systems using GAN-DRL-based approach

Active distribution grids can experience voltage fluctuations and violations due to the high penetration of variable distributed energy resources (DERs). These problems might occur because of the uncertain and variable generation natures of these resources, especially solar photovoltaic resources, during panel shadowing scenarios. Volt-VAR control (VVC) is an efficient method that controls the reactive power set-points of the inverters to regulate the voltage of distribution grids. Although several VVC approaches have been proposed recently, the performance of these approaches degrades significantly if behind-the-meter solar generation data are unobservable/missing. Therefore, it is necessary to impute missing/unobservable PV data accurately to be utilized in VVC approaches. Further, this paper proposes a model-free, data-driven, centrally trained, and decentrally executed multi-agent deep reinforcement learning-based VVC architecture to regulate the voltage of distribution networks. A generative adversarial network (GAN) is incorporated to impute the unobservable PV data accurately, which improves the performance of the proposed control architecture. The proposed multi-agent-soft-actor–critic algorithm (MASAC)-based VVC technique utilizes the actual PV dataset as well as the imputed dataset from the GAN framework to learn the optimal coordinated control policy for controlling the optimal reactive power set-points of PV inverters. The effectiveness of the proposed approach is analyzed on a modified IEEE 34-bus test case with added PV inverters. The results are compared and analyzed with a base case model with no VVC and VVC with a local droop control approach, genetic algorithm optimization, and a centralized soft actor–critic-based approach. Moreover, the performance of the proposed approach is compared with that of a multi-agent VVC framework without using the PV generation data and load information as the system state. The results illustrate that the proposed method with more state input improves the voltage profile and reduces the power loss of the network across various loading and PV generation scenarios.

14 SOLAR ENERGY↗
DOE OSTI2022

ICP etching of GaN microstructures in a Cl 2 –Ar plasma with subnanometer-scale sidewall surface roughness

In this work, substrate temperature, RF power, and ICP power were investigated for their effects on GaN micropillar sidewall roughness and etch characteristics. Elevated substrate temperature was shown to improve the sidewall etch morphology at low RF powers (reduced physical bombardment) and low ICP powers (lower plasma densities). Increased lateral etching is observed with both increased ICP power and substrate temperature, which both act to increase the chemical component of the etch. Etch conditions with a high chemical driving force resulted in faceting along the a-plane on the sidewalls. This faceting produced extremely smooth surfaces with root-meansquare roughness (R q ) as low as 0.20 nm which is comparable to typical epitaxy-ready surfaces and smaller than the a-plane lattice spacing of 0.3186 nm. The smooth surfaces produced in this study enable possibilities for laser facets or for new device structures that require high quality surfaces for GaN regrowth.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Optimizing performance and yield of vertical GaN diodes using wafer scale optical techniques

Abstract To improve the manufacturing of vertical GaN devices for power electronics applications, the effects of defects in GaN substrates need to be better understood. Many non-destructive techniques including photoluminescence, Raman spectroscopy and optical profilometry, can be used to detect defects in the substrate and epitaxial layers. Raman spectroscopy was used to identify points of high crystal stress and non-uniform conductivity in a substrate, while optical profilometry was used to identify bumps and pits in a substrate which could cause catastrophic device failures. The effect of the defects was studied using vertical P-i-N diodes with a single zone junction termination extention (JTE) edge termination and isolation, which were formed via nitrogen implantation. Diodes were fabricated on and off of sample abnormalities to study their effects. From electrical measurements, it was discovered that the devices could consistently block voltages over 1000 V (near the theoretical value of the epitaxial layer design), and the forward bias behavior could consistently produce on-resistance below 2 mΩ cm 2 , which is an excellent value considering DC biasing was used and no substrate thinning was performed. It was found that high crystal stress increased the probability of device failure from 6 to 20%, while an inhomogeneous carrier concentration had little effect on reverse bias behavior, and slightly (~ 3%) increased the on-resistance (R on ). Optical profilometry was able to detect regions of high surface roughness, bumps, and pits; in which, the majority of the defects detected were benign. However a large bump in the termination region of the JTE or a deep pit can induce a low voltage catastrophic failure, and increased crystal stress detected by the Raman correlated to the optical profilometry with associated surface topography.

42 ENGINEERING↗
DOE OSTI2023

Using machine learning with optical profilometry for GaN wafer screening

Abstract To improve the manufacturing process of GaN wafers, inexpensive wafer screening techniques are required to both provide feedback to the manufacturing process and prevent fabrication on low quality or defective wafers, thus reducing costs resulting from wasted processing effort. Many of the wafer scale characterization techniques—including optical profilometry—produce difficult to interpret results, while models using classical programming techniques require laborious translation of the human-generated data interpretation methodology. Alternatively, machine learning techniques are effective at producing such models if sufficient data is available. For this research project, we fabricated over 6000 vertical PiN GaN diodes across 10 wafers. Using low resolution wafer scale optical profilometry data taken before fabrication, we successfully trained four different machine learning models. All models predict device pass and fail with 70–75% accuracy, and the wafer yield can be predicted within 15% error on the majority of wafers.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Potential for neutron and proton transmutation doping of GaN and Ga 2 O 3

As the potential applications of GaN and Ga 2 O 3 are limited by the inadequacy of conventional doping techniques, specifically when uniform selective area p-type doping is required, the potential for transmutation doping of these materials is analyzed. All transmuted element concentrations are reported as a function of time for several common proton and neutron radiation sources, showing that previously published results considered a small subset of the dopants produced. A 40 MeV proton accelerator is identified as the most effective transmutation doping source considered, with a 2.25 × 10 17 protons per cm 2 fluence yielding net concentrations of uncompensated p-type dopants of 7.7 × 10 15 and 8.1 × 1015 cm -3 for GaN and Ga 2 O 3 , respectively. Furthermore, it is shown that high energy proton accelerator spectra are capable of producing dopants required for magnetic and neutron detection applications, although not of the concentrations required for current applications using available irradiation methods.

36 MATERIALS SCIENCE↗
DOE OSTI2020

High voltage GaN p-n diodes formed by selective area regrowth

GaN p-n diodes were formed by selective area regrowth on freestanding GaN substrates using a dry etch, followed by post-etch surface treatment to reduce etch-induced defects, and subsequent regrowth into wells. Etched-and-regrown diodes with a 150 gm diameter achieved 840 V operation at 0.5 A/cm 2 reverse current leakage and a specific on-resistance of 1.2 mΩ•cm 2 . Etched-and-regrown diodes were compared with planar, regrown diodes without etching on the same wafer. Both types of diodes exhibited similar forward and reverse electrical characteristics, which indicate that etch-induced defectivity of the junction was sufficiently mitigated soas not to be the primary cause for leakage. An area dependence for forward and reverse leakage current density was observed, suggesting that the mesa sidewall provided a leakage path.

42 ENGINEERING↗
DOE OSTI2020

Effects of cation stoichiometry on surface morphology and crystallinity of ZnGeN 2 films grown on GaN by metalorganic chemical vapor deposition

ZnGeN 2 films were grown on GaN-on-sapphire templates via metalorganic chemical vapor deposition. Energy dispersive X-ray spectroscopy was used to estimate the Zn/(Zn+Ge) composition ratio in the films. This ratio decreased with increase in growth temperature but increased with increase in total reactor pressure or Zn/Ge precursor flow rate ratio. Systematic mapping of these key growth parameters has allowed us to identify the growth window to achieve ZnGeN 2 with stoichiometric cation composition. Compositional and statistical analyses performed on data acquired from atom probe tomography provided insight into the local compositional homogeneity. The cations Zn and Ge did not demonstrate segregation or clustering at the sub-nanometer level. Based on X-ray diffraction 2θ-ω scan profiles and transmission electron microscope nano-diffraction patterns, the films with near-stoichiometric cation ratios were single crystalline with planar surfaces, whereas zinc-rich or zinc-poor films were polycrystalline with nonplanar surfaces. The growth direction of the single crystalline ZnGeN 2 films on GaN templates was along the c-axis. Room temperature Raman spectra showed features associated with the phonon density of states, indicating the presence of cation disorder in the lattice. A cathodoluminescence peak associated with transitions involving deep level defects was observed around 640 nm. The intensity of this peak increased by almost 2.5 times as the temperature was reduced to 77 K from room temperature. A similar peak was observed in photoluminescence spectra collected at 80 K.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

p-type conductivity and damage recovery in implanted GaN annealed by rapid gyrotron microwave annealing

We demonstrate p-type activation of GaN doped by Mg ion implantation, and in situ during metalorganic chemical vapor deposition through sequential short-duration gyrotron microwave heating cycles at temperatures of 1200–1350 °C. GaN is implanted with 1019 cm−3 Mg ions, capped with AlN, and annealed under 3 MPa N2 overpressure in 5 s heating cycles for less than 60 s total using a high-power gyrotron microwave heating source. Through I–V characterization, photoluminescence spectroscopy, and Raman spectroscopy, we study the evolution of electrical properties, optically active point defects, and material strain in response to implantation and annealing. For Mg-implanted samples, increasing annealing temperature is characterized by an increase in the PL substitutional Mg-related peak (UVL) relative to the shallow-donor vacancy-related peak (GL2). Through comparison of implanted and in situ doped samples, it is demonstrated that the origin of compensating VN lies primarily in implantation rather than degradation from the annealing process. Transmission line measurements and diode I–V measurements show a sheet resistance of 1083 kΩ/□ and a hole concentration of 1.23 × 1015 cm−3, respectively, in the Mg-implanted material annealed at 1350 °C. We conclude that temperature-cycled gyrotron annealing at 1350 °C decreases implant-induced compensating point defects and activates Mg to obtain selective p-type conduction.

Meyers, V. (ORCID:0000000176731553)↗
DOE OSTI2021

High Mg activation in implanted GaN by high temperature and ultrahigh pressure annealing

Not provided.We demonstrate high p-type conductivity and hole concentrations >10 18 cm -3 in Mg-implanted GaN. The implantation was performed at room temperature and by post-implantation annealing at 1 GPa of N 2 and in a temperature range of 1200–1400 °C. The high pressure thermodynamically stabilized the GaN surface without the need of a capping layer. We introduce a “diffusion budget,” related to the diffusion length, as a convenient engineering parameter for comparing samples annealed at different temperatures and for different times. Although damage recovery, as measured by XRD, was achieved at relatively low diffusion budgets, these samples did not show p-type conductivity. Further analyses showed heavy compensation by the implantation-induced defects. Higher diffusion budgets resulted in a low Mg ionization energy (~115 meV) and almost complete Mg activation. For even higher diffusion budgets, we observed significant loss of Mg to the surface and a commensurate reduction in the hole conductivity. High compensation at low diffusion budgets and loss of Mg at high diffusion budgets present a unique challenge for shallow implants. A direct control of the formation of compensating defects arising from the implantation damage may be necessary to achieve both hole conductivity and low Mg diffusion.

36 MATERIALS SCIENCE↗
DOE OSTI2021

A first-principles understanding of point defects and impurities in GaN

Attaining control over the electrical conductivity of gallium nitride through impurity doping is one of the foremost achievements in semiconductor science. Yet, unwanted contaminants and point defects continue to limit device performance, and experimental techniques alone are insufficient for elucidating the behavior of these unintentionally incorporated species. Methodological advancements have made first-principles calculations more powerful than ever and capable of quantitative predictions, though care must still be taken in comparing results from theory and experiment. In this Tutorial, we explain the basic concepts that define the behavior of dopants, unintentional impurities, and point defects in GaN. We also describe how to interpret experimental results in the context of theoretical calculations and also discuss how the properties of defects and impurities vary in III-nitride alloys. Lastly, we examine how the physics of defects and impurities in GaN is relevant for understanding other wide-bandgap semiconductor materials, such as the II–IV-nitrides, boron nitride, and the transition metal nitrides.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Epitaxial Sc x Al 1- x N on GaN exhibits attractive high-K dielectric properties

Epitaxial Sc x Al 1- x N thin films of ~100 nm thickness grown on metal polar GaN substrates are found to exhibit significantly enhanced relative dielectric permittivity (ε r ) values relative to AlN. ε r values of ~17–21 for Sc mole fractions of 17%–25% ( x = 0.17–0.25) measured electrically by capacitance–voltage measurements indicate that Sc x Al 1- x N has the largest relative dielectric permittivity of any existing nitride material. Since epitaxial Sc x Al 1- x N layers deposited on GaN also exhibit large polarization discontinuity, the heterojunction can exploit the in situ high-K dielectric property to extend transistor operation for power electronics and high-speed microwave applications.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Edge termination in vertical GaN diodes: Electric field distribution probed by second harmonic generation

We characterized the electric field distribution of GaN-on-GaN p–n diodes with partially compensated ion-implanted edge termination (ET) using an electric field induced second harmonic generation technique (EFISHG). The distributed electric field from the anode to the outer edge of the ET demonstrates the effectiveness of the ET structure. However, EFISHG also shows that its effectiveness is strongly dependent on the acceptor charge distribution in the ET's partially compensated layer (PC). A generally lower amount of acceptor charge can be inferred from the measured electric field distribution resulting from excessive ion implantation energy or dose during ET fabrication and causing lower than optimal breakdown voltage. Localized field crowding can be observed when the remaining acceptors uncompensated by the implant in the PC layer are nonuniformly distributed around the periphery of the devices. Important information can be obtained from these direct electric field measurements and used for optimizing the device design and fabrication process.

Cao, Yuke↗
DOE OSTI2022

Indium as a surfactant: Effects on growth morphology and background impurity in GaN films grown by ammonia-assisted molecular beam epitaxy

We report on the improvement of the surface morphology of c-plane GaN films grown at high growth rates (~1 µm/h) using ammonia molecular beam epitaxy through a series of growth optimizations as well as the introduction of indium as a surfactant. The indium surfactant was expected to help with the adatom mobility and, thus, provide smoother growth surfaces. Through a combination of varying V/III ratios, In flux, and growth temperatures, an optimal condition for surface morphology, characterized by atomic force microscopy, was achieved. At higher Ga fluxes for fast growth rates (~1 µm/h and beam equivalent pressures of ~5 × 10 –7 Torr), higher ammonia flows were necessary to preserve the surface morphology. In addition, indium was an effective surfactant—reducing the roughness and improving the overall surface morphology. However, excessive indium causes the surface morphology to degrade, potentially due to the enhancement of the Ga desorption from the surface as a result of the reaction of indium with ammonia for high indium fluxes. The indium surfactant also resulted in a reduction of background Si impurity concentrations in the film. These effects allow for the growth of thick drift layers with low background dopant concentrations for vertical GaN power devices.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Dynamics of carrier injection through V-defects in long wavelength GaN LEDs

The efficiency of high-power operation of multiple quantum well (QW) light emitting diodes (LEDs) to a large degree depends on the realization of uniform hole distribution between the QWs. In long wavelength InGaN/GaN QW LEDs, the thermionic interwell hole transport is hindered by high GaN barriers. However, in polar LED structures, these barriers may be circumvented by the lateral hole injection via semipolar {$10\bar{1}1$} QWs that form on the facets of V-defects. The efficiency of such carrier transfer depends on the transport time since transport in the semipolar QWs is competed by recombination. In this work, we study the carrier transfer from the semipolar to polar QWs by time-resolved photoluminescence in long wavelength (green to red) LEDs. We find that the carrier transfer through the semipolar QWs is fast, a few tens of picoseconds with the estimated room temperature ambipolar diffusion coefficient of ∼5.5 cm 2 /s. With diffusion much faster than recombination, the hole transport from the p-side of the structure to the polar QWs should proceed without a substantial loss, contributing to the high efficiency of long wavelength GaN LEDs.

Crystallographic defects↗
DOE OSTI2026

AlGaN/GaN Core-Shell Nanofin Diodes on Si Substrates for Vertical GaN Electronics

We demonstrate plasma-assisted MBE selective-area growth of GaN/AlGaN core-shell structures on Si(111) as a path to vertical GaN devices. The selective-area growth enables the full vertical and uniform core-shell fin structures to be realized in a continuous growth without breaking vacuum. TEM images and EDS mapping of the core-shell structures show well-aligned crystal structure and sharp heterointerfaces. Dislocation filtering was observed in STEM imaging of the fin structure. P-N junction diodes fabricated with GaN/AlGaN core-shell structures reveal ideality factors as low as 1.5, and the reverse-bias leakage is consistent with trap-assisted space-charge-limited conduction. While practical challenges exist, including material-related leakage, growth-related challenges and optimization, this demonstration of p-n junction by this method may provide a path to vertical Superjunction device concepts in GaN and related materials.

Xiong, Juncheng [University of Notre Dame, IN (Uni↗
DOE OSTI2023

Selective area doping of GaN toward high-power applications

Abstract Selective area doping in GaN, especially p-type, is a critical and inevitable building block for the realization of advanced device structures for high-power applications, including, but not limited to, current-aperture vertical electron transistors, junction termination extensions, junction barrier Schottky diodes, junction field-effect transistors (JFETs), vertical-channel JFETs, U-shaped metal–oxide–semiconductor field-effect transistors (U-MOSFETs), and Fin MOSFETs. This paper reviews and summarizes some of the recent advances in the fields of selective area etching and regrowth, ion implantation, and polarity-dependent doping that may lead to the practical realization of GaN-based power devices.

Physics↗