Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Mask”

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 145 records · Page 8
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

New Approaches to Low-Cost Scalable Doping of Interdigitated back Contact Silicon Solar Cells (Final Report)

The goal of this project was to develop novel approaches to patterning of dopants in the rear fingers of interdigitated back contact (IBC) Si solar cells to reduce the cost of manufacturing this high-efficiency-potential cell architecture. The work throughout this project can be divided into four categories: (a) development of dopant patterning technique using laser scribed Si contacts masks that are mechanically aligned with ~10 μm resolution to the underlying Si substrate; (b) measuring dopant spreading profiles in the isolation region between n- and p-type dopant fingers during plasma-enhanced chemical vapor deposition (PECVD) of doped hydrogenated amorphous silicon (a-Si:H) via shadow masks, and dopant desorption and re-adsorption during high-temperature annealing; (c) understanding the role of dopant compensation on the shunt resistance in contaminated isolation regions through analysis of defect-enhanced compensation; and (d) simulation and fabrication of passivated two-sided grid and back-contact solar cells to demonstrate the use of direct dopant patterning in cell fabrication. For the passivated two-sided grid solar cells, masked deposition was used to demonstrate an improvement in the blue response of the cell by creating a shallow front emitter. During development of the masked PECVD patterning process, we measured 3-D dopant profiles using secondary ion mass spectrometry. After deposition, in the masked region, the phosphorus dopant tail was >100 µm at concentrations >10 19 cm -3 while the boron dopant tail was shorter. During high-temperature crystallization of doped a-Si:H films to polycrystalline Si (poly-Si), phosphorus atoms spread by desorbing from the poly-Si surface and re-adsorbing onto intrinsic poly-Si on adjacent wafers that were separated by several millimeters. These contamination mechanisms resulted in a decrease in resistivity from ~10 5 Ω·cm for intrinsic poly-Si to ~10 -1 Ω·cm for contaminated poly-Si. Mitigation strategies for each contamination mechanism were developed to maintain a resistivity of ~10 5 Ω·cm between doped fingers. During fabrication of the 209 cells created during this project, it was found that despite contamination of the IBC gap through the abovementioned mechanisms, high shunt resistances and FF ~75% were still reached. Investigation into this led to the discovery of defect-enhanced compensation which exists within highly defective poly-Si when net doping concentrations reach the value of defect density (~10 18 cm -3 for many poly-Si films). Simulations guided us in the fabrication of IBCs and the cells fabricated were able to meet the year-end goals for BP 2 and 3 of 15% and 17% IBC cell efficiency, as well as the BP 4 goal of a 1% absolute increase in efficiency for PERC-like devices. However, the most efficient cell created during this project of 18.6% fell short of the 21% final project target. While the champion device fell short in V oc and J sc , many devices fabricated were able to reach the necessary goals of ~40 mA/cm 2 , ~700 mV, and ~75% FF required for a 21% device. The results generated from this project were disseminated through 11 conference presentations and proceedings. Presentations included oral talks at 2019 IEEE PVSC, 2019 MRS Fall Meeting, 2020 PVSEC-30 and 2021 IEEE PVSC, as well as poster presentations at 2020 IEEE PVSC and 2021 SiPV. The project resulted in 2 peer reviewed publications – one published in IEEE Journal of Photovoltaics and one in ACS Applied Energy Materials. The information gained in this project will aid in development of improved processes for fabrication of high-efficiency solar cells and other areas of the semiconductor device industry as well. The IBC cells will also pave the way for higher efficiency tandem devices. Through further manufacturing of high-efficiency solar cells, more of the world’s energy demands can be met through renewable sources, helping to stave off the worst effects that may come about from global climate change.

14 SOLAR ENERGY↗
DOE OSTI2021

Spectroscopic Fourth-order Coronagraph for the Characterization of Terrestrial Planets at Small Angular Separations from Host Stars

We propose a new approach for high-contrast imaging at the diffraction limit using segmented telescopes in a modest observation bandwidth. This concept, named “spectroscopic fourth-order coronagraphy,” is based on a fourth-order coronagraph with a focal-plane mask that modulates the complex amplitude of the Airy disk along one direction. While coronagraphs applying the complex amplitude mask can achieve the theoretical limit performance for any arbitrary pupils, the focal-plane mask severely limits the bandwidth. Here, focusing on the fact that the focal-plane mask modulates the complex amplitude along one direction, we noticed that the mask can be optimized for each spectral element generated by a spectrograph. We combine the fourth-order coronagraph with two spectrographs to produce a stellar spectrum on the focal plane and reconstruct a white pupil on the Lyot stop. Based on the wave-front analysis of an optical design applying an Offner-type imaging spectrograph, we found that the achievable contrast of this concept is 10{sup −10} at 1.2–1.5 times the diffraction limit over the wavelength range of 650–750 nm for the entrance pupil of the LUVOIR telescope. Thus, this coronagraph concept could bring new habitable planet candidates not only around G- and K-type stars beyond 20–30 pc but also around very nearby M-type stars. This approach potentially promotes the characterization of the atmospheres of nearby terrestrial planets with future on- and off-axis segmented large telescopes.

47 OTHER INSTRUMENTATION↗
DOE OSTI2023

Graphics processor unit with opportunistic inter-path reconvergence

A graphics processing unit and methods for comping and executing instructions with opportunistic inter-path reconvergence are provided. A graphics processing unit may access computer executable instructions mapped to code blocks of a control flow for a warp. The code blocks may include an immediate dominator block and an intermediate post dominator block. The graphics processing unit may store a first thread mask associated with the first code block. The first thread mask may include a plurality of bits indicative of the active or non-active status for the threads of the warp, respectively. The graphics processing unit may a second thread mask corresponding to an intermediate code block between the immediate dominator block and intermediate post dominator block. The graphics processing unit may execute, with threads indicated as active by the first thread mask, instructions of the intermediate code block with a first operand or a second operand depending on the second thread mask.

Kulkarni, Milind↗
DOE OSTI2024

SAM-I-Am: Semantic boosting for zero-shot atomic-scale electron micrograph segmentation

Image segmentation is a critical enabler for tasks ranging from medical diagnostics to autonomous driving. However, the correct segmentation semantics — where are boundaries located? what segments are logically similar? — change depending on the domain, such that state-of-the-art foundation models can generate meaningless and incorrect results. Moreover, in certain domains, fine-tuning and retraining techniques are infeasible: obtaining labels is costly and time-consuming; domain images (micrographs) can be exponentially diverse; and data sharing (for third-party retraining) is restricted. To enable rapid adaptation of the best segmentation technology, we propose the concept of semantic boosting: given a zero-shot foundation model, guide its segmentation and adjust results to match domain expectations. Here, we apply semantic boosting to the Segment Anything Model (SAM) to obtain microstructure segmentation for transmission electron microscopy. Our booster, SAM-I-Am, serves as a post-processing engine that extracts geometric and textural features of various intermediate masks to perform mask removal and mask merging operations. We demonstrate a zero-shot performance increase of (absolute) +21.35%, +12.6%, +5.27% in mean IoU, and a -9.91%, -18.42%, -4.06% drop in mean false positive masks across images of three difficulty classes over vanilla SAM (ViT-L).

36 MATERIALS SCIENCE↗
DOE OSTI2022

Unveiling the influence of selective-area-regrowth interfaces on local electronic properties of GaN p-n junctions for efficient power devices

Here, we report correlated nanoscale mapping of the structure, composition, and properties of regrown GaN p-n junctions to identify how etching and non-planar regrowth processes limit diode performance via the introduction of unintentional dopants and defect states. p-GaN was selectively regrown in n-GaN trenches with SiO 2 masks of variable mask-to-trench-width ratio. Dilute Al layers were periodically introduced during regrowth as markers of the growth interface. Correlated nanoscale mapping of doping, conductivity, and dopant complexes was achieved with atom probe tomography (APT), scanning spreading resistance microscopy (SSRM), and cathodoluminescence (CL) spectroscopy, respectively. The Al marker layers, detected by APT, enabled reconstruction of the faceted growth interface and correlation of the dopant concentration with position and time. The p-GaN growth rate is proportional to the mask-to-trench width ratio while the dopant incorporation rate is invariant. At trench edges, magnesium incorporation is suppressed, and oxygen incorporation enhanced, due to preferential incorporation on the semi-polar growth surface, leading to compensation and less abrupt p-n junctions; the SiO 2 mask is a source of oxygen. Residual etch damage below the regrowth interface induces n-type and p-type conductivity, creating leakage pathways. The non-uniform Mg incorporation is driven by crystal anisotropy and is thus inherent to non-planar regrowth, but can be mitigated by engineering the regrowth interface and process parameters. The unprecedented integration of spatially resolved mapping of dopants, impurities, conductivity, and carrier type is a powerful approach to discriminating distinct factors that limit the performance of regrown diodes, enabling the rational optimization of process and device design.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Deep learning for morphological identification of extended radio galaxies using weak labels

Abstract The present work discusses the use of a weakly-supervised deep learning algorithm that reduces the cost of labelling pixel-level masks for complex radio galaxies with multiple components. The algorithm is trained on weak class-level labels of radio galaxies to get class activation maps (CAMs). The CAMs are further refined using an inter-pixel relations network (IRNet) to get instance segmentation masks over radio galaxies and the positions of their infrared hosts. We use data from the Australian Square Kilometre Array Pathfinder (ASKAP) telescope, specifically the Evolutionary Map of the Universe (EMU) Pilot Survey, which covered a sky area of 270 square degrees with an RMS sensitivity of 25–35 $\mu$ Jy beam $^{-1}$ . We demonstrate that weakly-supervised deep learning algorithms can achieve high accuracy in predicting pixel-level information, including masks for the extended radio emission encapsulating all galaxy components and the positions of the infrared host galaxies. We evaluate the performance of our method using mean Average Precision (mAP) across multiple classes at a standard intersection over union (IoU) threshold of 0.5. We show that the model achieves a mAP $_{50}$ of 67.5% and 76.8% for radio masks and infrared host positions, respectively. The network architecture can be found at the following link: https://github.com/Nikhel1/Gal-CAM

Astronomy & Astrophysics↗
DOE OSTI2024

Mechanisms of Three-Dimensional Solid-Phase Epitaxial Crystallization of Strontium Titanate

Strontium titanate (SrTiO 3 , STO) is a complex metal oxide with a cubic perovskite crystal structure. Due to its easily described and understood crystal structure in the cubic phase, STO is an ideal model system for exploring the mechanistic details of solid-phase epitaxy (SPE) in complex oxides. SPE is a crystallization approach that aims to guide crystal growth at low homologous temperatures to achieve targeted microstructures. Beyond planar thin films, SPE can also exploit the addition of a chemically inert, noncrystallizing, amorphous obstacle in the path of crystallization to generate complex three-dimensional structures. The introduction of this mask fundamentally alters the SPE process, inducing a transition from two- to three-dimensional geometries and from vertical to lateral crystal growth under the influence of the crystal/mask/amorphous boundary. Using a combination of molecular dynamics simulations and experiments, we identify several unique phenomena in the nanoscale growth behaviors in both conventional (unmasked) and masked SPE. Examining conventional SPE of STO, we find that crystallization at the interface is strongly correlated to, and potentially driven by, density fluctuations in the region of the amorphous STO near the crystalline/amorphous interface with a strong facet dependence. In the masked case, we find that the crystalline growth front becomes nonplanar near contact with the mask. We also observe a minimum vertical growth requirement prior to lateral crystallization. Both phenomena depend on the relative bulk and interfacial free energies of the three-phase (crystal/mask/amorphous) system.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Patchy nanoparticles by atomic stencilling

Stencilling, in which patterns are created by painting over masks, has ubiquitous applications in art, architecture and manufacturing. Modern, top-down microfabrication methods have succeeded in reducing mask sizes to under 10 nm, enabling ever smaller microdevices as today’s fastest computer chips. Meanwhile, bottom-up masking using chemical bonds or physical interactions has remained largely unexplored, despite its advantages of low cost, solution-processability, scalability and high compatibility with complex, curved and three-dimensional (3D) surfaces. Here we report atomic stencilling to make patchy nanoparticles (NPs), using surface-adsorbed iodide submonolayers to create the mask and ligand-mediated grafted polymers onto unmasked regions as ‘paint’. We use this approach to synthesize more than 20 different types of NP coated with polymer patches in high yield. Polymer scaling theory and molecular dynamics (MD) simulation show that stencilling, along with the interplay of enthalpic and entropic effects of polymers, generates patchy particle morphologies not reported previously. These polymer-patched NPs self-assemble into extended crystals owing to highly uniform patches, including different non-closely packed superlattices. We propose that atomic stencilling opens new avenues in patterning NPs and other substrates at the nanometre length scale, leading to precise control of their chemistry, reactivity and interactions for a wide range of applications, such as targeted delivery, catalysis, microelectronics, integrated metamaterials and tissue engineering.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Emission spectroscopy with coded apertures for enhanced dimensionality

A coded aperture is used to demonstrate emission spectroscopy from multiple one-dimensional measurement locations simultaneously with a single camera. The coded aperture mask has several columns of periodic apertures, each with a unique spatial frequency. Light transmitted through all mask columns is detected through an imaging spectrometer. Dispersed light from the various mask columns overlaps on the spectrometer camera but is separated using Fourier-domain filtering using the known spatial frequencies of the mask. As the coded aperture is placed at an image plane, each Fourier-filtered spectrogram comes from a unique one-dimensional measurement location. This technique represents a significant increase in the amount of spatially and spectrally resolved emission data available using a single emission spectrometer and camera at the expense of some spatial resolution due to the Fourier filtering. This instrument is particularly useful for studying transient, non-repeating events. Megahertz-rate emission spectroscopy from five one-dimensional measurement locations is demonstrated with explosive fireballs using a single camera. Optical design parameters and instrument performance characteristics are discussed.

42 ENGINEERING↗
DOE OSTI2020

Isolating p- and n-Doped Fingers With Intrinsic Poly-Si in Passivated Interdigitated Back Contact Silicon Solar Cells

Polycrystalline silicon on silicon oxide (poly-Si/SiO x ) passivating contacts enable ultra high efficiency interdigitated back contact silicon solar cells. To prevent shunt between n- and p-type doped fingers, an insulating region is required between them. We evaluate the use of intrinsic poly Si for this isolation region. Interdigitated fingers were formed by plasma deposition of doped hydrogenated amorphous silicon through mechanically aligned shadow masks, on top of a full-area intrinsic amorphous silicon layer. High temperature annealing then crystallized the a-Si:H to poly Si and drove in the dopants. Two mechanisms were identified which cause contamination of the intrinsic poly Si gap during processing. During deposition of doped fingers, we show using secondary ion mass spectrometry and conductivity measurements that the intrinsic gap becomes contaminated by doped a-Si:H tails several nanometers thick to concentrations of ~10 20 cm -3 . Another source of contamination occurs during high-temperature annealing, where dopants desorb from doped regions and readsorb onto intrinsic a Si:H. Both pathways reduce the resistivity of the intrinsic gap from ~10 5 Ω·cm to ~10 -1 Ω·cm. We show that plasma etching of the a-Si:H surface before crystallizing with a capping layer can eliminate the contamination of the intrinsic poly-Si, maintaining a resistivity of ~10 5 Ω·cm. Lastly, this demonstrates masked plasma deposition as a dopant patterning method for Si solar cells.

14 SOLAR ENERGY↗
DOE OSTI2021

Improved Localization Precision and Angular Resolution of a Cylindrical, Time-Encoded Imaging System From Adaptive Detector Movements

To the first order, the localization precision and angular resolution of a cylindrical, time-encoded imaging (c-TEI) system is governed by the geometry of the system. Improving either measure requires increasing the mask radius or decreasing the detector diameter, both of which are undesirable. Here, we propose an alternative option of repositioning the detector within the mask to increase the detector-to-mask distance in the direction of a source, thereby improving the localization precision and angular resolution in that direction. Since the detector-to-mask distance only increases for a small portion of the field of view (FOV), we propose implementing adaptive imaging where one leverages data collected during the measurement to optimize the system configuration. This article utilizes both simulations and experiments to set upper bounds on the potential gain from adaptive detector movements for one and two sources in the FOV. When only one source is present, adaptive detector movements can improve the localization precision and angular resolution by 20% for a source at 90 cm and by 32% for a far-field source. When two sources are present, adaptive detector movements can improve localization precision and angular resolution by up to 50% for sources that are ~10° apart (90 cm from the system). We experimentally verify these results through maximum likelihood estimation of the source position(s) and image reconstruction of point sources that are close together. As a demonstration of an adaptive imaging algorithm, we image a complex arrangement of special nuclear material at the Zero Power Physics Reactor facility at Idaho National Laboratory.

46 - INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AN↗
DOE OSTI2023

Adaptive methods of generating complex light arrays

Structured light arrays of various shapes have been a cornerstone in optical science, driven by the complexities of precise and adaptable generation. This study introduces an approach using a spatial light modulator (SLM) as a generator for these arrays. By projecting a holographic mask onto the SLM, it functions simultaneously as an optical convolution device, focusing mechanism, and structured light beam mask. Our approach offers unmatched versatility, allowing for the experimental fabrication of traditional beam arrays like azimuthal Laguerre–Gaussian (LG), Bessel–Gaussian (BG), and Hermite–Gauss (HG) in the far-field. Notably, it has enabled a method of generating Ince–Gauss (IG) and LG radial mode beam arrays using a convolution solution. Our system provides exceptional control over array periodicity and intensity distribution, bypassing the Talbot self-imaging phenomenon seen in traditional setups. We provide an in-depth theoretical discussion, supported by empirical evidence, of our far-field results. This method has vast potential for applications in optical communication, data processing, and multi-particle manipulation. It paves the way for rapid generation of structured light with high spatial frequencies and complex shapes, promising transformative advances in these domains.

Optics↗
DOE OSTI2022

Dynamically Downscaled (WRF) 1km, Hourly Meteorological Conditions 1987-2020. East/Taylor Watersheds

This dataset contains meteorological output from the Weather Research and Forecasting (WRF) version 3.8.1. This dataset has been created to 1) investigate hydrometeorological processes impacting the East River and water-delivery to the Critical Zone, and 2) provide meteorological forcing data for distributed Earth-science modeling applications in the East River watershed. Variables have a 1 kilometer spatial resolution and hourly temporal resolution and encompass a rectangular region encompassing the East and Taylor River watersheds, Colorado, near the town of Crested Butte. WRF was forced using Climate Forecast System Reanalysis (CFSR) lateral boundary conditions. Each .zip file contains one "water year" of data (October 1 -- September 30; i.e. water year 2017 starts October 1, 2016 and ends September 30, 2017). Each zip folder contains 12 netcdf (.nc) files containing one month of hourly data each and are approximately 250mb. Model timestamps are in UTC time.The files contain the following data variables:EAST_MASK: binary mask of the watershed regionTAYLOR_MASK: binary mask of the watershed regionGLW downwelling longwave radiation (w/m2)HR_PRCP: Hourly Precipitation Rate (mm/hr). Includes all hydrometeors (solid+liquid). HFX NoahMP LSM total grid-cell modelled sensible heat flux (w/m2) [positive towards atmosphere]LH NoahMP LSM total grid-cell modelled latent heat flux (w/m2) [positive towards atmosphere; can be converted to ET]PSFC Surface Barometric Pressure (hPa)Q2 Two-meter specific humidity (kg/kg)SWDOWN Downwelling shortwave solar radiation (w/m2)SWNORM Terrain-normal downwelling shortwave radiation (w/m2)T2 Two-meter air temperature (deg K)U10 10-m U-component of wind velocity (m/s)V10 10m V-component of wind velocity (m/s)XLAT Latitude of grid-center point XLONG Longitude of grid-center point XTIME Model timestamp, **in UTC**

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Addressing Issues with Working Memory in Video Object Segmentation

Contemporary state-of-the-art video object segmentation (VOS) models compare incoming unannotated images to a history of image-mask relations via affinity or cross-attention to predict object masks. We refer to the internal memory state of the initial image-mask pair and past image-masks as a working memory buffer. While the current state of the art models perform very well on clean video data, their reliance on a working memory of previous frames leaves room for error. Affinity-based algorithms include the inductive bias that there is temporal continuity between consecutive frames. To account for inconsistent camera views of the desired object, working memory models need an algorithmic modification that regulates the memory updates and avoid writing irrelevant frames into working memory. A simple algorithmic change is proposed that can be applied to any existing working memory-based VOS model to improve performance on inconsistent views, such as sudden camera cuts, frame interjections, and extreme context changes. The resulting model performances show significant improvement on video data with these frame interjections over the same model without the algorithmic addition. Our contribution is a simple decision function that determines whether working memory should be updated based on the detection of sudden, extreme changes and the assumption that the object is no longer in frame. By implementing algorithmic changes, such as this, we can increase the real-world applicability of current VOS models.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

The FastrSHWFS Project Development Motivation, Analysis, & Test Results

The recent 2020 Decadal Survey of Astronomy and Astrophysics listed habitable exoplanet imaging with future extreme adaptive optics (AO) on 30m-class telescopes as a key priority in the coming decade. However, there is a current 100x contrast gap between the best systems today and what is needed to enable this goal. Astronomical AO is a required approach to enable ground-based diffraction-limited imaging of exoplanets on future extremely large telescopes. Time lag between the end of an exposure and the application of deformable mirror commands is a major contributor to the error budget in many AO systems, and detector read time is often a large component of this lag. We present two designs for a modified Shack Hartmann wavefront sensor (SHWFS), named Focal plane Actualized Shifted Technique Realized for a SHWFS (fastrSHWFS), to reduce the time lag component. This design steers the spot pattern at the focal plane into a rectangular or linear array with a custom aspect ratio, reducing readout time. The mask with focus yields aberrated results while the mask with tip/tilt only yields some defined spots. This essay outlines the current SHWFS concept, our fastrSHWFS theoretical solution to addressing time lag, reflection and quality analysis of printed mask designs, and results from testing both masks on the High Contrast Testbed at Lawrence Livermore National Lab. This work follows the test of a previous fastrSHWFS design.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Preventing Reverse Engineering of Critical Industrial Data with DIOD

Business analytics augmented by artificial intelligence and machine learning (AI/ML) have revolutionized the role of data in the modern world. In recent years, businesses have incorporated data into their decision-making process for better prediction, risk-assessment, content creation, etc. While such businesses often seek to leverage the full use of their data through third-party AI/ML services, they are often hampered by the risks of data leaks, reverse-engineering, stolen technology, etc. that often have disastrous consequences for businesses and their stakeholders alike. Thus, there arises a need for data masking prior to its transmission that obfuscates proprietary information while preserving the information relevant for AI/ML applications. In order to meet the needs of industrial data which are significantly different from those of data warehouses, previous work proposed an efficient time and space-scalable data masking paradigm known as the deceptive infusion of data (DIOD) methodology. The present work expands upon this work by leveraging existing reverse-engineering capabilities to facilitate the decomposition of industrial data into its proprietary and AI/ML-relevant parts, referred to as fundamental and inference metadata respectively. Both sets of metadata are further obfuscated in accordance with the DIOD methodology to create the DIOD rendition of the industrial data, which is rendered immune to reverse-engineering by discarding proprietary information and only preserving AI/ML-relevant information. Additionally, constraints of the original DIOD manuscript are relaxed using mutual information by configuring the methodology to the target AI/ML application to unlock the full potential of the DIOD methodology. As an example, data from a nuclear reactor is transformed into that from a nonlinear spring-mass system with different levels of data masking as required by the generic system and the target application.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Birefringent Glass‐Engraved Quasi‐Linear Nanograting Metasurface Based on Self‐Organizing Process for Large Aperture High Power Laser Applications

All-glass metasurface “nanograting” structures that exhibit birefringence in the formed layer are reported. The key enabler of this work is ion beam processing at an angle sufficiently off-normal incidence, inducing self-assembly of a deposited metal layer into quasi-linear metallic features that can function as an etching mask. As a result, a fused silica metasurface, monolithic to the underlying substrate, is demonstrated at 375 nm wavelength to exhibit a phase delay angle of 30° between the principal axes. The capability of an angled etch mask replenishment process is also demonstrated for achieving deeper etch depth and for increasing the grating period, another first – to the best of the knowledge. This is the first display of a technology capable of fabricating glass-engraved near-linear grating structure with a feature-to-feature period as small as 118.6 nm. Furthermore, this technology has the potential to generate grating-like structures with periods as small as 12.4 nm, as demonstrated here with reactive ion beam processing assisted mask assembly. Furthermore, these structures are shown to have reflectivity < 0.4% across the wavelength band 350 nm – 1000 nm. Such a technology can enable laser-durable grating structures for the deep-UV and even down to soft X-ray wavelengths.

Ray, Nathan J. [Lawrence Livermore National Labora↗
DOE OSTI2023

Multiscale graph neural network autoencoders for interpretable scientific machine learning

The goal of this work is to address two limitations in autoencoder-based models: latent space interpretability and compatibility with unstructured meshes. This is accomplished here with the development of a novel graph neural network (GNN) autoencoding architecture with demonstrations on complex fluid flow applications. To address the first goal of interpretability, the GNN autoencoder achieves reduction in the number nodes in the encoding stage through an adaptive graph reduction procedure. Further, this reduction procedure essentially amounts to flowfieldconditioned node sampling and sensor identification, and produces interpretable latent graph representations tailored to the flowfield reconstruction task in the form of so-called masked fields. These masked fields allow the user to (a) visualize where in physical space a given latent graph is active, and (b) interpret the time-evolution of the latent graph connectivity in accordance with the time-evolution of unsteady flow features (e.g. recirculation zones, shear layers) in the domain. To address the goal of unstructured mesh compatibility, the autoencoding architecture utilizes a series of multi-scale message passing (MMP) layers, each of which models information exchange among node neighborhoods at various lengthscales. The MMP layer, which augments standard single-scale message passing with learnable coarsening operations, allows the decoder to more efficiently reconstruct the flowfield from the identified regions in the masked fields. Analysis of latent graphs produced by the autoencoder for various model settings are conducted using unstructured snapshot data sourced from large-eddy simulations in a backward-facing step (BFS) flow configuration with an OpenFOAM-based flow solver at high Reynolds numbers.

97 MATHEMATICS AND COMPUTING↗