Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “time scale separation”

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

Intermediate scattering functions of a rigid body monoclonal antibody protein in solution studied by dissipative particle dynamic simulation

In the past decade, there was increased research interest in studying internal motions of flexible proteins in solution using Neutron Spin Echo (NSE) as NSE can simultaneously probe the dynamics at the length and time scales comparable to protein domain motions. However, the collective intermediate scattering function (ISF) measured by NSE has the contributions from translational, rotational, and internal motions, which are rather complicated to be separated. Widely used NSE theories to interpret experimental data usually assume that the translational and rotational motions of a rigid particle are decoupled and independent to each other. To evaluate the accuracy of this approximation for monoclonal antibody (mAb) proteins in solution, dissipative particle dynamic computer simulation is used here to simulate a rigid-body mAb for up to about 200 ns. The total ISF together with the ISFs due to only the translational and rotational motions as well as their corresponding effective diffusion coefficients is calculated. The aforementioned approximation introduces appreciable errors to the calculated effective diffusion coefficients and the ISFs. For the effective diffusion coefficient, the error introduced by this approximation can be as large as about 10% even though the overall agreement is considered reasonable. Thus, we need to be cautious when interpreting the data with a small signal change. In addition, the accuracy of the calculated ISFs due to the finite computer simulation time is also discussed.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2020

Predicting Excitation Energies of Twisted Intramolecular Charge-Transfer States with the Time-Dependent Density Functional Theory: Comparison with Experimental Measurements in the Gas Phase and Solvents Ranging from Hexanes to Acetonitrile

Electronically excited states characterized by intramolecular charge transfer play an essential role in many biological processes and optical devices. The ability to make quantitative ab initio predictions of the relative energetics involved is a challenging yet desirable goal, especially for large molecules in solution. In this work, we present a data set of 61 experimental measurements of absorption and emission processes, both in the gas phase and in solvents representing a broad range of polarities, which involve intramolecular charge transfer mediated by a nonzero, "twisted" dihedral angle between one or more donor and acceptor subunits. Among a variety of density functionals investigated within the framework of linear-response theory, the "optimally tuned" LRC-ωPBE functional, which utilizes a system-specific yet nonempirical procedure to specify the range-separation parameter, emerges as the preferred choice. For the entire set of excitation energies, involving changes in dipole moment ranging from 4 to >20 Debye, the mean signed and absolute errors are 0.02 and 0.18 eV, respectively (compared, e.g., to -0.30 and 0.30 for PBE0, 0.44 and 0.47 for LRC-ωPBEh, 0.83 and 0.83 for ωB97X-V). We analyze the performance of polarizable continuum solvation models available in Q-Chem that partition the solvent response into fast and slow time scales, and clear trends emerge when measurements corresponding to the four small 4-(dimethylamino)benzonitrile (DMABN)-like molecules and a charged species are excluded. We make the case that the large errors found only for small molecules in the gas phase and weak solvents cannot be expected to improve via the optimal tuning procedure, which enforces a condition that is exact only in the well-separated donor-acceptor limit, and present empirical evidence implicating the outsized importance for small donor-acceptor systems of relaxation effects that cannot be accounted for by the linear-response time-dependent density functional theory within the adiabatic approximation. Finally, we demonstrate the utility of the optimally tuned density functional approach by targeting the charge-transfer states of a large biomimetic model system for light-harvesting structures in Photosystem II.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Measuring the neutrino-oxygen neutral current quasielastic cross section using the accelerator neutrino neutron interaction experiment

The Accelerator Neutrino Neutron Interaction Experiment (ANNIE) is a 26-ton gadolinium-doped water Cherenkov detector located on-axis to Fermilab’s Booster Neutrino Beam (BNB). ANNIE is uniquely positioned to perform high-statistics measurements of neutrino-nucleus interactions in water, benefiting from a large neutrino flux due to a short (100-meter) baseline. A central focus of ANNIE’s physics program is the measurement of both charged current (CC) and neutral current (NC) cross sections on water, including neutral current quasielastic (NCQE) and CC-inclusive channels. The NCQE measurement is particularly critical for constraining uncertainties in rare-event searches such as the Diffuse Supernova Neutrino Background (DSNB), where atmospheric $\nu$NCQE interactions constitute a significant and poorly constrained background. This dissertation presents a measurement of the flux-averaged neutrino-oxygen neutral current quasielastic ($\nu$NCQE) cross section using $2.573 \times 10^{20}$~POT of BNB exposure from the 2022 and 2023 beam years. The $\nu$NCQE interaction is identified through the primary $\gamma$-rays produced by nuclear de-excitation of the residual $^{15}$N$^*$ or $^{15}$O$^*$ nucleus following nucleon knockout from $^{16}$O. A dedicated Monte Carlo (MC) re-tuning campaign was conducted using an americium-beryllium (AmBe) calibration source, Michel electrons from stopped muons, and throughgoing dirt muons originating upstream of the detector. This multi-sample approach provided a wide-ranging $\mathcal{O}(\text{MeV})$--$\mathcal{O}(\text{GeV})$ dataset for tuning the simulated detector response, which was subsequently validated against AmBe neutron and Michel electron data for use in the $\nu$NCQE analysis. A dedicated laser calibration campaign was carried out to reduce timing uncertainties across the PMT system, enabling reconstruction of the BNB bunch substructure with sufficient resolution to serve as a background rejection tool. By selecting events in-time with individual neutrino bunches, beam-correlated $\nu$NCQE events are separated from diffuse and accelerator-induced backgrounds, notably skyshine neutrons and externally-originating events, that would otherwise dominate traditional charge-based selections within a small-scale, surface-level, short-baseline detector. A data-driven estimation of the skyshine neutron and external background rates was performed and incorporated into the systematic uncertainty budget. The flux-averaged $\nu$NCQE cross section on oxygen is measured to be $1.57 \pm 0.06\,(\text{stat.})$ $^{+0.91}_{-0.67}\,(\text{syst.})$ $\times 10^{-38}\ \text{cm}^{2}$. A full systematic budget is constructed by propagating uncertainties in the secondary hadronic interaction modeling, background cross section normalizations, detector response, neutrino flux, and the primary $\gamma$-ray emission probabilities from oxygen nuclear de-excitation. An idealized de-excitation model, constructed from existing measurements in the literature is developed to benchmark the predictions of the \textsc{GENIE} event generator. A comparison reveals that \textsc{GENIE} systematically overpredicts the primary $\gamma$-ray emission probability from oxygen de-excitation by a factor of $1.49\times$ for $E_\gamma > 6$~MeV and $3.07\times$ in the $3$--$6$~MeV band. This comparison motivates the dominant systematic uncertainty in this analysis, where a conservative uncertainty of $^{+39.9\%}_{-0\%}$ on the primary $\gamma$-ray signal prediction is assigned. The ANNIE result is consistent with and complementary to existing flux-averaged $\nu$NCQE cross section measurements from T2K and Super-Kamiokande, providing an independent measurement with a different detector, neutrino beam, and analysis methodology. Looking ahead, an upgrade to the ANNIE DAQ infrastructure enabling continuous extended readout will allow a complementary $\nu$NCQE neutron multiplicity measurement, directly relevant to constraining the NCQE background in DSNB searches, competitive with the recent T2K measurement at SK-Gd. The planned Super-SANDI upgrade, deploying a large Water-based Liquid Scintillator (WbLS) volume, will further extend ANNIE's reach to hadronic final states and exclusive NC channels, and enable joint measurements with liquid argon detectors sharing the BNB beamline ahead of DUNE and Hyper-Kamiokande.

Doran, Steven [Iowa State U.]↗
DOE OSTI2022

Precise stellarator quasi-symmetry can be achieved with electromagnetic coils

Magnetic fields with quasi-symmetry are known to provide good confinement of charged particles and plasmas, but the extent to which quasi-symmetry can be achieved in practice has remained an open question. Recent work [M. Landreman and E. Paul, Phys. Rev. Lett. 128, 035001, 2022] reports the discovery of toroidal magnetic fields that are quasi-symmetric to orders-of-magnitude higher precision than previously known fields. We show that these fields can be accurately produced using electromagnetic coils of only moderate engineering complexity, that is, coils that have low curvature and that are sufficiently separated from each other. Our results demonstrate that these new quasi-symmetric fields are relevant for applications requiring the confinement of energetic charged particles for long time scales, such as nuclear fusion. The coils’ length plays an important role for how well the quasi-symmetric fields can be approximated. For the longest coil set considered and a mean field strength of 1 T, the departure from quasi-symmetry is of the order of Earth’s magnetic field. Additionally, we find that magnetic surfaces extend far outside the plasma boundary used by Landreman and Paul, providing confinement far from the core. Simulations confirm that the magnetic fields generated by the new coils confine particles with high kinetic energy substantially longer than previously known coil configurations. In particular, when scaled to a reactor, the best found configuration loses only 0.04% of energetic particles born at midradius when following guiding center trajectories for 200 ms.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2021

Performing In Situ Closed-Cell Gas Reactions in the Transmission Electron Microscope

Gas reactions studied by in situ electron microscopy can be used to capture the real-time morphological and microchemical transformations of materials at length scales down to the atomic level. In situ closed-cell gas reaction (CCGR) studies performed using (scanning) transmission electron microscopy (STEM) can separate and identify localized dynamic reactions, which are extremely challenging to capture using other characterization techniques. For these experiments, we used a CCGR holder that utilizes microelectromechanical systems (MEMS)-based heating microchips (hereafter referred to as "E-chips"). The experimental protocol described here details the method for performing in situ gas reactions in dry and wet gases in an aberration-corrected STEM. This method finds relevance in many different materials systems, such as catalysis and high-temperature oxidation of structural materials at atmospheric pressure and in the presence of various gases with or without water vapor. Here, several sample preparation methods are described for various material form factors. During the reaction, mass spectra obtained with a residual gas analyzer (RGA) system with and without water vapor further validates gas exposure conditions during reactions. Integrating an RGA with an in situ CCGR-STEM system can, therefore, provide critical insight to correlate gas composition with the dynamic surface evolution of materials during reactions. In situ/operando studies using this approach allow for detailed investigation of the fundamental reaction mechanisms and kinetics that occur at specific environmental conditions (time, temperature, gas, pressure), in real-time, and at high spatial resolution.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2023

Stream lamination and rapid mixing in a microfluidic jet for X-ray spectroscopy studies

Microfluidic mixers offer new possibilities for the study of fast reaction kinetics down to the microsecond time scale, and methods such as soft X-ray absorption spectroscopy are powerful analysis techniques. These systems impose challenging constraints on mixing time scales, sample volume, detection region size and component materials. The current work presents a novel micromixer and jet device which aims to address these limitations. The system uses a so-called ‘theta’ mixer consisting of two sintered and fused glass capillaries. Sample and carrier fluids are injected separately into the inlets of the adjacent capillaries. At the downstream end, the two streams exit two micron-scale adjoining nozzles and form a single free-standing jet. The flow-rate difference between the two streams results in the rapid acceleration and lamination of the sample stream. This creates a small transverse dimension and induces diffusive mixing of the sample and carrier stream solutions within a time scale of 0.9 microseconds. The reaction occurs at or very near a free surface so that reactants and products are more directly accessible to interrogation using soft X-ray. We use a simple diffusion model and quantitative measurements of fluorescence quenching (of fluorescein with potassium iodide) to characterize the mixing dynamics across flow-rate ratios.

Huyke, Diego A. (ORCID:0000000283356613)↗
DOE OSTI2024

Mineral Deposition on the Rough Walls of a Fracture

Modeling carbonate growth in fractures and pores is important for understanding carbon sequestration in the environment or when supersaturated solutions are injected into rocks. Here, we study the simple but nontrivial problem of calcite growth on fractures with rough walls of the same mineral using kinetic Monte Carlo simulations of attachment and detachment of molecules and scaling approaches. First, we consider wedge-shaped fracture walls whose upper terraces are in the same low-energy planes and show that the valleys are slowly filled by the propagation of parallel monolayer steps in the wedge sides. The growth ceases when the walls reach these low-energy configurations so that a gap between the walls may not be filled. Second, we consider fracture walls with equally separated monolayer steps (vicinal surfaces with roughness below 1 nm) and show that growth by step propagation will eventually clog the fracture gap. In both cases, scaling approaches predict the times to attain the final configurations as a function of the initial geometry and the step-propagation velocity, which is set by the saturation index. The same reasoning applied to a random wall geometry shows that step propagation leads to lateral filling of surface valleys until the wall reaches the low-energy crystalline plane that has the smallest initial density of molecules. Thus, the final configurations of the fracture walls are much more sensitive to the crystallography than to the roughness or the local curvature. The framework developed here may be used to determine those configurations, the times to reach them, and the mass of deposited mineral. Effects of transport limitations are discussed when the fracture gap is significantly narrowed.

calcite↗
DOE OSTI2021

Three practical workflow schedulers for easy maximum parallelism

Runtime scheduling and workflow systems are an increasingly popular algorithmic component in HPC because they allow full system utilization with relaxed synchronization requirements. There are so many special-purpose tools for task scheduling, one might wonder why more are needed. Use cases seen on the Summit supercomputer needed better integration with MPI and greater flexibility in job launch configurations. Preparation, execution, and analysis of computational chemistry simulations at the scale of tens of thousands of processors revealed three distinct workflow patterns. A separate job scheduler was implemented for each one using extremely simple and robust designs: file-based, task-list based, and bulk-synchronous. Comparing to existing methods shows unique benefits of this work, including simplicity of design, suitability for HPC centers, short startup time, and well-understood per-task overhead. All three new tools have been shown to scale to full utilization of Summit, and have been made publicly available with tests and documentation. This work presents a complete characterization of the minimum effective task granularity for efficient scheduler usage scenarios. Here, these schedulers have the same bottlenecks, and hence similar task granularities as those reported for existing tools following comparable paradigms.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Plasmon-induced excitation energy transfer in silver nanoparticle dimers: A real-time TDDFTB investigation

Here, using real-time quantum dynamics calculations, we perform theoretical investigations of light-induced interactions and electronic excitation transfer in a silver nanoparticle dimer. Real-time time-dependent density functional tight-binding (RT-TDDFTB) calculations provide details of the quantum dynamical processes at an electronic/atomistic level with attosecond resolution. The computational efficiency of RT-TDDFTB allows us to examine electronic dynamics up to picosecond time scales. With time scales varying over six orders of magnitude, we provide insight into interactions between the nanoparticle and laser and between nanoparticles. Our results show that the coupling between nanoparticle monomers is dependent on the separation distance between the nanoparticles in the dimer. As the interparticle distance is varied, the dipole–dipole interactions and electronic excitation transfer mechanisms are markedly different. At large distances (from 50 to 20 Å), the energy transfer from NP1 to NP2 becomes more efficient as the interparticle distance decreases. The total dipole moment of the Ag 14 nanoparticle dimer increases linearly at an interparticle distance of 20 Å and reaches its maximum after 1.2 ps. The electronic excitation transfer is also the most efficient at 20 Å. At short distances, back-transfer effects reduce the ability of the dimer and NP1 to accept energy from the incident electric field. We attribute the distance-dependent features of the nanoparticle dimer to the beating between the laser acting on NP1 and the back transfer from NP2 to NP1.

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

Equilibrium Separation of Siloxanes in Metal–Organic Frameworks

We present an in silico assessment of metal–organic frameworks (MOFs) for the equilibrium separation of linear and cyclic siloxanes. Using a combination of configurational bias/continuous fractional component Monte Carlo (CB/CFCMC) simulations and Ideal Adsorbed Solution Theory (IAST), we investigated the adsorption of both equimolar and nonequimolar mixtures of linear and cyclic siloxanes in a selection of synthesizable MOFs with medium to large pore volumes. We showed that configurational entropy effects drive the preferential adsorption of linear siloxanes over cyclic siloxanes. Based on synthesizability metrics we identified ZIF-70 as a promising adsorbent for the separation of linear and cyclic siloxanes. Self-diffusivities of linear and cyclic siloxanes in ZIF-70 calculated from molecular dynamics simulations show that equilibrium can be reached on reasonable time scales. We explored vacuum-temperature swing adsorption (VTSA) as a potential method for the recovery of adsorbed linear siloxanes from ZIF-70, achieving a 75% reduction in adsorbed phase concentration. Additionally, we demonstrated that supercritical CO 2 offers an alternative desorption strategy by displacing adsorbed linear siloxanes in ZIF-70 at pressures above 200 bar driven by entropy effects.

Adsorption↗
DOE OSTI2023

Holistic fleet optimization incorporating system design considerations

The methodology described in this article enables a type of holistic fleet optimization that simultaneously considers the composition and activity of a fleet through time as well as the design of individual systems within the fleet. Often, real-world system design optimization and fleet-level acquisition optimization are treated separately due to the prohibitive scale and complexity of each problem. Importantly, this means that fleet-level schedules are typically limited to the inclusion of predefined system configurations and are blind to a rich spectrum of system design alternatives. Similarly, system design optimization often considers a system in isolation from the fleet and is blind to numerous, complex portfolio-level considerations. In reality, these two problems are highly interconnected. To properly address this system-fleet design interdependence, we present a general method for efficiently incorporating multi-objective system design trade-off information into a mixed-integer linear programming (MILP) fleet-level optimization. This work is motivated by the authors' experience with large-scale DOD acquisition portfolios. However, the methodology is general to any application where the fleet-level problem is a MILP and there exists at least one system having a design trade space in which two or more design objectives are parameters in the fleet-level MILP.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Ligand Control of Ultrafast Hot-Carrier Cooling in CdSe Quantum Dots by a Coherent Nonadiabatic Mechanism

Using the results from global modeling of the broadband two-dimensional electronic spectrum and an analysis of the concurrent damping of excited-state vibronic coherences, we show herein that a coherent nonadiabatic mechanism converts the electronic excitation energy of the X3 (1P 3/2 –1P e ) exciton state to ligand vibrations on the <50 fs time scale in oleate- and hexadecylamine-capped CdSe quantum dots. A comparison of the rates for the two ligands suggests that this process is promoted by mid-frequency vibrations of the ligands due to modulation of their electron-donating tendency. An intramolecular vibrational redistribution process then follows on the ∼200 fs time scale with both ligands, which accompanies thermalization in the band-edge X1 (1S 3/2 –1S e ) state and dephasing of the spectator ligand vibrations. These findings suggest that charge-separated intermediates associated with photoinduced charge transfer or triplet–triplet excitation energy transfer will be produced with retention of phase coherence in the vibrations that modulate the charge-transfer character of surface-bound organic acceptors.

14 SOLAR ENERGY↗
DOE OSTI2025

Real-Time Characterization of Salt Aerosols Generated from Static and Sparged Molten Salt

The formation of radionuclide-bearing aerosols in the respirable size range has the potential to significantly influence offsite dose consequences and is, therefore, an important consideration in nuclear facility safety assessments. Molten salt reactor (MSR) developers will likely need to demonstrate an understanding of the conditions under which radionuclide-bearing aerosols may be generated from their reactor under normal operating and accident conditions, as well as the characteristics and transport behavior of these aerosols, to demonstrate to the U.S. Nuclear Regulatory Commission (NRC) that the facility can be operated safely. Recent reviews of the literature identified a lack of experimental data describing the mechanisms of formation and properties (size, concentration, and composition) of salt aerosol particles that are produced from molten salts. Experiments that identify the conditions that lead to radionuclide-bearing salt aerosol releases and quantify the characteristics of salt aerosols formed by different mechanisms are high-priority needs to support MSR licensing. This report describes tests that were conducted within the Argonne Salt Aerosol Test Stand (a sealed vessel and measurement system) to generate salt aerosols from static and sparged molten salts and measure their size and concentration in real-time. The results provide insight into salt aerosol formation by the vapor condensation and bubble bursting mechanisms and inform the potential radiological consequences of aerosol formation from molten fuel salt. Videos of the salt surface were taken during salt sparge tests to observe surface bubble behavior. The data in this report can be used to develop mechanistic source term and accident progression models for MSRs. The real-time salt aerosol characterization technique used in this study will be employed in future integral effects tests that are conducted at an engineering scale to simulate realistic MSR accidents and in future separate effects tests to address additional variables that may impact salt aerosol characteristics (e.g., presence of fission products in salt and humidity in atmosphere).

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2021

Generation of Subauroral Longitudinally Extended Emissions Following Intensifications of the Poleward Boundary of the Substorm Bulge and Streamer Production

Observations showing the development of Subauroral Longitudinally Extended Emigerssions on a global scale are presented. It is demonstrated that they occur as a separation of an east-west arc-like band of luminosity that detaches away from the equatorward edge of the auroral distribution following episodes of auroral streamer production. They persist for time intervals on the order of ~30 min and devolve into patchy disjointed segments before they fade. Emissions are seen in both 557.7 nm OI and 391.4 nm N + 2 1NG lines, and 630.0 nm emissions are often observed equatorward. The 391.4 nm emissions are typically weaker and fade away more quickly than the 557.7 nm emissions, suggesting that an auroral precipitation source is initially present but is rapidly depleted as the forms age. All cases are associated with enhanced Subauroral Polarization Stream (SAPS) flows, and one event shows clear association with large-scale Giant Undulations (GUs) and the formation of spur-like forms that fold equatorward under the main oval. A model is proposed for the production of subauroral longitudinally extended forms in which nonlinear growth of SAPS-induced surface waves on the plasmapause results in a disruption of the boundary. It is suggested that hot plasma of plasma sheet origin becomes entrained in the plasmasphere to produce transient precipitation-associated auroral emissions that may decay into STEVE emissions and that cold dense plasma from the plasmasphere becomes entrained onto open drift paths to feed long-lived drainage plumes. This process may occur quasi periodically during intense substorms and storm-time conditions.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

The Effect of the Flux Separability Approximation on Multigroup Neutron Transport

The angular dependence of flux-weighted multigroup cross sections is commonly neglected when generating multigroup libraries. The error of this flux separability approximation is typically not isolated from other error sources due to a lack of availability of library generation and corresponding solvers that cannot relax this approximation. These errors can now be isolated and quantified with the availability of a multigroup Monte Carlo transport and multigroup library-generation capability in the OpenMC Monte Carlo transport code. This work will discuss relevant details of the OpenMC implementation, provide an example case useful for detailing the type of errors one can expect from making the flux separability approximation, and end with more realistic problems which show the impact of the approximation and highlight how it can strongly arise from an energy-dependent resonance absorption effect. Since the angle-dependence is intrinsically linked to the energy group structure, these examples also show that relaxing the flux separability approximation with angle-dependent cross sections could be used to reduce either the fine-tuning required to set a multigroup energy structure for a specific reactor type or the number of energy groups required to obtain a desired level of accuracy for a given problem. This trade-off could increase the costs of generating multigroup cross sections, and has the potential to require more memory for storing the multigroup library during the transport calculations, but it can significantly reduce the computational time required since the runtime of a discrete ordinates or method of characteristics neutron transport solver scales roughly linearly with the number of groups.

42 ENGINEERING↗
DOE OSTI2020

Fracture resistance of hierarchical Cu–Mo nanocomposite thin films

We report the fracture behavior of high-temperature co-sputtered Cu/Mo nanocomposite thin films was investigated through in situ 3-point bend testing of pre-notched microbeams in SEM. The as-synthesized nanocomposites present hierarchical nano/microstructures composed of a matrix of phase-separated Cu–Mo with nanoscale ligaments dispersed with sub-micron scale Cu-rich islands containing Mo nano-precipitates. Results show a significant crack growth resistance in the hierarchical nanocomposite, several times higher than that measured in Cu/Mo nanoscale multilayers. Based on electron microscopy characterization, three mechanisms of crack growth resistance in the hierarchical structures are proposed: crack bridging by the Cu-rich layer, crack deflection via shear along the Cu/Mo interface, and multiple cracking. This work demonstrates an approach to increase toughness in high strength nanocomposites through interface micro-structure design.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Atomistic Origin of Microsecond Carrier Lifetimes at Perovskite Grain Boundaries: Machine Learning-Assisted Nonadiabatic Molecular Dynamics

The polycrystalline nature of perovskites, stemming from their facile solution-based fabrication, leads to a high density of grain boundaries (GBs) and point defects. However, the impact of GBs on perovskite performance remains uncertain, with contradictory statements found in the literature. We developed a machine learning force field, sampled GB structures on a nanosecond time scale, and performed nonadiabatic (NA) molecular dynamics simulations of charge carrier trapping and recombination in stoichiometric and doped GBs. The simulations reveal long, microsecond carrier lifetimes, approaching experimental data, stemming from charge separation at the GBs and small NA coupling, 0.01–0.1 meV. Stoichiometric GBs exhibit transient trap states, which, however, are not particularly detrimental to the carrier lifetime. Halide dopants form interstitial defects in the bulk, but have a stabilizing influence on the GB structure by passivating undersaturated Pb atoms and reducing the transient trap state formation. On the contrary, excess Pb destabilizes GBs, allowing formation of persistent midgap states that trap charges. Still, the charge carrier lifetime reduces relatively little, because the midgap states decouple from the bands, and charges are more likely to escape back into bands upon a GB structural fluctuation. The atomistic study into the structural dynamics of perovskite GBs and its influence on charge carrier trapping and recombination provides valuable insights into the complex properties of perovskites and the intricate role of GBs in the material performance.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Parallel physics-informed neural networks via domain decomposition

Here we develop a distributed framework for the physics-informed neural networks (PINNs) based on two recent extensions, namely conservative PINNs (cPINNs) and extended PINNs (XPINNs), which employ domain decomposition in space and in time-space, respectively. This domain decomposition endows cPINNs and XPINNs with several advantages over the vanilla PINNs, such as parallelization capacity, large representation capacity, efficient hyperparameter tuning, and is particularly effective for multi-scale and multi-physics problems. Here, we present a parallel algorithm for cPINNs and XPINNs constructed with a hybrid programming model described by MPI + X, where X ∈ {CPUs, GPUs}. The main advantage of cPINN and XPINN over the more classical data and model parallel approaches is the flexibility of optimizing all hyperparameters of each neural network separately in each subdomain. We compare the performance of distributed cPINNs and XPINNs for various forward problems, using both weak and strong scalings. Our results indicate that for space domain decomposition, cPINNs are more efficient in terms of communication cost but XPINNs provide greater flexibility as they can also handle time-domain decomposition for any differential equations, and can deal with any arbitrarily shaped complex subdomains. To this end, we also present an application of the parallel XPINN method for solving an inverse diffusion problem with variable conductivity on the United States map, using ten regions as subdomains.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗