Search NASA⌕ Search

Engineering topics

Zhou, Fei

Publications and source records attributed to Zhou, Fei.

At least 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Spectroscopy-guided discovery of three-dimensional structures of disordered materials with diffusion models

Spectroscopy techniques such as x-ray absorption near edge structure (XANES) provide valuable insights into the atomic structures of materials, yet the inverse prediction of precise structures from spectroscopic data remains a formidable challenge. In this study, we introduce a framework that combines generative artificial intelligence models with XANES spectroscopy to predict three-dimensional atomic structures of disordered systems, using amorphous carbon (a-C) as a model system. In this work, we introduce a new framework based on the diffusion model, a recent generative machine learning method, to predict 3D structures of disordered materials from a target property. For demonstration, we apply the model to identify the atomic structures of a-C as a representative material system from the target XANES spectra. We show that conditional generation guided by XANES spectra reproduces key features of the target structures. Furthermore, we show that our model can steer the generative process to tailor atomic arrangements for a specific XANES spectrum. Finally, our generative model exhibits a remarkable scale-agnostic property, thereby enabling generation of realistic, large-scale structures through learning from a small-scale dataset (i.e. with small unit cells). Our work represents a significant stride in bridging the gap between materials characterization and atomic structure determination; in addition, it can be leveraged for materials discovery in exploring various material properties as targeted.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Ice Phase Classification Made Easy with Score-Based Denoising

Accurate identification of ice phases is essential for understanding various physicochemical phenomena. However, such classification for structures simulated with molecular dynamics is complicated by the complex symmetries of ice polymorphs and thermal fluctuations. For this purpose, both traditional order parameters and data-driven machine learning approaches have been employed, but they often rely on expert intuition, specific geometric information, or large training data sets. In this work, we present an unsupervised phase classification framework that combines a score-based denoiser model with a subsequent model-free classification method to accurately identify ice phases. Further, the denoiser model is trained on perturbed synthetic data of ideal reference structures, eliminating the need for large data sets and labeling efforts. The classification step utilizes the smooth overlap of atomic position (SOAP) descriptors as the atomic fingerprint, ensuring Euclidean symmetries and transferability to various structural systems. Our approach achieves a remarkable 100% accuracy in distinguishing ice phases of test trajectories using only seven ideal reference structures of ice phases as model inputs. This demonstrates the generalizability of the score-based denoiser model in facilitating phase identification for complex molecular systems. The proposed classification strategy can be broadly applied to investigate structural evolution and phase identification for a wide range of materials, offering new insights into the fundamental understanding of water and other complex systems.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Score-based denoising for atomic structure identification

We propose an effective method for removing thermal vibrations that complicate the task of analyzing complex dynamics in atomistic simulation of condensed matter. Our method iteratively subtracts thermal noises or perturbations in atomic positions using a denoising score function trained on synthetically noised but otherwise perfect crystal lattices. The resulting denoised structures clearly reveal underlying crystal order while retaining disorder associated with crystal defects. Purely geometric, agnostic to interatomic potentials, and trained without inputs from explicit simulations, our denoiser can be applied to simulation data generated from vastly different interatomic interactions. The denoiser is shown to improve existing classification methods, such as common neighbor analysis and polyhedral template matching, reaching perfect classification accuracy on a recent benchmark dataset of thermally perturbed structures up to the melting point. Demonstrated here in a wide variety of atomistic simulation contexts, the denoiser is general, robust, and readily extendable to delineate order from disorder in structurally and chemically complex materials.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Machine Learning of Performance Space Mapping for the DPD Simulation of Drug Delivery to Endothelial Cells

Despite huge effort over the years, the design of functionalised nanocarriers (NCs) for targeted drug delivery to endothelial cells is still to be completely unveiled. Dissipative particle dynamics (DPD) simulation is used to study the adhesion of NCs to endothelial cells under the influence of series of parameters such as shape, size, and ligand density of the NCs. However, preparing a performance space mapping that illustrates the penetration depths of NCs as a function of variations in their properties requires simulations of all possible NCs with the above-mentioned properties, which is not feasible. This challenge was addressed by leveraging a Gaussian process regression (GPR)-informed active learning strategy and an extensive exploration of numerous samples, each representing different properties of NCs. The performance space mapping reveals that NCs with rod and disc shapes exhibit superior penetration capabilities compared to those with a spherical shape. Furthermore, it demonstrates that smaller-sized rod-shaped NCs and larger-sized disc-shaped NCs tend to achieve better penetration. When considering smaller NCs, the influence of ligand density appears to be limited. On the contrary, for larger NCs, an increase in ligand density correlates with greater penetration depth, underscoring its substantial role in shaping their penetration capabilities.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Accelerating discrete dislocation dynamics simulations with graph neural networks

Discrete dislocation dynamics (DDD) is a widely employed computational method to study plasticity at the mesoscale that connects the motion of dislocation lines to the macroscopic response of crystalline materials. However, the computational cost of DDD simulations remains a bottleneck that limits its range of applicability. Here, we introduce a new DDD-GNN framework in which the expensive time-integration of dislocation motion is entirely substituted by a graph neural network (GNN) model trained on DDD trajectories. As a first application, we demonstrate the feasibility and potential of our method on a simple yet relevant model of a dislocation line gliding through an array of obstacles. We show that the DDD-GNN model is stable and reproduces very well unseen ground-truth DDD simulation responses for a range of straining rates and obstacle densities, without the need to explicitly compute nodal forces or dislocation mobilities during time-integration. Our approach opens new promising avenues to accelerate DDD simulations and to incorporate more complex dislocation motion behaviors.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Molecular basis for antiviral activity of two pediatric neutralizing antibodies targeting SARS-CoV-2 Spike RBD

Neutralizing antibodies (NAbs) hold great promise for clinical interventions against SARS-CoV-2 variants of concern (VOCs). Understanding NAb epitopedependent antiviral mechanisms is crucial for developing vaccines and therapeutics against VOCs. Here we characterized two potent NAbs, EH3 and EH8, isolated from an unvaccinated pediatric patient with exceptional plasma neutralization activity. EH3 and EH8 cross-neutralize the early VOCsandmediatestrong Fc-dependent effector activity in vitro. Structural analyses of EH3 and EH8 in complex with the receptor-binding domain (RBD) revealed the molecular determinantsoftheepitope-drivenprotectionandVOCevasion.WhileEH3represents the prevalent IGHV3-53 NAb whose epitope substantially overlaps with the ACE2 binding site, EH8 recognizes a narrow epitope exposed in both RBD-up and RBD-down conformations. When tested in vivo, a single-dose prophylactic administration of EH3 fully protected stringent K18-hACE2 micefrom lethal challenge with Delta VOC. Our study demonstrates that protective NAbs responses converge in pediatric and adult SARS-CoV-2 patients.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2022

Neural Phase Simulation

The Neural Phase Simulation (NPS) is a package of codes for simulating microstructure evolution and accelerated molecular dynamics with deep neural-networks based surrogate models. NPS is designed to offer quantitatively accurate and computationally efficient simulation capabilities by leveraging modern machine-learning techniques. The primary intended use cases of NPS are training neural network surrogate models, though performing simulations on a single node is also supported. The NPS surrogate models can be trained from ground truth simulation methods, which are supposed to be accurate but expensive, such as molecular dynamics, phase field methods, kinetic Monte Carlo and discrete dislocation dynamics.

Zhou, Fei↗
DOE OSTI2021

Structural basis and mode of action for two broadly neutralizing antibodies against SARS-CoV-2 emerging variants of concern

Emerging variants of concern for the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) can transmit more efficiently and partially evade protective immune responses, thus necessitating continued refinement of antibody therapies and immunogen design. Here, we elucidate the structural basis and mode of action for two potent SARS-CoV-2 spike (S)-neutralizing monoclonal antibodies, CV3-1 and CV3-25, which remain effective against emerging variants of concern in vitro and in vivo. CV3-1 binds to the (485-GFN-487) loop within the receptor-binding domain (RBD) in the “RBD-up” position and triggers potent shedding of the S1 subunit. In contrast, CV3-25 inhibits membrane fusion by binding to an epitope in the stem helix region of the S2 subunit that is highly conserved among β-coronaviruses. Thus, vaccine immunogen designs that incorporate the conserved regions in the RBD and stem helix region are candidates to elicit pan-coronavirus protective immune responses.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Dominant effect of anharmonicity on the equation of state and thermal conductivity of MgO under extreme conditions

Understanding heat transfer in the earth mantle region is of great scientific interest but highly challenging due to dissimilar conditions from our daily lives. Thermal conductivity of highly pressured and thermally activated materials can be completely different from that at ambient condition. In this study we calculate the equation of state and thermal conductivity of cubic crystalline MgO, a major component of the mantle, in such an environment. It is shown that the material properties are not accurately captured unless rigorous treatment of anharmonicity of phonon vibrational motion is explicitly applied. To predict renormalized phonon dispersion relations, interatomic force constants of high-order expansion are required, especially at elevated temperature. The anharmonicity plays an increasingly substantial role for regulating volume expansion as temperature rises, because the Gibbs free energy of MgO considerably deviates from what was predicted by the quasiharmonic approximation. We find that the lattice thermal conductivity of MgO and the heat transfer of the earth's mantle are reliably estimated only when both the temperature-dependent renormalization of phonon frequencies and four-phonon scattering effects are taken into consideration. We map the characteristics at Geotherm into contour diagrams in which equilibrium volumes and the thermal conductivities are described as functions of temperature and pressure. Furthermore, the thermal conductivity values of MgO in this study are smaller than those in previous theoretical studies due to the additional incorporation of four-phonon scatterings in the calculations. Specifically, deviation between the values in previous studies and in this study becomes larger as thermodynamic variables reach those near the core-mantle boundary. For instance, thermal conductivities of 52.15 and 23.67 Wm -1 K -1 are estimated at ambient and extreme conditions (T = 4000K and P = 135 GPa), respectively.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Solubility-Dependent Protective Effects of Binary Alloys for Lithium Anode

Recently, alloys have been widely utilized as protection layers for homogenized Li growth. However, the protective mechanisms for different alloys are not clearly understood. In this work, guided by the binary phase diagrams, Mg, Au, Zn, Al, Fe, and Cu protective layers have been selected for study of their protective abilities and mechanisms. The selected metals can be classified into three categories according to the nature of their associated alloy protective layers: (1) metals with high solubility in Li while not forming any intermediate line compounds, e.g., Mg, which is good for fast Li stripping; (2) metals with limited solid solubility but forming stable line compounds with Li upon the lithiation, e.g., Au, Zn, and Al, where the formed stable line compounds Li x Au, Li x Zn, and Li x Al will hamper Li kinetics; (3) metals with negligible mutual solubility and no line compound formation with Li, e.g., Fe and Cu, which block the Li stripping path and therefore result in the highest polarization voltage.

25 ENERGY STORAGE↗
DOE OSTI2019

Compressive sensing lattice dynamics. I. General formalism

Ab initio calculations have been successfully used for evaluating lattice dynamical properties of solids within the (quasi)harmonic approximation (i.e., assuming noninteracting phonons with infinite lifetimes), but it remains difficult to treat anharmonicity in all but the simplest compounds. We detail a systematic information-theory-based approach to deriving ab initio anharmonic force constants: Compressive sensing lattice dynamics (CSLD). The non-negligible terms that are necessary to reproduce the first-principles calculated interatomic forces are automatically selected by minimizing the ℓ 1 norm (sum of absolute values) of the scaled force constants. Here, by using efficient sampling of the configuration space using a modest number of atomic configurations with quasirandom displacements, CSLD is well suited for deriving accurate anharmonic potentials for complex multicomponent compounds with large unit cells. We demonstrate the power and generality of CSLD by calculating the phonon lifetimes and thermal transport properties of type-I Si clathrates.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2019

Compressive sensing lattice dynamics. II. Efficient phonon calculations and long-range interactions

We apply the compressive sensing lattice dynamics method to calculate phonon dispersion for crystalline solids. While existing methods such as frozen phonon, small displacement, and linear response are routinely applied for phonon calculations, they are considerably more expensive or cumbersome to apply to certain solids, including structures with large unit cells or low symmetry, systems that require more expensive electronic structure treatment, and polar semiconductors/insulators. In the latter case, we propose an approach based on a corrected long-range force constant model with proper treatment of the acoustic sum rule and the symmetric on-site force constant matrix. Here, our approach is demonstrated to be accurate and efficient for these systems through case studies of NaCl, CeO 2 , Y 3 Al 5 O 12 , and La 2 Fe 14 B.

36 MATERIALS SCIENCE↗
DOE OSTI2019

Atomic Layer-Deposited Al{sub 2}O{sub 3} Interlayer for Improved Tribological and Anti-corrosion Properties of TiN Hard Coating on 316L Stainless Steel

A dense Al{sub 2}O{sub 3} interlayer was inserted into TiN to enhance the tribological and anti-corrosion properties of TiN hard coating via a hybrid process of magnetron sputtering and atomic layer deposition (ALD). The influence of the ALD-Al{sub 2}O{sub 3} layer on the microstructure, mechanical, tribological and anti-corrosion properties of TiN was examined. The TiN-Al{sub 2}O{sub 3} coating exhibits low friction coefficient and specific wear rate, due to the structure change and columnar refinement caused by the addition of amorphous ALD-Al{sub 2}O{sub 3} interlay which improves the mechanical properties of hard coating. The improved corrosion properties of TiN are attributed to the fact that the insulating ALD-Al{sub 2}O{sub 3} used as a sealing layer can prevent the corrosive chlorine ions migrating into the lower TiN layer and block the charge transport in coatings. Moreover, raising the insertion position of ALD-Al{sub 2}O{sub 3} layer away from the steel substrate results in a further increase in the anti-corrosion property of TiN-Al{sub 2}O{sub 3} composite coating.

36 MATERIALS SCIENCE↗
DOE OSTI2019

Establishment and Characterization of Two Novel Cholangiocarcinoma Cell Lines

Background: Cholangiocarcinoma (CCA) is a rare, aggressive and highly lethal tumor. The disease is very difficult to diagnose and multi-modality treatments are ineffective. To improve our understanding of the biological characteristics of CCA, and facilitate the identification of valid treatments, an in-depth characterization of two novel Chinese patient-derived primary CCA cell lines was performed. Methods: Two CCA cell lines were developed and labelled ZJU-0826 and -1125. The two cell lines were characterized with respect to phenotypic, molecular, biomarker, functional and histological properties. Results: Two novel cell lines were cultured for 2 years, and maintained for more than 100 passages. They retained their typical biliary epithelial morphology and ultrastructure. The population doubling times of ZJU-0826, and -1125 were 63.84 h and 44.73 h, respectively. The cells exhibited near-triploid karyotypes with complex structural aberrations. ZJU-1125 cells had mutations in TP53 exons. Short tandem repeats genotyping confirmed the human origin and difference between lines. An immunophenotype analysis showed that ZJU-0826 is positive for CD44, CD29, Pdx1, CD236, FoxA1, FoxA2, and Nanog, and ZJU-1125 positive for CD44, CD29, CD133, Pdx1, FoxA1, FoxA2, and Nanog. ZJU-1125 had greater invasion ability in vitro and tumorigenicity than those of ZJU-0826. Conclusions: Our results confirm the validity of the ZJU-0826 and -1125 as representative models for the elucidation of the molecular pathogenesis of perihilar CCA, and intrahepatic CCA in both in vitro and in vivo studies, respectively.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2019

Crystal structure of bacterial cytochrome bc 1 in complex with azoxystrobin reveals a conformational switch of the Rieske iron–sulfur protein subunit

Cytochrome bc 1 complexes (cyt bc 1 ), also known as complex III in mitochondria, are components of the cellular respiratory chain and of the photosynthetic apparatus of non-oxygenic photosynthetic bacteria. They catalyze electron transfer (ET) from ubiquinol to cytochrome c and concomitantly translocate protons across the membrane, contributing to the cross-membrane potential essential for a myriad of cellular activities. This ET-coupled proton translocation reaction requires a gating mechanism that ensures bifurcated electron flow. In this study, we report the observation of the Rieske iron–sulfur protein (ISP) in a mobile state, as revealed by the crystal structure of cyt bc 1 from the photosynthetic bacterium Rhodobacter sphaeroides in complex with the fungicide azoxystrobin. Unlike cyt bc 1 inhibitors stigmatellin and famoxadone that immobilize the ISP, azoxystrobin causes the ISP-ED to separate from the cyt b subunit and to remain in a mobile state. Analysis of anomalous scattering signals from the iron–sulfur cluster of the ISP suggests the existence of a trajectory for electron delivery. This work supports and solidifies the hypothesis that the bimodal conformation switch of the ISP provides a gating mechanism for bifurcated ET, which is essential to the Q-cycle mechanism of cyt bc 1 function.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2019

Compressive sensing lattice dynamics

A comprehensive package to study lattice dynamics from first-principles. The interatomic force constants were fitted using the compressive sensing algorithm.

Zhou, Fei↗