SEARCH · Search NASA

Results for “linear chain model”

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.

SourceResearch connectionFrom yearThrough year

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

At least 199 records · Page 11

All sources NASA NTRS DOE OSTI

DOE OSTI2021

Data-driven optimization of mixed-integer bi-level multi-follower integrated planning and scheduling problems under demand uncertainty

The coordination of interconnected elements across the different layers of the supply chain is essential for all industrial processes and the key to optimal decision-making. Yet, the modeling and optimization of such interdependent systems are still burdensome. Here we address the simultaneous modeling and optimization of medium-term planning and short-term scheduling problems under demand uncertainty using mixed-integer bi-level multi-follower programming and data-driven optimization. Bi-level multi-follower programs model the natural hierarchy between different layers of supply chain management holistically, while scenario analysis and data-driven optimization allow us to retrieve the guaranteed feasible solutions of the integrated formulation under various demand considerations. We address the data-driven optimization of this challenging class of problems using the DOMINO framework, which was initially developed to solve single-leader single-follower bi-level optimization problems to guaranteed feasibility. This framework is extended to solve single-leader multi-follower stochastic formulations and its performance is characterized by well-known single and multi-product process scheduling case studies. Through our data-driven algorithmic approach, we present guaranteed feasible solutions to linear and nonlinear mixed-integer bi-level formulations of simultaneous planning and scheduling problems and further characterize the effects of the scheduling level complexity on the solution performance, which spans over several hundred continuous and binary variables, and thousands of constraints.

42 ENGINEERING↗

DOE OSTI2025

Magnetic properties of metastable honeycomb KCoAsO 4

We present comprehensive neutron scattering data on polycrystalline samples of a new metastable honeycomb material KCoAsO 4 . Below 𝑇 𝑁 = 14 K, the system orders into a zigzag antiferromagnetic state, with spins ordered into alternating ferromagnetic chains similar to the isostructural sister compound KNiAsO 4 . In the case of KCoAsO 4 , we find the moments are robustly canted out of plane closer to the crystallographic 𝑐 axis. A combination of weak interlayer coupling, lattice strain, and inhomogeneities lead to the coexistence of two magnetic ordering wave vectors 𝑘 1 = (1.5 0 0) and 𝑘 2 = (0.5 0 0.5), where the two structures differ only in their layer stacking. Inelastic data show the presence of a spin orbital mode at 24 meV, supporting a pseudospin $\tilde{𝑆}$ = 1/2 Kramer's doublet ground state of the Co 2+ ions. We model the low energy excitations using both a conventional XXZ Hamiltonian and generalized Kitaev Heisenberg Hamiltonian within a linear spin wave limit. While either model can qualitatively reproduce the observed spectra, the lack of fine features and observation of disorder prevent a clear-cut determination of the low energy Hamiltonian. In the case of an XXZ-type model, a large easy-axis anisotropy is necessary to reproduce the gapped spectra and canting of the magnetic moments. For the generalized Kitaev model, despite the large canting of the moments away from the honeycomb layers, we find a noticeable if nondominant Kitaev term persists. In conclusion, the contrast of the magnetic properties of KCoAsO 4 to other cobalt honeycombs highlights the sensitivity of the low energy magnetic properties of Co 2+ to fine details of its crystalline environment.

Honeycomb lattice↗

DOE OSTI2020

Low-temperature spin dynamics in the TmFeO 3 orthoferrite with a non-Kramers ion

We investigate the magnetic dynamics of the orthorhombic perovskite TmFeO 3 at low temperatures, below the spin reorientation transition at TSR≈80 K, by means of time-of-flight neutron spectroscopy. We find that the magnetic excitation spectrum combines two emergent collective modes associated with different magnetic sublattices. The Fe subsystem orders below TN~632 K into a canted antiferromagnetic structure and exhibits sharp, high-energy magnon excitations. We describe them using linear spin-wave theory, and reveal a pronounced anisotropy between in- and out-of-plane exchange interactions, which was mainly neglected in previous reports on the spin dynamics in orthoferrites. At lower energies, we find two crystalline electrical field (CEF) excitations of Tm3+ ions at energies of ~2 and 5 meV. In contrast to the sister compound YbFeO 3 , where the Yb 3+ ions form quasi-one-dimensional chains along the c axis, the Tm excitations show dispersion along both directions in the (0KL) scattering plane. Analysis of the neutron scattering polarization factor reveals a longitudinal polarization of the 2 meV excitation. To evaluate the effect of the CEF on the Tm 3+ ions, we perform point-charge model calculations, and their results quantitatively capture the main features of Tm single-ion physics, such as energies, intensities, and polarization of the CEF transitions, and the type of magnetic anisotropy.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DOE OSTI2022

Robust estimation of charge carrier diffusivity using transient photoluminescence microscopy

Transient microscopy has emerged as a powerful tool for imaging the diffusion of excitons and free charge carriers in optoelectronic materials. In many excitonic materials, extraction of diffusion coefficients can be simplified because of the linear relationship between signal intensity and local excited state population. However, in materials where transport is dominated by free charge carriers, extracting diffusivities accurately from multidimensional data is complicated by the nonlinear dependence of the measured signal on the local charge carrier density. To obtain accurate estimates of charge carrier diffusivity from transient microscopy data, statistically robust fitting algorithms coupled to efficient 3D numerical solvers that faithfully relate local carrier dynamics to raw experimental measurables are sometimes needed. Here, we provide a detailed numerical framework for modeling the spatiotemporal dynamics of free charge carriers in bulk semiconductors with significant solving speed reduction and for simulating the corresponding transient photoluminescence microscopy data. To demonstrate the utility of this approach, we apply a fitting algorithm using a Markov chain Monte Carlo sampler to experimental data on bulk CdS and methylammonium lead bromide (MAPbBr 3 ) crystals. Parameter analyses reveal that transient photoluminescence microscopy can be used to obtain robust estimates of charge carrier diffusivities in optoelectronic materials of interest, but that other experimental approaches should be used for obtaining carrier recombination constants. Additionally, simplifications can be made to the fitting model depending on the experimental conditions and material systems studied. Our open-source simulation code and fitting algorithm are made freely available to the scientific community.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2023

Gapless spinons and a field-induced soliton gap in the hyperhoneycomb Cu oxalate framework compound [(C 2 H 5 ) 3 NH] 2 Cu 2 (C 2 O 4 ) 3

Here, we report a detailed study of the specific heat and magnetic susceptibility of single crystals of a spin-liquid candidate: the hyperhoneycomb Cu oxalate framework compound [(C 2 H 5 ) 3 NH] 2 Cu 2 (C 2 O 4 ) 3 . The specific heat shows no anomaly associated with a magnetic transition at low temperatures down to T ~ 180 mK in zero magnetic field. We observe a large linear-in-T contribution to the specific heat γT, γ = 98 (1) mJ / mol K 2 , at low temperatures, indicative of the presence of fermionic excitations despite the Mott insulating state. The low-T specific heat is strongly suppressed by applied magnetic fields H, which induce an energy gap, Δ(H), in the spin-excitation spectrum. We use the four-component relativistic density-functional theory (DFT) to calculate the magnetic interactions, including the Dzyaloshinskii-Moriya antisymmetric exchange, which causes an effective staggered field acting on one copper sublattice. The magnitude and field dependence of the field-induced gap, Δ (H) ∝ H 2/3 , are accurately predicted by the soliton mass calculated from the sine-Gordon model of weakly coupled antiferromagnetic Heisenberg chains with all parameters determined by our DFT calculations. Thus our experiment and calculations are entirely consistent with a model of [(C 2 H 5 ) 3 NH] 2 Cu 2 (C 2 O 4 ) 3 in which anisotropic magnetic exchange interactions due to Jahn-Teller distortion cause one copper sublattice to dimerize, leaving a second sublattice of weakly coupled antiferromagnetic chains. We also show that this model quantitatively accounts for the measured temperature-dependent magnetic susceptibility. Thus [(C 2 H 5 ) 3 NH] 2 Cu 2 (C 2 O 4 ) 3 is a canonical example of a one-dimensional spin-1/2 Heisenberg antiferromagnet and not a resonating-valence-bond quantum spin liquid, as previously proposed.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DOE OSTI2025

Photoproduction of the Σ + hyperon using linearly polarized photons with CLAS

Measurements of the polarization observables Σ, P, T, O x , O z for the reaction $\vec{γ}$ p → $K^0_S$Σ + using a linearly polarized photon beam of energy 1.1 to 2.1 GeV are reported. The measured data provide information on a channel that has not been studied extensively, but is required for a full coupled-channel analysis in the nucleon resonance region. Observables have been simultaneously extracted using likelihood sampling with a Markov-Chain Monte Carlo process. Angular distributions in bins of photon energy E γ are produced for each polarization observable. T, O x , and O z are first time measurements of these observables in this reaction. The extraction of Σ extends the energy range beyond a previous measurement. The measurement of P, the recoil polarization, is consistent with previous measurements. The measured data are shown to be significant enough to affect the estimation of the nucleon resonance parameters when fitted within a coupled-channels model.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2024

Development of (NO)Fe(N 2 S 2 ) as a Metallodithiolate Spin Probe Ligand: A Case Study Approach

The ubiquity of sulfur–metal connections in nature inspires the design of bi- and multimetallic systems in synthetic inorganic chemistry. Common motifs for biocatalysts developed in evolutionary biology include the placement of metals in close proximity with flexible sulfur bridges as well as the presence of π-acidic/delocalizing ligands. This Account will delve into the development of a (NO)Fe(N 2 S 2 ) metallodithiolate ligand that harnesses these principles. The Fe(NO) unit is the centroid of a N 2 S 2 donor field, which as a whole is capable of serving as a redox-active, bidentate S-donor ligand. Its paramagnetism as well as the ν(NO) vibrational monitor can be exploited in the development of new classes of heterobimetallic complexes. We offer four examples in which the unpaired electron on the {Fe(NO)} 7 unit is spin-paired with adjacent paramagnets in proximal and distal positions. First, the exceptional stability of the (NO)Fe(N 2 S 2 )-Fe(NO) 2 platform, which permits its isolation and structural characterization at three distinct redox levels, is linked to the charge delocalization occurring on both the Fe(NO) and the Fe(NO) 2 supports. This accommodates the formation of a rare nonheme {Fe(NO)} 8 triplet state, with a linear configuration. A subsequent FeNi complex, featuring redox-active ligands on both metals (NO on iron and dithiolene on nickel), displayed unexpected physical properties. Our research showed good reversibility in two redox processes, allowing isolation in reduced and oxidized forms. Various spectroscopic and crystallographic analyses confirmed these states, and Mössbauer data supported the redox change at the iron site upon reduction. Oxidation of the complex produced a dimeric dication, revealing an intriguing magnetic behavior. The monomer appears as a spin-coupled diradical between {Fe(NO)} 7 and the nickel dithiolene monoradical, while dimerization couples the latter radical units via a Ni 2 S 2 rhomb. Magnetic data (SQUID) on the dimer dication found a singlet ground state with a thermally accessible triplet state that is responsible for magnetism. A theoretical model built on an H 4 chain explains this unexpected ferromagnetic low-energy triplet state arising from the antiferromagnetic coupling of a four-radical molecular conglomerate. For comparison, two (NO)Fe(N 2 S 2 ) were connected through diamagnetic group 10 cations producing diradical trimetallic complexes. Antiferromagnetic coupling is observed between {Fe(NO)} 7 units, with exchange coupling constants (J) of -3, -23, and -124 cm –1 for Ni II , Pd II , and Pt II , respectively. This trend is explained by the enhanced covalency and polarizability of sulfur-dense metallodithiolate ligands. A central paramagnetic trans-Cr(NO)(MeCN) receiver unit core results in a cissoid structural topology, influenced by the stereoactivity of the lone pair(s) on the sulfur donors. This {Cr(NO)} 5 radical bridge, unlike all previous cases, finds the coupling between the distal Fe(NO) radicals to be ferromagnetic (J = 24 cm –1 ). The stability and predictability of this S = 1/2 moiety and the steric/electronic properties of the bridging thiolate sulfurs suggest it to be a likely candidate for the development of novel molecular (magnetic) compounds and possibly materials. The role of synthetic inorganic chemistry in designing synthons that permit connections of the (NO)Fe(N 2 S 2 ) metalloligand is highlighted as well as the properties of the heterobi- and polymetallic complexes derived therefrom.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

NASA NTRS1996

Cometary Nuclei and Tidal Disruption: The Geologic Record of Crater Chains on Callisto and Ganymede

Prominent crater chains on Ganymede and Callisto are most likely the impact scars of comets tidally disrupted by Jupiter and are not secondary crater chains. We have examined the morphology of these chains in detail in order to place constraints on the properties of the comets that formed them and the disruption process. In these chains, intercrater spacing varies by no more than a factor of 2 and the craters within a given chain show almost no deviation from linearity (although the chains themselves are on gently curved small circles). All of these crater chains occur on or very near the Jupiter-facing hemisphere. For a given chain, the estimated masses of the fragments that formed each crater vary by no more than an order of magnitude. The mean fragment masses for all the chains vary by over four orders of magnitude (W. B. McKinnon and P. M. Schenk 1995, Geophys. Res. Lett. 13, 1829-1832), however. The mass of the parent comet for each crater chain is not correlated with the number of fragments produced during disruption but is correlated with the mean mass of the fragments produced in a given disruption event. Also, the larger fragments are located near the center of each chain. All of these characteristics are consistent with those predicted by disruption simulations based on the rubble pile cometary nucleus model (in which nuclei are composed on numerous small fragments weakly bound by self-gravity), and with those observed in Comet D/Shoemaker-Levy 9. Similar crater chains have not been found on the other icy satellites, but the impact record of disrupted comets on Callisto and Ganymede indicates that disruption events occur within the Jupiter system roughly once every 200 to 400 years.

Schenk, Paul M.↗

NASA NTRS2019

Differences Between OCO‐2 and GOME‐2 SIF Products From a Model‐Data Fusion Perspective

Space-borne retrievals of solar-induced chlorophyll fluorescence (SIF) over land surfaces have recently become a resource for studying and quantifying the broad scale dynamics of gross carbon uptake (gross primary productivity—GPP) across ecosystems. To prepare for the assimilation of SIF data in terrestrial biosphere models, we examine how differences between SIF products (due to differences in acquisition characteristics and processing chain) may affect the optimization of model parameters and the resultant GPP estimate. We compare recent daily mean SIF products (one from the Orbiting Carbon Observatory-2 [OCO-2] and two from the Global Ozone Monitoring Experiment–2 [GOME-2], GlobFluo [GF] and NASA-v28 [N28], missions), averaged at 0.5° × 0.5° spatial resolution and 16-day temporal resolution, at the biome level. Phase differences between these products are relatively small. A first-order correction of the difference in spectral sampling between the two instruments shows that OCO-2 and N28 are consistent in terms of magnitude and amplitude, while GF is twice as large as the others. Using a bias-blind toy data assimilation framework, we analyze how biases between SIF products, and between model and products, can be partially alleviated by optimizing the slope and intercept parameters of a linear GPP-SIF operator. As observation biases can transfer to biases in other optimized process-based parameters and to modeled carbon fluxes— thereby resulting in unidentified inaccurate parameter values—we argue that potential SIF biases should be treated cautiously in real-world experiments in order to achieve realistic and reliable future simulations.

Gross primary production↗

NASA NTRS1995

A correlated ab initio study of linear carbon-chain radicals CnH (n = 2-7)

Linear carbon-chain radicals CnH for n = 2-7 have been studied with correlation consistent valence and core-valence basis sets and the coupled cluster method RCCSD(T). Equilibrium structures, rotational constants, and dipole moments are reported and compared with available experimental data. The ground state of the even-n series changes from 2 sigma+ to 2 pi as the chain is extended. For C4H, the 2 sigma+ state was found to lie only 72 cm-1 below the 2 pi state in the estimated complete basis set limit for valence correlation. The C2H- and C3H- anions have also been characterized.

NASA Discipline Exobiology↗

NASA NTRS2001

A novel reagentless sensing system for measuring glucose based on the galactose/glucose-binding protein

The galactose/glucose-binding protein (GBP) is synthesized in the cytoplasm of Escherichia coli in a precursor form and exported into the periplasmic space upon cleavage of a 23-amino-acid leader sequence. GBP binds galactose and glucose in a highly specific manner. The ligand induces a hinge motion in GBP and the resultant protein conformational change constitutes the basis of the sensing system. The mglB gene, which codes for GBP, was isolated from the chromosome of E. coli using the polymerase chain reaction (PCR). Since wild-type GBP lacks cysteines in its structure, introducing this amino acid by site-directed mutagenesis ensures single-label attachment at specific sites with a sulfhydro-specific fluorescent probe. Site-directed mutagenesis by overlap extension PCR was performed to prepare three different mutants to introduce a single cysteine residue at positions 148, 152, and 182. Since these residues are not involved in ligand binding and since they are located at the edge of the binding cleft, they experience a significant change in environment upon binding of galactose or glucose. The sensing system strategy is based on the fluorescence changes of the probe as the protein undergoes a structural change on binding. In this work a reagentless sensing system has been rationally designed that can detect submicromolar concentrations of glucose. The calibration plots have a linear working range of three orders of magnitude. Although the system can sense galactose as well, this epimer is not a potential interfering substance since its concentration in blood is negligible. Copyright 2001 Academic Press.

Non-NASA Center↗

DOE OSTI2023

Kinetic model-based group contribution method for derived cetane number prediction of oxygenated fuel components and blends

A four-step autoignition model was used to derive an expression for the ignition delay (ID) measured in ignition quality testers (IQT) with the derived cetane number (DCN) determined from this ID using the ASTM D6890 standard correlation. The model predicts DCN for individual compounds and blends as a function of each compound's global initiation and net chain branching rate constants. Expressions for these values were determined assuming they could be related to the functional groups present in each compound. Measurement data for 125 compounds and 94 binary and ternary blends (including oxygenates: alcohols, aldehydes, esters, ethers, and ketones), from literature and from measurements performed in our lab, were used to obtain the dependence of the measured ignition delay on each functional group. The new kinetic model-based group contribution method was able to predict the ignition delay of both pure compounds and blends, with an average DCN error of 4.4 (19%) and 2.8 (11%), respectively. Here, the blend model was also used to develop an ID mixing rule by incorporating existing IQT ignition delay data for each compound. Use of the mixing rule gave an average DCN error of 3.6 (16%) for blends. Both the blend model and mixing rule were found to be superior compared to standard linear by volume fraction or mole fraction mixing rules commonly used to estimate the DCN of mixtures.

42 ENGINEERING↗

NASA NTRS2002

Human motion planning based on recursive dynamics and optimal control techniques

This paper presents an efficient optimal control and recursive dynamics-based computer animation system for simulating and controlling the motion of articulated figures. A quasi-Newton nonlinear programming technique (super-linear convergence) is implemented to solve minimum torque-based human motion-planning problems. The explicit analytical gradients needed in the dynamics are derived using a matrix exponential formulation and Lie algebra. Cubic spline functions are used to make the search space for an optimal solution finite. Based on our formulations, our method is well conditioned and robust, in addition to being computationally efficient. To better illustrate the efficiency of our method, we present results of natural looking and physically correct human motions for a variety of human motion tasks involving open and closed loop kinematic chains.

Non-NASA Center↗

DOE OSTI2021

Quasi-one-dimensional uniform spin- 1 2 Heisenberg antiferromagnet KNaCuP 2 O 7 probed by P 31 and Na 23 NMR

Here we present the structural and magnetic properties of KNaCuP 2 O 7 investigated via x-ray diffraction, magnetization, specific heat, and P 31 and Na 23 NMR measurements and complementary electronic structure calculations. The temperature-dependent magnetic susceptibility and P 31 NMR shift could be modeled very well by the uniform spin- 1 2 Heisenberg antiferromagnetic chain model with a nearest-neighbor interaction J / k B ≃ 58.7 K . The corresponding mapping using first-principles electronic structure calculations leads to J DFT / k B ≃ 59 K with negligibly small interchain couplings, further confirming that the system is indeed a one-dimensional uniform spin- 1 2 Heisenberg antiferromagnet. The diverging trend of NMR spin-lattice relaxation rates ( 1 / T 1 31 and 1 / T 1 23 ) implies the onset of a magnetic long-range ordering at around T N ≃ 1 K . From the value of T N , the average interchain coupling is estimated to be J ' / k B ≃ 0.28 K . Moreover, the NMR spin-lattice relaxation rates show the dominant contributions from uniform ( q = 0 ) and staggered ( q = ± π / a ) spin fluctuations in the high- and low-temperature regimes, respectively, mimicking one-dimensionality of the spin lattice. We have also demonstrated that 1 / T 1 31 in high temperatures varies linearly with 1 / H , reflecting the effect of spin diffusion on the dynamic susceptibility. The temperature-dependent unit cell volume could be described well using the Debye approximation with a Debye temperature of Θ D ≃ 294 K , consistent with the heat capacity data.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DOE OSTI2021

Correlated Response of Protein Side-Chain Fluctuations and Conformational Entropy to Ligand Binding

The heterogeneous fast side-chain dynamics of proteins plays crucial roles in molecular recognition and binding. Site-specific NMR experiments quantify these motions by measuring the model-free order parameter (O axis 2 ) on a scale of 0 (most flexible) to 1 (least flexible) for each methyl-containing residue of proteins. In this work, we have examined ligand-induced variations in the fast side-chain dynamics and conformational entropy of calmodulin (CaM) using five different CaM–peptide complexes. O axis 2 of CaM in the ligand-free (O axis,U 2 ) and ligand-bound (O axis,B 2 ) states are calculated from molecular dynamics trajectories and conformational energy surfaces obtained using the adaptive biasing force (ABF) method. ΔO axis 2 = O axis,B 2 – O axis,U 2 follows a Gaussian-like unimodal distribution whose second moment is a potential indicator of the binding affinity of these complexes. The probability for the binding-induced O axis,U 2 → O axis,B 2 transition decreases with increasing magnitude of ΔO axis 2 , indicating that large flexibility changes are improbable for side chains of CaM after ligand binding. A linear correlation established between ΔOaxis2 and the conformational entropy change of the protein makes possible the determination of the conformational entropy of binding of protein–ligand complexes. The results not only underscore the functional importance of fast side-chain fluctuations but also highlight key motional and thermodynamic correlates of protein–ligand binding.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2026

Novel insight into the kinetics of amide bond glycolysis for nylon-6 depolymerization

Chemical recycling of nylon-6 to short-chain oligomers and monomer ε-caprolactam via catalytic glycolysis is a potential solution for plastic waste remediation. Here, in this work, the kinetics of amide bond glycolysis (with ethylene glycol) in nylon-6 and the model compound N-phenethyl-3-phenylpropanamide (M1) were evaluated at 473 K in the presence of the cyclic amidine catalyst 1,5,7-triazabicyclo[4.4.0]dec-5-ene. Rates of polymer glycolysis were determined by the time-dependent shift in molecular weight distribution, whereas rates of M1 glycolysis were determined using liquid chromatography. The similarity of the first-order rate constants for glycolysis of nylon-6 and M1 at 473 K with 0.03 M amidine catalyst (5 mol% relative to amide bonds), 1.22 × 10 −5 s −1 and 2.18 × 10 −5 s −1 , respectively, confirmed the suitability of M1 as a model compound for nylon-6 glycolysis. Similar rates of glycolysis in the presence of other cyclic amidine catalysts as well as sodium methoxide revealed little influence of base strength. Glycolysis rates were unexpectedly non-linear in catalyst loading and deactivation occurred with long reaction times, presumably by non-selective decomposition of products as detected by liquid chromatography.

Depolymerization rate and rate constant↗

DOE OSTI2023

ZeoNet: 3D convolutional neural networks for predicting adsorption in nanoporous zeolites

Zeolites are one of the most widely used materials in the chemical industry due to their nanometer-sized pores that can adsorb and react upon molecules selectively. With hundreds of known framework topologies and hundreds of thousands of computationally predicted structures, the ability to rapidly predict zeolite performance allows researchers to prioritize their efforts on the most promising structures for a given application. Although the accuracy of forcefield-based atomistic simulations has advanced significantly in the past two decades, these simulations can be computationally expensive, especially for long-chain, complex molecules. Here, we present ZeoNet, a representation learning framework using convolutional neural networks (ConvNets) and 3D volumetric representations for predicting adsorption in zeolites. ZeoNet was trained on the task of predicting Henry's constants for adsorption, k H , of n-octadecane in more than 330 000 known and predicted zeolite materials. Employing a 3D grid based on the distances to solvent-accessible surfaces, a volumetric representation that can be generated efficiently, the best-performing ZeoNet achieved a correlation coefficient r 2 = 0.977 and a mean-squared error MSE = 3.8 in ln k H , which corresponds to an error of 9.3 kJ mol -1 in adsorption free energy. In comparison, a model based on hand-designed geometric features has values of r 2 = 0.783 and MSE = 35.7. ZeoNet is also relatively efficient and can process ≈8 structures per second on an Nvidia RTX 2080TI GPU, orders of magnitude faster than forcefield-based simulations. A systematic analysis was conducted to investigate how the choice of ConvNet architectures, the linear dimension (L) and spatial resolution (Δd) of the distance grids, batch size, optimizer, and learning rate impact the model performance. We found that ConvNets based on the ResNet architecture offer the best tradeoff between expressiveness and efficiency. The performance for all models reaches a plateau at L = 30–45 Å and depends less sensitively on grid resolution, with a small benefit around Δd = 0.30–0.45 Å. Finally, saliency maps were visualized to identify which regions of the materials contributed the most to model predictions. It was found, interestingly, that the predictions are driven primarily by the accessible pore volume rather than the region occupied by the framework atoms.

36 MATERIALS SCIENCE↗

DOE OSTI2020

On decoupling the integrals of cosmological perturbation theory

ABSTRACT Perturbation theory (PT) is often used to model statistical observables capturing the translation and rotation-invariant information in cosmological density fields. PT produces higher order corrections by integration over linear statistics of the density fields weighted by kernels resulting from recursive solution of the fluid equations. These integrals quickly become high dimensional and naively require increasing computational resources the higher the order of the corrections. Here, we show how to decouple the integrands that often produce this issue, enabling PT corrections to be computed as a sum of products of independent 1D integrals. Our approach is related to a commonly used method for calculating multiloop Feynman integrals in quantum field theory, the Gegenbauer Polynomial x-Space Technique. We explicitly reduce the three terms entering the 2-loop power spectrum, formally requiring 9D integrations, to sums over successive 1D radial integrals. These 1D integrals can further be performed as convolutions, rendering the scaling of this method Nglog Ng with Ng the number of grid points used for each fast Fourier transform. This method should be highly enabling for upcoming large-scale structure redshift surveys where model predictions at an enormous number of cosmological parameter combinations will be required by Monte Carlo Markov Chain searches for the best-fitting values.

Slepian, Zachary↗