SEARCH · Search NASA
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.
Prediction of bond dissociation energies for charged molecules with a graph neural network enabled by global molecular features and reaction difference features between products and reactants.
Facet-dependent surface charging of metal oxides in water dominates the ion transport behavior across the interface, in turn impacting many natural and industrial processes such as adsorption, the formation and stabilization of nanoparticle suspensions, corrosion, and heterogeneous catalysis. Here we investigated the pH-dependent surface chemistry of two low-index MgO single crystal surfaces, namely MgO(100) and MgO(111), using vibrational sum frequency generation (vSFG) spectroscopy. This allowed us to evaluate facet-dependent pH effects on the hydration and hydroxylation at the solid/aqueous interface and point-of-zero charge (PZC) values. The MgO system is complicated by its thermodynamic instability with respect to Mg(OH) 2 in water at ambient conditions. Here, for both hydroxylated MgO(100) and MgO(111) surfaces, the PZC is found to be around pH ∼ 12, which compares well with reported values for MgO single crystal and nanoparticle surfaces. However, structure specific differences in the molecular water hydrogen bonding network near the surface are evident at mildly acidic pH. To our knowledge, this is the first account of the PZC values for the MgO(111) single-crystal surface, an electrostatically unstable MgO termination that is prone to reconstruction.
Graph Neural Networks (GNNs) have revolutionized molecular property prediction by leveraging graph-based representations, yet their opaque decision-making processes hinder broader adoption in drug discovery. This study introduces the Activity-Cliff-Explanation-Supervised GNN (ACES-GNN) framework, designed to simultaneously improve predictive accuracy and interpretability by integrating explanation supervision for activity cliffs (ACs) into GNN training. ACs, defined by structurally similar molecules with significant potency differences, pose challenges for traditional models due to their reliance on shared structural features. By aligning model attributions with chemist-friendly interpretations, the ACES-GNN framework bridges the gap between prediction and explanation. Validated across 30 pharmacological targets, ACES-GNN consistently enhances both predictive accuracy and attribution quality for ACs compared to unsupervised GNNs. Our results demonstrate a positive correlation between improved predictions and accurate explanations, offering a robust and adaptable framework to better understand and interpret ACs. This work underscores the potential of explanation-guided learning to advance interpretable artificial intelligence in molecular modeling and drug discovery.
The mechanisms that underly the adaptation of enzyme activities and stabilities to temperature are fundamental to our understanding of molecular evolution and how enzymes work. Here, we investigate the molecular and evolutionary mechanisms of enzyme temperature adaption, combining deep mechanistic studies with comprehensive sequence analyses of thousands of enzymes. We show that temperature adaptation in ketosteroid isomerase (KSI) arises primarily from one residue change with limited, local epistasis, and we establish the underlying physical mechanisms. This residue change occurs in diverse KSI backgrounds, suggesting parallel adaptation to temperature. Furthermore, we identify residues associated with organismal growth temperature across 1005 diverse bacterial enzyme families, suggesting widespread parallel adaptation to temperature. We assess the residue properties, molecular interactions, and interaction networks that appear to underly temperature adaptation.
Molecular dynamics simulations offer detailed insights into atomic motions but face timescale limitations. Enhanced sampling methods have addressed these challenges but even with machine learning, they often rely on pre-selected expert-based features. Here, in this work, we present a Graph Neural Network-State Predictive Information Bottleneck (GNN-SPIB) framework, which combines graph neural networks and the state predictive information bottleneck to automatically learn low-dimensional representations directly from atomic coordinates. Tested on three benchmark systems, our approach predicts essential structural, thermodynamic and kinetic information for slow processes, demonstrating robustness across diverse systems. The method shows promise for complex systems, enabling effective enhanced sampling without requiring pre-defined reaction coordinates or input features.
Interatomic potential development using machine learning (ML) approaches has attracted a lot of attention in recent years because these potentials can effectively describe the structural and dynamical properties of complex materials at the atomistic level. Here, in this work, we present the development of a neural network (NN) deep ML interatomic potential for Fe-Si alloys, and we demonstrate the effectiveness of the NN-ML potential in predicting the structures and energies of liquid and crystalline phases of Fe-Si alloys in comparison with the results from ab initio molecular dynamics simulations or experimental data. The developed NN-ML potential is also used to perform molecular dynamics simulations to study the structures of Fe-Si alloys with various compositions under rapid solidification conditions. The short-ranged orders in the rapidly solidified Fe-Si alloys are also analyzed by a cluster alignment method.
Abstract The interaction of condensed phase systems with external electric fields is of major importance in a myriad of processes in nature and technology, ranging from the field-directed motion of cells (galvanotaxis), to geochemistry and the formation of ice phases on planets, to field-directed chemical catalysis and energy storage and conversion systems including supercapacitors, batteries and solar cells. Molecular simulation in the presence of electric fields would give important atomistic insight into these processes but applications of the most accurate methods such as ab-initio molecular dynamics (AIMD) are limited in scope by their computational expense. Here we introduce Perturbed Neural Network Potential Molecular Dynamics (PNNP MD) to push back the accessible time and length scales of such simulations. We demonstrate that important dielectric properties of liquid water including the field-induced relaxation dynamics, the dielectric constant and the field-dependent IR spectrum can be machine learned up to surprisingly high field strengths of about 0.2 V Å −1 without loss in accuracy when compared to ab-initio molecular dynamics. This is remarkable because, in contrast to most previous approaches, the two neural networks on which PNNP MD is based are exclusively trained on molecular configurations sampled from zero-field MD simulations, demonstrating that the networks not only interpolate but also reliably extrapolate the field response. PNNP MD is based on rigorous theory yet it is simple, general, modular, and systematically improvable allowing us to obtain atomistic insight into the interaction of a wide range of condensed phase systems with external electric fields.
Classical molecular dynamics (MD) simulations represent a very popular and powerful tool for materials modeling and design. The predictive power of MD hinges on the ability of the interatomic potential to capture the underlying physics and chemistry. There have been decades of seminal work on developing interatomic potentials, albeit with a focus predominantly on capturing the properties of bulk materials. Such physics-based models, while extensively deployed for predicting the dynamics and properties of nanoscale systems over the past two decades, tend to perform poorly in predicting nanoscale potential energy surfaces (PESs) when compared to high-fidelity first-principles calculations. These limitations stem from the lack of flexibility in such models, which rely on a predefined functional form. Machine learning (ML) models and approaches have emerged as a viable alternative to capture the diverse size-dependent cluster geometries, nanoscale dynamics, and the complex nanoscale PESs, without sacrificing the bulk properties. Here, in this study, we introduce an ML workflow that combines transfer and active learning strategies to develop high-dimensional neural networks (NNs) for capturing the cluster and bulk properties for several different transition metals with applications in catalysis, microelectronics, and energy storage, to name a few. Our NN first learns the bulk PES from the high-quality physics-based models in literature and subsequently augments this learning via retraining with a higher-fidelity first-principles training data set to concurrently capture both the nanoscale and bulk PES. Our workflow departs from status-quo in its ability to learn from a sparsely sampled data set that nonetheless covers a diverse range of cluster configurations from near-equilibrium to highly nonequilibrium as well as learning strategies that iteratively improve the fingerprinting depending on model fidelity. All the developed models are rigorously tested against an extensive first-principles data set of energies and forces of cluster configurations as well as several properties of bulk configurations for 10 different transition metals. Our approach is material agnostic and provides a methodology to transfer and build upon the learnings from decades of seminal work in molecular simulations on to a new generation of ML-trained potentials to accelerate materials discovery and design.
Mechanical forces generated by myosin II molecular motors drive diverse cellular processes, most notably shape change, division and locomotion. These forces may be transmitted over long range through the cytoskeletal medium – a disordered, viscoelastic network of biopolymers. The resulting cell size scale force chains can in principle mediate mechanical interactions between distant actomyosin units, leading to self-organized structural order in the cell cytoskeleton. Inspired by such force transmission through elastic structures in the cytoskeleton, we consider a percolated fiber lattice network, where fibers are represented as linear elastic elements that can both bend and stretch, and the contractile activity of myosin motors is represented by force dipoles. Then, by using a variety of metrics, we show how two such contractile force dipoles interact with each other through their mutual mechanical deformations of the elastic fiber network. As a prelude to two-dipole interactions, we quantify how forces propagate through the network from a single anisotropic force dipole by analyzing clusters of nodes connected by highly strained bonds, as well as through the decay rate of strain energy with distance from a force dipole. We show that predominant fiber bending screens out force propagation, resulting in reduced and strongly network configuration-dependent dipole interactions. On the other hand, stretching-dominated networks support longer-ranged inter-dipole interactions that recapitulate the predictions of linear elasticity theory. By characterizing the differences between tensile and compressive force propagation in the fiber network, we show how inter-dipole interaction depends on the dipoles’ mutual separation and orientation. The resulting elastic interaction energy may mediate a force between multiple distant dipoles, leading to their self-organization into ordered configurations. In conclusion, this provides a potential pathway for active mechanical force-driven structural order in elastic biopolymer networks.
For nearly a century, introduction of nanoparticles to elastomers has yielded extraordinarily tough nanocomposites that are critical to technologies from actuators to tires. The mechanisms by which this reinforcement occurs have nevertheless remained a central open question in material science. One widely debated hypothesis posits that strong interactions between polymer and particles induce "glassy bridges" that cement particles into a cohesive percolating network that resists elongation. Here, molecular dynamics simulations show that glassy particle shells do not primarily provide elongational cohesion. Instead, they amplify an underlying mechanism wherein competition between filler and elastomer networks causes the elastomer's volume to increase on deformation. This induces contributions from the elastomer's bulk modulus, which is of order 1000 times larger than its Young's modulus. These findings establish a unified understanding of low-strain reinforcement in filled elastomers as emanating from volumetric competition between coexisting particulate and elastomeric networks. This reframes and unifies our understanding of low-strain reinforcement, provides a clear-cut diagnostic for the presence of glassy bridging, and offers a new design principle for tough elastomeric nanocomposites.
Poly(ethylene glycol) (PEG)‐based materials, like PEG‐diacrylate (PEGDA), are prized for their hydrophilic, inert properties, and leveraged in hydrogels and as membrane mimics. While network chemistry is often tuned for selective transport and antifouling, fundamental understanding of water dynamics at the network surface and the impact on bulk transport is limited. We utilize Overhauser dynamic nuclear polarization (ODNP) to measure nanoscale water diffusivity near a tethered spin label at the water‐polymer surface and compare it to bulk water diffusivity from pulsed field gradient (PFG) NMR. Via active ester chemistry, spin labels and varied sidechain chemistries are introduced, modulating network hydrophilicity. Tuning network hydration through crosslinker content and functional groups further impacts water diffusivity. Results show rapid nanoscale water transport at the polymer surface, reflecting network volumetric water content, with further modulation by sidechain functionality. These findings demonstrate PEGDA's utility as a membrane mimic and the critical impact of network chemistry on water transport.
Existing models of thermoset deconstruction facilitated by incorporating cleavable comonomers rely on a mean-field reverse gel point paradigm, which predicts network dissolution once cleavable bonds reach a critical stoichiometric threshold, but does not account for where those bonds reside within the network architecture. Using reactive coarse-grained molecular dynamics simulations coupled with graph-theoretic analysis, we extend this stoichiometric picture to show that deconstructability is governed by the curing-imprinted network topology rather than stoichiometry alone. This topological organization is hierarchical: at the local scale, the elastic effectiveness of cross-link junctions determines which cross-links constitute the load-bearing scaffold; at the mesoscale, the cross-linking rate kinetically templates that scaffold into topologically modular communities─densely cross-linked clusters connected by sparse bridging strands that sustain network connectivity. Using betweenness centrality to identify nodes that disproportionately lie on intercommunity shortest paths, we demonstrate that effective deconstruction of the network into macromolecular fragments requires cleavable comonomers to intercept these high-centrality bridging strands. We further find that under uniform, disassortative comonomer incorporation, this topological requirement provides a mechanistic basis for extending the reverse gel point to incorporate network topology. We also show that modularity imposes a fundamental limit on fragment uniformity that persists even when the centrality requirement is met. Finally, we demonstrate that chain stiffness provides a nearly independent lever to suppress mechanically redundant cross-links and raise the glass transition temperature without significantly altering the deconstruction outcome. Together, these findings reframe the thermoset design space around network topology and provide actionable guidelines for engineering thermoset copolymers with predictable deconstructability and targeted thermomechanical performance.
Amorphous titanium dioxide (a-TiO 2 ) is widely used as a coating material in applications such as electrochemistry and self-cleaning surfaces where its interface with water has a central role. However, little is known about the structures of the a-TiO 2 surface and aqueous interface, particularly at the microscopic level. In this work, we construct a model of the a-TiO 2 surface via a cut-melt-and-quench procedure based on molecular dynamics simulations with deep neural network potentials (DPs) trained on density functional theory data. After interfacing the a-TiO 2 surface with water, we investigate the structure and dynamics of the resulting system using a combination of DP-based molecular dynamics (DPMD) and ab initio molecular dynamics (AIMD) simulations. Both AIMD and DPMD simulations reveal that the distribution of water on the a-TiO 2 surface lacks distinct layers normally found at the aqueous interface of crystalline TiO 2 , leading to an ~10 times faster diffusion of water at the interface. Bridging hydroxyls (Ti 2 –O b H) resulting from water dissociation decay several times more slowly than terminal hydroxyls (Ti–O w H) due to fast Ti–O w H 2 → Ti–O w H proton exchange events. These results provide a basis for a detailed understanding of the properties of a-TiO 2 in electrochemical environments. Moreover, the procedure of generating the a-TiO 2 -interface employed here is generally applicable to studying the aqueous interfaces of amorphous metal oxides.
A series of self-semi-interpenetrating network (ssIPN) thin films based on PIM-1 structure were developed by end-cross-linking telechelic PIM-1 oligomers end-capped with curable carboxylic acid groups to form model networks, which are penetrated by linear high-molecular-weight PIM-1 chains. ssIPN films with systematically varied network content ranging from 10 to 30 wt % were comprehensively examined on their microstructure and gas permeation properties. Fresh PIM-1 ssIPN films exhibited gas separation performances close to those of as-cast linear PIM-1, where the films closely followed the upper bound trade-off line, gaining as much selectivity as they lose gas permeability as the network content increases. This indicates that gas permeability is largely preserved despite cross-linking. Wide-angle X-ray scattering supported this observation, with peaks shifting toward lower d-spacing as oligomer content increased, suggesting tighter chain packing with higher cross-linkable oligomer loading. In physical aging studies over two months, all ssIPN films outperform linear PIM-1 for H 2 /CH 4 and O 2 /N 2 separations. Notably, the PIM-1 ssIPN with 10 wt % network content showed the best physical aging resistance, with negligible permeability loss even after one month. Finally, these findings highlight a promising macromolecular strategy for enhancing the physical aging resistance of microporous polymer membranes for gas separation.
The semitransparent and colorful properties of organic solar cells (OSCs) attract intensive academic interests due to their potential application in building integrated photovoltaics, wearable electronics, and so forth. The most straightforward and effective method to tune these optical properties is varying the componential ratio in the blend film. However, the increase in device transmittance inevitably sacrifices the photovoltaic performance because of severe carrier recombination that originates from discontinuous charge-transport networks in the blend film. In this work, a strategy is proposed via the molecular-doping strategy to overcome these shortcomings. It is discovered that p-doping is able to release the trapped holes in segregated polymer domains leading to short-circuit current enhancement, while n-doping is more effective to fill the bandgap states producing a higher fill factor. More importantly, either type of doping improves the photovoltaic performance in the semitransparent photovoltaic devices. These discoveries provide a new pathway to breaking the compromise between the photovoltaic performance and optical transmittance in semitransparent OSCs, and hold promise for their future commercialization.
Solvation thermodynamics in molten salt is accurately and efficiently predicted by combining ab initio molecular dynamics (AIMD) simulations, deep neural network interatomic potentials (NNIP), and quasichemical theory (QCT).
Widely known as “green solvents,” ionic liquids (ILs) and deep eutectic solvents (DESs) have been used as substitutes for traditional organic solvents in separation science. To achieve better separations using ILs and DESs, this work aimed to predict the infinite dilution activity coefficients (IDACs) of molecular solutes in these solvents with the state-of-the-art factorization-machine-based neural network (DeepFM). DeepFM combines the benefits of factorization machines and deep neural networks to learn both low-order and high-order interactions among features. The IDAC prediction model was established with 52,372 experimental IDAC datapoints including 260 solvents (252 ILs and 8 DESs) and 112 molecular solutes collected at various temperatures from 288.15 to 428.15 K. Chemical information describing the ILs, DESs, and molecular solutes was included in the IDAC prediction model, including chemical functional groups, molecular weights, and Abraham solvation parameters. The IDAC prediction model showed an improved accuracy compared with alternative models; additionally, DESs were included in the IDAC prediction model for the first time. Here, the model will reduce the energy and resources needed to optimize the selection of ILs and DESs for specific separations, which will promote the development of these green solvents for sustainable chemical processes.
The diffusion of small molecular penetrants through polymeric materials represents an important fundamental problem, relevant to the design of materials for applications such as coatings and membranes. Polymer networks hold promise in these applications because dramatic differences in molecular diffusion can result from subtle changes in the network structure. In this paper, we use molecular simulation to understand the role that cross-linked network polymers have in governing the molecular motion of penetrants. By considering the local, activated alpha relaxation time of the penetrant and its long-time diffusive dynamics, we can determine the relative importance of activated glassy dynamics on penetrants at the segmental scale vs entropic mesh confinement on penetrant diffusion. We vary several parameters, such as the cross-linking density, temperature, and penetrant size, to show that cross-links primarily affect molecular diffusion through the modification of the matrix glass transition, with local penetrant hopping at least partially coupled to the segmental relaxation of the polymer network. This coupling is very sensitive to the local activated segmental dynamics of the surrounding matrix, and we also show that penetrant transport is affected by dynamic heterogeneity at low temperatures. To contrast, only at high temperatures and for large penetrants or when the dynamic heterogeneity effect is weak, does the effect of mesh confinement become significant, even though penetrant diffusion more broadly empirically follows similar trends as established models of mesh confinement-based transport.