Atomistic Simulations Research Articles

Feature selection for high-dimensional neural network potentials with the adaptive group lasso

Neural network potentials are a powerful tool for atomistic simulations, allowing to accurately reproduce ab initio potential energy surfaces with computational performance approaching classical force fields. A central component of such potentials is the transformation of atomic positions into a set of atomic features in a most efficient and informative way. In this work, a feature selection method is introduced for high dimensional neural network potentials, based on the adaptive group lasso (AGL) approach. It is shown that the use of an embedded method, taking into account the interplay between features and their action in the estimator, is necessary to optimize the number of features. The method’s efficiency is tested on three different monoatomic systems, including Lennard–Jones as a simple test case, Aluminium as a system characterized by predominantly radial interactions, and Boron as representative of a system with strongly directional components in the interactions. The AGL is compared with unsupervised filter methods and found to perform consistently better in reducing the number of features needed to reproduce the reference simulation data at a similar level of accuracy as the starting feature set. In particular, our results show the importance of taking into account model predictions in feature selection for interatomic potentials.

Molecular dynamics simulations of lipid-protein interactions in SLC4 proteins.

The SLC4 family of secondary bicarbonate transporters is responsible for the transport of HCO3¯, CO32¯, Cl¯, Na+, K+, NH3 and H+ necessary for regulation of pH and ion homeostasis. They are widely expressed in numerous tissues throughout the body and function in different cell types with different membrane properties. Potential lipid roles in SLC4 function have been reported in experimental studies, focusing mostly on two members of the family: AE1 (Cl¯/HCO3¯ exchanger) and NBCe1 (Na+-CO32¯cotransporter). Previous computational studies of the outward facing (OF) state of AE1 with model lipid membranes revealed enhanced protein-lipid interactions between cholesterol (CHOL) and phosphatidylinositol bisphosphate (PIP2). However, the protein-lipid interactions in other members of the family and other conformation states are still poorly understood and this precludes the detailed studies of a potential regulatory role for lipids in the SLC4 family. In this work, we performed coarse-grained and atomistic molecular dynamics simulations on three members of the SLC4 family with different transport modes: AE1, NBCe1 and NDCBE (a Na+-CO32¯/Cl¯ exchanger), in model HEK293 membranes consisting of CHOL, PIP2, phosphatidylcholine (POPC), phosphatidylethanolamine (POPE), phosphatidylserine (POPS), and sphingomyelin (POSM). The recently resolved inward-facing (IF) state of AE1 was also included in the simulations. Lipid-protein contact analysis of the simulated trajectories was performed with the ProLint server, which provides a multitude of visualization tools for illustration of areas of enhanced lipid-protein contact and identification of putative lipid binding sites within the protein matrix. We observed enrichment of CHOL and PIP2 around all proteins with subtle differences in their distribution depending on the protein type and conformation state. Putative binding sites were identified for CHOL, PIP2, POPC, and POSM in the three studied proteins and their potential roles in the SLC4 transport function, conformational transition and protein dimerization were discussed.

Unveiling Core-Shell Structure Formation in a Ni3Fe Nanoparticle with In Situ Multi-Bragg Coherent Diffraction Imaging.

Solid-state reactions play a key role in materials science. The evolution of the structure of a single 350 nm Ni3Fe nanoparticle, i.e., its morphology (facets) as well as its deformation field, has been followed by applying multireflection Bragg coherent diffraction imaging. Through this approach, we unveiled a demixing process that occurs at high temperatures (600 °C) under an Ar atmosphere. This process leads to the gradual emergence of a highly strained core-shell structure, distinguished by two distinct lattice parameters with a difference of 0.4%. Concurrently, this transformation causes the facets to vanish, ultimately yielding a rounded core-shell nanoparticle. This final structure comprises a Ni3Fe core surrounded by a 40 nm Ni-rich outer shell due to preferential iron oxidation. Providing in situ 3D imaging of the lattice parameters at the nanometer scale while varying the temperature, this study─with the support of atomistic simulations─not only showcases the power of in situ multireflection BCDI but also provides valuable insights into the mechanisms at work during a solid-state reaction characterized by a core-shell transition.

Allosteric Cholesterol Site in Glycine Receptors Characterized through Molecular Simulations.

Glycine receptors are pentameric ligand-gated ion channels that conduct chloride ions across postsynaptic membranes to facilitate fast inhibitory neurotransmission. In addition to gating by the glycine agonist, interactions with lipids and other compounds in the surrounding membrane environment modulate their function, but molecular details of these interactions remain unclear, in particular, for cholesterol. Here, we report coarse-grained simulations in a model neuronal membrane for three zebrafish glycine receptor structures representing apparent resting, open, and desensitized states. We then converted the systems to all-atom models to examine detailed lipid interactions. Cholesterol bound to the receptor at an outer-leaflet intersubunit site, with a preference for the open and desensitized versus resting states, indicating that it can bias receptor function. Finally, we used short atomistic simulations and iterative amino acid perturbations to identify residues that may mediate allosteric gating transitions. Frequent cholesterol contacts in atomistic simulations clustered with residues identified by perturbation analysis and overlapped with mutations influencing channel function and pathology. Cholesterol binding at this site was also observed in a recently reported pig heteromeric glycine receptor. These results indicate state-dependent lipid interactions relevant to allosteric transitions of glycine receptors, including specific amino acid contacts applicable to biophysical modeling and pharmaceutical design.

Interatomic Interaction Models for Magnetic Materials: Recent Advances

Interatomic Interaction Models for Magnetic Materials: Recent Advances

Enhanced pyroptosis induction with pore-forming gene delivery for osteosarcoma microenvironment reshaping

Osteosarcoma is the most common malignant bone tumor without efficient management for improving 5-year event-free survival. Immunotherapy is also limited due to its highly immunosuppressive tumor microenvironment (TME). Pore-forming gasdermins (GSDMs)-mediated pyroptosis has gained increasing concern in reshaping TME, however, the expressions and relationships of GSDMs with osteosarcoma remain unclear. Herein, gasdermin E (GSDME) expression is found to be positively correlated with the prognosis and immune infiltration of osteosarcoma patients, and low GSDME expression was observed. A vector termed as LPAD contains abundant hydroxyl groups for hydrating layer formation was then prepared to deliver the GSDME gene to upregulate protein expression in osteosarcoma for efficient TME reshaping via enhanced pyroptosis induction. Atomistic molecular dynamics simulations analysis proved that the hydroxyl groups increased LPAD hydration abilities by enhancing coulombic interaction. The upregulated GSDME expression together with cleaved caspase-3 provided impressive pyroptosis induction. The pyroptosis further initiated proinflammatory cytokines release, increased immune cell infiltration, activated adaptive immune responses and create a favorable immunogenic hot TME. The study not only confirms the role of GSDME in the immune infiltration and prognosis of osteosarcoma, but also provides a promising strategy for the inhibition of osteosarcoma by pore-forming GSDME gene delivery induced enhanced pyroptosis to reshape the TME of osteosarcoma.

Atomistic molecular dynamics simulation and COSMO-SAC approach for enhanced 1,3-propanediol extraction with imidazolium-based ionic liquids.

1,3-Propanediol (1,3-PDO) is a key chemical in various industries, including pharmaceuticals and material sciences, and is projected to see significant market growth. However, the current challenges in its downstream processing, particularly in terms of cost and efficiency, highlight the need for innovative solutions. Our study delves into using ionic liquids (ILs) as a potential alternative, aiming to address these critical separation challenges more sustainably and efficiently. In this study, we utilized molecular dynamics (MD) simulations and the COSMO-SAC to examine 1,3-propanediol (1,3-PDO) extraction using four imidazolium-based ionic liquids with 1-butyl-3-methylimidazolium [Bmim] cation and with different anions bis(pentafluoroethanesulfonyl)imide [NPF2]-, bis(trifluoromethylsulfonyl)imide [NTF2]-, thiocyanate [SCN]-, and trifluoromethanesulfonate [TFO]-. Molecular dynamics simulations, incorporating analysis of radial distribution functions (RDF) and spatial distribution functions (SDF), revealed that [Bmim][SCN] and [Bmim][TFO] exhibit enhanced interactions with 1,3-PDO. Notably, [Bmim][SCN] formed the most hydrogen bonds, averaging 1.639 per molecule, due to its coordinating [SCN]- anion. This was in contrast to the fewer hydrogen bonds formed by non-coordinating anions in [Bmim][NPF2] and [Bmim][NTF2]. In ternary systems, [Bmim][SCN] and [Bmim][TFO] demonstrated superior selectivity for 1,3-PDO extraction compared to the other ionic liquids, with selectivity values around 29. These findings, supported by COSMO-SAC predictive modeling, highlight the potential of [Bmim][SCN] as a promising candidate for 1,3-PDO extraction, emphasizing the importance of anion selection in optimizing ionic liquid properties for this application. In our study, we employed MD simulations, incorporating the OPLS-AA force field, and COSMO-SAC to investigate the extraction of 1,3-PDO using imidazolium-based ionic liquids: [Bmim][NTF2], [Bmim][NPF2], [Bmim][SCN], and [Bmim][TFO]. The MD simulations were conducted using LAMMPS software, focusing on elucidating the RDF, SDF, and hydrogen bonding. Analysis of the distribution coefficient (β) and selectivity (S) for the ternary mixture was also conducted. These aspects of the simulation were analyzed using TRAVIS and VMD software. Additionally, the COSMO-SAC model was employed to determine the activity coefficients of 1,3-PDO in the ionic liquids, with molecular optimization conducted using Gaussian16 and sigma profile calculations performed using COSMO-SAC.

In-situ TEM characterization and atomistic simulation of cavity generation and interaction in tungsten at 800 °C under dual W2+/He+ irradiation

In this work, we have investigated the role of pre-existing nano-scale cavities and irradiation sequence in helium clustering, and in the evolution of the cavity population in tungsten at 800 °C, by performing in-situ Transmission Electron Microscopy (TEM) characterizations under dual-beam W2+(600 keV)/He+(10 keV) irradiation to a total dose of 0.351 dpa, and supported by atomistic simulations. Pre-existing cavities, induced by single W2+ irradiation to 0.256 dpa, were seen to attract induced defects from subsequent irradiations, resulting in enhanced cavity size and a reduced number density in sequential irradiation when compared to concurrent irradiation. Such cavities developed faceted geometries and saturated in number density at an additional 0.051 dpa induced by He+ exposure in the sequential irradiation. In contrast, concurrent irradiations generate a denser cavity distribution that showed no sign of saturation or faceting up to the total irradiation damage of 0.351 dpa. In addition, Molecular Dynamics and Statics simulations were performed using LAMMPS software. These showed the pre-existing cavities to be trapping sites for helium atoms. Despite the large binding energy of ∼6 eV of those cavities, He-He interactions still occurred beyond the effective radius of such sinks, leading to additional He-He clustering not seeded by neighbouring cavities. These results point to the importance of selecting the irradiation conditions to simulate synergistic ion effects on cavity formation and evolution in plasma-facing tungsten-base materials.

Investigating the role of network former interactions on charge carrier diffusivity in glasses

Ionic transport is a critical property for the glass industry, since emerging applications such as sensors, batteries, and electric melting are based on the phenomenon. Short-range interactions (anion-charge carrier) have not been able to explain the total activation barrier observed experimentally, and, as such, it is critical to understand the larger role of all ions in a glass, not just the carrier and the ‘site’ ions. This research focuses on the role of network formers and their impact on diffusion in glasses, something that current models lack an explicit explanation of. Atomistic simulations with randomly generated parameters for the cation potentials and classical simulations were used to determine the diffusion coefficients and activation energies for synthetic network formers. Using this database, explainable machine learning algorithms were employed to explore network former interactions and determine which parameters are the most influential for ion diffusion. Results suggest that the bond length of the cations changes the geometry of the structure contributing the greatest to cation-modifier interactions.

Effect of crystal facets in plasmonic catalysis

While the role of crystal facets is well known in traditional heterogeneous catalysis, this effect has not yet been thoroughly studied in plasmon-assisted catalysis, where attention has primarily focused on plasmon-derived mechanisms. Here, we investigate plasmon-assisted electrocatalytic CO2 reduction using different shapes of plasmonic Au nanoparticles - nanocube (NC), rhombic dodecahedron (RD), and octahedron (OC) - exposing {100}, {110}, and {111} facets, respectively. Upon plasmon excitation, Au OCs doubled CO Faradaic efficiency (FECO) and tripled CO partial current density (jCO) compared to a dark condition, with NCs also improving under illumination. In contrast, Au RDs maintained consistent performance irrespective of light exposure, suggesting minimal influence of light on the reaction. Temperature experiments ruled out heat as the main factor to explain such differences. Atomistic simulations and electromagnetic modeling revealed higher hot carrier abundance and electric field enhancement on Au OCs and NCs than RDs. These effects now dominate the reaction landscape over the crystal facets, thus shifting the reaction sites when comparing dark and plasmon-activated processes. Plasmon-assisted H2 evolution reaction experiments also support these findings. The dominance of low-coordinated sites over facets in plasmonic catalysis suggests key insights for designing efficient photocatalysts for energy conversion and carbon neutralization.

Atomistic modeling of nucleation kinetics of Guinier–Preston zones in Al–Cu alloys: Two formation scenarios and prediction of the time-temperature-transformation diagram

Metastable precipitates of solute atoms, known as Guinier–Preston (GP) zones, play a significant role in the age hardening of Al–Cu alloys. In this study, we utilized the classical nucleation theory (CNT) to obtain time-temperature-transformation (TTT) diagrams for steady-state nucleation over a wide temperature range. The CNT model, which is based on atomistic simulations using artificial neural network potentials, was constructed to predict the nucleation kinetics of GP zones in the Al–Cu system with solute concentrations ranging from 1.3 to 2.5 at% (3.0 to 5.7 wt%). The nose temperature at which the steady-state nucleation time for GP zone formation is the shortest was calculated. The nose temperature increased and the minimum nucleation time decreased as the solute concentration increased. These predictions are comparable to previous theoretical results and are in good qualitative agreement with experimental observations. Furthermore, two formation scenarios of double-layer GP (GP2) zones, considering synchronous and asynchronous attachments of solute atoms to clusters, were compared in terms of nucleation efficiency. This provides new insights into the nucleation pathways of the GP2 zone in Al–Cu alloys.

Molecular Autonomous Pathfinder Using Deep Reinforcement Learning.

Diffusion in solids is a slow process that dictates rate-limiting processes in key chemical reactions. Unlike crystalline solids that offer well-defined diffusion pathways, the lack of similar structural motifs in amorphous or glassy materials poses great challenges in bridging the slow diffusion process and material failures. To tackle this problem, we propose an AI-guided long-term atomistic simulation approach: molecular autonomous pathfinder (MAP) framework based on deep reinforcement learning (DRL), where the RL agent is trained to uncover energy efficient diffusion pathways. We employ a Deep Q-Network architecture with distributed prioritized replay buffer, enabling fully online agent training with accelerated experience sampling by an ensemble of asynchronous agents. After training, the agents provide atomistic configurations of diffusion pathways with their energy profile. We use a piecewise nudged elastic band to refine the energy profile of the obtained pathway and the corresponding diffusion time on the basis of transition-state theory. With the MAP framework, we demonstrate atomistic diffusion mechanisms in amorphous silica with time scales comparable to experiments.

Innovative modeling of monolayer puckered arsenene: Bridging quantum mechanics and finite element analysis

Current study presents a novel hybrid approach combining finite element modeling and density functional theory calculations to investigate the mechanical properties of monolayer puckered arsenene. The multiscale analysis in this study leverages finite element analysis as a distinctive approach, complementing the nano‐scale capabilities of density functional theory and molecular dynamics by overcoming limitations faced by these two methods in representing complex scenarios. Furthermore, finite element analysis demonstrates computational efficiency for larger structures, making it suitable for systems where atomistic simulations may be impractical. This hybrid methodology offers a unique framework for accurately predicting key properties, including elastic modulus and buckling force, by synergistically integrating the strengths of both computational techniques. In addition to demonstrating the effectiveness of our approach in accurately capturing material behavior, our findings shed light on fundamental aspects of nanoscale mechanics, with implications for various applications in nanotechnology, materials science, and structural engineering. By providing a deeper understanding of the mechanical response of 2D materials, our research contributes to advancing the field of nanoscale materials engineering and informs the design of innovative nanostructures with tailored mechanical properties.

Detection of dislocation motion in atomistic simulations of nanocrystalline materials

A method for identifying dislocation motion in atomistic simulations is presented. While identifying static and moving dislocations within a single crystal or a combination of such is well established, the method described here is tailored to identify dislocation motion by correlating the displacements of individual atoms. This facilitates the identification of dislocation motion in complex structural arrangements, and allows the specific contribution to plastic deformation, due to dislocation motion, to be separated from that of other mechanisms. The method is applied to test cases in crystals and grain boundaries (GBs), in which irradiation-induced creep (IIC) was induced. It is shown that the method singles out the moving dislocations from among the dislocation forest at GBs, thus identifying the specific reactions driving the distortion at any given time. This enables the study of dislocation processes in the presence of realistic obstacles, and the study of the effects of microstructure on dislocation mobility. As an example of such a study, the method is applied to rule out intragranular slip, and to estimate the contribution of dislocation motion to strain, in a NC undergoing IIC.

Simulating CO2 diffusivity in rigid and flexible Mg-MOF-74 with machine-learning force fields

The flexibility of metal–organic frameworks (MOFs) affects their gas adsorption and diffusion properties. However, reliable force fields for simulating flexible MOFs are lacking. As a result, most atomistic simulations so far have been carried out assuming rigid MOFs, which inevitably overestimates the gas adsorption energy. Here, we show that this issue can be addressed by applying a machine-learning potential, trained on quantum chemistry data, to atomistic simulations. We find that inclusion of flexibility is particularly important for simulating CO2 chemisorption in MOFs with coordinatively unsaturated metal sites. Specifically, we demonstrate that the diffusion of CO2 in a flexible Mg-MOF-74 structure is about one order of magnitude faster than in a rigid one, challenging the rigid-MOF assumption in previous simulations.

Spatially-resolved cluster dynamics modeling of irradiation growth

We develop here a spatially resolved, three-dimensional continuum model coupling cluster dynamics (SR-CD) and crystal plasticity to investigate irradiation growth in zirconium. The model uses scale separation to divide the population of the irradiation cluster into mobile and immobile families. Small interstitial and vacancy clusters are modeled using anisotropic reaction–diffusion equations. Among the immobile clusters, an atomistically-informed vacancy cluster to vacancy loop transition is taken into account. The coupling between the evolution equation of CD and the plastic deformation of the material is two-fold, with stress-informed bias factors and local inelastic strains computed from the evolution of the evolving cluster population. The numerical implementation of the model utilizes the finite element method to analyze both single-crystal and polycrystalline samples. The growth strains that are computed align well with the experimental data provided by Carpenter for single-crystal Zr. Furthermore, the transformation of a vacancy cluster into a complete vacancy loop, occurring at a size of 14 nm, is in agreement with experimental observations and atomistic simulations. The density, size, and growth rate of the dislocation loops, denoted as 〈c〉 and 〈a〉, also exhibit good agreement with transmission electron microscopy (TEM) analysis of irradiated Zr and its alloys. Our findings demonstrate that there is a spatial correlation between the growth of these dislocation loops and growth strains, significantly influenced by the crystal size. To explain the expansion of the 〈a〉 axis and the contraction of the 〈c〉 axis in irradiated Zr, it is necessary to consider the diffusion anisotropy difference (DAD) of mobile interstitial species. We show that the PWR Kearns parameters, specifically fr = 0.63, ft = 0.32, fa = 0.05, confer enhanced irradiation resistance to Zr along the principal directions when compared to single crystals. Additionally, reducing the grain size to nanograins further enhances the resistance to irradiation-induced growth, particularly along the direction with the highest volume fraction of basal poles [0001].

Mo‐Si alloys studied by atomistic computer simulations using a novel machine‐learning interatomic potential: Thermodynamics and interface phenomena

Mo‐Si alloys studied by atomistic computer simulations using a novel machine‐learning interatomic potential: Thermodynamics and interface phenomena

A study of hydrophobic domain formation of polymeric drug precipitation inhibitors in aqueous solution

Despite the widespread use of polymers as precipitation inhibitors in supersaturating drug formulations, the current understanding of their mechanisms of action is still incomplete. Specifically, the role of hydrophobic drug interactions with polymers by considering possible supramolecular conformations in aqueous dispersion is an interesting topic. Accordingly, this study investigated the tendency of polymers to create hydrophobic domains, where lipophilic compounds may nest to support drug solubilisation and supersaturation. Fluorescence spectroscopy with the environment-sensitive probe pyrene was compared with atomistic molecular dynamics simulations of the model drug fenofibrate (FENO). Subsequently, kinetic drug supersaturation and thermodynamic solubility experiments were conducted. As a result, the different polymers showed hydrophobic domain formation to a varying degree and the molecular simulations supported interpretation of fluorescence spectroscopy data. Molecular insights were gained into the conformational structure of how the polymers interacted with FENO in solution phase, which apart from nucleation and crystal growth effects, determined drug concentrations in solution. Notable was that even at the lowest polymer concentration of 0.01 %, w/v, there were polymer-specific solubilisation effects of FENO observed and the resulting reduction in apparent drug supersaturation provided relevant knowledge both from a mechanistic and practical perspective.

Regulating interfacial diffusion of nanocrystal‐in‐glass composites: Insights from atomistic simulation

AbstractBy combining the spectroscopic quality of a crystalline phase with the outstanding thermal and chemical stability of the matrix glass, nanocrystal‐in‐glass composites (NGCs) exhibit excellent optical properties. However, strong interfacial diffusion occurs during high‐temperature encapsulation, leading to the erosion of the nanocrystals and detrimental effects on the final performance. The control of interfacial diffusion has so far relied on intuition and experience, with the underlying physical mechanisms remaining largely unknown. In this study, through atomistic simulations and experimental verification, we have developed atomic models of NGCs, revealing the dynamic characteristics of the interface and elucidating how the glass structure can regulate the interface diffusion. Finally, we demonstrate how a tunable interface can be achieved through compositional adjustments of the glass matrix.

Conformational Dynamics in Corynebacterium glutamicum Diaminopimelate Epimerase: Insights from Ligand Parameterization, Atomistic Simulation, and Markov State Modeling.

The microbial enzyme diaminopimelate epimerase (DapF), a vital enzyme in the lysine biosynthetic pathway, catalyzes the conversion of L, L-diaminopimelate (L, L-DAP) to D, L-diaminopimelate (D, L-DAP) using a catalytic cysteine dyad with one cysteine in thiol state and another in thiolate. Under oxidizing conditions, the catalytic cysteines of apo DapF form a disulfide bond that alters the structure and function of DapF. Given its potential as a target for antimicrobial resistance treatments, understanding DapF's functional dynamics is imperative. In the present work, we employ microsecond-scale all-atom molecular dynamics simulations of product-bound DapF and apo-DapF under oxidized and reduced conditions. We employ a polarized charge model for the ligand and the active site residues, which was necessary to preserve the electrostatic environment in the active site and retain the ligand in the active site. The product-bound DapF and apo-DapF in oxidized and reduced conditions exhibit a closed, semi-open, and open conformation, respectively, as identified using the internal coordinates of the dimeric enzyme and the principal component analysis. The conformational switch is guided by the dynamic catalytic (DC) loop, loop II, and loop III movements in the active site. The time scale of the close-to-open conformational transition is estimated to be 0.8 μs through Markov state modeling (MSM) and transition path theory (TPT). The present study explains the role of various active site residues and loops in ligand binding and protein dynamics in the DapF enzyme under different redox conditions. Such information will be helpful in future inhibitor design studies targeting the DapF enzyme.