Browse Articles
Discover research articles across all indexed journals
A theoretical model of gas diffusivity in graphene nanochannels
Gas diffusion in graphene nanochannels is pivotal for applications such as gas sensing and membrane separation, where nanoscale confinement introduces unique transport phenomena. Unlike bulk-phases, diffusion in graphene nanochannels is significantly influenced by adsorption, which modifies density distributions and alters diffusivity behavior. In this study, molecular dynamics simulations are combined with a theoretical framework to comprehensively investigate gas diffusion under varying pressures and channel heights. A modified Chapman–Enskog model, derived from atomistic Lennard-Jones potential parameters, is proposed to account for the effects of confinement. Simulation results reveal that gas diffusivity decreases with increasing gas-phase pressure and decreasing channel height due to enhanced density in the nanochannels. Interestingly, for ultra-narrow channels (h ≲ 0.7 nm), the diffusivity correction factor exhibits non-monotonic behavior, initially decreasing but subsequently increasing due to overlapping repulsive potential fields. The proposed model integrates adsorption effects through density predictions based on the Boltzmann distribution and effectively predicts gas diffusivities with relative errors of less than 13%, even under strong confinement. These findings highlight the critical interplay between adsorption and confinement in shaping gas transport within graphene nanochannels. The theoretical model provides a predictive tool for designing graphene-based gas separation and sensing devices, offering fundamental insights for optimizing their performance.
Stretching and twisting of double-stranded RNA under forces: Unwinding mechanism and base-pair dependent elasticity
We used all-atom molecular dynamics simulations to investigate the mechanical response of double-stranded RNA (dsRNA) by applying various forces. We used the helical rise and helical twist, as well as a newly defined helical diameter, to characterize the stretching and twisting of dsRNA. The results indicate that dsRNA unwinds when stretched, accompanied by a linear increase in helical rise and helical diameter. Then, we utilized the normal modes, which are linear combinations of helical modes, to elucidate the underlying mechanism of dsRNA unwinding from an energetic perspective. On the other hand, we employed a stiffness matrix based on a rigid base pair model to examine the base-pair dependence of twist elasticity for dsRNA, as well as stretch elasticities with respect to the helical rise and helical diameter. The results show that the force induces variations in the local elasticities and their couplings of dsRNA, which are closely related to the distributions of base pairs. The mean stretch and twist elasticities can be considered as constants within the measurement uncertainties; however, their couplings demonstrate a slight linear dependency on applied force.
Iterative charge equilibration for fourth-generation high-dimensional neural network potentials
Machine learning potentials allow performing large-scale molecular dynamics simulations with about the same accuracy as electronic structure calculations, provided that the selected model is able to capture the relevant physics of the system. For systems exhibiting long-range charge transfer, fourth-generation machine learning potentials need to be used, which take global information about the system and electrostatic interactions into account. This can be achieved in a charge equilibration step, but the direct solution of the set of linear equations results in an unfavorable cubic scaling with system size, making this step computationally demanding for large systems. In this work, we propose an alternative approach that is based on the iterative solution of the charge equilibration problem (iQEq) to determine the atomic partial charges. We have implemented the iQEq method, which scales quadratically with system size, in the parallel molecular dynamics software LAMMPS for the example of a fourth-generation high-dimensional neural network potential (4G-HDNNP) intended to be used in combination with the n2p2 library. The method itself is general and applicable to many different types of fourth-generation MLPs. An assessment of the accuracy and the efficiency is presented for a benchmark system of FeCl3 in water.
Solvation free energies from neural thermodynamic integration
We present a method for computing free-energy differences using thermodynamic integration with a neural network potential that interpolates between two target Hamiltonians. The interpolation is defined at the sample distribution level, and the neural network potential is optimized to match the corresponding equilibrium potential at every intermediate time step. Once the interpolating potentials and samples are well-aligned, the free-energy difference can be estimated using (neural) thermodynamic integration. To target molecular systems, we simultaneously couple Lennard-Jones and electrostatic interactions and model the rigid-body rotation of molecules. We report accurate results for several benchmark systems: a Lennard-Jones particle in a Lennard-Jones fluid, as well as the insertion of both water and methane solutes in a water solvent at atomistic resolution using a simple three-body neural-network potential.
Quantum corrections to the kinetic energy and the <i>ab initio</i>-based prediction of the thermodynamic properties and vapor–liquid equilibria of hydrogen
The ability of ab initio-based intermolecular potentials to predict the vapor–liquid-equilibria (VLE) and thermodynamic properties of hydrogen is investigated via Monte Carlo simulation. The combination of a simplified ab initio atomic potential (SAAP) and first order Feynman–Hibbs (FH-1) interactions closely reproduces the VLE phase envelope, providing a good estimate of the critical point. The SAAP + FH-1 combination also improves the prediction of other thermodynamic properties. However, the accurate determination of enthalpy, heat capacity, isothermal compressibility, isochoric pressure coefficient, and isobaric thermal expansion coefficient requires the addition of a quantum correction to the kinetic energy (QCKE). The QCKE is a post-simulation contribution to the thermodynamic properties of quantum fluids and, as such, can be used to improve the accuracy of any predictions using an intermolecular potential. The addition of QCKE to the SAAP + FH-1 potential results in values for the thermodynamic properties that are close to reference data for hydrogen at temperatures greater than 40 K and pressures up to 100 MPa.
Slow dynamical modes from static averages
In recent times, efforts have been made to describe the evolution of a complex system not through long trajectories but via the study of probability distribution evolution. This more collective approach can be made practical using the transfer operator formalism and its associated dynamics generator. Here, we reformulate in a more transparent way the result of Devergne et al. [Adv. Neural Inform. Process. Syst. 37, 75495–75521 (2024)] and show that the lowest eigenfunctions and eigenvalues of the dynamics generator can be efficiently computed using data easily obtainable from biased simulations. We also show explicitly that the long time dynamics can be reconstructed by using the spectral decomposition of the dynamics operator.
Morphology regulation during mechanochemistry synthesis activating nanostructured aluminum lithium storage behavior
Aluminum (Al) is a potential anode material for lithium-ion batteries due to its high theoretical capacity and low volume expansibility. However, scalable fabrication of nanostructured Al still faces a great challenge. In addition, the lithium storage performance of Al anode materials always encounters a severe strike within a dozen discharge/charge cycles, and such an abnormal behavior of the Al anode material remains enigmatic. Herein, a mechanochemistry method without using any solvent is developed to achieve scalable production of Al nanoparticles and the morphology of the obtained Al nanoparticles could be regulated using Ketjen black (KB). KB with a chain-like structure could regulate the Al crystal growth process and the aggregation of Al nanoparticles during the solid-phase reaction, shortening the electron transfer path among Al crystals, ultimately activating the lithium storage behavior of nanostructured Al. Initial discharge/charge capacities of 630.6 and 402.0 mA h g−1 were achieved at 50 mA g−1; unfortunately, the nanostructured Al still suffered from rapid deterioration of lithium storage performance. Comprehensive analysis demonstrated that the raised energy barrier of LiAl formation and the slow lithium diffusion kinetics in the Al matrix may be the main factors destroying the lithium storage performance of the Al anode material. This work provided more evidence for illustrating the lithium storage behavior of the Al anode.
Constrained dipole moment density functional theory for the calculation of the charge-transfer energy in non-covalent complexes
The original constrained dipole moment density functional theory allows one to control the magnitude of the molecular dipole moment in a variational and non-empirical way. In this work, we extend this methodology to control the three Cartesian components of the molecular dipole moment. The new theoretical development is suitable for the calculation of the charge-transfer energy contributions to the total interaction energies in non-covalent complexes. To test the reliability of the theoretical development, we form three sets of non-covalent complexes from the literature with a total of fifty-one systems. The former set of complexes includes many different types of non-covalent interactions, the second set consists of prototypical non-covalent complexes and three biologically relevant interactions between DNA base pairs, and the third set comprises halogen bonding complexes. We determined the charge-transfer energy contributions and the total interaction energies of all these complexes. The calculated charge-transfer energies are in very good agreement with the ones calculated using the fragment-based Hirshfeld methodology, which has been proven to be reliable. Nevertheless, the new procedure relies on the molecular dipole moment, which is observable, while the fragment-based Hirshfeld methodology relies on a definition of a population analysis.
Radical scavenging rate constants determined by spin relaxation times of electron spin polarized radicals as measured by free induction decay signals
Radical scavenging reaction rate constants were measured by monitoring the free induction decay (FID) of the unpaired electron of radicals by using laser-synchronized pulsed-EPR. This method probes large electron spin magnetization arisen from dynamic electron spin polarization (DEP), which remarkably enhances the EPR signal. DEP decays with the longitudinal spin relaxation time, which was observed by using the FID detection method. In the presence of a radical scavenger, the DEP decay time depends on both spin-lattice relaxation and the chemical reaction with the scavenger, the latter of which reduces radical concentration. The plots of DEP decay rates against the radical scavenger concentrations gave the pseudo-first-order reaction rate constants for various radicals. This procedure was applied to determine the radical scavenging reaction rate constants of hydroxycyclohexyl, 2-hydroxypropyl, diphenylphosphinoyl, and α,α-dimethoxybenzyl radicals. The measured rate constants show good agreement with the previously reported values determined by using another method monitoring electron spin echo (ESE) decays of the radicals. We discussed the advantageous and disadvantageous characters of the FID detection method with respect to the existing ESE method in the viewpoints of signal intensity, selectivity of radicals, the simpleness of the measurements, and so on.
Energy decomposition analysis method with the DFT-in-xTB embedding strategy for intermolecular interactions in large systems
In this work, an energy decomposition analysis (EDA) method, termed DM-EDA(EB), is introduced to explore intermolecular interactions in large systems by employing a DFT-in-xTB embedding scheme. DM-EDA(EB) integrates density matrix-based EDA (DM-EDA) with the GFNn-xTB method to decompose the total interaction energy into electrostatic, exchange–repulsion, polarization, and correlation terms. Test cases demonstrate that DM-EDA(EB) can accurately analyze total interaction energies in large systems with the computational efficiency comparable to GFNn-xTB. Notably, by using the appropriate partition strategy, DM-EDA(EB) is able to provide quantificational knowledge of individual interactions in large assemblies.
Simulating many-body open quantum systems by harnessing the power of artificial intelligence and quantum computing
Simulating many-body open quantum systems (OQSs) is challenging due to the intricate interplay between the system and its environment, resulting in strong quantum correlations in both space and time. This Perspective presents an overview of recently developed theoretical methods using artificial intelligence (AI) and quantum computing (QC) to simulate the dynamics of these systems. We briefly introduce the dissipaton-embedded quantum master equation in second quantization, which provides a single master equation suitable for representation by neural quantum states or quantum circuits. The promising performance of AI- and QC-based approaches is demonstrated through preliminary research on simulating the quantum dissipative dynamics of many-body OQSs. We also discuss the limitations and future developments of these methods, which hold promise for overcoming the computational challenges associated with many-body OQS dynamics.
Non-equilibrium coexistence between a fluid and a hotter or colder crystal of granular hard disks
Non-equilibrium phase coexistence is commonly observed in both biological and artificial systems, yet understanding it remains a significant challenge. Unlike equilibrium systems, where free energy provides a unifying framework, the absence of such a quantity in non-equilibrium settings complicates their theoretical understanding. Granular materials, driven out of equilibrium by energy dissipation during collisions, serve as an ideal platform to investigate these systems, offering insights into the parallels and distinctions between equilibrium and non-equilibrium phase behavior. For example, the coexisting dense phase is typically colder than the dilute phase, a result usually attributed to greater dissipation in denser regions. In this article, we demonstrate that this is not always the case. Using a simple numerical granular model, we show that a hot solid and a cold liquid can coexist in granular systems. This counterintuitive phenomenon arises because the collision frequency can be lower in the solid phase than in the liquid phase, consistent with equilibrium results for hard-disk systems. We further demonstrate that kinetic theory can be extended to accurately predict phase temperatures even at very high packing fractions, including within the solid phase. Our results highlight the importance of collisional dynamics and energy exchange in determining phase behavior in granular materials, offering new insights into non-equilibrium phase coexistence and the complex physics underlying granular systems.
Adsorption behavior analysis of CNCl on transition metal-doped fluorinated diamanes: A first-principles study
Cyanogen chloride (CNCl) is a toxic chemical that poses significant risks to human health and the environment; therefore, its level must be accurately monitored. Herein, the adsorption of CNCl by transition metal-doped fluorinated diamanes (F-diamanes) has been extensively studied via first-principles calculations. Key parameters such as adsorption energies, charge transfer amounts, bandgaps, sensitivity, densities of states, projected density of states, charge density differences, and recovery time were systematically analyzed. Results reveal that monometallic doping significantly enhances CNCl adsorption, with increases in adsorption energy by 164%–368% and charge transfer by 1234%–1571%, particularly in the AuFD-CNCl, AgFD-CNCl, and CuFD-CNCl systems, which demonstrated improved sensing performances. Similarly, bimetallic co-doping further strengthened adsorption, with energy enhancements of 277%–309% and charge transfer increases of 1238%–1505%. Au-CuFD-CNCl, Au-AgFD-CNCl, and Cu-AgFD-CNCl systems also showed superior sensing performances. Meanwhile, the recovery time of CNCl molecules on the AuFD, AgFD, Au-CuFD, and Au-AgFD surfaces was drastically reduced to acceptable levels at 279–412 K, leading to their desorption. Therefore, these four systems exhibited excellent reversibility properties, suggesting their applicability in gas-sensing applications. This work can facilitate the applications of doped F-diamanes in environmental conservation, energy storage, and chemical engineering.
Challenges in determining the thermal conductivity of core–shell nanowires by atomistic simulation
In the present work, we investigate the thermal conductivity (κ) of different core–shell nanowires using molecular dynamics simulation and Green–Kubo (EMD), imposing a temperature gradient (NEMD) and Müller-Plathe (rNEMD) approaches. We show that in GaAs@InAs nanowires, the interface effect becomes more significant than the nanowire cross-sectional geometry. In particular, κ decreases as the interface area increases, reaching a minimum, and then increases when the interface strain relaxes. This is particularly important for thermoelectric applications, where minimization of κ is desired. In particular, the different methods can predict minima at different core diameters without special considerations. In addition, the NEMD approach and, to a lesser extent, rNEMD tend to overestimate the κ values, which cannot be corrected with the methods available in the literature. By analyzing the temperature and length dependence, (I) we show that interfacial scattering primarily involves phonon–phonon interactions, which mainly affect low-energy modes, a mechanism that effectively reduces κ at low temperatures. (II) The Langevin thermostat tends to pump low-energy modes in the NEMD approach, but this effect decreases with longer nanowires. (III) Energy exchanges in rNEMD stimulate high-energy phonons, derived from the saturation of κ at a much shorter nanowire length than NEMD. These findings highlight the challenges of accurately determining κ of ultrathin core–shell nanowires, where only the EMD approach provides precise results. With the recognition of non-equilibrium contributions to the overestimation of κ by NEMD and rNEMD, these methods can still provide valuable insights for a comprehensive understanding of the underlying thermal transport mechanisms.
Polariton-induced Purcell effects via a reduced semiclassical electrodynamics approach
Recent experiments have demonstrated that polariton formation provides a novel strategy for modifying local molecular processes when a large ensemble of molecules is confined within an optical cavity. Herein, a numerical strategy based on coupled Maxwell–Schrödinger equations is examined for simulating local molecular processes in a realistic cavity structure under collective strong coupling. In this approach, only a few molecules, referred to as quantum impurities, are treated quantum mechanically, while the remaining macroscopic molecular layer and the cavity structure are modeled using dielectric functions. When a single electronic two-level system embedded in a Lorentz medium is confined in a two-dimensional Bragg resonator, our numerical simulations reveal a polariton-induced Purcell effect: the radiative decay rate of the quantum impurity is significantly enhanced by the cavity when the impurity frequency matches the polariton frequency, while the rate can sometimes be greatly suppressed when the impurity is near resonance with the bulk molecules forming strong coupling. In addition, this approach demonstrates that the cavity absorption of light exhibits Rabi-splitting-dependent suppression due to the inclusion of a realistic cavity structure. Our simulations also identify a fundamental limitation of this approach—an inaccurate description of polariton dephasing rates into dark modes. This arises because the dark-mode degrees of freedom are not explicitly included when most molecules are modeled using simple dielectric functions. As the polariton-induced Purcell effect alters molecular radiative decay differently from the Purcell effect under weak coupling, this polariton-induced effect may facilitate understanding the origin of polariton-modified photochemistry under electronic strong coupling.
A molecular beam study of olefin adsorption on ultrathin ionic liquid films on Pt(111)
We investigate fundamental aspects concerning selectivity in hydrogenation reactions for Solid Catalyst with Ionic Liquid Layer (SCILL)-type systems. Some of us recently reported that the adsorption behavior of 1,3-butadiene and 1-butene on Pt(111) can be tuned with ultrathin layers of the ionic liquid (IL) 1,3-dimethylimidazolium bis(trifluoromethylsulfonyl)imide ([C1C1Im][Tf2N]): Increasing the IL coverage leads to increased blocking of olefin adsorption sites. Notably, a smaller IL amount is needed to prevent 1-butene adsorption as compared to 1,3-butadiene adsorption, leading to a selectivity window, in which 1,3-butadiene still can be adsorbed while 1-butene cannot. With the aim to evaluate whether this is a general behavior, we study an IL with a longer alkyl chain at the cation, 1-methyl-3-octylimidazolium bis(trifluoromethylsulfonyl)imide ([C8C1Im][Tf2N]), and an IL with a different anion, 1-methyl-3-octylimidazolium hexafluorophosphate ([C8C1Im][PF6]). Indeed, we observe a selectivity window for all ILs and thus demonstrate that this concept also applies for other ILs with different chain lengths and anions. Nevertheless, there are pronounced differences for the three ILs in terms of the IL coverages required for full blocking and the width of the selectivity window. Different explanations are discussed, e.g., the structure of the IL layer and the interaction strengths of olefins and ILs with the substrate.
Thermal conductivity of the layered titanate K0.8Li0.27Ti1.73O4 explored by a deep learning interatomic potential
The theoretical prediction of thermal conductivity in many layered oxides remains challenging, primarily due to their structural complexity and low symmetry. The traditional Boltzmann transport equation method is highly accurate but limited by the low-order phonon scattering model, which makes it difficult to resolve the high-order scattering effects of low symmetry layered materials. The classical molecular dynamics calculation is efficient but lacks accuracy due to the missing multi-component potential function. In this study, we develop a strategy to predict the thermal conductivity of K0.8Li0.27Ti1.73O4 (KLTO), a model of layered oxides by machine-learning using a deep neural network model to acquire the interatomic potential of KLTO. The deep learning potential (DLP) is in excellent agreement with density functional theory in predicting atomic force, energy, and elastic properties. In addition, the calculated out-of-plane thermal conductivity values based on the DLP (0.37 W m−1 K−1) are close to experimental results (0.28 W m−1 K−1). This machine-learning framework for constructing interatomic potentials can be extended to other layered materials, offering a promising approach for advancing the theoretical study of such systems.
A molecular dynamics study on coalescence-induced jumping of moving and static droplets
In this paper, molecular dynamics simulations are employed to investigate the coalescence-induced jumping behavior of moving and stationary droplets at the nanoscale on superhydrophobic surfaces. The results show that the initial velocity of the droplets significantly influences the coalescence time and jumping characteristics. As the initial velocity increases, the coalescence time decreases, and the horizontal velocity increases, suggesting that controlling the initial velocity can adjust droplet motion behavior. In terms of energy conversion, the total energy conversion rate remains relatively constant at lower initial velocities but increases significantly as the velocity rises. This is primarily due to the reduced coalescence time and viscous dissipation caused by the increased initial kinetic energy, allowing more energy to be converted into the kinetic energy of jumping. The energy conversion rate in the horizontal direction increases with initial velocity, while in the vertical direction, it tends to decrease. This study deepens the understanding of coalescence-induced jumping phenomena at the nanoscale and provides a theoretical basis for engineering applications, showing that droplet behavior can be effectively modulated by controlling the initial velocity.
Structural and dynamical properties of aqueous NaCl brines confined in kaolinite nanopores
Quantifying thermodynamics, structural, and dynamical properties of brine confined in clay pores is critical for a variety of geo-energy applications, including underground hydrogen storage (UHS) and carbon capture and sequestration (CCS). Atomistic molecular dynamics simulations are applied here to study aqueous NaCl brines within 10-Å kaolinite slit pores. NaCl concentrations are chosen at 5, 10, 12.5, and 15 wt. %, all below the solubility limit and high enough to provide statistically relevant information. The distribution of the ions within the nanopores is found not to be homogeneous. Explicitly, Na+ cations, preferentially attracted to the siloxane surface, accumulate in regions with low water density, whereas Cl− anions, attracted to the gibbsite surface of kaolinite, are found within the hydration layers. Confinement affects the properties of ions, with ion pairing being more pronounced within the pore than in bulk aqueous solutions at similar temperatures, pressures, and compositions. Conversely, the ions affect the properties of confined water. For example, the lifetime of water–water hydrogen bonds in confinement is shortened within the hydration shells; increasing salinity from 5 to 12.5 wt. % reduces the likelihood of water density fluctuations near the kaolinite surfaces, although when the NaCl concentration rises from 12.5 to 15 wt. %, Cl− anions enhance the likelihood of density fluctuations for the hydration layer near the gibbsite surface. The simulated molecular trajectories are studied further to extract diffusion coefficients. While confinement in the kaolinite nanopore reduces the mobility of all species, non-monotonic trends are observed as a function of salt concentration. The trends seem associated with the likelihood of ion pairing. Furthermore, the diffusion coefficients for the cations are predicted to be higher than those for the anions, which is contrary to what is typically observed in bulk brines. Because density fluctuations are correlated with properties such as the solubility of gases in confined water, our observations may have important implications for geo-energy applications such as UHS and CCS.
Real-time propagation of adaptive sampling selected configuration interaction wave functions
We have developed a new time propagation method, time-dependent adaptive sampling configuration interaction (TD-ASCI), to describe the dynamics of a strongly correlated system. We employ the short iterative Lanczos method as the time-integrator, which provides a unitary, norm-conserving, and stable long-time propagation scheme. We used the TD-ASCI method to evaluate the time-domain correlation functions of molecular systems. The accuracy of the correlation function was assessed by Fourier transforming into the frequency domain to compute the dipole-allowed absorption spectra. The Fourier transform (FT) has been carried out with a short-time signal of the correlation function to reduce the computation time, using an efficient alternative FT scheme based on the ESPRIT signal processing algorithm. We have applied the TD-ASCI method to prototypical strongly correlated molecular systems and compared the absorption spectra to spectra evaluated using the equation of motion coupled cluster method with a truncation at the singles, doubles, and triples level.