Journal of Chemical Physics最新文献

Electrochemical devices often charge both through Faradaic reactions and electric double layer formation. Here, we study these coupled processes in a model system of a long electrolyte-filled pore subject to a small suddenly applied potential, close to the equilibrium potential Ψeq at which there is no net Faradaic charge transfer. Specifically, we solve the coupled Poisson-Nernst-Planck and Frumkin-Butler-Volmer equations by asymptotic approximations, using the pore's small inverse aspect ratio as the small parameter. In the early time limit, the reaction-diffusion equations yield an extended Faradaic transmission line model that includes a voltage source, Ψeq, biasing the Faradaic reactions, captured by the resistance RF. In the long-time limit, the pore exhibits a nontrivial potential of zero total charge, Ψpztc=Ψeq1-Ẑ(0)/RF, where Ẑ(0) is the experimentally accessible zero-frequency impedance of the system. This expression provides a new means to measure the Faradaic contribution to Ψpztc experimentally.

Accurately capturing long-range interactions is critical for molecular dynamics simulations based on machine learning interatomic potentials. We recently proposed the sum-of-Gaussians neural network (SOG-Net), which learns long-range energy contributions directly from energy and force data such that the long-range tail of different decay rates can be well fitted. In this work, we incorporate the SOG-Net with a short-range descriptor of the Cartesian atomic cluster expansion, resulting in the CACE-SOG model, to show that the SOG-Net is a general module that can be coupled with different short-range descriptors. We also study new technical developments in the SOG-Net, including improved extrapolation accuracy, handling of different charge states, and faster convergence. We evaluated the CACE-SOG model across a diverse set of systems, including molecular dimers, aqueous salt solutions, charged ionic clusters, and liquid-vapor and Pt(111) interfacial water systems, and compared it with the CACE-based latent Ewald summation and the CACE-only methods. These results demonstrate that the SOG-Net is promising for accurately learning long-range interatomic interactions.

In this work, we present calculations of the basins and electronic population at the relativistic level, with expressions derived from a total Lagrangian density associated with the modified Dirac Hamiltonian [Zapata-Escobar and Maldonado, J. Chem. Phys. 163, 024310 (2025)]. Calculations were carried out using a local implementation for the numerical construction of the basins and the corresponding integrations of the electronic population. The basins' evaluation and electronic population calculations were obtained on both relativistic and nonrelativistic expressions, reproducing 99.9% of the total charge in the HX (X = F, Cl, Br, I, At) and SnH4 molecules. We found that the relativistic effects cause a contraction of the electron density around the nuclei of the heavy atoms, while for the hydrogen atom, the electronic population decreases, leading to a decrease in the dipole moment. In addition, we present the continuity equation from the modified Dirac Hamiltonian in order to interpret the scalar field used to obtain the basin with this scheme of calculation. Finally, we propose a new Lagrangian density associated with the Dirac Hamiltonian, from which it is possible to obtain an expression for the zero-flux condition, thereby recovering the nonrelativistic basin expression within the quantum theory of atoms in molecules approach directly from the relativistic formulation without adding a heuristic term.

We propose a method that combines the quantum-classical mapping approach to surface hopping with the dissipative quantum dynamics of the Lindblad master equation. Like conventional surface-hopping methods, our approach is based on classical trajectories coupled to the dynamics of a quantum subsystem. However, instead of evolving the subsystem wavefunction according to the time-dependent Schrödinger equation, we use stochastic quantum trajectories derived from secular Redfield theory. This approach enables the simulation of open quantum systems coupled simultaneously to Markovian quantum baths and anharmonic, non-Markovian classical degrees of freedom. Applications to the spin-boson model and to the cavity-enhanced fluorescence of an electronically nonadiabatic molecule show excellent agreement with fully quantum-mechanical benchmarks.

We study polymorph selection in a model of charged colloids, with a focus on the higher-order structure prior to and during nucleation. Specifically, we carry out molecular dynamics simulations of a repulsive Yukawa system with a slightly softened (Weeks-Chandler-Andersen) core. We consider the case where the interaction is long-ranged and the BCC crystal is stable, and also intermediate- and short-ranged cases where the FCC crystal is stable. We use two methods for structure identification, the topological cluster classification (TCC) [A. Malins et al., J. Chem. Phys. 139, 234506 (2013)] and the bond orientational order parameter analysis of Lechner and Dellago [J. Chem. Phys. 129, 114707 (2008)]. Under conditions of high supersaturation appropriate to experiments with colloids, we find that the system forms a precursor state in which the particles are hexagonally ordered. That is to say, the precursors are indistinguishable from an HCP crystal using the bond orientational order parameters. This ordering occurs at state points when the body-centered cubic crystal is the stable phase and also when the face-centered cubic crystal is stable. In all cases, the stable polymorph forms from the precursor phase in a second stage. Although at freezing the fluid is much more ordered when the interactions are short-ranged (when FCC is stable), at the supersaturations where nucleation occurs in our simulations, the higher-order structure of the metastable fluids is almost identical for the long-, short-, and intermediate-ranged systems when measured with the TCC.

Collisions of atoms and molecules with metal surfaces create electronic excitations in the metal, leading to nonadiabatic energy dissipation, inelastic scattering, and sticking. Mixed quantum-classical molecular dynamics simulation methods, such as molecular dynamics with electronic friction, are able to capture nonadiabatic energy loss during dynamics at metal surfaces. Hydrogen atom scattering from semiconductors, on the other hand, exhibits strong adsorbate-surface energy transfer only when the projectile kinetic energy exceeds the bandgap of the substrate. Electronic friction fails to describe this effect. Here, we report a first-principles parameterization of a simple Haldane-Anderson Hamiltonian model of hydrogen atom gas-surface scattering on Ge(111)c(2 × 8), for which hyperthermal scattering experiments have been reported. We subsequently perform independent-electron surface hopping and Ehrenfest dynamics simulations on this model and validate these results through numerically exact quantum-dynamical simulations using the hierarchical equation of motion approach. While mean-field dynamics yield weak nonadiabatic energy loss that is independent of the initial kinetic energy, independent electron surface hopping simulations qualitatively agree with the experimental observation that nonadiabatic energy dissipation only occurs if the initial kinetic energy exceeds the bandgap of the surface.

The binding energy and the vibrational stretching frequency of the probe molecule CO adsorbed on the low-index CeO2 surfaces [(100), (110), and (111)] were benchmarked using the coupled-cluster singles, doubles, perturbative triples [CCSD(T)] method, employing an embedded cluster approach. Using the same methodology as for the top configuration of CO on the (111) surface [J. Vázquez Quesada et al., J. Chem. Phys. 161, 224707 (2024)], the best theoretical estimate for the CO frequency on the CeO2(100) surface (CO bridge configuration) obtained at the CCSD(T)/def2-TZ/QZVPP level of theory and under low-coverage conditions (2193 cm-1) is 17 cm-1 larger than the experimental value (1 ML coverage saturation), which is in agreement with previous estimates for the CO adsorption on the CeO2(111) surface (12 cm-1). For the (110) surface, theoretical and experimental data compare differently. The CCSD(T)/def2-TZ/QZVPP values are -7 cm-1 (top configuration) and -21 cm-1 (tilt-x configuration) lower than the two experimental features measured at 2170 cm-1 (negative feature) and 2160 cm-1 (positive feature). MP2 predictions suggest the existence of a case of multiple-configuration dynamics with various almost isoenergetic configurations in a low-coverage situation. The CO harmonic vibrational frequencies were not semi-empirically scaled but explicitly corrected for anharmonic effects, which amount to 25-26 cm-1 with all tested methods. CO adsorption energies of -0.40 ± 0.07 eV, -0.17 ± 0.07 eV, and -0.20 ± 0.07 eV for the (100), (110) (top), and (110) (tilt-x) adsorption sites, respectively, are obtained at the CCSD(T)/def2-TZ/QZVPP level of theory. These results agree well with those proposed for the (111) surface (-0.22 ± 0.07 eV) [J. Vázquez Quesada et al., J. Chem. Phys. 161, 224707 (2024)] and confirm the physisorption character of the adsorption of CO on the three low-index surfaces of CeO2.

A fluctuating charge model (FCM) is developed to consider two-dimensional networks of boron nitride. In the FCM, the charge on each atom site is controlled by parameters linked to the atom's electronegativity and the interactions with other atoms (the coordination environment). The charge held on each atom site is a strong function of the local (first shell) coordination environment. The site charges are shown to be in excellent agreement with those extracted from independent density-functional theory-based calculations. The behavior of the site charges is investigated as a function of the network topology and site disorder. In the first case, specific defects (both site and topological) are introduced, and the spatial "decay" of the local charge to bulk values is assessed. In the second, highly disordered (amorphous) networks are generated, and the distribution of site charges is studied as a function of the degree of topological and site disorder (characterized by the fraction of six-membered rings and mean boron-nitrogen coordination numbers, respectively). Domains of high and low charges are observed to form across a wide range of topological disorder.

Inelastic x-ray scattering spectroscopy (IXS) is a modern synchrotron based vibrational spectroscopy, which measures transitions in the terahertz vibrations. This paper presents the first IXS on a metallo-enzyme derived sample, namely, the iron-molybdenum cofactor (FeMoco) of nitrogenase. We have also measured the corresponding far infrared absorption spectrum (far IR spectrum) of FeMoco and compared these data to previously published nuclear resonant vibrational spectra (NRVS) of FeMoco and far infrared spectra of the model Mo-dioxo-homocitrate complex K2[MoO2(R,S-H2homocitrate)2]·2H2O. Although NRVS offers several clear advantages, IXS can probe vibrations not involving iron motion and is still sensitive to heavier elements. This provides a useful method to study molybdenum vibrations, such as the ν(Mo-O) mode at 530 cm-1 in FeMoco, which is observable by IXS but not by NRVS. The comparison of the vibrational frequencies of FeMoco with those of the Mo-complex supports an earlier proposal that the homocitrate ligand in FeMoco is protonated. This work, which introduces IXS to bioinorganic chemistry, concludes by examining both the opportunities and the challenges of applying IXS to other biologically relevant systems.

We develop two multi-defect machine learning interatomic potentials (MLIPs) trained at the BLYP-D2 and PBE-D3 density functional theories using the DeepMD-kit, allowing for the investigation of structural and thermodynamic properties of nitric acid over a wide range of concentrations via molecular dynamics (MD) simulations. We directly compute the degree of dissociation, α, and pKa from MD simulations, revealing that HNO3 behaves as a weaker acid at higher concentrations, noting that our standard-state pKa value is in excellent agreement with the experimental one. In general, good agreement is observed with experimental results such as α and density outside the training dataset, with only modest deviations at low-to-medium concentrations. We benchmark our custom multi-defect DeepMD MLIPs against foundational models MACE-MP0 and MACE-OFF23. The foundation models capture some aspects of HNO3/NO3- solvation in concentrated nitric acid but show noticeable density errors and miss subtle structural features relevant to spectroscopy, whereas the bespoke DeepMD MLIPs yield more compact solvation shells, reproduce density-concentration trends, and run ∼12-15× faster than MACE-MP0. Although classical FFs are still more efficient and match experimental densities better, they lack chemical reactivity and thus cannot predict α or pKa, underscoring the need for system-specific reactive MLIPs beyond universal MLIPs.