Journal of Chemical Theory and Computation最新文献

Machine learning interatomic potentials (MLIPs), also known as machine learning force fields (MLFFs), offer scalable means of simulating complex systems and processes at ab initio level accuracy. One such process is the critical yet still poorly understood formation of the solid electrolyte interphase (SEI) at the anode of a Li-ion battery (LIB) during the first charge cycle, where electrochemical reduction of the electrolyte leads to the generation of decomposition products. MLIPs are uniquely poised to atomistically describe these electrochemical processes, as they are not as affected by the same limitations in bonding and electron transfer as classical force fields. Nonetheless, training MLIPs to run accurate dynamics of a condensed phase with two different oxidation states, such as in electrochemistry, is challenging for many architectures. In this work, we show that by using MPNICE, a message passing MLIP architecture with iterative charge equilibration, we are able to accurately (within 1 kcal/mol) train models along two potential energy surfaces (reduced and unreduced) for LIB-relevant electrolyte systems. Importantly, we demonstrate strategies for sampling and training to examples of anion radicals of these species, which often are not centered on any atom (off-center radicals, or OCRs). We additionally discuss well-known limitations of global charge equilibration (Qeq) algorithms in erroneously delocalizing charge, and test methods to alleviate the impact on resulting dynamics. Simulations using these models reveal new insights into electrolyte reduction and considerations for the realistic simulation of electron transfer processes in the condensed phase.

Chondroitin sulfate A (CSA) is a negatively charged linear glycosaminoglycan that plays a vital role in many biological processes. Research on CSA has been challenging due to its size, chemical heterogeneity, and multitude of binding partners. To address these issues, we developed a model of CSA for coarse-grained molecular dynamics simulations based on the Martini 3 force field. We demonstrate that this model is capable of reproducing atomistic properties of the repeating CSA disaccharide unit, including its molecular volume, bonded interactions, and structural polymer properties of CSA chains of different lengths. In particular, for biologically relevant long chains and despite using an explicit solvent, the computational cost is significantly reduced, relative to the cost equivalent atomistic simulations would require. The compatibility of the model with the Martini Go̅ protein model was tested by retrieving the force–response relationship of the CSA–malaria adhesin VAR2CSA complex. Importantly, we explored the influence of electrostatics on CSA aggregation. We show that the default Martini 3 parameters lead to overaggregation. We provide at least three different strategies to alleviate this issue, making use of a bigger bead for sodium cations, reflecting their hydration shell, partial ionic charges as a mean-field resource to take into account electronic polarizability, and, optionally, particle mesh Ewald summation as a more robust treatment of long-range electrostatics. Our model enables predictive modeling of CSA and potentially other chondroitin sulfates with the Martini 3 force field. In addition, this model provides insights for the further development of coarse-grained models of highly charged systems.

Drug losses during infusions can be caused by complex adsorption and absorption phenomena at the interface between aqueous formulations and polymer-based medical devices, most commonly made of plasticized poly(vinyl chloride) (PVC). Despite their success in describing adsorption phenomena, molecular simulation methods are inherently limited by their length and time scales, which precludes their application to the study of absorption processes. Here, we extended this molecular understanding using the Martini 3 coarse-grained framework to explore absorption within plasticized PVC matrices containing either DEHT or TOTM plasticizers. We used a top-down approach to refine intermolecular interactions. After optimizing the solute–water and solute–PVC nonbonded interactions to reproduce experimental partitioning data, we combined potential of mean force (PMF), free-energy perturbation (BAR), and long equilibrium simulations to map both the thermodynamic and kinetic aspects of sorption. This work establishes a transferable coarse-grained framework for modeling drug–polymer interactions over extended time scales, bridging atomistic insights with experimental observations and paving the way toward full infusion-system simulations including excipients and complex polymer formulations.

In this work, we discuss decoherence, frustrated hops and internal consistency in surface-hopping-based methodologies. We demonstrate that moving away from an independent-trajectory picture is the strategy which allows us to propose a robust and reliable surface-hopping scheme. Based on the exact factorization and on the idea of coupled trajectories, we consider the swarm of trajectories, that mimics the nuclear dynamics in nonadiabatic processes, as a unique entity. In this way, imposing energy conservation of the swarm and allowing the trajectories to share energy when hops occur clearly indicates the route toward a new surface hopping scheme. Encouraging results are reported, in terms of electronic and vibrational time-dependent properties on the photodynamics of fulvene and 4-(dimethyloamino)benzonitrile, modeled with full-dimensional linear vibronic coupling Hamiltonians.

Seniority is a useful way of organizing Hilbert space for strongly correlated systems. The exact zero-seniority wave function, doubly occupied configuration interaction (DOCI), provides accurate results (given the right orbitals) for many strongly correlated electronic systems but has a combinatorial computational cost. In many cases, pair coupled cluster doubles provide a polynomial-cost approximation that closely reproduces the energies of DOCI, but it breaks down in some cases and, as shown herein, it does not provide particularly good density matrices. In this work, we demonstrate that by using the Jordan–Wigner transformation to turn the seniority zero problem back into a Fermionic one, we can provide mean-field variational results of DOCI quality for the Hubbard model and a few small molecular dissociation examples, with polynomial cost, both for the energies and for density matrices, all while being protected from collapse. This success is rooted in the proof we provide, showing that the Hartree–Fock wave function on the Jordan–Wigner-transformed Hamiltonian transforms back to variational coupled cluster doubles in the seniority zero representation, but restricted to have determinant rather than permanent amplitude coefficients, without compromising its overall accuracy.

The hydrated electron, an excess electron in liquid water, plays a crucial role in a plethora of chemical processes, motivating extensive research efforts to characterize its structure, dynamics, and reactivity in solution. Recent theoretical approaches to understanding this intriguing object have involved ab initio simulations based on density functional theory (DFT). Although DFT allows for the study of hydrated electron reactivity and quantum mechanical behavior, it is well-known that anionic systems can suffer from significant density-driven errors (DDEs). Density-corrected DFT (DC-DFT) provides a framework to mitigate such errors; the method reduces DDEs by replacing the self-consistent (SC) density associated with a given density functional with the Hartree–Fock (HF) density. Since HF densities tend to be more localized than DFT SC densities, the DC-DFT scheme significantly improves errors in calculations where the SC density is spuriously delocalized. Here, we investigate how the use of density correction affects the calculated properties of the DFT-simulated (PBEh) hydrated electron, a particularly challenging diffuse anionic system to simulate. First, we analyze charge delocalization in a system consisting of a model octahedral hydrated electron water cluster (the so-called Kevan structure) along with a spatially separated sulfur atom. We show that the use of density correction indeed reduces DDEs in comparison to a standard DFT global hybrid functional. We then propagate molecular dynamics trajectories of the hydrated electron using DC-DFT, where we find that DC further localizes electron density in the cavity region, a signature of reduced charge delocalization. Unfortunately, the decreased radius of gyration of the spin density and corresponding tightening of the local solvation structure from density correction causes predicted observables to deviate further from experimental measurements than when density correction is not employed. We argue that DC’s worse agreement with experiment results from the removal of a fortuitous cancellation of errors that is intrinsic to the PBEh functional. This indicates that the difficulties with DFT to simulate hydrated electrons are primarily due to the inherent approximations in DFT rather than to density-driven errors.

Interpretable reaction coordinates are essential for understanding rare conformational transitions in molecular dynamics. The Atomistic Mechanism of Rare Events in Molecular Dynamics (AMORE-MD) framework enhances the interpretability of deep-learned reaction coordinates by connecting them to atomistic mechanisms, without requiring any a priori knowledge of collective variables, pathways, or end points. Here, AMORE-MD employs the ISOKANN algorithm to learn a neural membership function χ representing the dominant slow process, from which transition pathways are reconstructed as minimum-energy paths aligned with the gradient of χ, and atomic contributions are quantified through gradient-based sensitivity analysis. Iterative enhanced sampling further enriches transition regions and improves coverage of rare events, enabling recovery of known mechanisms and chemically interpretable structural rearrangements at atomic resolution for the Müller–Brown potential, alanine dipeptide, and the elastin-derived hexapeptide VGVAPG.

Vibrational spectra convey a wealth of structural and dynamical information; however, their reliable assignment and interpretation often benefit from the integration of complementary spectroscopic techniques and require the support of accurate quantum chemical calculations. The harmonic approximation is frequently insufficient for quantitative spectroscopy, while fully anharmonic treatments rapidly become computationally prohibitive for large and flexible molecular systems, in particular, for biomolecules. In this framework, we introduce a general perturb-then-diagonalize approach that relies on a three-class partitioning of normal modes into primary, auxiliary, and spectator subsets and combines numerical strategies based on analytical Hessians and analytical gradients. Accurate anharmonic contributions are explicitly included for the modes of primary interest, while the influence of external modes is accounted for through finite differences of analytical gradients, avoiding the much more expensive evaluation of Hessians. Several case studies demonstrate the robustness, ease of use, and accuracy of the proposed approach across a broad range of molecular systems, including situations in which vibrational and rotational spectroscopic data provide complementary information. When combined with a dual-level strategy in which accurate methods are employed for harmonic terms and less expensive methods for anharmonic contributions, the present framework enables vibrational spectra of near-spectroscopic accuracy for biomolecules and other chemically rich systems. More complex environments can be addressed by coupling the method with multilayer approaches.

Accurately describing the electronic properties of heavy-element molecular systems in complex environments is essential for advancing technologies such as optoelectronics and solar cells. However, achieving accurate predictions remains challenging because both relativistic and electron correlation effects must be considered equally, along with interactions involving other species in the complex environment (e.g., solvent). This paper extends our real-time time-dependent Dirac-Kohn–Sham (rt-TDDKS) implementation in PyBERTHA-RT to include environmental effects using the “uncoupled” Frozen-Density-Embedding (FDE) scheme, where only the active subsystem evolves dynamically in time. This adaptation utilizes existing FDE functionality within the PyEmbed module of the PyADF scripting framework. The native Python APIs of PyBERTHA-RT and PyADF provide an ideal environment for development, enhancing readability and reusability. We demonstrate that the FDE potential maintains the numerical stability of the active subsystem’s density matrix propagation. Illustrative results for lead halides (PbCl₂ and PbI₂) in γ-butyrolactone (GBL) solution show the effects of increasing solvent molecules on absorption spectra. This case study demonstrates the new implementation’s applicability to realistic systems, offering a basis for studying electron dynamics in heavy-element molecules in complex environments under linear and nonlinear regimes, relevant to perovskite precursor chemistry.

Highly emissive organic molecular crystals find applications in several areas, such as organic electronics, solar cells, and sensors. Understanding the excited-state mechanisms underlying these applications is essential for optimizing and controlling them effectively. Exciton models coupled with nonadiabatic dynamics, particularly quantum dynamics, provide crucial insights into photochemical and photophysical processes in molecular crystals. Nevertheless, there remains a lack of general tools and automated workflows to facilitate such simulations. In this paper, we present a computational strategy to investigate the photoactivated dynamics of organic molecular crystals, bridging methodologies traditionally used for molecular systems and materials science, with a particular focus on the interplay between local excitations and charge transfer (CT) processes. We have implemented an interface between the fromage and Overdia programs, enabling the construction of vibronic Hamiltonians for molecular crystals within an excited-state ONIOM(QM:QM′) framework, incorporating long-range electrostatics through a RESP-based Ewald summation. Fragment-based diabatization provides a route to quantum dynamics simulations in weak-to-intermediate coupling regimes. The method was applied to the photophysics of dibenzo[g,p]chrysene (DBC) crystals using time-dependent DFT. The fromage/Overdia interface was employed to compute the couplings of local excitations and CT states for 18 unique DBC dimers in the crystal and to quantify the influence of electrostatic embedding, which was found to be modest (10–20%). Simulations on π-stacked dimers reproduced the small red shift observed experimentally from solution to crystal, attributed to electronic interactions among fixed monomers rather than crystal electrostatics. Quantum dynamics simulations revealed ultrafast population transfer from bright local excitations to CT states. This approach establishes a robust framework linking molecular and solid-state excited-state dynamics, with potential applications for studying excitations, defects, and impurities in molecular crystals.