Journal of Chemical Theory and Computation最新文献

UGA-SSMRPT2, the spin-free perturbative analogue of Mukerjee's State-Specific Multireference Coupled Cluster Theory (MkMRCC), is known to be successful for size-extensive and intruder-free construction of dissociation curves. This work demonstrates that UGA-SSMRPT2 is also an accurate and computationally inexpensive framework for computing the excitation energies. The method achieves near-chemical accuracy for the vast majority of π → π*, n → π*, charge-transfer, valence-Rydberg, and Rydberg excited states commonly used for benchmarking electronic structure theories for excited states. Our results demonstrate that UGA-SSMRPT2 excitation energies lie within 0.20 eV of EOM-CCSD and/or well-established theoretical best estimates, often surpassing the popular MRPT2 approaches like NEVPT2, CASPT2, and MCQDPT while typically requiring smaller active spaces. Its state-specific formulation circumvents the well-known intruder-state problem and eliminates the need for empirical parameters, such as IPEA shifts in CASPT2. This work proposes UGA-SSMRPT2 as a robust and scalable approach for modeling challenging electronically excited states.

The Global Natural Orbital Functional (GNOF) provides a straightforward approach to capture most electron correlation effects without needing perturbative corrections or limited active spaces selection. In this work, we evaluate both the original GNOF and its modified variant, GNOFm, on a set of twelve 5- and 6-membered molecular rings, systems characterized primarily by dynamic correlation. This reference set is vital as it comprises essential substructures of more complex molecules. We report complete-basis-set limit correlation energies for GNOF, GNOFm, and the benchmark CCSD(T) method. Across the Dunning basis sets, both functionals deliver a balanced and accurate description of the molecular set, with GNOFm showing small but systematic improvements while preserving the overall robustness of the original formulation. These results confirm the reliability of the GNOF family and its ability to capture dynamic correlation effects.

In this work, we implement a local pair natural orbital-based coupled-cluster method through the full treatment of quadruple excitations (CCSDTQ). The domain-based local pair natural orbital (DLPNO) approach, which has successfully been applied to lower levels of coupled-cluster theory, is utilized in our algorithm, and thus our algorithm is called DLPNO-CCSDTQ. For simplicity in the working equations and in the implementation, we t₁-dress the two-electron integrals as well as Fock matrix elements. Our method can recover CCSDTQ-CCSDT and CCSDTQ-CCSDT(Q) energy differences on the order of 0.01-0.05 kcal mol^-1, even at a loose quadruples natural orbital (QNO) occupation number cutoff of 3.33 × 10^-6. To highlight the capabilities of our code and its potential future applications, we showcase computations that would be intractable with canonical CCSDTQ, such as the benzene dimer, (H₂O)₁₇, and adamantane. With sufficient computing resources, computations up to 15 heavy atoms (40 atoms overall) may be feasible for fully bonded 3D systems.

We generalize the interpolative separable density fitting (ISDF) method, used for compressing the four-index electron repulsion integral (ERI) tensor, to incorporate adaptive real space grids for potentially highly localized single-particle basis functions. To do so, we employ a fast adaptive algorithm, the recently introduced dual-space multilevel kernel-splitting method, to solve the Poisson equation for the ISDF auxiliary basis functions. The adaptive grids are generated by using a high-order accurate, black-box procedure that satisfies a user-specified error tolerance. Our algorithm relies on the observation, which we prove, that an adaptive grid resolving the pair densities appearing in the ERI tensor can be straightforwardly constructed from one that resolves the single-particle basis functions, with the number of required grid points differing only by a constant factor. We find that the ISDF compression efficiency for the ERI tensor with highly localized basis sets is comparable to that for smoother basis sets compatible with uniform grids. To demonstrate the performance of our procedure, we consider several molecular systems with all-electron basis sets that are intractable using uniform grid-based methods. Our work establishes a pathway for scalable many-body electronic structure simulations with arbitrary smooth basis functions, making simulations of phenomena such as core-level excitations feasible on a large scale.

CO₂ from flue gas is central to mitigating fossil-fuel-derived emissions, where adsorbent performance directly dictates process energy efficiency and process cost. Although machine learning (ML) has emerged as a powerful tool for accelerating adsorbent discovery, its predictive accuracy is fundamentally limited by the physical reliability of the underlying training data, a manifestation of the "garbage in, garbage out" (GIGO) problem. Most existing CO₂ adsorption databases rely on Lennard-Jones (LJ) force fields, whose deficiencies in describing CO₂-CO₂ and CO₂-framework interactions, particularly at high pressures, introduce systematic bias into the ML models. To address this, we developed a physically accurate van der Waals force field based on an Exp-PE potential and constructed a high-fidelity CO₂ adsorption database. Building on this data set, we introduce quadrupole-responsive descriptors that explicitly capture the anisotropic electrostatics of CO₂, leading to improved ML predictive accuracy. This framework identifies high-performing COF/MOF adsorbents, including COF-50 (ΔN_CO2 = 13.58 mol/kg) and COF-364 (ΔN_CO2 = 12.43 mol/kg), whose working capacities exceed those of current reported porous materials.

Understanding how charge distributions on aggregated chains change with microstructure under constant electrochemical potential is crucial for elucidating the behavior of polymeric organic mixed ionic-electronic conductors (OMEICs), yet it remains difficult to study. To address this challenge, we introduce a methodology to perform classical atomistic simulations of doped semiconductors at a constant electrochemical potential. The method allows individual polymer chains to be oxidized and reduced, taking into account their individual redox potentials and the externally tunable electrochemical potential. The implementation follows a grand-canonical molecular dynamics (GC-MD) scheme, with the local modulation of the redox potential being described by a QM/MM Hamiltonian. Applied to a semicrystalline polymer with ordered layered and lamellar structures, the method reproduces the experimentally observed minimal structural changes over the electrochemical potentials and charging levels considered. Near the redox potential, charging levels fluctuate more strongly, and variations in the interlamellar angle (defined by the normal of adjacent lamellae) are most pronounced. Moreover, analysis of the local environment reveals no detectable correlation between a chain's redox reaction and the charge distribution of neighboring chains, except at the most negative potentials, where redox events occur preferentially in more positively charged surroundings. Lastly, examination of individual chains shows minimal chain-chain charge correlation, and the single-chain conformation remains closely linked to its redox behavior. Overall, this work provides a robust framework for investigating charge distributions in dynamically doped systems and offers new conceptual routes for studying polymer structural responses under constant electrochemical potentials.

A major challenge in light-matter simulations is bridging the disparate time and length scales of electrodynamics and molecular dynamics. Current computational approaches often rely on heuristic approximations of either the electromagnetic (EM) or the material component, hindering the exploration of complex light-matter systems. Herein, MaxwellLink─a modular, open-source Python framework─is developed for the massively parallel, self-consistent propagation of classical EM fields interacting with a large heterogeneous molecular ensemble. The package utilizes a robust TCP/UNIX socket interface to couple EM solvers with a wide range of molecular drivers. In this initial release, MaxwellLink supports EM solvers spanning from single-mode cavities to full-feature three-dimensional finite-difference time-domain (FDTD) engines and molecules described by multilevel open quantum systems, force-field and first-principles molecular dynamics, and nonadiabatic real-time Ehrenfest dynamics. With the socket-based architecture, users can seamlessly switch between levels of theory of either the EM solver or molecules without modifying the counterpart. Moreover, the EM engine and molecular drivers scale independently across multiple high-performance computing (HPC) nodes, facilitating large-scale simulations previously inaccessible to existing numerical schemes. The versatility and accuracy of this code are further demonstrated through applications including superradiance, radiative energy transfer, vibrational strong coupling in Bragg resonators, and plasmonic heating of molecular gases. By providing a unified, extensible engine, MaxwellLink potentially offers a powerful platform for exploring emerging phenomena across the research fronts of spectroscopy, quantum optics, plasmonics, and polaritonics.

Three-body interactions in water play a crucial role in accurately modeling its structural and thermodynamic properties. These interactions consist of a polarization term that decays as an inverse power of the intermolecular separations R_ab and a term that is usually assumed to describe exchange interactions and decay exponentially. Due to the complexity of fitting the latter term at large R_ab, it is often damped or truncated beyond a certain distance, also because the computational cost of including three-body effects in molecular simulations scales as N³ with the number of molecules, compared to the N² scaling of two-body interactions. Here, investigations of the impact of long-range three-body exchange interactions on the results of such simulations have been performed by systematically extending the average R_ab of trimers included. It is demonstrated that these long-range effects are important for accurately describing the density of liquid water, ρ(T), as a function of temperature, but are essentially negligible for several other properties of water. The effects of three-body damping onset on ρ(T) are larger than they would have been with an exponential decay; however, it is shown here that the decay is dominated by exponential components only at fairly small R_ab, while for large R_ab, the nonpolarization three-body effects decay as 1/R_abⁿ. These findings are rationalized by calculations with the symmetry-adapted perturbation theory. Another reason for the importance of three-body effects is their N³ scaling. Clearly, long-range three-body exchange interactions should be included in high-accuracy water models. It is shown that the reason these interactions have such large effects on ρ(T) is their extreme anisotropy affecting the structure of liquid water. Our work also sheds light on discrepancies between the theory and experiment for ρ(T).