Journal of Computational Chemistry最新文献

Seventeen electronic states of the dication VH²⁺ were characterized by the SA-CASSCF/icMRCI methodology using very extended basis sets; 11 were described for the first time. Potential energy curves were constructed and the associated spectroscopic parameters evaluated. Triplet and quintet states correlating with the V²⁺ + H channel are thermodynamic stable. For states dissociating into the channel V⁺ + H⁺, avoided crossings at large distances give rise to thermodynamic metastability but do not affect the characterization of the bound region. Configuration state functions with the 3σ orbital /doubly occupied give rise to covalent contributions to the bonding; the major contribution, however, comes from the electrostatic charge-induced dipole interaction. This explains the shape and proximity of the potential energy curves beyond their equilibrium distances. Dipole moment functions and vibrationally averaged dipole moments quantify the polarity of the molecule. Spin–orbit couplings give rise to complex and dense regions of very close-lying Ω states.

We report a direct application of the molecular tailoring approach-based (MTA-based) method to calculate the individual hydrogen bond (HB) energy in molecular crystal. For this purpose, molecular crystals of nitromalonamide (NMA) and salicylic acid (SA) were taken as test cases. Notably, doing a correlated computation using a large molecular crystal structure is difficult. Among 15 density functional theory functionals, the B3LYP provides accurate estimates of HB energies closed to the CCSD(T) ones using the 6–311 + G(d,p) basis set for all atoms. The direct application of the MTA-based method to these crystal structures is although straightforward. For instance, the calculated energy suggests that three intramolecular HBs in NMA crystal are of stronger strength (7.3–17.0 kcal/mol) than the intermolecular ones (2.7–4.0 kcal/mol). On the other hand, intermolecular HB in SA crystal is moderately stronger (9.9 kcal/mol) than intramolecular one (8.1 kcal/mol). However, these energy calculations by the MTA-based method are very expensive. For instance, the time needed to evaluate the energy of all seven HBs in NMA crystal (having molecules within maximum of 15 unit cells) is 122,681 min (~2.7 months). In view of this, we assessed our recently proposed linear-scaling Fragments-in-Fragments (Frags-in-Frags) method for estimating the single-point energies of parent molecular crystal and fragments of the MTA-based method. It has been found that the estimated HB energies by the Frags-in-Frags method are in excellent linear agreement with their MTA-based counterparts (R² = 0.9993). Furthermore, root mean square deviation is 0.12 kcal/mol. Mean and maximum absolute errors are 0.10 and 0.5 kcal/mol, respectively, and the standard deviation is 0.14 kcal/mol. Importantly, the Frags-in-Frags method is computationally efficient; it needs only 18,289 min (~12.7 days) for the estimation of energy of all HBs in NMA crystal and 3499 min (~2.4 days) for all HBs in SA crystal.

Hydrogen gas (H₂) can be produced via entirely solar-driven photocatalytic water splitting (PWS). A promising set of organic materials for facilitating PWS are the so-called inverted singlet-triplet, INVEST, materials. Inversion of the singlet (S₁) and triplet (T₁) energies reduces the population of triplet states, which are otherwise destructive under photocatalytic conditions. Moreover, when INVEST materials possess dark S₁ states, the excited state lifetimes are maximized, facilitating energy transfer to split water. In the context of solar-driven processes, it is also desirable that these INVEST materials absorb near the solar maximum. Many aza-triangulenes possess the desired INVEST property, making it beneficial to describe an approach for systematically and efficiently predicting the INVEST property as well as properties that make for efficient photocatalytic water splitting, while exploring the large chemical space of the aza-triangulenes. Here, we utilize machine learning to generate post hoc corrections to simplified Tamm–Dancoff approximation density functional theory (sTDA-DFT) for singlet and triplet excitation energies that are within 28–50 meV of second-order algebraic diagrammatic construction, ADC(2), as well as the singlet-to-triplet, ΔE_S1T1, gaps of PWS systems. Our Δ-ML model is able to recall 85% of the systems identified by ADC(2) as candidates for PWS. Further, with a modest database of ADC(2) excitation energies of 4025 aza-triangulenes, we identified 78 molecules suitable for entirely solar-driven PWS.

Voltage-dependent anion channel (VDAC) is the primary conduit for regulated passage of ions and metabolites into and out of a mitochondrion. Calculating the solvation free energy for VDAC is crucial for understanding its stability, function, and interactions within the cellular environment. In this article, numerical schemes for computing the total solvation free energy for VDAC—comprising electrostatic, ideal gas, and excess free energies plus the nonpolar energy—are developed based on a nonuniform size modified Poisson–Boltzmann ion channel (nuSMPBIC) finite element solver along with tetrahedral meshes for VDAC proteins. The current mesh generation package is also updated to improve mesh quality and accelerate mesh generation. A VDAC Solvation Free Energy Calculation (VSFEC) package is then created by integrating these schemes with the updated mesh package, the nuSMPBIC finite element package, the PDB2PQR package, and the OPM database, as well as one uniform SMPBIC finite element package and one Poisson–Boltzmann ion channel (PBIC) finite element package. With the VSFEC package, many numerical experiments are made using six VDAC proteins, eight ionic solutions containing up to four ionic species, including ATP⁴⁻ and Ca²⁺, two reference states, different boundary values, and different permittivity constants. The test results underscore the importance of considering nonuniform ionic size effects to explore the varying patterns of the total solvation free energy, and demonstrate the high performance of the VSFEC package for VDAC solvation free energy calculation.

Accurate prediction of physicochemical properties, such as electronic energy, enthalpy, free energy, and average vibrational frequencies, is critical for optimizing lithium-ion battery (LIB) performance. Traditional methods like density functional theory (DFT) are computationally expensive and inefficient for large-scale screening. In this study, we apply active learning on top of graph neural networks (GNNs) to efficiently predict these properties. By focusing on uncertain data points, active learning reduces training data size while maintaining high accuracy. Applied to the LIBE and MPcules datasets, the model achieved an R-squared (R²) values of 0.9977 with a mean absolute error (MAE) of 9.66 Ha for electronic energy and an R² values of 0.957 with an MAE of 13.94 cm⁻¹ for average vibrational frequencies. SHapley Additive exPlanations (SHAP) provided insights into key features influencing predictions, such as atomic number and spin multiplicity. This approach enhances both predictive accuracy and model interpretability, offering a scalable solution for LIB electrolyte discovery.

Multi-copper oxidases (MCOs) are enzymes of significant interest in biotechnology due to their efficient catalysis of oxygen reduction to water, making them valuable in sustainable energy production and bio-electrochemical applications. This study employs time-dependent density functional theory (TDDFT) to investigate the electronic structure and spectroscopic properties of the Type 1 (T1) copper site in Azurin, which serves as a model for similar sites in MCOs. Four model complexes of varying complexity were derived from the T1 site, including 3 three-coordinate models and 1 four-coordinate model with axial methionine ligation, to explore the impact of molecular branches and axial coordination. Calculations using ωB97X-D3 functional, def2-TZVP basis set, and conductor-like polarizable continuum model (CPCM) solvation model reproduced key experimental spectral features, with increased model complexity improving agreement, particularly for the ~400 cm⁻¹ band splitting in resonance Raman spectra. This work enhances our understanding of T1 copper sites' electronic properties and spectra, bridging the gap between simplified models and complex proteins. The findings contribute to the interpretation of spectroscopic data in blue copper proteins and may inform future studies on similar biological systems.