Journal of Computational Chemistry

The image shows an isosurface of Hirshfeld-I “atoms in molecules” for Ti-doped CeO₂, taken at an electron density of 0.03e/A³, as presented by Danny E. P. Vanpoucke, Patrick Bultinck, and Isabel Van Driessche on page 405. The cubic Ce_0.75Ti_0.25O₂ unit cell is shown along the 111 direction. The different atoms are still clearly distinguishable at this iso-surface level, and show the Ti atom in the corners to be much smaller than the Ce atoms on the sides. In this issue, this implementation of the Hirshfeld- I method for solids is published back to back with a Comment from Thomas A. Manz and the authors' Reply.

Many approximate accelerated stochastic simulation methods exist that can provide a means of employing Kinetic Monte Carlo techniques to model chemical kinetics. The drawback in using such methods is that sometimes the population of the reactant species present during the simulations becomes negative, leading to unphysical results. This is due to the manner in which the change in time during the reactions is handled by the stochastic simulation methods in such systems. The work by Shantanu Kadam and Kumar Vanka on page 394 discusses a means of avoiding this problem by combining the representative reaction approach (RRA) based methodology with the stochastic simulation algorithm (SSA) and the binomial method.

Vibration–rotation energy levels of the H₂O₂, D₂O₂, and HOOD molecules were predicted to “spectroscopic” accuracy using the ab initio state-of-the-art potential energy surface.

A fully ab initio technique is presented for the computation of atomic anisotropic displacement parameters, DebyeWaller factors and dynamical X-ray structure factors of crystalline materials, which is based on the description of the harmonic lattice dynamics and the electron charge distribution of the system. This scheme is here applied to crystalline silicon at 298 K; an overall agreement of 0.47% is found between the predicted and the experimental structure factors.

The artificial neural network is a computational technique that can be used to predict different variables from molecular descriptors. In this case, four molecular descriptors have been used to develop an artificial neural network that can predict the flash point of 95 esters. The results indicate that the artificial neural network has a good predictive capability with an average error less than 2.35%.

To fabricate nanoelectronic devices, electron transport behavior in functional molecules has been studied over the years. This article investigates electron transport in triangular zigzag graphene nanoribbons (ZGNRs) by the first-principles theory. The results show a prominent molecular rectification property in the triangular ZGNRs because of their unique electronic states. The influences of asymmetric structure and anchoring groups on transport property are also discussed in detail.

Quantum mechanical computations of solute properties with explicit solvent molecules and molecular dynamics clusters are increasingly popular. To reduce the number of clusters and computational time needed, the structure of the solvation shell and averaged but structured solvent density are explored. The algorithm appears to be quite general; for nuclear magnetic resonance and Raman spectra taken as examples, it provided about 90% savings in the total computational times.

Cluster-continuum model calculations show that the hydration of CO₂ follows a stepwise mechanism and a labile H₃O⁺ intermediate is involved in the reaction. Furthermore, if a proton transfer from the H₃O⁺ moiety to the carbonyl oxygen occurs, the hydration pathway will be followed. If there is a proton transfer from the H₃O⁺ intermediate to the outer water phase through the water bridge, the dissociation product of HCO₃⁻ will be formed.

Long-range correction (LC), which is required to give accurate orbital energies, also drastically improves Diels–Alder reaction energy diagrams in density functional theory (DFT) calculations. Based on LC-DFT calculations, on the intrinsic reaction coordinates of Diels–Alder reactions proceeding only with HOMO and LUMO, corresponding global hardness responses, the halves of HOMO-LUMO gaps, behave almost constant from reactants to transition states and then rapidly increase with reaction enthalpies to products.

When one-step perturbation is applied to deal with the perturbation from a polar moiety to a nonpolar one and vice versa in a polar environment in molecular dynamics simulations, a soft-core nonpolar reference state is the best reference state, because the properly sized cavity can be sampled in such a simulation, but post molecular dynamics simulation rotational and translational sampling has to be applied additionally to sample the optimal orientations for the polar solute molecule.

In some cases, the stochastic simulation of chemical kinetics is accompanied by the appearance of negative numbers. To solve this problem, two new algorithms based on the representative reaction approach have been proposed. These two novel methods outperform other binomial distribution-based algorithms that have been used in the past to solve the problem of negative populations.

The Hirshfeld-I method is extended to solids, allowing for the partitioning of a solid density into constituent atoms. The use of precalculated density grids makes the implementation code independent, and the use of pseudopotential-based electron density distributions is shown to give qualitatively the same results as all electron densities. Results for some simple solids/periodic systems like cerium oxide and graphene are presented.

The issues raised in the preceding comment are addressed. It is shown why Hirshfeld-I is, from a theoretical point of view, a good method for defining AIM and obtaining charges. Charges for a set of ionic systems are calculated using the presented method and shown to be chemically feasible. Comparison of pseudodensity to all-electron based results shows the pseudodensities to be sufficient to obtain all-electron quality results. Timing results for systems containing hundreds of atoms are presented.