Test and modification of the van der Waals' radii employed in the default PCM model

Abstract

ABSTRACT:High level ab initio calculations at the B3LYP/6-311ϩϩG(d,p) and MP2(full)/6-311ϩϩG(d,p) levels employing PCM/UA0 model with different van der Waals' radii for the systems that contain lithium atoms have been carried out, in order to see if the van der Waal's radius for lithium atom employed in the default PCM/UA0 model is proper or not. Comparative analysis indicated that the van der Waals' radius for alkali metals, especially for lithium atom in the default PCM/UA0 model within the Gaussian 03 package, is too small, which causes erroneous redundant imaginary frequencies (RIFs) in the characterization of Li-containing compounds from moderate to big size. A new set of van der Waals' atomic radii based on QTAIM, proposed by Bader, was suggested for a better choice in the characterization of compounds containing alkali metals, for which it can effectively avoid the erroneous RIFs for corresponding geometries of these Li-containing systems.

A comparative study of crystallographic van der Waals radii

“…Finally, modern theoretical studies, mainly quantum chemical calculations, combined with experimental work, are still worth further effort, although the results from past practices have not always been encouraging [21,26,[35][36][37].…”

Balancing the atomic waistline: Isodensity‐based scrf radii for main‐group elements and transition metals

“…contour is defined as IDSCRF radius; if the latter is reached first, the distance from the nucleus to the ρ = 0.0010 a.u. contour is defined as IDSCRF radius (this latter value is the same as Bader's radius30, 38). However, for some elements, the ∇ 2 ρ = 0.0010 a.u.…”

“…The accuracy of the IDSCRF radii with respect to the geometric structures, energetics, and curvatures of the Hessian matrices for these systems are addressed in the following sections. Li, Be, and Mg use atomic radii as their compounds display more covalent than ionic character (for Li, see Ref 30…”

“…This may be true for most cases, however, it is not for several important systems; Li‐containing30 and simple organometallic compounds (i.e., ClMBr, RMX, M = Mg, Zn etc.) suffering from blatant inaccuracies.…”

“…The aim of this work is to introduce a novel set of radii toward more accurately calculating the solvent effect, an approach not based on any training set. Herein, we reevaluate the radii for main‐group elements from our previous work30 and extend the application of Bader‐determined radii to include transition metals, toward remedying the RIFs and geometric errors, and eventually predicting more accurate solvation free energy.…”