Are exponential‐type basis sets preferable to Gaussians?

Abstract

ABSTRACT:The relative merits of exponential-type and Gaussian basis sets are discussed. It is argued that contracted Gaussian basis sets are not just an expedient replacement for exponential-type basis sets but have desirable properties of their own that make them appropriate choices for molecular electronic structure calculation, apart from the ease of calculation of their multicenter integrals. A variety of studies are quoted showing that appropriately chosen contracted Gaussian basis sets can produce satisfactory results even for electronic properties that depend strongly on the values of the wave function in regions close to and far from the nuclei. Further, it is argued that the long-range behavior of the wave function in regions of consequence for the calculation of energies and electronic properties can actually be represented better by such basis sets than by exponential-type basis sets.

Method for Slater-Type Density Fitting for Intermolecular Electrostatic Interactions with Charge Overlap. I. The Model

“…As shown in great detail by Klopper and Kutzelnigg, this can be done quite successfully . The arguments in favor of using GTOs rather than STOs in quantum chemical calculations are therefore strong …”

“…69 The arguments in favor of using GTOs rather than STOs in quantum chemical calculations are therefore strong. 70 On the other hand, given their exponential decay, Slater-type functions…”

Benchmark all-electron ab initio quantum Monte Carlo calculations for small molecules

“…The results agree very well with an earlier study based on numerical atomic orbitals (NAO) on a radial grid. 2 While the relative merits of the GTO and STO representations of orbitals in quantum chemistry are still under debate, 12 attempts have been made to reduce the computational effort of working with STO basis sets, [13][14][15][16] leading to the development of several STO-based electronic structure codes. [17][18][19] Of these, the only code active development and available for general use is ADF, 19 offering a full state-of-the-art implementation of Hartree-Fock (HF) and density functional theory (DFT) electronic structure calculations for molecules and, via its sister program BAND, for periodic systems.…”

Fe(CO)4 flexible as a two‐level system avoided conical intersection

“…The final basis set is as follows: (14 s 11 p 6 d 1 f )/[10 s 8 p 3 d 1 f ] for Fe and (11 s 6 p 1 d )/[5 s 3 p 1 d ] for C, O. Contracted Gaussian basis sets are not just an expedient replacement for Slater orbital basis sets (used in most accurate calculations of iron carbonyls 5, 9), but have desirable properties of their own that make them appropriate choices for molecular modeling 33. Similar basis sets were used in several calculations of iron carbonyls 7, 9, 10, 20, 21, 32, 34, 35.…”