Atomic Radii in Crystals

TLDR

Bragg’s 1920 atomic radii were refined using a larger crystal dataset. The authors discuss why the radii agree with experimental distances and relate them to established ionic and metallic radii. The empirical radii sum to bond lengths with an average 0.12 Å deviation across 1200 cases, accurately describing covalent, metallic, and ionic bonds and matching calculated outer‑shell radii.

Abstract

A set of empirical atomic radii has been set up, such that the sum of the radii of two atoms forming a bond in a crystal or molecule gives an approximate value of the internuclear distance. These radii give fair agreement with experiment in over 1200 cases of bonds in all types of crystals and molecules, with an average deviation of about 0.12 Å. The radii are similar to a set suggested by W. L. Bragg in 1920, but refined by consideration of many more crystals. They hold for covalent, metallic, and ionic binding equally well. These radii agree remarkably well with calculated radii of maximum radial charge density in the outermost shells of the atoms, as taken from previously unpublished calculations of D. Liberman, J. T. Waber, and D. T. Cromer, of the Los Alamos Laboratory, by relativistic self-consistent field calculation, using the exchange correction suggested in 1951 by the present author. There is discussion of the probable reason for this agreement, and discussion of the relation of these radii to the ionic radii of Pauling, Zachariasen, and others, and the tetrahedral and metallic radii of Pauling.