[soori71]

I want to calculate the atomic packing factor for covalently bonded solids with significant ionic character, like crystalline SiC, Si3N4, hydrogenated amorphous SiC (denoted by a-SiC:H), a-SixNy:H, a-SixCyNz:H, a-Si:H etc. Now Si3N4 is more ionic in character than SiC, which in turn is also not completely covalent. Thus one may not use pure ionic radii or pure covalent radii available in internet. From literature I find bond lengths for atom pairs: Si-C (1.87-1.89 Angstroms), Si-N ((1.70-1.76 Angstroms), and Si-H (1.50-1.55 Angstroms), for the same or similar materials. But I am unable t find well defined and self consistent data for the radii of individual atoms (call it ions or bonding atoms) composing these kind of compounds.

The covalent radii from http://web.mit.edu/course/3/3.091/www3/pt/pert5.html are for example consistent for Si-C bond length, but Si radius is smaller for Si-H bond, bigger for Si-N bond.

The idea is to use it in the following formula for atomic packing factor:

APF = (total volume of atoms per mole)/(Molar volume of compound);
Total volume of atoms per mole = (4/3)*Pi*(x*r[sub-A]^3 + y*r[sub-B]^3)   where r[sub-A] and r[sub-B] are the ionic or covalent radii mentioned above, forming the compound AxBy.

I need self consistent data for radius of Si, C, N and H for calculating APF of say, a typical composition: SixCyNzHw, assuming Si, C, N and H have respectively, coordination numbers of 4, 4, 3 and 1. Can any of you please help?

[soori71]

Dear members,

If I sound like asking for data (of course, for any pointers to a reliable data set on atomic radii for calculating APF I would be highly thankful) and not scientific clarification, I will put my question differently:

How can I calculate the atomic packing factor for a compound of any given composition, provided density and chemical composition have been determined experimentally, thus molar volume is known? Note that possible compounds can fall within the triangular boundary whose vertices are pure ionic, covalent and metallic compounds. Ok, may be we need not immediately jump to the center of this triangle, but approach from the boundaries, say starting from the line connecting ionic and covalent vertices.

(The compositions of course would not be something out of wild, but practically realizable compositions, example a glass or a pyrolytic residue, where one can get an amorphous single phase of compositions, that are stoichiometrically far from thermodynamically stable crystalline compounds.)

The radii data from http://www.ccdc.cam.ac.uk/products/csd/radii/ also does not explain the Si-N bond length of 1.70 - 1.76 Angstroms.