What does covalent bonding in salts actually look like?

[Tiiba]

I learned some time ago that in table salt, the valence electron spends only about 2/3 of its life in chlorine, and goes home on the weekends. This still feels weird to me, because every representation of an ionic compound shows the ions as completely separate.

So what does the covalent bonding actually look like? Does a sodium share its electron with one specific chloride, or with every chloride in its vicinity? Chlorides and sodiums? And what if it's something complex, like ammonium nitrate?

(Edited, because I misremembered.)

 

[hilbert2]

A simple model of a perfectly ionic crystal could be made by filling lattice positions with spheres of constant charge density and then calculating the electric potential energy from Coulomb's law. This kind of model doesn't result in as low an energy as the actual binding energy of the crystal, so it's deduced that the bonds are only partially ionic.

 

[DrDu]

Btw, the decomposition of the total binding into ionic and covalent parts is non-unique and highly non-trivial. Where does one atom begin and where does the other end? One of the best solutions to this problem is due to Bader with his atoms in molecules theory:
http://www.chemistry.mcmaster.ca/esam/Chapter_7/intro.html
So it appears that highly ionic compounds like LiF are really nearly 100% ionic.