Once you take gravity and the electrons out of it, the only thing governing these structures really is the ordinary Coulomb electrical force – electromagnetism (and quantum mechanics) is behind essentially all solid-state physics and chemistry, but this is an even simpler situation that removes the quantum effects of the electrons and most of the dynamics. Actually, back when I was a practicing physicist, I wrote a paper that extended the usual approach for calculating these static energies to a new class of materials, so this all brings back memories :-)

Anyway, it’s reasonably well known that the best crystal structure, as far as the electrostatic energy goes, is body-centered cubic – just slightly higher Madelung constant than the close-packed structures (face-centered cubic and hexagonal close packed). So this isn’t exactly a surprise, but it is a very nice clean demonstration of how it should work (don’t have to worry about all those pesky electrons and their other bonding-related effects). And presumably this setup should also be able to investigate the dynamics and the liquid-solid transition (the article mentions some of this): given that we know exactly what the forces are, this is an excellent testing-ground for matching up theory and experiment. There’s certainly been a lot of theoretical studies of one-component plasmas of this sort, but nothing beats a real experiment!