[Wien] [Wien2k User Group] Lattice stability in bcc, fcc and hcp and other crystal structures

[Stefaan Cottenier]

> I assume U is in the hcp Zr ( we know the lattice parameters of Zr in hcp
> from ICSD) 
> The atomic positions are 0.3333333 0.6666667 and 0.25 
> So I imagine U to have the same lattice parameters and then I do min_lapw. 
> I optimize only the internal coordinates; I try to minimize the forces
> assuming my lattice parameters are same. 
> Can I make such an assumption? Please pardon me if I am doing any basic
> mistake in assuming this. 

This is more or less nonsense. You quoted Peter Blaha's hcp-paper. Check 
there how much variation there is in a and c lattice parameters across 
the periodic table. And now you are going to extrapolate from Zr (Z=40) 
to U (Z=92)...!?

Moreover, there is nothing to do for min_lapw in a hcp cell (you need 
positions with symmetry-undetermined freedom in the positions, i.e. 
'x'-values in the coordinates for that point group in the 
crystallographic tables).

> I am fairly a new user of Wien2k and I would like
> to know what would be the best computational strategy. 
> 1. min_lapw only 
> 2. volume optimization only 
> 3. both 
> 
> If I go for 1 alone, I assume that my lattice parameters are optimized 
> If I go for 2 alone, I need not do min_lapw 
> If I go for 3, it would take a large amount of computational time and
> effort. 
> 
> My limited knowledge in Wien2k makes me go for 1 only that is going for
> min_lapw. 
> Since these phases are complicated, especially beta-U, which have 8
> nonequivalent positions, I cannot go for 3. 
> Is my strategy ok? 

Option 3 is definitely the best. If you can't afford that, take option 
2. Option 1 is only useful if you have at least a good guess about the 
cell volume and shape (from experiment or from other calculations). The 
influence of cell volume and shape (lattice parameters and angles) is 
much larger than the influence of internal coordinates.

When you select the problem you want to solve, it is wise to select one 
that can be solved with the resources you have available...

Stefaan