[Wien] Problem creating struct files from VASP contcar files

[B. Yanchitsky]

> I'm new to using wien 2k, so I hope I'm not missing a simple solution to this.
>  I'm trying to convert VASP contcar files for low symmetry structures, like
> monoclinic, into struct files.  A sample contcar file is pasted below.
> 
> GSS check structure #: 81
>     4.100000000000000
>      1.0062816128952770    0.5260273706595530    0.5121801204340521
>     -0.4948479628567789    1.0535466565671180    0.4948479628567789
>     -0.4941014924612246   -1.0520547413191060    0.4941014924612247
>    6   3
> Direct
>   0.3333333333333428  0.7607588622281583  0.8969830898152509
>   0.6694622253955771  0.7390086361365178  0.5980793830199076
>   0.9972044412710943  0.7390086361365178  0.2619504909576591
>   0.6694622253955771  0.2609913638634822  0.0713828423756695
>   0.9972044412710943  0.2609913638634822  0.7352539503134352
>   0.3333333333333428  0.2392411377718417  0.4363502435180777
>   0.3333333333333428  0.0000000000000000  0.6666666666666572
>   0.6680618367231830  0.5000000000000000  0.8340309183616057
>   0.9986048299434884  0.5000000000000000  0.4993024149717371
> 
> The first line (after the title) is the scale factor, the next three lines
> define the basis vectors, and the nine atom positions are listed in direct
> coordinates.  Using StructGen, I entered a=5.105, b=5.187, c=5.178 and
> alpha=112.8, beta=120.5, gamma=78.6.  The atom positions were entered in
> direct coords as above.
> 
> When I run init_lapw it keeps identifying it as a triclinic structure.  I have
> also tried placing a single atom at 0,0,0 to see if it would show up as
> monoclinic, but it doesn't.
> 

Well, sgroup does perform a correct job, only you need to specify all
the numerical values with maximal possible precision. The value for a is
5.1071519 (not 5.105) the value for b is 5.1856647 (not 5.187)
for c is 5.1781746, alpha, beta and gamma are 112.81407 120.46110 78.66372.
On these numbers and single atom at (0,0,0) sgroup is able to
find monoclinic base centred cell, so, what you need is accurately 
specify all the atomic positions and parameters of the unit cell.

> The atom positions were entered in direct coords as above.

Probably that's wrong, because the positions must be specified as fractions
of the lattice vectors (not cartesian coordinates), transformation
is needed.

Regards,
Bogdan Yanchitsky