[Wien] Deformation of Diamond

[Kurt Lejaeghere]

Citeren "Hui Wang" <hwang at imr.ac.cn>:

> Dear wien2k master:
>     I am running wien2k09.2 on Suse linux OS, with ifort 11 and mkl  
> 10, inter mpi.
>     Here, there are 3 problem confusing me:
>     (1)I wanna calculate the bulk modulus of Diamond. Diamond have  
> cubic sysmetry, and there are 8 atoms in unit cell, 2 atoms in  
> primitive cell. My question is: I wanna calculate the unit cell with  
> 8 atoms, I don't wanna calculate the primitive cell with 2 atoms,  
> because "x sgroup" will automatically change to primitive cell with  
> 2 atoms if I set up a unit cell with 8 atoms, so how can I calculate  
> the unit cell with 8 atoms ?

You can explicitly make the atoms inequivalent. If the struct-file for  
example says 'C ...', you can change this to C1, C2, ... (make sure  
the same amount of characters is on each line, so replace C + space by  
C + number). Sgroup will then be no problem anymore.

>     (2)I wanna calculate the elastic constant such as C11 of  
> Diamond, so I wanna deform the unit cell with 8 atoms in x axes, for  
> instance: lattice constant is a=b=c=3.56 Ang, I wanna calculate the  
> unit cell with 8 atoms when a=3.54, a=3.55, a=3.56, a=3.57, a=3.58  
> (b=c=3.56 all the time). When I calculate a=3.54, the sgroup will  
> find a new spacegroup "141 (I 41/a m d) [origin choice 2]", and "x  
> sgroup" find:
> =====================================
> Bravais lattice: Tetragonal body centred
>   a             b            c
>  4.73936977    4.73936977   6.72248100
>      alpha         beta         gamma
>  90.00000000    90.00000000   90.00000000
> =====================================
> I think this result is too ridiculus, far away from my requirment.  
> If I don't accept the "x sgroup" result, there are errors when "x  
> symetry" ends. So what shoud I do when I wanna deform Diamond along  
> x axes ?

I have less experience with this, but isn't it logical that if you  
change your cell only in one dimension, the space group will change  
(it will at least not be cubic anymore)?

>     (3)I used to use vasp to calculate bulk modulus of Diamond, when  
> I use Murnaghan fit formula or Mosrse fit formula to get the bulk  
> modulus, the energy unit is eV and lattice constant is Angstron.  
> Here my questions is: wien2k's energy unit is Ry and lattice  
> constant unit is Bohr, so do I need to convert Ry to eV and Angstron  
> to Bohr when I use Murnaghan fit and Morse fit ?

A routine 'eosfit' exists that makes use of the volumes in bohr^3 and  
the energies in Ry (more information can be found in the user's  
guide). It can e.g. be used for Murnaghan and Birch-Murnaghan fits.

With kind regards

Kurt Lejaeghere
Ghent University