Cubic or tetragonal/orthorhombic supercell

Alloy Theoretic Automated Toolkit (ATAT)
1

Hi,

Below is an example of the coordinate system and lattice vectors of "bestsqs8.out" for 100 atoms of a bcc system:

3.600000 0.000000 0.000000
0.000000 3.600000 0.000000
0.000000 0.000000 3.600000
-0.500000 1.500000 1.500000
-0.500000 2.500000 -2.500000
-6.500000 -0.500000 -0.500000

The resultant supercell (for vasp) is not cubic. (it seems to be orthorhombic).

Now is there anyway to instruct mcsqs to generate a cubic supercell?
If not, then what is the best k-points selection for it?
and finally what is the best constant for (no. of kpoints)x(no. atoms)?

Thanks

2

Dear Axel,
I have the same question. I need a SQS alloy that has the similiar lattice vectors with its parent lattice (rndstr.in). For instant, a cubic structure in rndstr.in,

5.409300 0.000000 0.000000
0.000000 5.409300 0.000000
0.000000 0.000000 5.409300
0.500000 0.500000 0.000000
0.000000 0.500000 0.500000
0.500000 0.000000 0.500000
0.250000 0.250000 0.250000 O
0.000000 0.000000 0.000000 Ca=0.25, Zn=0.75

I expect to produce a SQS (bestsqs.out) with similiar lattice vectors with rndstr.in, as follows,

5.409300 0.000000 0.000000
0.000000 5.409300 0.000000
0.000000 0.000000 5.409300
2.000000 0.000000 0.000000
0.000000 2.000000 0.000000
0.000000 0.000000 2.000000
0.250000 0.250000 0.250000 O
0.000000 0.000000 0.000000 Ca
… …
… …

However, I usually get a strange lattice vectors. Maybe, is there a command to constrain lattice vectors so that keeping same shape with parent lattice (rndstr.in) ?

Many thanks,

Li

3

The file sqscell.out holds all the supercells the mcsqs will go through while running. Make one by hand containing the supercells you want it to go through and run mcsqs with an added -rc argument. With the -rc argument the mcsqs skips the generation of sqscell.out, and uses the file you’ve provided instead.

4

Thanks very much. It works well and I could find a few orthogonal cells such as tetragonal and orthorhombic. However, I could not find any cubic among them.
I chose the tetragonal one.
The reason I am looking for cubic supercell is that it is much easier to calculate elastic constants when the supercell is cubic (only 3 independent elastic constants).

Or since the symmetry of the alloy to be considered is BCC, it does not matter if the supercell is tetragonal? and I can deform the system with three types of strains and fine the three independent elastic constants?

Thanks

5

Actually the sqs will rarely have exact cubic symmetry even if the underlying lattice is cubic.
It is not a problem, however.
The trick is to symmetrize the resulting elastic constants, see:
https://dx.doi.org/10.1103/PhysRevB.72.144109
or
https://dx.doi.org/10.1016/j.calphad.2013.06.006
and references cited herein.