My Community

Hi all,

I am currently trying to perform a standard two-step relaxation of a 45-atom SQS alloy. The convergence is extremely slow, and also produces a lot of inaccurate results. The forces in the structure after relaxation are large. At this time, I am guessing that the initial structure is very far from the equilibrium structure, which is why convergence is difficult. Does anyone have any tips on tags to use for difficult relaxations? Or just tips in general for relaxing SQS structures?

Thank you in advance.

Hi,

Thanks for reaching out. Without further information, it is very difficult to assess what is going on in your calculation. Therefore, I would ask you to upload the input and output files according to our forum posting guidelines. Based on your description, this could be related to Pulay stress but it's hard to know for sure.

Here are some hints for hard to converge systems:

1 ) First check whether you need to use spin-polarized calculations (ISPIN=2). For that look at the configuration of the valence states POTCAR file (V_RHFIN).
2 ) Check whether your energy cut-off ENCUT is large enough (as Manuel wrote this has to do with the Pulay stress - be careful VASP writes "Pullay" instead of Pulay).
3 ) Check if you have enough k-points.
4 ) Try ALGO=ALL for problematic systems.
5 ) Set LASPH=.TRUE. to ensure that the non-spherical contributions from PAW spheres are added correctly.
6 ) For PAW potentials set LMAXMIX=6 to ensure correct mixing of the high-l components.
7 ) Reduce the mixing parameters:
AMIX = 0.2
BMIX = 0.0001
AMIX_MAG = 0.8
BMIX_MAG = 0.0001
8 ) The previous step maybe only needs to be done until a pre-convergence is reached. After that proceed with the default mixing tags and set ISTART=1 to make a continuation calculation. Do not forget to copy the WAVECAR file.

Please try the above things before you send us any files.