Convergence of the individual components of the stress tensor
Posted: Tue Oct 30, 2012 8:57 am
Dear all,
I'm trying to optimize a cell under external pressure, setting the PSTRESS tag to values ranging from 0 to 15 kBa.
I noticed, that although the total pressure, i.e., the average of the xx, yy and zz components appears to be converged, the individual components seems to be more scattered. For example:
PSTRESS = 3 and EDIFFG = -0.02 results in the following output:
xx yy zz: 2.86453 3.10009 2.89908
and
PSTRESS = 3 and EDIFFG = -0.00001 :
xx yy zz: 3.09167 3.33564 3.09940
The cutoff energy is large enough. Could k-point sampling be insufficient?
Thank you in advance,
Kind regards
Monica
I'm trying to optimize a cell under external pressure, setting the PSTRESS tag to values ranging from 0 to 15 kBa.
I noticed, that although the total pressure, i.e., the average of the xx, yy and zz components appears to be converged, the individual components seems to be more scattered. For example:
PSTRESS = 3 and EDIFFG = -0.02 results in the following output:
xx yy zz: 2.86453 3.10009 2.89908
and
PSTRESS = 3 and EDIFFG = -0.00001 :
xx yy zz: 3.09167 3.33564 3.09940
The cutoff energy is large enough. Could k-point sampling be insufficient?
Thank you in advance,
Kind regards
Monica