I am running a project with YPO4. With default parameters (including RMTs and LOs) the calculated charge error was about 0.18e out of total (correct) 172e - which is a lot.
So, I went to the forum, read the posts on RMTs of Si, found no posts on phosphates, messed around with LO's, lo's, LAPW, number of points, rgkmax, gmaxvr, none of which helped noteworthy. Finally, I optimized the RMTs in a primitive steepest-descent style and got an improvement: the error got reduced to about 0.008, confirming the idea that RMTs must be tested for convergence/optimized.
The next plateau was with the further improvement. I need it for a larger system, and this phosphate is a mere test case. I tried rgkmax, lmaxi, and eventually lmaxo and lmaxapw. The best result so far 0.002 with lmaxo = lmaxapw = 6 and lmaxi = 3, albeit the latter likely makes almost no effect.
So, I just leave it here for the Elk people of the future. The question is: why is the default lmaxapw so large? Shouldn't this cause the problem where the higher-l plane waves do not have anything to stick to at muffin-tin surface, due to lmaxo not being enough?
Anyway, the input is below, maybe someone would like to mess around more.
To improve on this, you may need to promote some core electrons to the valence, but I think the error is negligible for most purposes.
lmaxpaw is large (8 by default) in order to converge the matching of the orbitals at the muffin-tin boundary. lmaxo can be smaller than this because the density is averaged over all orbitals and is thus more 'spherical'.
Regards,
Kay.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear Users,
I am running a project with YPO4. With default parameters (including RMTs and LOs) the calculated charge error was about 0.18e out of total (correct) 172e - which is a lot.
So, I went to the forum, read the posts on RMTs of Si, found no posts on phosphates, messed around with LO's, lo's, LAPW, number of points, rgkmax, gmaxvr, none of which helped noteworthy. Finally, I optimized the RMTs in a primitive steepest-descent style and got an improvement: the error got reduced to about 0.008, confirming the idea that RMTs must be tested for convergence/optimized.
The next plateau was with the further improvement. I need it for a larger system, and this phosphate is a mere test case. I tried rgkmax, lmaxi, and eventually lmaxo and lmaxapw. The best result so far 0.002 with lmaxo = lmaxapw = 6 and lmaxi = 3, albeit the latter likely makes almost no effect.
So, I just leave it here for the Elk people of the future. The question is: why is the default lmaxapw so large? Shouldn't this cause the problem where the higher-l plane waves do not have anything to stick to at muffin-tin surface, due to lmaxo not being enough?
Anyway, the input is below, maybe someone would like to mess around more.
Dear Andrew,
I used the following input file:
...which gave a charge error of 0.04 e.
To improve on this, you may need to promote some core electrons to the valence, but I think the error is negligible for most purposes.
lmaxpaw is large (8 by default) in order to converge the matching of the orbitals at the muffin-tin boundary. lmaxo can be smaller than this because the density is averaged over all orbitals and is thus more 'spherical'.
Regards,
Kay.