[Wien] small organic molecules (12 atoms): ? convergence criteria and number of points

[Jorissen Kevin]

There's usually no need to change NPT.  Leave it at the default value of 781 unless you have a very specific reason for changing it.
K-points have to be converged; it's impossible to make a statement for your case without knowing the dimension of your unit (super) cell and its content.  If it's rather big, your k-list is probably sufficient for your purpose.
Rkmax 3 and 4 are generally meaningful only if you have H in the unit cell and therefore very small RMT, as is the case in your molecule, apparently.  So I guess this would not be a problem for your case.
As for convergence, even though you choose only one convergence criterium, it is recommended to make sure the other quantities behave 'reasonable'.  Sometimes either charge or energy appears to be converged very well while the other still fluctuates significantly (rare, but it happens).  In that case you should continue SCF before using the results.
If you go for charge, I'd say 0.01 is making life *too* easy for you :-).  I'd add at least one more '0' before doing any analysis (preferentially more than one, but you said you're doing only a preliminary calculation).
 
I'm not the pope; these are not rules carved in stone, but just guidelines.
 
 
Hope this helps,
 
 
Kevin Jorissen
 
EMAT - Electron Microscopy for Materials Science   (http://webhost.ua.ac.be/emat/)
Dept. of Physics
 
UA - Universiteit Antwerpen
Groenenborgerlaan 171
B-2020 Antwerpen
Belgium
 
tel  +32 3 2653249
fax + 32 3 2653257
e-mail kevin.jorissen at ua.ac.be
 

________________________________

Van: wien-admin at zeus.theochem.tuwien.ac.at namens Komolov
Verzonden: di 21-9-2004 14:56
Aan: wien at zeus.theochem.tuwien.ac.at
Onderwerp: [Wien] small organic molecules (12 atoms): ? convergence criteria and number of points


Dear all
I am calculating electronic structure of isolated benzene molecules (C6H6)
 
I have managed to run SCF cycles with the values of 
convergence paramerters: charge convergence 0.01 electron charge,
and NPT (struct gen) 61and
number of points for kgen 61
(RKMAX was 3 and 4 and 5)
 
Under these conditions covergence was reached after 10-15 cycles. These values seem to provide faster calculations, but I suspect they also serve for inaccuracy. 
 
I have read a recommendation on 381*2 NPT value for TiC example and a faq on that for larger supercell even 1 point for kgen may be enough. 
 
Please recommend:
Which one of the convergence parameters shall I choose for small organic molecules (such as benzene) and how much should it be?
How namy points in structgen /NPT as they are to be assigned for each atom/ shall I choose for this case?
How many points as an input for kgen shal I choose?
                 
/in order to make, let's say, a realistic calculation (not necessary a perfect one for now) 
 
----Thanks in advance-----
Alexei Komolov
 
 
 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/ms-tnef
Size: 7651 bytes
Desc: not available
Url : http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20040922/243a3d73/attachment.bin