[Pw_forum] question on ecutwfc and ecutrho

[Stefano de Gironcoli]

When using US pseudopotentials cutoff for rho is usually much largere 
than 4 times the cutoff for wfc.

ecutwfc and ecutrho are properties of the pseudopotentials not of the 
particular structure.

Choose a simple structure and compute etot vs ecutwfc keeping ecutrho at 
the default value.
look for convergence of total energy down to about a mryd/atom.
at this point the choice (ecutwfc, ecutrho=4ecutwfc) is a good set but 
probably ecutwfc could be chosen
smaller with no loss of accuracy.
fix ecutrho to the determined value and progressively reduce ecutwfc to 
see up to where you can go.

stefano
 From your data I cannot tell, it is usually better to change one 
parameter at a time.
Probably ecutrho=288 is fine  and and 32-35 is ok for ecutwfc while 24 
is not.

Xunlei Ding wrote:

>Dear all,
>
>I am now doing a calculation on a adsorption system.
>To determine the ecutwfc and ecutrho,
>I calculate the scf energies of a (3*2*5) slab with two adsorbed CO2.
>
>unit:ryd
>ecutwfc    ecutrho    total energy
>24           288      -2725.94125538              case1
>32           140      -2726.54980713              case2
>32           200      -2726.06221809              case3
>35           288      -2726.04364368              case4
>
>Question1: It is said that the ecutrho should be about 4*ecutwfc, but
>why the difference between case2 and case3 is so large?
>
>Question2: It the difference between case3 and case4 small? (The
>adsorption energy is less than 0.5eV. )
>
>Question3: Can I determine the suitable values for ecutwfc and ecutrho
>from these data?
>
>Thank you!
>
>Yours, 
>Ding
>
>
>_______________________________________________
>Pw_forum mailing list
>Pw_forum at pwscf.org
>http://www.democritos.it/mailman/listinfo/pw_forum
>  
>