[Wien] FW: FW: [Wien2k Users] ENE Warning

[Stefaan Cottenier]

> I am considering f-states in case of actinides like U and Pu. 

For U and Pu, the 5f states are above -0.5 Ry in the free atom. They are 
automatically treated as valence states, and it makes no sense not to do 
so. The 4f states are at -25 to -30 Ry. They are excellent core states, 
and it makes no sense to treat them as valence states.

Stefaan


> -----Original Message-----
> From: wien-bounces at zeus.theochem.tuwien.ac.at
> [mailto:wien-bounces at zeus.theochem.tuwien.ac.at] On Behalf Of Stefaan
> Cottenier
> Sent: Friday, December 18, 2009 1:47 PM
> To: A Mailing list for WIEN2k users
> Subject: Re: [Wien] FW: FW: [Wien2k Users] ENE Warning
> 
> 
>> I do have f-states in my case. 
>> I want to know if it is advisable to take these f-states in the valence
> and
>> do the calculation. 
>> Would it be more accurate? 
>> Moreover, during lstart, we give -6.0 Ry as the separation energy between
>> core and valence. 
>>
>> If we put -4.5 Ry, then does it mean that we include f-states in the
> valence
>> as well or when we put -8.0 Ry?
> 
> No-one can answer this without seeing your case.outputst, or at least 
> knowing which element with f-electrons you consider.
> 
>> I am not studying the system under high pressure. 
>> I guess you have come to this conclusion by seeing RKmax (Please mention
> if
>> I am wrong) 
>> Rkmax is 9.50 in my case. 
> 
> No, because you had f linearization energies of 1.6 Ry. This might 
> indicate a very high fermi energy, which could be due to either a wrong 
> case.struct or to high pressure.
> 
>> Actually, I would like to be more accurate in the total energy till the
>> second decimal place. So I kept on changing the RKmax until it gave the
> same
>> total energy till the second decimal place. 
> 
> Not useful, as the accuracy in the total energy itself is not needed for 
> almost all properties. What you need is accuracy in total energy 
> *differences*. Repeat your calculations for a slightly different lattice 
> constant, and observe at which rkmax the energy difference between both 
> cases converges to the precision you want. This will happen at a much 
> smaller rkmax than your present one.
> 
> Stefaan
> 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien


-- 
Stefaan Cottenier
Center for Molecular Modeling (CMM)
Ghent University
Technologiepark 903 (2nd floor)
BE-9052 Zwijnaarde
Belgium

http://molmod.Ugent.be
email: Stefaan . Cottenier /at/ UGent . be