[Wien] case.inst file -- Europium

[Stefaan Cottenier]

Anurag Chaudhry wrote:
> Thank you Dr. Cottenier,
> 
> I followed your suggestion and could plot 5d states as well as 4f for 
> Europium.
> 
> I would like to now proceed to do a constrained calculation as 
> prescribed in Madsen & Novak notes to determine the effective U 
> parameter for Europium 4f states.  In this regard I have a question.
> 
> The SrI2_Eu.inc (for Eu doped SrI2) file says something like the 
> following for Eu
> 
> 14 0.00     NUMBER OF ORBITALS (EXCLUDING SPIN), SHIFT
> 1,-1,2               ( N,KAPPA,OCCUP)
> 2,-1,2               ( N,KAPPA,OCCUP)
> 2, 1,2               ( N,KAPPA,OCCUP)
> 2,-2,4               ( N,KAPPA,OCCUP)
> 3,-1,2               ( N,KAPPA,OCCUP)
> 3, 1,2               ( N,KAPPA,OCCUP)
> 3,-2,4               ( N,KAPPA,OCCUP)
> 3, 2,4               ( N,KAPPA,OCCUP)
> 3,-3,6               ( N,KAPPA,OCCUP)
> 4,-1,2               ( N,KAPPA,OCCUP)
> 4, 1,2               ( N,KAPPA,OCCUP)
> 4,-2,4               ( N,KAPPA,OCCUP)
> 4, 2,4               ( N,KAPPA,OCCUP)
> 4,-3,6               ( N,KAPPA,OCCUP)
> .....
> while the SrI2_Eu.inst file has the following information --
> 
> Eu
> Xe 4
> 4, 3,3.0  N
> 4, 3,0.0  N
> 4,-4,3.0  N
> 4,-4,0.0  N
> 5, 2,1.0  N
> 5, 2,0.0  N
> 6,-1,1.0  N
> 6,-1,1.0  N
> .....
> 
> According to the Madsen & Novak notes I should include the 4f states in 
> the core. And they recommend doing two calculations - one with 4 up 3.5 
> dn and another with 4 up 2.5 dn constrained f electrons.  Is this right 
> for Europium ??

I leave this for the Madsen & Novak to comment upon (is the fact that 
you go over half a shell here an issue? If it is, then do these 
calculations for a slightly lighter lanthanide -- U should not depend 
too much on it).

However, I remember to have tried such constrained calculations for 
lanthanides a few years ago, and obtained a U that was only half as 
large as would be meaningful. I'm not sure whether these constrained 
calculations are appropriate for f elements.

> For Ni in NiO the notes indicate the headers of the files should look like
> 
> NiO.incup_+.50: NiO.incdn_+.50 
> 
> 7 1.00          7 1.00 
> 
> 1,-1,2          1,-1,2 
> 
> 2,-1,2          2,-1,2 
> 
> 2, 1,2          2, 1,2 
> 
> 2,-2,4          2,-2,4 
> 
> 3,-1,2          3,-1,2 
> 
> 3, 2,4          3, 2,4   <= line added 
> 
> 3,-3,5          3,-3,4      <= line added
> 
> 
> for a calculation with 4.5 up 4 dn electrons.  The notes mention the 
> above changes to the files for the case of 8.5 constrainted d electrons 
> but I am not able to understand how this is achieved ??  To me it 
> appears that we have 9 spin-up and 8 spin-dn electrons.

The case.inc file has always all electrons (up+dn), look for instance at 
the earlier lines with 2 1s electrons, 6 2p electrons, etc. When it is 
split in a case.incup and case.incdn, the number of electrons in the 
d-shell is made such that after a division by 2 the desired number is 
obtained.

Stefaan


> 
> 
> Thank you very much for prompt and precise replies earlier.
> 
> 
> regards,
> 
> Anurag
> 
> 
> 
> 
> 
> ------------------------------------------------------------------------
> *From:* Stefaan Cottenier <stefaan.cottenier at ugent.be>
> *To:* Anurag Chaudhry <anuragchaudhry at yahoo.co.uk>
> *Sent:* Sunday, 7 December, 2008 23:53:19
> *Subject:* Re: [Wien] case.inst file -- Europium
> 
> 
> Have a look at the section on tetra in the usersguide. If you put a line 
> in case.int which corresponds to the 5d states (how this line looks like 
> can only be decided based on the case.qtlup/dn file for your compound, 
> see usersguide), then the 5d DOS will be given as one of columns in 
> case.dos*. It's integral is given in case.outputtup/dn as well. 
> Inspecting the latter for E=E_fermi gives you the 5d occupation. It will 
> never be really empty, for Eu inside a solid.
> 
> Stefaan
> 
> Citeren Anurag Chaudhry <anuragchaudhry at yahoo.co.uk 
> <mailto:anuragchaudhry at yahoo.co.uk>>:
> 
>  > Hello Dr. Cottenier,
>  >
>  > Thanks for your answer.  I have a follow up question now that I  
> managed to run a spin polarized calculation for Eu doped SrI2.  I  can 
> see from the DOS the Eu 4f states (spin up and spin down) but is  there 
> an output file which mentions about the occupation of orbitals  after 
> the SCF run ?  I wanted to check that the 5d states of Eu are  empty.
>  >
>  > In the file case.int which is the input for TETRA how do I specify  
> if I have to plot Eu 4f states ?
>  >
>  > Thank you.
>  >
>  > regards,
>  > Anurag
>  >
>  >
>  >
>  > ________________________________
>  > From: Stefaan Cottenier <stefaan.cottenier at ugent.be 
> <mailto:stefaan.cottenier at ugent.be>>
>  > To: wien at zeus.theochem.tuwien.ac.at 
> <mailto:wien at zeus.theochem.tuwien.ac.at>
>  > Sent: Wednesday, 3 December, 2008 22:32:21
>  > Subject: Re: [Wien] case.inst file -- Europium
>  >
>  >
>  >> From the periodic table I see that Eu has an electronic
>  >> configuration of [Xe]4f76s2  so I do not understand why do see the
>  >> 5d orbitals having a single electron whereas there are only 6
>  >> electrons in the 4f.  can someone please explain ?
>  >
>  > The occupation in case.inst need not to agree with the free atom
>  > configuration. It is only meant to give a reasonable starting value to
>  > lstart. The scf cycle will later take care of where exactly to put the
>  > electrons. Some free configurations (even when they are the ones that
>  > experimentally appear) make lstart crash. Therefore, the default
>  > inputs have been verified not to lead to a crash.
>  >
>  > Stefaan