[Wien] open shell case [Si and Ge]
Peter Blaha
pblaha at theochem.tuwien.ac.at
Fri Nov 3 14:47:39 CET 2017
I don't know what exactly you are doing, but certainly this is not correct.
The energy of a single free Si atom should be around -579.7 Ry (PBE),
not twice as large. So your energy is for 2 atoms ???
(You should NOT take the Si crystal structure (which contains 2 Si) and
increase the lattice parameters !!!! Instead, create a new case with a P
lattice and 12 Ang lattice parameter and a single Si atom at (0,0,0)
(and yes, an FCC cell with a sqrt2 larger lattice parameter would be
even more efficient, but forget it at the moment). Take 1k, TEMP 0.002;
If you just do x lstart for a Si atom, with the default case.inst
and with a "spin-compensated" inst file (see below)
Si
Ne 3
3,-1,1.0 N
3,-1,1.0 N
3, 1,1.0 N
3, 1,1.0 N
3,-2,0.0 N
3,-2,0.0 N
the energies in case.outputst are: -579.733781 and -579.677533,
respectable, and thus the spin-polarized solution is MUCH more stable.
It is also obvious from the eigenvalues: the difference between p and p*
eigenvalues (p3/2 and p1/2) is only 2 mRy, since Si is such a light
element. However, the difference between spin-up and dn eigenvalues is
more like 0.1 Ry, and this is what you see in the total energies.
If you now do
runsp -it (use iterative diagonalization, which is MUCH faster) you
obtain a spin-polarized solution with M=2 and and energy which is almost
identical to the one of lstart: -579.73486013
(in fact it is even lower, because my 12 Ang is too small and there is
still an attractive Si-Si interaction).
runsp_c -it then does it non-spinpolarized and the corresponding
energy is -579.67824905. Again a bit lower then the "correct" energy of
the free atom from lstart.
Next step would be to increase the 12 Ang to eg. 15 and repeat the
procedure until the energies are converged with respect to a0.
I'd expect that the runsp energy is a few mRy higher (less negative !)
then the one from lstart, because runsp neglects spin-orbit for 3p, i.e.
the 2 mRy I was mentioning earlier.
For heavier elements, the lstart energies become much lower than the
lapw-energies and thus cohesive energies would be inconsistent.
In any case, neglecting spin-polarization for Si "enhances" the
atomization energies by 0.8 eV / Si atom
--------------------------
Now we come to the O2 molecule: Why did you choose a B lattice ???
For simplicity choose a P lattice, but now make c by about 3 Ang longer
than a (guess why: you want to have similar distances between O2 and the
next one in x,y and z direction !).
PS: The most "efficient" calculations can always be done with a F-cell,
since this is the "closest" packing, which yields the best ratio between
distance-of-atoms and cell volume. But at that level of discussion I
recommend a P cell.
Calculations for free atoms or molecules will always be quite expensive.
But you gain significant time (factor 10 or so !!!) by using iterative
diagonalization: runsp -it
On 11/03/2017 11:38 AM, chin Sabsu wrote:
> Sorry, it was sent partly my mistake.
> I am repeating here:
>
>
>
>
>
>
> I could manage to calculate Si atomization energy for both sp and
> without -sp case.
>
> Si_atom_nsp Ene=-1157.44977178 Ry= -15747.912121276 eV
>
> Si_atom_sp Ene= -1157.44961225 Ry= -15747.909950759 eV
>
> So the Si should note be treated as open Shell case. Right?
>
>
> and the Cohesive Energy I obtained is: -8.037453341 eV per f.u.
>
>
>
> But I am not able to run O2 molecule.
>
> Below is my O2 structure.
>
>
> The optimization is still running for last 24 hrs (1 kpont). Only 19 scf
> cycles are completed (13+6)
> min_lapw -j "runsp_lapw -ec 0.00001 -cc 0.00001 -NI"
>
> Please advice me how to treate O2 calculations.
>
>
> I still ahve doubt why for O2 bulk we consider a=b=~10Ang and C=~15Ang
> and for O2 atomization energy a=b=c~12 to ~15 Ang.
>
>
> O2
>
> B LATTICE,NONEQUIV.ATOMS: 2 107
> I4mm
> MODE OF CALC=RELA
> unit=ang
> 21.213203 21.213203 40.000000 90.000000 90.000000
> 90.000000
> ATOM -1: X=0.00000000 Y=0.00000000 Z=0.00000000
> MULT= 1 ISPLIT=-2
> O 1 NPT= 781 R0=0.00010000 RMT= 1.1100 Z:
> 8.0
> LOCAL ROT MATRIX: 1.0000000 0.0000000 0.0000000
> 0.0000000 1.0000000 0.0000000
> 0.0000000 0.0000000 1.0000000
> ATOM -2: X=0.00000000 Y=0.00000000 Z=0.94293024
> MULT= 1 ISPLIT=-2
> O 2 NPT= 781 R0=0.00010000 RMT= 1.1100 Z:
> 8.0
> LOCAL ROT MATRIX: 1.0000000 0.0000000 0.0000000
> 0.0000000 1.0000000 0.0000000
> 0.0000000 0.0000000 1.0000000
> 8 NUMBER OF SYMMETRY OPERATIONS
> 1 0 0 0.00000000
> 0 1 0 0.00000000
> 0 0 1 0.00000000
> 1
> -1 0 0 0.00000000
> 0-1 0 0.00000000
> 0 0 1 0.00000000
> 2
> 0-1 0 0.00000000
> 1 0 0 0.00000000
> 0 0 1 0.00000000
> 3
> 0 1 0 0.00000000
> -1 0 0 0.00000000
> 0 0 1 0.00000000
> 4
> 1 0 0 0.00000000
> 0-1 0 0.00000000
> 0 0 1 0.00000000
> 5
> -1 0 0 0.00000000
> 0 1 0 0.00000000
> 0 0 1 0.00000000
> 6
> 0-1 0 0.00000000
> -1 0 0 0.00000000
> 0 0 1 0.00000000
> 7
> 0 1 0 0.00000000
> 1 0 0 0.00000000
> 0 0 1 0.00000000
> 8
>
>
>
>
>
> On Friday, 3 November 2017 2:23 PM, Stefaan Cottenier
> <Stefaan.Cottenier at UGent.be> wrote:
>
>
> Hello Gerhard,
>
> I'm reasoning by analogy now, not by deep understanding, but indeed, if
> you would apply Hund's rule to the relativistic p-shell, it looks to me
> it would be as you describe. The p-1/2 would then play the same role as
> an s-orbital for Hund's rules.
>
> Anyway, this is not relevant for wien2k free atom calculations because
> (1) Si and Ge are too light for the fully relativistic quantum numbers
> to apply, and (2) the calculations are done at the scalar-relativistic
> level.
>
> Best,
> Stefaan
>
>
>> -----Oorspronkelijk bericht-----
>> Van: Wien [mailto:wien-bounces at zeus.theochem.tuwien.ac.at
> <mailto:wien-bounces at zeus.theochem.tuwien.ac.at>] Namens
>> Fecher, Gerhard
>> Verzonden: vrijdag 3 november 2017 9:29
>> Aan: A Mailing list for WIEN2k users <wien at zeus.theochem.tuwien.ac.at
> <mailto:wien at zeus.theochem.tuwien.ac.at>>
>> Onderwerp: Re: [Wien] open shell case [Si and Ge]
>>
>> Dear Stefaan
>> What would happen if you treat all electrons relativistic, I would
> expect that
>> the two p valence electrons occupy the p1/2,+1/2 and p1/2,-1/2 states
>> leaving the 4 p3/2 states unoccupied and there would be no
> spinpolarisation
>>
>> I did not check, maybe I am wrong.
>>
>> Ciao
>> Gerhard
>>
>> DEEP THOUGHT in D. Adams; Hitchhikers Guide to the Galaxy:
>> "I think the problem, to be quite honest with you, is that you have never
>> actually known what the question is."
>>
>> ====================================
>> Dr. Gerhard H. Fecher
>> Institut of Inorganic and Analytical Chemistry Johannes Gutenberg -
>> University
>> 55099 Mainz
>> and
>> Max Planck Institute for Chemical Physics of Solids
>> 01187 Dresden
>> ________________________________________
>> Von: Wien [wien-bounces at zeus.theochem.tuwien.ac.at
> <mailto:wien-bounces at zeus.theochem.tuwien.ac.at>] im Auftrag von
>> Stefaan Cottenier [Stefaan.Cottenier at UGent.be
> <mailto:Stefaan.Cottenier at UGent.be>]
>> Gesendet: Freitag, 3. November 2017 06:52
>> An: A Mailing list for WIEN2k users
>> Betreff: Re: [Wien] open shell case [Si and Ge]
>>
>> Following Hund's rules, the atomic ground state for Si and Ge is spin-
>> polarized ('open shell'):
>>
>> https://www.webelements.com/silicon/atoms.html
>>
>> Hence, yes, you need -sp to find a meaningful value for the free atom
> total
>> energy.
>>
>> Stefaan
>>
>>
>> Van: Wien [mailto:wien-bounces at zeus.theochem.tuwien.ac.at
> <mailto:wien-bounces at zeus.theochem.tuwien.ac.at>] Namens
>> Gavin Abo
>> Verzonden: vrijdag 3 november 2017 5:42
>> Aan: wien at zeus.theochem.tuwien.ac.at
> <mailto:wien at zeus.theochem.tuwien.ac.at>
>> Onderwerp: Re: [Wien] open shell case [Si and Ge]
>>
>> Feel free to correct me if I'm wrong, but I think Si and Ge have an even
>> number of electrons (or paired electrons).
>>
>> For Si, 4 electrons in the 3s^2 3p^2. For Ge, 4 electrons in the 4s^2
> 4p^2. [1]
>>
>> This making them closed shell [2].
>>
>> In Prof. Blaha's example for free atoms [3], spin-polarized
> (runsp_lapw) is
>> used for open shell (or non-closed shell [4]) Li (1 electron in 2s^1)
> and B (3
>> electron in 2s^2 2p^1) and non-spin polarized is used for closed shell
> Be (2
>> electron in 2s^2).
>>
>> I don't know why it is fluctuating, diverging? Is the box you used
> containing
>> the free atom large enough? I remember that 12x12x12 angstrom usually
>> might be large enough [5].
>>
>> [1]
>> https://en.wikipedia.org/wiki/Electron_configurations_of_the_elements_(da
>> ta_page)
>> [2] https://www.mail- <https://www.mail-/>
>> archive.com/wien at zeus.theochem.tuwien.ac.at
> <mailto:wien at zeus.theochem.tuwien.ac.at>/msg11320.html
>> [3] https://www.mail- <https://www.mail-/>
>> archive.com/wien at zeus.theochem.tuwien.ac.at
> <mailto:wien at zeus.theochem.tuwien.ac.at>/msg16691.html
>> [4] http://susi.theochem.tuwien.ac.at/reg_user/faq/cohesive.html
>> [5] https://www.mail- <https://www.mail-/>
>> archive.com/wien at zeus.theochem.tuwien.ac.at
> <mailto:wien at zeus.theochem.tuwien.ac.at>/msg12815.html
>> On 11/2/2017 7:06 PM, chin Sabsu wrote:
>> Dear Peter Sir,
>>
>> Do Si [S^23P^4] and Ge [4S^24P^2] are also an open shell case?
>>
>> Because my Si atomization energy is fluctuating without -sp switch
> after 55
>> scf cycles also.
>>
>>
>>
>> Sincerely
>>
>> Chin
>> _______________________________________________
>> Wien mailing list
>> Wien at zeus.theochem.tuwien.ac.at <mailto:Wien at zeus.theochem.tuwien.ac.at>
>> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>> SEARCH the MAILING-LIST at: http://www.mail- <http://www.mail-/>
>> archive.com/wien at zeus.theochem.tuwien.ac.at
> <mailto:wien at zeus.theochem.tuwien.ac.at>/index.html
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at <mailto:Wien at zeus.theochem.tuwien.ac.at>
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>
>
>
>
>
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>
--
P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300 FAX: +43-1-58801-165982
Email: blaha at theochem.tuwien.ac.at WIEN2k: http://www.wien2k.at
WWW: http://www.imc.tuwien.ac.at/TC_Blaha
--------------------------------------------------------------------------
More information about the Wien
mailing list