[Wien] ELF
fabien.tran at vasp.at
fabien.tran at vasp.at
Fri Nov 4 22:20:46 CET 2022
It is on my to-do list to reimplement ELF in VASP in a more proper way.
In any case, it seems that ELF was buggy in a certain number of codes
(WIEN2k, VASP, Quantum Espresso).
On 04.11.2022 22:06, Kateryna Foyevtsova wrote:
> Hello,
>
> since I see that VASP results are being discussed here, I'd like to
> bring your attention to my communication with the VASP developers in
> April this year:
>
> https://www.vasp.at/forum/viewtopic.php?f=3&t=18484
>
> where they admitted that "the current implementation depends strongly
> on the choice of the POTCAR. You can test this by using the GW POTCAR
> and will find contributions from Ni 3d electrons. We believe that with
> this pseudo-potential dependent ELF is not a good measure for
> electronic localisation and thus want to replace this (very old)
> implementation in a future release."
>
> which means that ELF calculated in the current version (and all
> earlier versions, apparently) of VASP can have errors and should not
> be used as a gold standard. I think they still have not fixed that bug
> according to
>
> https://www.vasp.at/forum/viewtopic.php?f=4&t=18628&p=22510&hilit=elf#p22510
>
> Best,
> Kateryna
>
>
> On 2022-11-04 13:11, reyhaneh ebrahimi wrote:
>> [CAUTION: Non-UBC Email]
>>
>> Dear Prof. Blaha
>> Thank you for your valuable answer to my Email.
>> I put my ELF graph, your ELF results, and Jiawang and Olivier' graph
>> for SnSe on one page to have a better comparison, see
>> "https://www.mediafire.com/file/kyfi46ppx6mhtnx/SnSe-final.jpg/file".
>> I also specified the plane which I did my ELF calculation on it.
>> Therefore, as you mentioned in your Email, the existing differences
>> between my graph, your graph, and Jiawang and Olivier' graph would be
>> due to the choice of different planes in these works.
>> Sincerely yours
>> Reyhaneh Ebrahimi
>>
>> On Fri, Nov 4, 2022 at 8:33 AM Peter Blaha <peter.blaha at tuwien.ac.at>
>> wrote:
>>
>>> Sorry: the links should be: SnSe, not SnGe
>>>
>>> http://www.wien2k.at/Depository/SnSe-f.png
>>> http://www.wien2k.at/Depository/SnSe-g.jpg
>>> http://www.wien2k.at/Depository/SnSe-t.png
>>>
>>> Am 04.11.2022 um 15:27 schrieb Peter Blaha:
>>>> Your picture for SnSe is probably in a different plane as compared
>>> to
>>>> the 4 pictures in the paper.
>>>> I produced 2 elf pictures, which resembles the planes in Fig. 6f
>>> and 6g.
>>>>
>>>> They look as expected. In the interstital identical (see eg. the 2
>>>
>>>> different blue features in 6f), but inside the spheres quite
>>> different
>>>> because of the pseudopotentials.
>>>>
>>>> I guess it is a general feature that inside spheres (and for
>>> heavier
>>>> atoms) the PP ELF is nonsense.
>>>>
>>>> You can download them at:
>>>>
>>>> http://www.wien2k.at/Depository/SnGe-f.png
>>>> http://www.wien2k.at/Depository/SnGe-g.jpg
>>>> http://www.wien2k.at/Depository/SnGe-t.png
>>>>
>>>>
>>>> Am 04.11.2022 um 00:13 schrieb reyhaneh ebrahimi:
>>>>> Dear Prof. Blaha
>>>>> I apologize. Let me make my previous Email a little more
>>> complete. As
>>>>> you mentioned in your Email, for SnS the sources of differences
>>>>> between the results of ELF using WIEN2k code and VASP code is due
>>> to
>>>>> the difference between all-electron and pseudopotentials
>>> calculations
>>>>> in these two codes. However, I am still not sure that the
>>> differences
>>>>> between my results for the ELF of SnSe and Jiawang and Olivier's
>>> paper
>>>>> are normal or not. I would be glad if you let me know your
>>> opinion
>>>>> about this subject.
>>>>> Sincerely yours
>>>>> Reyhaneh Ebrahimi
>>>>>
>>>>> On Thu, Nov 3, 2022 at 1:13 PM Peter Blaha
>>> <peter.blaha at tuwien.ac.at
>>>>> <mailto:peter.blaha at tuwien.ac.at>> wrote:
>>>>>
>>>>> Good to hear that this has been resolved.
>>>>>
>>>>> PS: I just did a SnSe calc. and compared with the VASP paper.
>>>>> Similarly,
>>>>> very good agreement in the interstitial, while inside the
>>> atomic
>>>>> cores
>>>>> there is the expected difference between all-electron and
>>>>> pseudopotentials.
>>>>>
>>>>> Am 03.11.2022 um 21:06 schrieb Kateryna Foyevtsova:
>>>>>> Dear Prof. Blaha,
>>>>>>
>>>>>> I think I know what's going on with ELF. Wien2k gets it
>>>>> correctly, but
>>>>>> Quantum Espresso has a bug which shows up in nspin=1
>>>>> calculations. In
>>>>>> the attached figure I compare the wien2k result with two
>>> QE
>>>>>> calculations: (1) one with nspin=1 switch and (2) one with
>>> nspin=2
>>>>>> switch. In both cases I am looking at the same
>>> non-magnetic
>>>>> solution
>>>>>> that has the same energy in the two QE calculations.
>>>>>>
>>>>>> Now you see that the difference between QE nspin=1 and
>>> nspin=2 is
>>>>>> dramatic whereas there should be none.
>>>>>>
>>>>>> The wien2k result looks very similar to the QE nspin=2
>>> result
>>>>> in the
>>>>>> interstitial region at 0.5,0.5,0.0, marked with a big
>>> purple "X".
>>>>> There
>>>>>> are differences close to atomic nuclei but this is
>>> expected given
>>>>> that
>>>>>> we are comparing an all-electron and a pseudo-potenial
>>> code.
>>>>>>
>>>>>> Thank you very much for helping me resolve this issue.
>>>>>>
>>>>>> Best,
>>>>>> Kateryna
>>>>>>
>>>>>> On 2022-11-02 12:21, Peter Blaha wrote:
>>>>>>> [CAUTION: Non-UBC Email]
>>>>>>>
>>>>>>> My result looks like the attached picture. I do get 0.8
>>> in the
>>>>> core
>>>>>>> region of Ni, but not larger than that. It is probably
>>> similar
>>>>> than
>>>>>>> yours.
>>>>>>> I have no idea why it is different from QE, except maybe
>>> that
>>>>> these
>>>>>>> are pseudopotential calc.
>>>>>>>
>>>>>>> As I said before, you should compare other compounds, and
>>> also
>>>>> compare
>>>>>>> with literature ELF calculations.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Am 01.11.2022 um 21:16 schrieb Kateryna Foyevtsova:
>>>>>>>> Dear Prof. Blaha,
>>>>>>>>
>>>>>>>> thank you for looking into this issue. I've tried the
>>> modified
>>>>>>>> create_rho.f and calculated the ELF of NdNiO2 again
>>> using
>>>>> create_elf.
>>>>>>>> I am getting a better agreement with QE, but it is not
>>> perfect
>>>>> as you
>>>>>>>> noted it too. My calculation was well converged and I
>>> used the
>>>>> same
>>>>>>>> k-grid and RKmax=7. The bandstructures from QE and
>>> wien2k agree
>>>>> very
>>>>>>>> well.
>>>>>>>>
>>>>>>>> I attach my comparison as a png file. I wonder whether
>>> you have
>>>>> any
>>>>>>>> idea about the possible reasons for the differences in
>>> ELF that
>>>>> the
>>>>>>>> two codes give? For example, at 0.5,0.5,0 the wien2k
>>> value is
>>>>> ~0.22
>>>>>>>> and the QE value is ~0.43.
>>>>>>>>
>>>>>>>> Thank you,
>>>>>>>> Kateryna
>>>>>>>>
>>>>>>>> On 2022-10-28 04:43, Peter Blaha wrote:
>>>>>>>>> [CAUTION: Non-UBC Email]
>>>>>>>>>
>>>>>>>>> Dear Kateryna ,
>>>>>>>>>
>>>>>>>>> In fact, I found a big difference between
>>> create_elf and
>>>>>>>>> x lapw0 (with VX_ELF); x lapw5 -exchange
>>>>>>>>>
>>>>>>>>> I traced it back to normalization errors in tau_w and
>>> tau_tf,
>>>>> which
>>>>>>>>> missed a factor of 2.
>>>>>>>>>
>>>>>>>>> The attached create_rho.f fixes the problem. It
>>> should be
>>>>> copied
>>>>>>>>> into SRC_trig; make
>>>>>>>>>
>>>>>>>>> Then you can use create_elf again.
>>>>>>>>>
>>>>>>>>> PS: I would always compare the ELF created with both
>>> methods as
>>>>>>>>> indicated above. Depending on the numerics, one or the
>>> other
>>>>> method
>>>>>>>>> may give smoother plots, but in any case, they should
>>> be very
>>>>> similar.
>>>>>>>>>
>>>>>>>>> PPS: The agreement to QE-ELF seems reasonable (but not
>>>>> perfect), but
>>>>>>>>> I've not converged my calculations.
>>>>>>>>>
>>>>>>>>> Thanks for the report
>>>>>>>>> Peter Blaha
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> I attach a pdf showing the differences. Also attached
>>> are my
>>>>> wien2k
>>>>>>>>>> >struct file and quantum espresso input file.
>>>>>>>>>
>>>>>>>>>> Both calculations were done without spin polarization
>>> and
>>>>> using PBE.
>>>>>>>>>
>>>>>>>>>> To me, the differences are big enough to question
>>> whether
>>>>> it is
>>>>>>>>>> >meaningful to use ELF at all if it depends on
>>> all-electron vs
>>>>>>>>>> >pseudopotential so strongly. Unless I am missing
>>> something or
>>>>>>>>>> doing >something wrong.
>>>>>>>>>
>>>>>>>>>> Thank you,
>>>>>>>>>> Kateryna
>>>>>>>>>
>>>>>>>>> _______________________________________________
>>>>>>>>> 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
>>>>> <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
>>>
>>>>>
>>>
>> <http://www.mail-archive.com/wien@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
>>>>> <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
>>>
>>>>>
>>>
>> <http://www.mail-archive.com/wien@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
>>>>> <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
>>>
>>>>>
>>>
>> <http://www.mail-archive.com/wien@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
>>>>> <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
>>>
>>>>>
>>>
>> <http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html>
>>>>>
>>>>> --
>>>>>
>>>
>> --------------------------------------------------------------------------
>>>>> Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060
>>> Vienna
>>>>> Phone: +43-1-58801-165300
>>>>> Email: peter.blaha at tuwien.ac.at
>>> <mailto:peter.blaha at tuwien.ac.at>
>>>>> WIEN2k: http://www.wien2k.at <http://www.wien2k.at>
>>>>> WWW: http://www.imc.tuwien.ac.at
>>> <http://www.imc.tuwien.ac.at>
>>>>>
>>>>>
>>>
>> -------------------------------------------------------------------------
>>>>> _______________________________________________
>>>>> 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
>>>>> <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
>>>
>>>>>
>>>
>> <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
>>>>
>>>
>>> --
>>>
>> --------------------------------------------------------------------------
>>> Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
>>> Phone: +43-1-58801-165300
>>> Email: peter.blaha at tuwien.ac.at WIEN2k: http://www.wien2k.at
>>> WWW: http://www.imc.tuwien.ac.at
>>>
>> -------------------------------------------------------------------------
>>> _______________________________________________
>>> 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
>> _______________________________________________
>> 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
>
> --
> Kateryna Foyevtsova, PhD
>
> Research Associate
> Stewart Blusson Quantum Matter Institute
> The University of British Columbia | Vancouver
> 261C - 2355 East Mall | Vancouver BC | V6T 1Z4 Canada
> Phone +1 (604) 822-8545
> foyevtsova at phas.ubc.ca
>
> www.sbqmi.ubc.ca
> _______________________________________________
> 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
More information about the Wien
mailing list