[gpaw-users] gpaw-users Digest, Vol 62, Issue 22
Toma Susi
toma.susi at univie.ac.at
Mon Mar 16 17:23:47 CET 2015
(Sorry about the long reply quote, I only get the digest emails.)
Dear Eric,
What Mathias wrote in his last email is correct: comparing to total energies of a ground state and core excited state calculation is the way to go for core level shifts in DFT. The absolute zero reference is not really meaningful; experimentally, because the sample workfunction is usually not known, the binding energies are not referred to the vacuum level, but to the Fermi level, which is aligned between the spectrometer and the sample.
However, for your particular case, I’m not so sure GPAW can work for you since it doesn’t do spin-orbit coupling, and thus cannot give you multipole-splitted energies. I did some tests for the phosphorus 2p level, but didn’t get very satisfactory results. (Note that VASP doesn’t do spin-orbit splitting either.)
Best,
Toma
> On 16 Mar 2015, at 17:09, gpaw-users-request at listserv.fysik.dtu.dk wrote:
>
> Send gpaw-users mailing list submissions to
> gpaw-users at listserv.fysik.dtu.dk
>
> To subscribe or unsubscribe via the World Wide Web, visit
> https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
> or, via email, send a message with subject or body 'help' to
> gpaw-users-request at listserv.fysik.dtu.dk
>
> You can reach the person managing the list at
> gpaw-users-owner at listserv.fysik.dtu.dk
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of gpaw-users digest..."
>
>
> Today's Topics:
>
> 1. Calculating core-level shifts between different systems
> (Eric Hermes)
> 2. Re: Calculating core-level shifts between different systems
> (Mathias Ljungberg)
> 3. Re: Calculating core-level shifts between different systems
> (Eric Hermes)
> 4. Re: Calculating core-level shifts between different systems
> (Eric Hermes)
> 5. Re: Calculating core-level shifts between different systems
> (Mathias Ljungberg)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Mon, 16 Mar 2015 09:45:24 -0500
> From: Eric Hermes <ehermes at chem.wisc.edu>
> To: gpaw-users <gpaw-users at listserv.fysik.dtu.dk>
> Subject: [gpaw-users] Calculating core-level shifts between different
> systems
> Message-ID: <5506EC84.5030303 at chem.wisc.edu>
> Content-Type: text/plain; charset=utf-8; format=flowed
>
> Hello,
>
> I wish to calculate the core-level shift of the Pd 3d 5/2 state in PdO,
> PdO2, and Pd-Te intermetallics with respect to metallic Pd in order to
> compare against experimental XPS spectra. The experimentalists indicate
> that the presence of Te in the intermetallic shifts the Pd 3d 5/2 state
> to higher binding energy, but my chemical intuition coupled with Bader
> analysis indicates the opposite. I have attempted to calculate these
> properties in VASP (using the ICORELEVEL tag), but the oxide shifts do
> not even semi-quantitatively agree with experiment (although they do
> agree qualitatively, as the shifts are ordered correctly). Part of the
> ambiguity here is that I am unsure what to take as the zero of energy in
> fully 3D periodic systems to properly compare binding energies.
>
> I believe that GPAW might be better able to calculate these values, as I
> can do 2D-periodic slab calculations which have a true zero reference
> energy. However, I am slightly confused by the information on this page:
> https://wiki.fysik.dtu.dk/gpaw/tutorials/xas/xas.html
> The citation Leetmaa2006 seems to be missing, and I cannot find what
> paper this refers to. I am unsure of what exactly is being calculated in
> the "Further considerations" section; is this how one compares to
> experimental XPS spectra? Can someone help guide me in doing these
> calculations within GPAW?
>
> Thank you,
> Eric Hermes
>
> --
> Eric Hermes
> J.R. Schmidt Group
> Chemistry Department
> University of Wisconsin - Madison
>
>
>
> ------------------------------
>
> Message: 2
> Date: Mon, 16 Mar 2015 16:21:14 +0100
> From: Mathias Ljungberg <mathias.ljungberg at fysik.su.se>
> To: Eric Hermes <ehermes at chem.wisc.edu>
> Cc: gpaw-users <gpaw-users at listserv.fysik.dtu.dk>
> Subject: Re: [gpaw-users] Calculating core-level shifts between
> different systems
> Message-ID: <210BCE78-32E2-461E-90D9-2DF5E2F9AB21 at fysik.su.se>
> Content-Type: text/plain; charset=us-ascii
>
> Hi Eric,
>
> On Mar 16, 2015, at 3:45 PM, Eric Hermes wrote:
>
>> Hello,
>>
>> I wish to calculate the core-level shift of the Pd 3d 5/2 state in PdO, PdO2, and Pd-Te intermetallics
>
>> with respect to metallic Pd in order to compare against experimental XPS spectra. The experimentalists indicate that the presence of Te in the intermetallic shifts the Pd 3d 5/2 state to higher binding energy, but my chemical intuition coupled with Bader analysis indicates the opposite. I have attempted to calculate these properties in VASP (using the ICORELEVEL tag), but the oxide shifts do not even semi-quantitatively agree with experiment (although they do agree qualitatively, as the shifts are ordered correctly). Part of the ambiguity here is that I am unsure what to take as the zero of energy in fully 3D periodic systems to properly compare binding energies.
>>
>
> I am not sure how well a 3d core hole would work in GPAW, you would have to try to create your own PAW dataset with such a core hole. Also there is no spin-orbit in GPAW.
>
>
>> I believe that GPAW might be better able to calculate these values, as I can do 2D-periodic slab calculations which have a true zero reference energy.
>
> I think this is not necessary. Just introduce an extra electron in the system to compensate for the core hole, this should work well for metals. You calculate the XPS position by total energy differences between the ground state and the state with a core hole (and an extra electron)
>
>> However, I am slightly confused by the information on this page:
>> https://wiki.fysik.dtu.dk/gpaw/tutorials/xas/xas.html
>> The citation Leetmaa2006 seems to be missing, and I cannot find what paper this refers to. I am unsure of what exactly is being calculated in the "Further considerations" section; is this how one compares to experimental XPS spectra? Can someone help guide me in doing these calculations within GPAW?
>>
>
> The tutorial mainly concerns XAS for small molecules. For molecules XPS should be computed by really removing the electron, but for a solid experience has shown that is it better add an electron to the system to preserve charge neutrality. For a metal the XPS is even experimentally referenced to the fermi level so there should be no problems there. I think the Leetmaa reference is not so useful for you anyway since it deals with liquid water. Maybe it should be removed from the tutorial...
>
> Best regards,
> Mathias Ljungberg
>
>> Thank you,
>> Eric Hermes
>>
>> --
>> Eric Hermes
>> J.R. Schmidt Group
>> Chemistry Department
>> University of Wisconsin - Madison
>>
>> _______________________________________________
>> gpaw-users mailing list
>> gpaw-users at listserv.fysik.dtu.dk
>> https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
>
>
>
>
> ------------------------------
>
> Message: 3
> Date: Mon, 16 Mar 2015 10:39:30 -0500
> From: Eric Hermes <ehermes at chem.wisc.edu>
> To: gpaw-users <gpaw-users at listserv.fysik.dtu.dk>
> Subject: Re: [gpaw-users] Calculating core-level shifts between
> different systems
> Message-ID: <5506F932.6080409 at chem.wisc.edu>
> Content-Type: text/plain; charset=windows-1252; format=flowed
>
> Miguel,
>
> Thank you for your quick reply. The way that I performed the calculation
> in VASP was to use 5 layers of the material (Pd, PdO, PdO2, and Pd3Te)
> fixed at their bulk geometry and a 15 angstrom vacuum gap. Because the
> slabs were symmetric, I did not use a dipole correction. I used
> ICORELEVEL=1 (the initial state approximation) with the primitive slab
> unit cell and a high K-point density. I produced the LOCPOT file with
> LVTOT=.TRUE., averaged the local potential over X and Y, and took the
> value of the local potential in the middle of the vacuum gap as my
> reference energy (Evac). The core level energies were computed as Evac -
> Ecore, where Ecore is the reported 3d state energy for the most buried
> Pd atom in my system. Relative to metallic Pd, I obtained a core level
> shift of 0.83 eV for PdO and 2.09 eV for PdO2, as compared to the
> experimental values of about 1.8 eV for PdO and 3.5 eV for PdO2. I also
> performed the same procedure in the Slater-Janak transition state
> approximation (ICORELEVEL=2, CLZ=0.5) with very large supercells of the
> slab and obtained shifts of 1.30 eV for PdO and 1.98 eV for PdO2.
>
> Is there something about these procedures that is incorrect? For now I
> will try the same procedure with HSE06 and see if it produces values
> that are in better agreement with experiment.
>
> Thanks,
> Eric
>
> On 3/16/2015 10:05 AM, Caro Miguel wrote:
>> Dear Eric,
>>
>> You can also use a 3D slab calculation with VASP (or plane waves in GPAW) including dipole corrections along the direction perpendicular to your slab. The Hartree potential will converge to the vacuum level if you leave enough vacuum between periodic replicas of the slab, and this is your absolute reference (subtract that from every other number). In a non-periodic calculation you still need to leave enough vacuum to prevent interaction with the box edges.
>>
>> You can check out our recent paper, and especially the more standard references therein: http://scitation.aip.org/content/aip/journal/jap/117/3/10.1063/1.4905915
>>
>> Note that in order to get the correct positioning for your states you might need to resort to a scheme beyond GGA/LDA, that is try a hybrid functional or GW.
>>
>> Regards,
>> Miguel
>>
>> Dr. Miguel Caro
>> Postdoctoral researcher
>> Department of Electrical Engineering and Automation, and
>> COMP Centre of Excellence in Computational Nanoscience
>> Aalto University, Finland
>> Email: mcaroba at gmail.com
>> Work: miguel.caro at aalto.fi
>> Website: http://mcaroba.dyndns.org
>>
>> ________________________________________
>> From: gpaw-users-bounces at listserv.fysik.dtu.dk [gpaw-users-bounces at listserv.fysik.dtu.dk] on behalf of Eric Hermes [ehermes at chem.wisc.edu]
>> Sent: 16 March 2015 16:45
>> To: gpaw-users
>> Subject: [gpaw-users] Calculating core-level shifts between different systems
>>
>> Hello,
>>
>> I wish to calculate the core-level shift of the Pd 3d 5/2 state in PdO,
>> PdO2, and Pd-Te intermetallics with respect to metallic Pd in order to
>> compare against experimental XPS spectra. The experimentalists indicate
>> that the presence of Te in the intermetallic shifts the Pd 3d 5/2 state
>> to higher binding energy, but my chemical intuition coupled with Bader
>> analysis indicates the opposite. I have attempted to calculate these
>> properties in VASP (using the ICORELEVEL tag), but the oxide shifts do
>> not even semi-quantitatively agree with experiment (although they do
>> agree qualitatively, as the shifts are ordered correctly). Part of the
>> ambiguity here is that I am unsure what to take as the zero of energy in
>> fully 3D periodic systems to properly compare binding energies.
>>
>> I believe that GPAW might be better able to calculate these values, as I
>> can do 2D-periodic slab calculations which have a true zero reference
>> energy. However, I am slightly confused by the information on this page:
>> https://wiki.fysik.dtu.dk/gpaw/tutorials/xas/xas.html
>> The citation Leetmaa2006 seems to be missing, and I cannot find what
>> paper this refers to. I am unsure of what exactly is being calculated in
>> the "Further considerations" section; is this how one compares to
>> experimental XPS spectra? Can someone help guide me in doing these
>> calculations within GPAW?
>>
>> Thank you,
>> Eric Hermes
>>
>> --
>> Eric Hermes
>> J.R. Schmidt Group
>> Chemistry Department
>> University of Wisconsin - Madison
>>
>> _______________________________________________
>> gpaw-users mailing list
>> gpaw-users at listserv.fysik.dtu.dk
>> https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
>
> --
> Eric Hermes
> J.R. Schmidt Group
> Chemistry Department
> University of Wisconsin - Madison
>
>
>
> ------------------------------
>
> Message: 4
> Date: Mon, 16 Mar 2015 10:47:12 -0500
> From: Eric Hermes <ehermes at chem.wisc.edu>
> To: gpaw-users <gpaw-users at listserv.fysik.dtu.dk>
> Cc: Mathias Ljungberg <mathias.ljungberg at fysik.su.se>
> Subject: Re: [gpaw-users] Calculating core-level shifts between
> different systems
> Message-ID: <5506FB00.70708 at chem.wisc.edu>
> Content-Type: text/plain; charset=windows-1252; format=flowed
>
> Mathias,
>
> I understand that to calculate the core binding energy for periodic
> systems it is sufficient and necessary to add an electron to the valence
> in the core-hole calculation. However, the issue is that in my
> understanding this binding energy is still relative to some
> system-dependent zero of energy at infinite separation. In non
> 3D-periodic systems, this zero is well defined as the potential
> infinitely far away from the system, but for 3D-periodic calculations
> there is no such thing as infinitely far away from the system, so this
> reference becomes ill-defined. Because I am attempting to calculate core
> binding energy shifts, I need to correct for this value, and it is
> easier to think about for 2D-periodic systems, which are better suited
> for slab calculations anyway than fully 3D-periodic systems.
>
> Do you believe that generating a core-hole setup for Pd would be
> problematic? I am unclear on the specifics of how setup generation works
> in GPAW. However, in VASP, the core-hole PAW pseudopotentials are
> generated on-the-fly during the course of the calculation from the
> ground-state pseudopotentials if ICORELEVEL=2 is specified. This allows
> one to perform a calculation in which the valence electrons relax in the
> presence of the core hole, but the other core electrons are fixed in
> their ground state configuration. Presumably this is also how GPAW would
> perform a core-hole calculation. Are there some differences that may
> cause problems that I should be aware of?
>
> Thanks for your reply,
> Eric
>
> On 3/16/2015 10:21 AM, Mathias Ljungberg wrote:
>> Hi Eric,
>>
>> On Mar 16, 2015, at 3:45 PM, Eric Hermes wrote:
>>
>>> Hello,
>>>
>>> I wish to calculate the core-level shift of the Pd 3d 5/2 state in PdO, PdO2, and Pd-Te intermetallics
>>> with respect to metallic Pd in order to compare against experimental XPS spectra. The experimentalists indicate that the presence of Te in the intermetallic shifts the Pd 3d 5/2 state to higher binding energy, but my chemical intuition coupled with Bader analysis indicates the opposite. I have attempted to calculate these properties in VASP (using the ICORELEVEL tag), but the oxide shifts do not even semi-quantitatively agree with experiment (although they do agree qualitatively, as the shifts are ordered correctly). Part of the ambiguity here is that I am unsure what to take as the zero of energy in fully 3D periodic systems to properly compare binding energies.
>>>
>> I am not sure how well a 3d core hole would work in GPAW, you would have to try to create your own PAW dataset with such a core hole. Also there is no spin-orbit in GPAW.
>>
>>
>>> I believe that GPAW might be better able to calculate these values, as I can do 2D-periodic slab calculations which have a true zero reference energy.
>> I think this is not necessary. Just introduce an extra electron in the system to compensate for the core hole, this should work well for metals. You calculate the XPS position by total energy differences between the ground state and the state with a core hole (and an extra electron)
>>
>>> However, I am slightly confused by the information on this page:
>>> https://wiki.fysik.dtu.dk/gpaw/tutorials/xas/xas.html
>>> The citation Leetmaa2006 seems to be missing, and I cannot find what paper this refers to. I am unsure of what exactly is being calculated in the "Further considerations" section; is this how one compares to experimental XPS spectra? Can someone help guide me in doing these calculations within GPAW?
>>>
>> The tutorial mainly concerns XAS for small molecules. For molecules XPS should be computed by really removing the electron, but for a solid experience has shown that is it better add an electron to the system to preserve charge neutrality. For a metal the XPS is even experimentally referenced to the fermi level so there should be no problems there. I think the Leetmaa reference is not so useful for you anyway since it deals with liquid water. Maybe it should be removed from the tutorial...
>>
>> Best regards,
>> Mathias Ljungberg
>>
>>> Thank you,
>>> Eric Hermes
>>>
>>> --
>>> Eric Hermes
>>> J.R. Schmidt Group
>>> Chemistry Department
>>> University of Wisconsin - Madison
>>>
>>> _______________________________________________
>>> gpaw-users mailing list
>>> gpaw-users at listserv.fysik.dtu.dk
>>> https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
>
> --
> Eric Hermes
> J.R. Schmidt Group
> Chemistry Department
> University of Wisconsin - Madison
>
>
>
> ------------------------------
>
> Message: 5
> Date: Mon, 16 Mar 2015 17:03:47 +0100
> From: Mathias Ljungberg <mathias.ljungberg at fysik.su.se>
> To: Eric Hermes <ehermes at chem.wisc.edu>
> Cc: gpaw-users <gpaw-users at listserv.fysik.dtu.dk>
> Subject: Re: [gpaw-users] Calculating core-level shifts between
> different systems
> Message-ID: <89034D45-D627-4EB9-BA77-B9E516E57E58 at fysik.su.se>
> Content-Type: text/plain; charset=windows-1252
>
>
> On Mar 16, 2015, at 4:47 PM, Eric Hermes wrote:
>
>> Mathias,
>>
>> I understand that to calculate the core binding energy for periodic systems it is sufficient and necessary to add an electron to the valence in the core-hole calculation. However, the issue is that in my understanding this binding energy is still relative to some system-dependent zero of energy at infinite separation. In non 3D-periodic systems, this zero is well defined as the potential infinitely far away from the system, but for 3D-periodic calculations there is no such thing as infinitely far away from the system, so this reference becomes ill-defined. Because I am attempting to calculate core binding energy shifts, I need to correct for this value, and it is easier to think about for 2D-periodic systems, which are better suited for slab calculations anyway than fully 3D-periodic systems.
>>
>
> Well, the truth is that the experimentalists don't measure XPS like that because the work function of a material is dependent on the actual surface, and to avoid taking this into account one refers the XPS to the fermi level (at least for metals where there is a well-defined fermi level). So what you want to compute it the energy it takes to lift the core electron to the fermi level which is exactly what the total energy difference Delta SCF or Delta Kohn Sham is accomplishing. See for example http://journals.aps.org/prb/abstract/10.1103/PhysRevB.91.081401
>
>
>> Do you believe that generating a core-hole setup for Pd would be problematic? I am unclear on the specifics of how setup generation works in GPAW. However, in VASP, the core-hole PAW pseudopotentials are generated on-the-fly during the course of the calculation from the ground-state pseudopotentials if ICORELEVEL=2 is specified. This allows one to perform a calculation in which the valence electrons relax in the presence of the core hole, but the other core electrons are fixed in their ground state configuration. Presumably this is also how GPAW would perform a core-hole calculation. Are there some differences that may cause problems that I should be aware of?
>>
>
> In GPAW you need to construct your own PAW dataset or "setup" with a core hole. This means that the core electrons will adjust to the hole, but when it is put into the actual calculation only the valence electrons can relax. It is also possible to construct setups with more electron in the valence but I don't think this is a good option in your case (there will be many electrons).
>
> Best regards,
> Mathias
>
>> Thanks for your reply,
>> Eric
>>
>> On 3/16/2015 10:21 AM, Mathias Ljungberg wrote:
>>> Hi Eric,
>>>
>>> On Mar 16, 2015, at 3:45 PM, Eric Hermes wrote:
>>>
>>>> Hello,
>>>>
>>>> I wish to calculate the core-level shift of the Pd 3d 5/2 state in PdO, PdO2, and Pd-Te intermetallics
>>>> with respect to metallic Pd in order to compare against experimental XPS spectra. The experimentalists indicate that the presence of Te in the intermetallic shifts the Pd 3d 5/2 state to higher binding energy, but my chemical intuition coupled with Bader analysis indicates the opposite. I have attempted to calculate these properties in VASP (using the ICORELEVEL tag), but the oxide shifts do not even semi-quantitatively agree with experiment (although they do agree qualitatively, as the shifts are ordered correctly). Part of the ambiguity here is that I am unsure what to take as the zero of energy in fully 3D periodic systems to properly compare binding energies.
>>>>
>>> I am not sure how well a 3d core hole would work in GPAW, you would have to try to create your own PAW dataset with such a core hole. Also there is no spin-orbit in GPAW.
>>>
>>>
>>>> I believe that GPAW might be better able to calculate these values, as I can do 2D-periodic slab calculations which have a true zero reference energy.
>>> I think this is not necessary. Just introduce an extra electron in the system to compensate for the core hole, this should work well for metals. You calculate the XPS position by total energy differences between the ground state and the state with a core hole (and an extra electron)
>>>
>>>> However, I am slightly confused by the information on this page:
>>>> https://wiki.fysik.dtu.dk/gpaw/tutorials/xas/xas.html
>>>> The citation Leetmaa2006 seems to be missing, and I cannot find what paper this refers to. I am unsure of what exactly is being calculated in the "Further considerations" section; is this how one compares to experimental XPS spectra? Can someone help guide me in doing these calculations within GPAW?
>>>>
>>> The tutorial mainly concerns XAS for small molecules. For molecules XPS should be computed by really removing the electron, but for a solid experience has shown that is it better add an electron to the system to preserve charge neutrality. For a metal the XPS is even experimentally referenced to the fermi level so there should be no problems there. I think the Leetmaa reference is not so useful for you anyway since it deals with liquid water. Maybe it should be removed from the tutorial...
>>>
>>> Best regards,
>>> Mathias Ljungberg
>>>
>>>> Thank you,
>>>> Eric Hermes
>>>>
>>>> --
>>>> Eric Hermes
>>>> J.R. Schmidt Group
>>>> Chemistry Department
>>>> University of Wisconsin - Madison
>>>>
>>>> _______________________________________________
>>>> gpaw-users mailing list
>>>> gpaw-users at listserv.fysik.dtu.dk
>>>> https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
>>
>> --
>> Eric Hermes
>> J.R. Schmidt Group
>> Chemistry Department
>> University of Wisconsin - Madison
>>
>
>
>
>
> ------------------------------
>
> _______________________________________________
> gpaw-users mailing list
> gpaw-users at listserv.fysik.dtu.dk
> https://listserv.fysik.dtu.dk/mailman/listinfo/gpaw-users
>
> End of gpaw-users Digest, Vol 62, Issue 22
> ******************************************
More information about the gpaw-users
mailing list