Dear all,
I'm performing some calculations of the dielectric function of CeO2. While at low energy I obtained a goo convergence, beyond 100eV I found a discontinuity in the imaginary part of the dielectric function passing from q=0 to q small (in the range of 0.1-0.2 blat). The only parameter that affect this behavior moving the curves towards the continuity i the muffin tin radius that I imposed to be smaller than default by means of rmtdelta=0.5 , 0.8, 1.1. This partially improve the continuity, but I would like have some suggestion from you to better understand the causes of this behavior.
APW does not expand excited orbitals all that well. This is because the linearisation energies are fixed to the occupied band energies.
To improve the accuracy of the excited states you can set
lorbcnd.true.
which adds several local orbitals at higher energy. However, this is already enabled with
highq.true.
which you have already set.
Alternatively, you can use LAPW or super-LAPW with (for example)
nxoapwlo1
which increases the order of the APWs and local-orbitals by 1. This switches on LAPW and thus the muffin-tin functions are matched at the muffin-tin boundary with their value and first derivative. This improves the convergence of the higher conduction bands. You can also try setting it to 2 which would correspond to super-LAPW, i.e. matching of the function and its first and second derivatives.
The planewave cut-off should also be increased at the same time because LAPW converges more slowly than APW. Try setting
rgkmax8.5
or so.
Finally, plot the band structure with and without LAPW. You should see some difference for the high-lying bands.
Regards,
Kay.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
APW does not expand excited orbitals all that well. This is because the linearisation energies are fixed to the occupied band energies.
To improve the accuracy of the excited states you can set
lorbcnd
.true.
which adds several local orbitals at higher energy. However, this is already enabled with
highq
.true.
which you have already set.
Alternatively, you can use LAPW or super-LAPW with (for example)
nxoapwlo
1
which increases the order of the APWs and local-orbitals by 1. This switches on LAPW and thus the muffin-tin functions are matched at the muffin-tin boundary with their value and first derivative. This improves the convergence of the higher conduction bands. You can also try setting it to 2 which would correspond to super-LAPW, i.e. matching of the function and its first and second derivatives.
The planewave cut-off should also be increased at the same time because LAPW converges more slowly than APW. Try setting
rgkmax
8.5
or so.
Finally, plot the band structure with and without LAPW. You should see some difference for the high-lying bands.
Dear all,
I'm doing some calculations on the ELF of CeO2 including the 4d and 4p electrons for the cerium.
I obtained a peak in the correct position around 120 eV (4d transition), comparable with that of NIST experimental database.
However, as the momentum increases, the peak shifts towards lower energies. This behavior is somewhat suspect because is different from the typical increase of the energy with the momentum.
I would like to know how to further test or increase the precision of my calculation, in particular at high momentum.
Try increasing the k-point grid and nempty.
Also, consider using meta-GGA (e.g. xctype 100 209 12), although I am not sure if that will work with +U. I shoud do fine without +U though.
Still, you are interested in super high energies, which are in principle not something you would analize with DFT. Consider CASPT2 or CCSDT methods in OpenMolcas with embedding AIMPs. The embedding data are likely to be available somewhere.
Best regards.
Andrew
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear Sir J.K. Dewhurst,
I want to know that, is it possible to use GGA as xctype with ALDA in TDDFT. Because in examples only xctype 3 is used with ALDA. Please clear my small doubt.
Thanks
Nikhil.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear all,
I'm performing some calculations of the dielectric function of CeO2. While at low energy I obtained a goo convergence, beyond 100eV I found a discontinuity in the imaginary part of the dielectric function passing from q=0 to q small (in the range of 0.1-0.2 blat). The only parameter that affect this behavior moving the curves towards the continuity i the muffin tin radius that I imposed to be smaller than default by means of rmtdelta=0.5 , 0.8, 1.1. This partially improve the continuity, but I would like have some suggestion from you to better understand the causes of this behavior.
Thanks in advance,
Andrea
Hi Andrea,
APW does not expand excited orbitals all that well. This is because the linearisation energies are fixed to the occupied band energies.
To improve the accuracy of the excited states you can set
which adds several local orbitals at higher energy. However, this is already enabled with
which you have already set.
Alternatively, you can use LAPW or super-LAPW with (for example)
which increases the order of the APWs and local-orbitals by 1. This switches on LAPW and thus the muffin-tin functions are matched at the muffin-tin boundary with their value and first derivative. This improves the convergence of the higher conduction bands. You can also try setting it to 2 which would correspond to super-LAPW, i.e. matching of the function and its first and second derivatives.
The planewave cut-off should also be increased at the same time because LAPW converges more slowly than APW. Try setting
or so.
Finally, plot the band structure with and without LAPW. You should see some difference for the high-lying bands.
Regards,
Kay.
Hi John,
thank you very much for the explanation, I will try your suggestions.
Best,
Andrea.
Da: John Kay Dewhurst jkdewhurst@users.sourceforge.net
Inviato: domenica 15 marzo 2020 09:45
A: [elk:discussion] 897820@discussion.elk.p.re.sourceforge.net
Oggetto: [elk:discussion] Dielectic function beyond 100eV
Hi Andrea,
APW does not expand excited orbitals all that well. This is because the linearisation energies are fixed to the occupied band energies.
To improve the accuracy of the excited states you can set
lorbcnd
.true.
which adds several local orbitals at higher energy. However, this is already enabled with
highq
.true.
which you have already set.
Alternatively, you can use LAPW or super-LAPW with (for example)
nxoapwlo
1
which increases the order of the APWs and local-orbitals by 1. This switches on LAPW and thus the muffin-tin functions are matched at the muffin-tin boundary with their value and first derivative. This improves the convergence of the higher conduction bands. You can also try setting it to 2 which would correspond to super-LAPW, i.e. matching of the function and its first and second derivatives.
The planewave cut-off should also be increased at the same time because LAPW converges more slowly than APW. Try setting
rgkmax
8.5
or so.
Finally, plot the band structure with and without LAPW. You should see some difference for the high-lying bands.
Regards,
Kay.
Dielectic function beyond 100eVhttps://sourceforge.net/p/elk/discussion/897820/thread/3e3c7feb00/?limit=50#149d
Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/elk/discussion/897820/
To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/
Dear all,
I'm doing some calculations on the ELF of CeO2 including the 4d and 4p electrons for the cerium.
I obtained a peak in the correct position around 120 eV (4d transition), comparable with that of NIST experimental database.
However, as the momentum increases, the peak shifts towards lower energies. This behavior is somewhat suspect because is different from the typical increase of the energy with the momentum.
I would like to know how to further test or increase the precision of my calculation, in particular at high momentum.
Thanks in advance,
Andrea
Last edit: Andrea Pedrielli 2021-03-04
Dear Andrea,
Try increasing the k-point grid and nempty.
Also, consider using meta-GGA (e.g. xctype 100 209 12), although I am not sure if that will work with +U. I shoud do fine without +U though.
Still, you are interested in super high energies, which are in principle not something you would analize with DFT. Consider CASPT2 or CCSDT methods in OpenMolcas with embedding AIMPs. The embedding data are likely to be available somewhere.
Best regards.
Andrew
Dear Sir J.K. Dewhurst,
I want to know that, is it possible to use GGA as xctype with ALDA in TDDFT. Because in examples only xctype 3 is used with ALDA. Please clear my small doubt.
Thanks
Nikhil.