Separate specie files for the most common rare earth ions
Brought to you by:
jkdewhurst,
sangeetasharma
Menu
▾
▴
Dear developers, Lars, Kay,
could you please add a separate specie files for the most common rare earth ions (from Ce to Yb), like it is done in VASP (Dy_3, Nd_3, etc.), which do not produce large charge leaks, contain correct energies for the valence states, are checked, optimized, etc. Not everybody need magnetism of the rare earths. Very often a simple R3+ ion works just fine for the geometry optimization, DOS, ELF, Thermodynamics (phonons), enthalpy of formation, elastic properties, etc.
I know how to change the species manually, but usually I face a significant charge leakage or some other issues and it prevents from a wider usage of the code.
I can show, if you would like to discuss, how nicely this approach works with VASP and how good is the correlation with experimenental data. I would like to do the same with Elk.
Thank you in advance.
Last edit: Vitaliy Romaka 2024-01-24
Dear Vitaliy,
Take a look at /elk/examples/ELNES/Pu/Pu+.in . Is that the sort of species file you have in mind?
Regards,
Kay.
Dear Kay,
I am in the train right now and can not have a look on the example file, but here is a very good description of what I mean, which is realized in VASP:
https://www.vasp.at/wiki/index.php/Available_PAW_potentials
In addition, special GGA potentials are supplied for Ce-Lu, in which f f electrons are kept frozen in the core, which is an attempt to treat the localized nature of f f electrons. The number of f electrons in the core equals the total number of valence electrons minus the formal valency. For instance: According to the periodic table, Sm has a total of 8 valence electrons, i.e., 6 f f electrons and 2 s s electrons. In most compounds, Sm adopts a valency of 3; hence 5 f f electrons are placed in the core when the pseudopotential is generated. The corresponding potential can be found in the directory Sm_3. The formal valency n is indicted by _n, where n is either 3 or 2. Ce_3 is, for instance, a Ce potential for trivalent Ce (for tetravalent Ce, the standard potential should be used).
Last edit: Vitaliy Romaka 2024-01-23
I wouldn't change the species file -- it probably won't remove the f-electrons from the valence.
Instead they can be moved down in energy by using DFT+U as a sort of projection operator. For example:
The f electrons are moved down to about 4.6 eV below the Fermi energy. See attached partial DOS.
I hope this is the effect you're after.
Regards,
Kay.
I try to avoid the DFT+U concept, this is not even an option.
I have changed those files and here is an example of the Er3+ specie file I used for Elk and its implementation for the half-Heusler ErNiSb (semiconductor). The specie file is not perfect, I know, I steel have some core leakage (larger that usual). So what I kindly ask is to create a set of such files that you may optimize, check for the correct energy values for other rare earths and include them into the package. It will help users to avoid a lot of linearization energy error messages, large core leakage, etc. If you do, I can test them for some real systems, where I have experimental data and VASP/AkaiKKR results for comparison. The results will be reported here, or in private, as you wish.
The frozen core concept for the rare earths is realized in VASP, and even in the AkaiKKR code.
Last edit: Vitaliy Romaka 2024-01-24
I'm not sure this is possible to do reliably. By adding the f-electrons to the core you are just weakening the effective potential. The f orbitals will still appear in the valence spectrum, but may be shifted up.
I ran FeGd with the f-electrons in core and the l=3 local-orbitals removed. Here is the elk.in:
This is the Gd_f_core.in species file:
~~~
'Gd' : spsymb
'gadolinium' : spname
-64.0000 : spzn
286649.2142 : spmass
0.250000E-06 2.8000 51.1390 700 : rminsp, rmt, rmaxsp, nrmt
21 : nstsp
1 0 1 2.00000 T : nsp, lsp, ksp, occsp, spcore
2 0 1 2.00000 T
2 1 1 2.00000 T
2 1 2 4.00000 T
3 0 1 2.00000 T
3 1 1 2.00000 T
3 1 2 4.00000 T
3 2 2 4.00000 T
3 2 3 6.00000 T
4 0 1 2.00000 T
4 1 1 2.00000 T
4 1 2 4.00000 T
4 2 2 4.00000 T
4 2 3 6.00000 T
4 3 3 3.00000 T <---- modified
4 3 4 4.00000 T <----
5 0 1 2.00000 F
5 1 1 2.00000 F
5 1 2 4.00000 F
5 2 2 1.00000 F
6 0 1 2.00000 F
1 : apword
0.1500 0 F : apwe0, apwdm, apwve
0 : nlx
5 : nlorb <---- modified
0 2 : lorbl, lorbord
0.1500 0 F : lorbe0, lorbdm, lorbve
0.1500 1 F
1 2 : lorbl, lorbord
0.1500 0 F : lorbe0, lorbdm, lorbve
0.1500 1 F
2 2 : lorbl, lorbord
0.1500 0 F : lorbe0, lorbdm, lorbve
0.1500 1 F
0 2 : lorbl, lorbord
0.1500 0 F : lorbe0, lorbdm, lorbve
-1.8262 0 T
1 2 : lorbl, lorbord
0.1500 0 F : lorbe0, lorbdm, lorbve
-0.9546 0 T
! removed the f local orbitals
3 2 : lorbl, lorbord
0.1500 0 F : lorbe0, lorbdm, lorbve
-0.3108 0 T
3 2 : lorbl, lorbord
0.1500 0 F : lorbe0, lorbdm, lorbve
0.1500 1 F
~~~
The partial DOS is attached. As you can see, the f-orbitals are still there around the Fermi energy.
There is more control over the valence states with pseudopotentials because they consist of projection operators for each l. The effective potential from the core electrons in Elk is strictly local.
Also the DFT+U operator as used above, is just such a projection operator: it uses U=J=-0.4 Ha which is clearly not physical, but it has roughly the same effect as a l=3 pseudopotential.
Regards,
Kay.
Thank you Kay for the explanation, I think I've got your point.