Dear support team,
I have a few questions on the basis sets of siesta given in https://www.simuneatomistics.com/siesta-pro/siesta-pseudos-and-basis-database/ :
(1) some materials do not have a file that gives their basis set (ex : Cu,Cr,Ni). Does this mean the default siesta basis (DZP, eshift=0.02Ry, splitnorm=0.15) is the optimised basis ?
(2) Some materials have a basis set, but only for some orbitals (ex : Pt is missing 5s and 5p). Does this again mean that the non-specified orbitals are at the default parameters of siesta ?
Thank you for your time.
You are correct some materials do no have an optimized basis set. This might be due to different reasons such as reasonably good default basis set or the need of customized basis set based on the structure to be considered. We are planning to improve our database in the future and advice on the best choice for the basis set.
In case of no basis set block in your input fiel we suggest to use an energyshift value in between 20-50 meV (0.00147 - 0.00367 Ry) which is smaller than the SIESTA default value. These values will generated a more extended basis set which often provides a better description of the system.
You can use the default splitnorm value.
The number of orbital to be included in the basis set block can vary and depends on the pseudopotential. In case of Pt the 5s are not considered explicitly in the basis set block.
Hope this will be helpful.
Thank you for your answer !
If I understood well : when (1) no basis block is specified in Simune basis, or (2) some orbitals are not explicitly defined, then I have to use an eshift of 20-50meV and keep the default splitnorm ?
That indeed should do in most cases.