My Community

I used optics to work out the EPS and found a strange phenomena:
0 NaN NaN
………………………………
0.4500000000000000 15.8730650828125700 82.4986122823371300
0.4548875000000000 15.6280745773936700 81.7180399923139000
0.4597750000000000 15.3872837289285700 81.0039784860626400
………………………………
Why the imaginary parts of the dielectric function(the last coulmb) are so large?And I see that of larger energy range seems OK.
I appreciate for any help.
<span class='smallblacktext'>[ Edited ]</span>

What kind of system are you dealing with? Is it metallic?

No, it's a insulator.(Badgap is about 4eV) I calculate again using VASP5.2. There's no NaN, but also very large value, and reach 0 abruptly when energy=0.

Can VASP5.2 deal with f electrons? I calculate the optical properties of a system containing 7 f electrons.

yes

Thank you. So the weird phenomenon may attribute to the divergency problem? I set NBANDS=190 for 94 valence electrons for considering ISPIN=2. It needs large memory……

Do you get a LDA gap of 4 eV?
Are you using the tetrahedron method?

I use GGA-PBE.
Yes, ISMEAR=-5