Page 1 of 1

Higher Diffusion barrier in unit cell

Posted: Mon Dec 01, 2014 3:31 pm
by dixitmuditk
Dear Prof. Graeme,
I am doing NEB at the LiFePO4 unit-cell only(four formula unit) and getting significantly higher barrier compared to reported barriers using super-cells. Also, The barriers using Spin polarized NEB are higher than that of NON spin polarized calculations. I feel this is due to image interactions. Please suggest if I am wrong.


Thanks!!

Re: Higher Diffusion barrier in unit cell

Posted: Tue Dec 02, 2014 1:51 pm
by graeme
A unit cell may be too small to avoid interaction between the images. The fact that spin polarization influences the barrier is expected. This is a wide-gap insulator in which the electron from the Li ion is localized on a neighboring Fe center. It is important to use a +U correction or a hybrid functional to avoid artificial electron delocalization. If you use +U, check the on-site occupancies to understand where the electrons are localized and that you have the correct ground state. You can use the MAGMOM tag for the Fe atoms to help localize the electron near the Li+ ion and find the ground state.

Re: Higher Diffusion barrier in unit cell

Posted: Thu Dec 04, 2014 11:03 am
by dixitmuditk
Thanks Prof. Graeme for your kind reply.