Page 1 of 1

probleme of calculation frequences

Posted: Wed Apr 08, 2009 6:40 pm
by rghomari
Hi user VASP.
I would like to study the adsorption of molecule on surface Rh(111) with two layers; the first layer contain 18 atoms of Rh.
so i used k-POINTS 3x3x1 and Ecut=400ev.
i want to calculate the frequence with relaxation the top layer and the molecule ;the second layed was fixed. i give 240 hours and didn't finish i restart with 240 hours and it didn't finish.
what i will do? can any body help me?
2- when i calculated the frequences of TS (NEB) , i find many imaginary frequences. why?
My INCAR file is :
System=
PREC=normal
ISMEAR=0
ISTAR=1
ENMAX=400
IBRION=5
NFREE=2
ISIF=2
POTIM=0.5
EDIFFG=1e-04
NSW=20
LREAL=Auto

Thank you.

probleme of calculation frequences

Posted: Fri Apr 10, 2009 3:44 pm
by admin
when calculating vibration spectra, please keep in mind that for each atom which is allowed to relax 6 displacements have to be regarded explicitely to construct the Hessian matrix. This leads to very long computing times. ---> please restrict the vibrations to the CO molecule and the nearest Rh neighbors. Usually, the frequencies do not change significantly if more neighbors / layers are accounted for the vibration as well (if you want to be sure, please test how adding additional neighbors as T T T in POSCAR changes the relevant frequencies).
ALSO, please decrease POTIM to 0.015 . the frequencies are calculated under the assumption that the atoms vibrate in the harmonic oscillator limit. A displacement of 0.5 Ã… is possibly much to large that this condition is matched.