Imaginary frequencies at transition state

[serenity_3]

Hello,
I understand there are some similar posts to this one, but I am not able to fix my problem: I ran a vibrational frequency calculation for the transition state structure obtained from the NEB calculation dimer. I find that there are two large imaginary frequencies. I am not sure what contributes to this unexpected result. I tried changing different flags (e.g POTIM), but I was not able to get rid of the second imaginary frequency. Any suggestion on how to fix the problem is appreciated. Thank you.

1 f = 34.616099 THz 217.499365 2PiTHz 1154.668774 cm-1 143.160618 meV
2 f = 32.913453 THz 206.801322 2PiTHz 1097.874605 cm-1 136.119041 meV
3 f = 31.624827 THz 198.704651 2PiTHz 1054.890696 cm-1 130.789718 meV
4 f = 30.502092 THz 191.650295 2PiTHz 1017.440263 cm-1 126.146459 meV
5 f = 13.852408 THz 87.037249 2PiTHz 462.066605 cm-1 57.288932 meV
6 f = 12.130810 THz 76.220125 2PiTHz 404.640254 cm-1 50.168975 meV
7 f = 7.175303 THz 45.083761 2PiTHz 239.342361 cm-1 29.674657 meV
8 f = 6.336087 THz 39.810807 2PiTHz 211.349102 cm-1 26.203937 meV
9 f = 5.091282 THz 31.989466 2PiTHz 169.826877 cm-1 21.055840 meV
10 f = 5.018189 THz 31.530211 2PiTHz 167.388764 cm-1 20.753552 meV
11 f = 4.324435 THz 27.171226 2PiTHz 144.247623 cm-1 17.884418 meV
12 f = 4.238026 THz 26.628302 2PiTHz 141.365326 cm-1 17.527059 meV
13 f = 3.905560 THz 24.539354 2PiTHz 130.275443 cm-1 16.152089 meV
14 f = 3.557582 THz 22.352949 2PiTHz 118.668174 cm-1 14.712972 meV
15 f = 3.407740 THz 21.411462 2PiTHz 113.669970 cm-1 14.093274 meV
16 f = 3.125111 THz 19.635653 2PiTHz 104.242488 cm-1 12.924415 meV
17 f = 2.982482 THz 18.739489 2PiTHz 99.484903 cm-1 12.334550 meV
18 f = 1.637520 THz 10.288840 2PiTHz 54.621780 cm-1 6.772235 meV
19 f = 1.592645 THz 10.006886 2PiTHz 53.124931 cm-1 6.586649 meV
20 f/i= 36.870124 THz 231.661820 2PiTHz 1229.854948 cm-1 152.482511 meV
21 f/i= 50.493946 THz 317.262820 2PiTHz 1684.296742 cm-1 208.826087 meV

Here is the INCAR file I used:

SYSTEM = NH2Decomp
NWRITE = 1
ISTART = 0

ENCUT = 400.0 eV
PREC = HIGH
EDIFF = 1E-08
EDFIFFG = -3E-04

LCORR = T
ISIF = 2

ISMEAR = 0
SIGMA = 0.1

IALGO = 38
LDIAG = T
LREAL = F

NBANDS = 206
LDIPOL = T
IDIPOL = 3
LWAVE = F
NSW=100
POTIM = 0.001
IBRION=5
NFREE=2
NELM = 200

[jonathan_lahnsteiner2]

Dear serenity_3,

The imaginary modes indicate that your structure is not in equilibrium and wants to relax along these vibrational modes. For a transition state there should be one such imaginary mode along which
the transition state will reach it's final or product state. Since you see two imaginary modes you could try to optimize the structure of your transition state to get rid of the second one. Therefore do a structure optimization on the transition state. You should also note that the static method you are applying is an approximation method and you might not be able to get rid of the second imaginary mode. This would be the case if finite temperature effects are important. In this case you could analyse your transition state with the help of a molecular dynamics run. Maybe this wiki pages can be of interest to you
https://www.vasp.at/wiki/index.php/Cate ... ion_states
And the pages about the blue moon ensemble:
Blue moon ensemble
Constrained molecular dynamics
INCREM

All the best Jonathan

[serenity_3]

Thank you!
I will follow your suggestions.