Getting rid of multiple imaginary frequencies after CI-NEB Calculation

[Natalie Austin]

Hello,

I ran a CI-NEB calculation on co2 adsorption on a metal from the physisorbed to chemisorbed state. After the NEB calculation completed I ran a VIBRATIONAL ANALYSIS on the structure from the NEB which would represent the transition state (I took the structure with the highest energy). I found 3 imaginary frequencies after the VIBRATIONAL ANALYSIS completed:

 VIB|                        NORMAL MODES - CARTESIAN DISPLACEMENTS
 VIB|
 VIB|                         1                    2                    3
 VIB|Frequency (cm^-1)  -373.959745          -155.894278           -82.219450
 VIB|Intensities           0.000000             0.000000             0.000000
 VIB|Red.Masses (a.u.)    15.645084            12.197575            14.794711
 VIB|Frc consts (a.u.)    -0.000783            -0.000018            -0.000002
  ATOM  EL             X     Y     Z        X     Y     Z        X     Y     Z
    56  O             0.04  0.11 -0.92    -0.02 -0.02  0.19     0.51  0.21  0.24
    57  C             0.08  0.06 -0.28    -0.25  0.09 -0.94     0.50  0.20 -0.10
    58  O             0.20  0.04  0.04    -0.04  0.02 -0.09     0.57  0.12 -0.01


 VIB|                         4                    5                    6
 VIB|Frequency (cm^-1)   233.171137           297.756764           384.024573
 VIB|Intensities           0.000000             0.000000             0.000000
 VIB|Red.Masses (a.u.)    14.149167            15.693032            15.822496
 VIB|Frc consts (a.u.)     0.000107             0.000316             0.000881
  ATOM  EL             X     Y     Z        X     Y     Z        X     Y     Z
    56  O            -0.07  0.27 -0.00    -0.30  0.74  0.08    -0.01 -0.19  0.06
    57  C            -0.17  0.65  0.13    -0.16  0.23  0.04    -0.17  0.11 -0.05
    58  O            -0.22  0.64 -0.05     0.13 -0.50  0.13     0.14  0.09  0.94


I wanted to know the best way to get rid all but one of these imaginary frequencies? The first thing I'm attempting to do is run a TRANSITION STATE calculation on the structure.
From what I read on this board another option would be to change some of the XYZ coordinates by using adding a fraction of the atomic displacements to the coordinates from one of the imaginary modes and then run TRANSITION STATE calculation on the new structure.

Or should I change the coordinates (using the atomic displacements) and then use the new structure to rerun the NEB and then perform the vibrational analysis again?
I've included my CI-NEB, TRANSITION STATE, and VIBRATIONAL ANALYSIS inputs.
 
Any advice on what else I could do to get rid of all but one of the imaginary frequencies (or if I'm taking the right steps already) would be appreciated.

Thanks,

Natalie

[S Ling]

Hi

I think running a TRANSITION STATE calculation would be a good idea, but you may need to use tighter convergence criteria for energies (EPS_SCF) and forces (MAX_FORCE). And also, you may specify a vector (see &DIMER_VECTOR subsection) for the dimer method, and a good initial guess could be the eigenvectors corresponding to your first imaginary vibrational mode (with the most negative frequency).

SL

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To post to this group, send email to cp...@googlegroups.com.
Visit this group at https://groups.google.com/group/cp2k.
For more options, visit https://groups.google.com/d/optout.

[Natalie Austin]

Thank you SL I will try you suggestion and let you know how it goes.

[S Ling]

Hi

In addition to what I have suggested in my previous post, you may double check whether the first imaginary mode with the most negative frequency is exactly the mode that you want to follow (based on your chemical intuition) by visualising the vibrations using MOLDEN. If not, then perhaps you should follow another mode.

SL

On 24 January 2016 at 22:33, Natalie Austin <natalie...@gmail.com> wrote:

Thank you SL I will try you suggestion and let you know how it goes.

--