CI-NEB calculations

[Nasir Ahmad Rajabi]

Hi dear CP2K users,

I'm quite new in CP2K; I am doing CI-NEB calculations to locate transition state, but I don't know why after 24 hours calculations in ARCHER, it dose not converged. 

I checked the output file and found these:

 *******************************************************************************

 RMS DISPLACEMENT =   0.00024     [ 0.00010]                        ( NO)

 MAX DISPLACEMENT =   0.00152     [ 0.00020]                        ( NO)

 RMS FORCE        =   0.00024     [ 0.00030]                               (YES)

 MAX FORCE        =   0.00152     [ 0.00080]                               ( NO)

 *******************************************************************************

I am wondering how to deal this problem; so could anyone with the same problem help me?

The input file is attached.

Thanks,

Nasir

[Conrad]

Hi Nasir,

I'm not familiar with NEB, but I think one of the first responses you'll get from someone more knowledgeable will be to request a larger section of your output file. 

Additionally, what is the system you are studying? How many atoms are there?

You ran on ARCHER, but how many cores did you use? 24 hours may not necessarily be unreasonable. One would need more information to make that judgement.

Best wishes,

Conrad

[Nasir Ahmad Rajabi]

Hi Conrad,

Thanks for the quick reply. I am working on a 15 atoms Rh complex as a test, so this kind of system is not large to not to be converged in 24 hours, I assume. 

The type of function that I'm using is PBE and the box size is large enough; I did plot the energy profile, it looks quite reasonable; and transition state structure is what I want as well.

I attached these two files for more clarification ( Energy Profile and Trajectory File).

Thanks,

Nasir

 BAND TYPE                     =                                          CI-NEB

 BAND TYPE OPTIMIZATION        =                                            DIIS

 STEP NUMBER                   =                                              63

 NUMBER OF NEB REPLICA         =                                              16

 DISTANCES REP =        0.413103        0.417538        0.427866        0.450718

                        0.490582        0.551013        0.624194        0.679018

                        0.698048        0.710425        0.414412        0.434663

                        0.466947        0.476874        0.472343

 ENERGIES [au] =     -156.225794     -156.226224     -156.226055     -156.225758

                     -156.225333     -156.224749     -156.223892     -156.222365

                     -156.218487     -156.209512     -156.203999     -156.205623

                     -156.207505     -156.208043     -156.208360     -156.206615

 BAND TOTAL ENERGY [au]        =                            -2499.46535599070148

 *******************************************************************************

 *******************************************************************************

 RMS DISPLACEMENT =   0.00023     [ 0.00010]                               ( NO)

 MAX DISPLACEMENT =   0.00116     [ 0.00020]                               ( NO)

 RMS FORCE        =   0.00023     [ 0.00030]                                     (YES)

 MAX FORCE        =   0.00118     [ 0.00080]                                    ( NO)

[Marcella Iannuzzi]

Dear Nasir,

have you considered the possibility that your convergence criteria are too tight?

Is there anything relevant that is still energy in the reaction path after 63 NEB iterations?

You could check the total energy of the band, or the energy and structure of the TS, the energy barriers.

If nothing is changing anymore, you can safely stop your NEB simulation.

regards 

marcella

[Nasir Ahmad Rajabi]

Hi Marcella,

Thanks for your comment. I would use a less strict SCF convergence criteria to see what happens.

Thanks, 

Nasir