Dear Expertise
I am trying to calculate the reaction rate, for example Li+ ion deposition or dissolution rate from Li metal surface to liquid electrolyte using CI-NEB method in VASP. Here are my steps:
1. Optimizing the initial and final geometry.
2. Determining and optimizing the intermediate images for transition state search using the CI-NEB method.
3. Determining the imaginary frequency (single) of the optimized intermediate images to confirm the transition state.
4. Determining the rate constant from the free energy difference between the transition state and initial state using Erying equation ( k = (kBoltz*T/h)exp(-DeltaGdagger/RT)).
Now, I want to know
1. Are my steps are okay?
2. When I am calculating the imaginary frequency should I calculate it for each of the intermediate images or only for the one with highest energy?
3. And for calculating the frequency my usual settings of INCAR are IBRION = 5; NEFREE= 2; POTIM= 0.02 etc. Are they ok?
Any help would be appreciated. Thank you.
Hello Arindam Sannyal,
Globally, your methodology seems good. Some details though:
The calculation of the imaginary frequency to characterize the transition state has to be performed on the geometry associated to the saddle point. From the CI-NEB method, you can perform a quasi-newton (IBRION=1) to make sure you caught the transition state (start from the structure that looks the closest to the transition state structurally and energetically - i.e. highest energy point; and run the quasi-newton calculation). Once the quasi-newton reached the saddle point, run the frequency calculation with your computational setup that looks fine. If you have the transition state, you should have one imaginary frequency.
In principle you can then use the Eyring equation. If you assume that the entropic contributions can be neglected, you can then identify the free enthalpy of activation with the energy of activation (vasp energies difference). Nonetheless, for such processes (that can be compared to adsorption/desorption processes) entropic contributions are usually not negligible. You can estimate them using statistical physics (loss of 1 translation degree of freedom upon deposition).
Dear Romain Réocreux
Thank you very much for the details. As you can assume I am new to VASP, additionally I would like know that for the frequency calculation say if I have more than one imaginary frequency or all positive frequency, it means my transition state is not well optimized, right? In that case should I perform additional optimization of saddle point?
And another query is that for frequency calculation particularly to which atom should I consider the degree of freedom (selective dynamics) and how many (to which direction)? Only the adsorbed/ desorbed atoms or as I considered for the quasi-newton (IBRION=1) run?
Thank you again.
Dear Arindam Sannyal,
You can only perform frequency calculations for fully optimized structures (with quasi-newton for example): do not freeze the atoms that are supposed to move in the course of the elementary step, otherwise you won't be able to compute the imaginary frequency.
If you have many imaginary frequencies, it means that you haven't caught the transition state (= first ordre saddle point), but rather a higher order saddle point (which means higher in energy and therefore irrelevant).
Good luck with transition states search!
Dear Romain Réocreux
Thank you very much for your prompt clarification.
Dear Romain Réocreux
I need another query regarding the nebbarrier.pl script file. Actually when I am compiling the script the neb.dat file gives distances between each images as zero. Would you please inform me why the distances between each images is zero for every case?
Thank you very much
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials
- Carbon Materials
- Graphite