Question around CI-NEBs and the saddle point

Vasp transition state theory tools

Moderator: moderators

Post Reply
BuddyMacaulay
Posts: 1
Joined: Mon Aug 23, 2021 6:58 pm

Question around CI-NEBs and the saddle point

Post by BuddyMacaulay »

Hey, I've been working on some NEBs for a group of fairly finnicky systems for a while now and have a few questions that I'm a little unsure about.

Generally, for most of the NEB's the results look nice and converge well, with the climbing image finding a visually reasonable Transition state.

1.
I'm about to run some NEB's on a process which i believe may have a fairly stable intermediate and not transition state at the midpoint.
Will running these calculations with a climbing image result in poor convergence/incorrect intermediates or will it work completely fine?
If i shouldn't use a climbing image, is there any suggestions as to anything else I can do with computational resources being somewhat limited.

2.
With another bit of work I'm trying to show the relative energies of a specific pathway alongside a different (expect higher energy) pathway . However obviously running a neb on the higher energy path results in the MEP being found. Was wondering if you had any suggestions to force this higher energy path to be found.

I'm assuming for both these cases, running two separate nebs of initial->midpoint then midpoint->final would likely be an idea although i'm not sure the midpoint for the higher energy pathway is very stable and I would be able to get a reasonably converged energy.

Thanks In advance for your suggests. I can send some images of my structures and expected TS/intermediates at a later point if that'd help.
- Buddy
graeme
Site Admin
Posts: 2256
Joined: Tue Apr 26, 2005 4:25 am
Contact:

Re: Question around CI-NEBs and the saddle point

Post by graeme »

Hi Buddy,

1. If you have an intermediate minimum, I would recommend finding that first and then running two NEB calculations (as you suggest). It's not that the climbing image will not work with an intermediate minimum, it just that you will need more images to resolve a multi-step path, and there are annoying situations such as where the path goes into an intermediate minimum and back out following the direction that it entered, causing a 180 deg turn in the path.

2. You would have to say more about what makes the higher energy path a well-defined path. If there is a higher energy saddle, the NEB can be used to find it. If there is no saddle corresponding to the higher-energy path, I imagine that there should be some way to define it, either by symmetry or some constraint. If so, you might be able to apply the constraint to the NEB. I might be able to give more specific suggestions if you provide more details.

Graeme
Post Reply