[ase-users] The Dimer method

[Michael Joseph Waters]

Hi Zainab,

I have used the dimer method extensively outside of ASE, but I'll try to answer your questions anyway,


  1.   The dimer method has no problem with periodic systems. I doubt the ASE implementation would have any issues with periodic systems.
  2.  In reaction coordinates, your displacement vector is for your total system, ie, 3*Natoms components. You're right, it's a good idea to only include the atoms that you think will move as an initial condition. Freeze any atom that you think will be sufficiently uninvolved in the reaction to speed up dimer convergence and help ensure you get the saddle you are interested in.
  3.  From this example (https://wiki.fysik.dtu.dk/ase/ase/dimer.html) the mask is just a flag set for each atom to indicate whether or not it is included in the dimer search. It seems redundant with the fix constraint, but I'm sure there is a good reason for both.
  4.  For most calculators, parallelization depends on the specific calculator that you use. Are you planning on doing this with an interatomic potential or ab initio technique?

Best,
-Mike


________________________________
From: ase-users-bounces at listserv.fysik.dtu.dk <ase-users-bounces at listserv.fysik.dtu.dk> on behalf of Alaithan, Zainab via ase-users <ase-users at listserv.fysik.dtu.dk>
Sent: Tuesday, January 14, 2020 4:15 PM
To: Ase-users <ase-users at listserv.fysik.dtu.dk>
Subject: [ase-users] The Dimer method

Dear ASE-users,

I have a converged neb transition state geometry that I would like to further optimise with the dimer method. I have several questions on this method.

1- I assume it works well with periodic systems ?

2- how to best define the displacement vector for a molecule rather that one atom? Should I include only the atoms involved in the reaction and make the displacement zero for the rest?

3- how is the dimer mask used in conjunction with the displacement vector?

4- am I right to assume that I don't need to parallelise this algorithm?

 I will appreciate any help.

Regards,
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20200114/7ab5301b/attachment-0001.html>