5
$\begingroup$

Background

Using Quantum Espresso examining a porous material, that has a ligand with an aromatic group. The aim is to rotate the aromatic group and determine an energy profile for the rotation. The initial (un-rotated) structure was optimized with 'vc-relax' and a subsequent SCF converges well. The aromatic group of the ligand is then rotated 90 degrees.

Problem

When trying to perform an SCF on the structure with a rotated aromatic core (ie. 90 degrees), the SCF does not converge and instead oscillates. I have played with a number of settings (ie. k-points, Ecut, mixing, smearing, etc), but with no-luck. The rotated input looks chemically sensible. However, it does occur to me, that the structure is not the lowest energy state, so essentially to perform an SCF on a transition state - which is likely to be unstable. Is this do-able and if so what is the best way to tackle the problem given the SCF is not converging?

$\endgroup$
2
  • 1
    $\begingroup$ +1 When you have such technical questions, it is recommended that you provide the input files you have used to run the calculations. In general, when facing SCF convergence problems and if you are sure that your structure is reasonable, then you are suggested to start with minimal input tags, find out which tags are causing the problem first and then you could work out a solution. As I said at first, share your inputs unless it is hard to address the problem. $\endgroup$ Dec 3, 2023 at 11:21
  • 1
    $\begingroup$ I'm not expert enough to answer fully, but I think you may be looking for the nudged elastic band (NEB) functionality of QE, neb.x. Here are slides from a 2019 tutorial about it. Probably you'll get a PES along the whole 180-degree rotation of the functional group. Otherwise, if you're confident that your 90-degree rotation is exactly where you want it, you can freeze the desired atoms in place by setting all three components of their respectiveif_pos values to zero in the ATOMIC_POSITIONS block, then run vc_relax as usual. $\endgroup$
    – elutionary
    Dec 7, 2023 at 17:20

0

You must log in to answer this question.

Browse other questions tagged .