Issue with visualizing multiple frequency results in ORCA output

General Discussion
1

Hello,

I’m using the following ORCA input:

! r2SCAN-3c
! OptTS Freq
! CPCM(water)

%geom
 recalc_hess 5
 calc_hess true
end

*xyzfile 0 2 F83_TS_calcfc.xyz

The output file contains multiple sets of frequencies. When I open the file, Avogadrolists them all in a single ‘Vibrational Modes’ list. However, while the initial frequencies calculated at the start of the job can be visualized, the subsequent ones calculated later in the run are not working.

2

Yeah, this is a known issue: Frequency Parsing in ORCA - #2 by ghutchis

It’s partly a bug in the parser (because the code can only store one Hessian right now) - should have been only showing the final vibrations.

It’s also that there’s a need to store multiple vibrations across multiple coordinate sets. That’s on the drawing board for the 2.1 release.

3

Hi, thanks. Indeed, it’s the same issue. To be honest, it’s quite problematic; I want to ensure the frequency I found relates to the specific transition state I’m interested in without having to redo the frequency computations.

4

I understand - I hope to make some progress over the next week. It’s the classic 80-20 problem. We concentrated on the key features, but there’s a lot of other important features.

If you’ve got a few example / test files, it would certainly help.

5

Yea, would be cool. This limitation is quite impactful for TS workflows :frowning: I can’t share specific production files, but the behaviour is easy to reproduce with standard TS optimisation setups with just calc_hess :slight_smile:

6

Hi there, any news? :slight_smile:

7

May tends to be busy - I have a plan. The first step was making sure that we could store properties for each set of coordinates, which I merged yesterday:

The next steps will be a bit trickier.

  • Modifying the ORCA parser to record multiple vibrations (fairly easy)
  • Exposing a new signal when the conformer changes
  • Making the vibration code aware of multiple vibrations
  • Tweaking the vibration code so it doesn’t obliterate the trajectory set while still allowing animations

The last one is the real tricky thing. Right now, the code writes a set of 3D coordinate frames for the vibration animations. That’s kinda useful because some people want to project along a vibrational mode. But everything else in your use case requires that the conformer sets, etc. are retained.