Skip to content

Friday, 10. January 2020

To Do

  • Read up on SAPT
  • Check submitted jobs
  • Check that N(SOOCF3)2 has converged
  • Submit optimisations for metals
  • Run HF theory with #p HF/aug-cc-pvtz scf=tight pop=full
  • Download Wine 5.0Rc4

SAPT config:

SAPT2+3 16 procs 128Gb ram aug-cc-pTZV


What I did

Today I moved on to the HF single point energy calculations and ran them at #p HF/aug-cc-pvtz scf=tight pop=full. I wrote a script that took the xyz coordinate file and turns it into a gaussian script.

The SCF wasn’t converging for nickel though, so we had to run #p HF/Gen scf=(tight,MaxCycle=300,QC) pop=full to compensate.

Palladium turned out to be a bit too good of a catalyst and ended up completely breaking the \(\ce{C#C}\) bonds almost instantly. we decided to run a series of jobs with Pd atom at increasing distances form the ethyne to try and see if we can get it to optimise stably enough to run single point calcs on.

So I stayed back until 7:30 trying to get gaussian to not crash, but once I left, it stated to just not converge instead.

I’m on the train now (very tired) and am trying to coax them back to life with QC… They’re not segfaulting straight away at least, so that’s a nice start

Okay, one just did 😑

What I really need to do

One of the biggest challenges facing me next is how to pull the useful MO data from the HF jobs, to be able to compare their energy shifts. while it’s pretty easy to to just grab the MOs and chuck them in a CSV, we only really care about the change of the HOMO and LUMO on the neutral species, but the list of MOs will include that of both molecules.

NBO will tell you which MOs correspond to which fragments, however this approach also seems to modify the energies respective to the fragments, so won’t give us the numbers we specifically need.

NTF2 is optimising, but it’s very slow -_- so we’re still waiting for that…