Geometry optimisation with ORCA, starting with an input file containing the initial geometry in 'xyz' format, leads to unusual connectivity - disappearance of multiple bonds and appearance of bonds between atoms which aren't actually connected in the input geometry.
DFT as well as HF, post-HF, and even semiempirical methods not aware of the concept of bonds. You can fix bond length but then the result will not correspond to the local minimum. There are multiple reasons why the results of calculations do not fit to initial geometries or your expectations. Maybe geometry, charge, or multiplicity do not fit to reality. May be used method incapable to describe the system correctly. Unfortunately, my orbuculum is not showing your output.
Hi Oleg B. Gadzhiev and Dmitry Sharapa
Thanks for confirming that connectivity is not dealt with by DFT methods employed in ORCA (4.1 and older) as expected. Further, I have realized that molecular visualization programs such as PyMOL are not adept for geometry optimization trajectory (.xyz file) generated by ORCA, hence the geometry artefacts.
The problem, however, seems (yes, it was a fluke) to have been done away with by using slightly modified input geometry together with newer ORCA (>5.0), viz. no geometry artefacts while visualizing in PyMOL.
- Badges
- Science topic
- Similar topics
- Philosophy
- Ethics