Sometimes when I try to do a geometry optimization in Gaussian, the bonds to an atom end up broken and it acts essentially as a free atom - or at least that's what it seems to me. It ends up in weird locations, not to mention how much slower computation is. Last time this happened was with a benzofurazan derivative. The oxygen atom from the small heterocyclic ring was abstracted from the molecule and ended up moving freely. I was trying to make a TD-DFT optimization of the first excited state in vacuum. The ground state was optimized successfully. Is there a way to fix this problem?
Dear Maria
1- use the command geom=connectivity to specify the exact bond order between atoms. If you could not justify this command line, i can do that for you if you send the input file to me.
2- use a larger basis function.
That's what I did. I took the LOG file and saved it as GJF, fixing the keywords.
Check whether molecular specification contains bond and angles information and if so, check its correctness.
Mrs/Miss Dimitrova,
At the Sofia University 'St. Kliment Ohridski', at the Faculty of Chemistry and Pharmacy have 5 groups of computational and theoretical chemistry. All of their group leasders can answered of your question. The names are Prof. Galabov, Prof. Ilieva, Prof. Vayssilov, Prof. Tadjer, and Assoc Prof. (A.) Ivanova. So, you shall save lot time if contact them, instead to wait somebodey to answer you via Internet.
If you eventually affraid for your scientific idea?! Do not warry! Everybody of the staff of this University have own indepemdent and different (from anothers) research profile and comntributions, so that nobody does not 'waits' of the students' ideas. :)))) Furthermore the staff of SU supports the research even the academic staffs of other so-called 'universities' up to 54 numberds for Bulgaria. Therefore, your intelectual ownership (if this is the case) is ensured.
NB! The last paragraph is based on personal experience of 18 years.
Dear Maria
1- use the command geom=connectivity to specify the exact bond order between atoms. If you could not justify this command line, i can do that for you if you send the input file to me.
2- use a larger basis function.
I have included geom=connectivity, so it might be the basis. Thank you for the suggestion.
Actually using geom=connectivity shows a bound in graphical view of gview software. Setting this option has no effect on optimization results.For testing this idea set a bound between to H atom with 20 angstrom distance (using gview bond option) .After optimization this distance will remains but gview shows a bound for this two H atom because of setting geom=connectivity.I believe that this keyword can solve your graphical issues but it has no effect on optimization results.
Best
Its more convenient but it can creates more complicate problems.optimization with freeze bond condition depends on bond length and easily changes with arbitrary input data. I think artificial forces on a bond creates incorrect and arbitrary results.
Best
Hy Hamed!
If your molecule should have a charge, but you don't determine it in the input file (charge and multiplicity) it will broke up some bonds to correct the charge to the default=0. Geom=connectivity keyword should be used as well.
I hope it will helps!
Sorry, I pick up the wrong name :)
The last one is for Maria Dimitrova!
If the ground stated was fine, might it be that you picked up a dissociative excited state ?
In some cases, it helps to use Opt=(MaxStep=5) or Maxstep=10 to prevent the system from exploding. If the system is not too large, Opt=(CalcFC,NoRaman,MaxStep=5) is also an option to see if the system really wants to explode or if the atom leaving is an artefac of wrong forces at the begining of the optimisation.
Krisztián Csankó, I had the charge and spin multiplicity set correctly, it is a neutral molecule. On other occasions it has happened to me that wrong charge would lead to bond breaking.
Ujval Gupta, I tried optimizing the structure at:
#p opt freq rb3lyp/3-21g* geom=connectivity
but the bond was still broken.
Paul Fleurat-Lessard, I will test your suggestion.
Thank you!
Hai,
1. Try to prepare input file using z-mat of your own numbering and try optimizing.
2. When numbering your molecule try to follow standard numbering system
3.when you write input file you can fix your bond lengths so that it wont break unless there is any weak interaction prevails.
bye...
Even at STO-3G, the same bond between oxygen and nitrogen was broken. When I tried entering the Z-matrix manually, the optimization completed at 6-31G* successfully. The problem must have been caused by the input. I used a .MOL file from ChemDraw before.
Thank you for the suggestions!
Hi friend,
you are welcome always. Go ahead with your research works.
bye
HI,
Its always worthy to have a zmatrix to do any calculation. Because if we wirte a zmatrix formation to construct a geometry and to optimize or to do any calculation, the entire molecule will be in our control and we can do the things in the way we want. So I always prefer to have zmatrix and also suggest my friends to follow the same!......
Visit our web page: smmg.pl and download my Hauusian program.
May be Hauusian can help you in generation of proper . gjf file
Michal
- Badges
- Science topic