Hello,
I am trying to get optimised geometries of brominated benzenes from Gaussian09 using CIS and TD methods. Both gives unexpected shoter C-Br bond lengths (around 1.86 A), even shorter than bond corresponding distances in the ground state (around 1.9 A) , Experimentally, bromobenzene degrade by loss of Br via photo-induced decomposition so shortening the C-Br bonds in the optimised first excited state, in reference to the ground state is not what you would expect.
If I use the newly added optics feature to Dmol3 to carry out optimisations for the first excited S1 state, the C-Br distances are elongated to 2.10 A which makes a sense. Turbomole gives similar results. I am just wondering if any one has experienced similar situation when using G09 to option geometries for halogenated hydrocarbons in the excited state.
Mr. Altarawneh,
You have received this deviation, because of most probably you have not performed a preliminary optimization of the geometry in ground state.
Please find as attachment an example of CH3Br. First, the geometry have been computed at a given suitable level of theory, and afterward this geometry has been used as initial one to obtain excited state properties.
Hello Bojidarka
Thank you for your answer, much appreciated. I started from optimised S0 state of bromobenzene and I still have a short bond length in the optimised S1 state when compared with the S1 state. My results are consistent with Chem. Phys. Lett 381, 2003, 617 for mono bromobenzene.
Optimised geometries for S1 brominated diphenyl ethers from Gaussian (as in ChemPhysChem 14, 2013, 1264) are very different from corresponding S1 geometries obtained by Turbomole (as in Chemosphere 88, 2012, 33). Turbomole (as well as Dmol3) gives significant elongation aromatic C-Br bonds (which what you would expect), whereas Gaussian give slightly shorter lengths. I could reproduce your values for CH3Br but the situation is a bit different for aromatic brominated compounds. I am not sure what is the source of discrepancy between Gaussian and Dmol3 for these particluar compounds.
Mr. Altarawneh,
Since the quentum chemistry is based on exact methods within the frame of given theoretical levels you should not obtain any difference, nevertheless which software package you have used.
In this respect please find as attachment the optimization of the C6H5Br, using Gaussian.
At first we can perform an equivalent of 'interlaboratory variability', meaning that I am attaching you a gaussian input file of C6H5Br. You can run it using your version of the Gaussian. You should obtain the shown results.
Than you could run calculations of the ES using Turbomole, using however the output geometries as they are given for the ground state at the same level of theory (the computations are in the gas-phase). This could be equivalent of something like 'inter software variability'. Because of there have numerouse studies correlating the eaccuracy of the theoretical methods, but there have not correlation studies between the different softwares as well as versions within the frame of one and the same software. In this way every calculation performed/reported with given software should be verify (or validated in terms of the analytical chemistry) and eventual deviations should be known.
Hi Bojidarka, Many thanks, I will run the output file from Gaussian into Turbomole, your help is greatly appreciated,
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology
- Research Papers