I've a few structures which are optimized by a semi empirical method, but by using these optimized structures, if I run the b3lyp calculation with 6-31g basis set, the structure will terminate with a Gaussian error, like berny optimization, error in internal coordinate system or some time link 9999.
Can anyone explain what is the main reason for this?
Dear Chethan,
With the amount of information you provide, I think it will be difficult for others to help diagnose your problem. There are a number of things you could check, such as if the semi empirical calculations really finishes do to finding a minimum, or do they exceed their default optimization cycles. If your molecule is a sugar then consider using AM1, which should treat the intramolecular hydrogen bonding well.
I would guess that the B3LYP run is having trouble due to the input geometry. From what I have read from others, Berny optimization may fail due "problems" with the coordinates, including having three atoms being collinear. However, there could be other geometric reasons.
If you are indeed modeling sugars, consider using B3LYP/6-311++G(2d,2p) [1].
1. Jenn-Huei Lii, Kuo-Hsiang Chen, Norman L. Allinger, Alcohols, ethers, carbohydrates, and related compounds. IV. carbohydrates, Journal of Computational Chemistry Volume 24, Issue 12, pages 1504–1513, September 2003.
DOI: 10.1002/jcc.10271
Best regards,
Karl
Dear Karl,
Thank you for your suggestion.
I am working on Isomerization of glucose to fructose using inorganic catalysts using G09, it is a 54 atoms system. Most of the time, three atoms will collinear and leads to error. But as explained before, I am using the optimized structure obtained from Semi-empirical (pm3) as a input structure for B3LYP calculation.
I hope i will get optimized structures with the suggested basis set.
Thanks
Chethan
Dear Sinisa
Thank you. I considered the changes should be made in my input file as well as in structure suggested by you and Karl. Insted of using Z-matrix, am using "opt" and i frozen the bond angles as well.
Thanks
chethan
- Badges
- Science topic