I am computing the optimized molecular structure of an organic molecule in the Gaussian 09 computational chemistry package.
As I have it now, I do not think that the optimized structure that I am getting from Gaussian is correct, in part due to the symmetry of some of the bonds. To try and correct this, I want to fix two of the bond angles in the molecule in order to help/force it to produce a more believable structure. Even though I want to fix these bond angles, I want to leave everything else variable and open to optimization, including those bond lengths.
How can I specify this in the input file?
Fixing the bond angle to a set value seems like an extreme solution to this problem. Perhaps I can just constrain the angle within a certain range, is this possible?
After a lot of work and playing with settings, I have figured out how to fix bond angles, but now my difficulty is fixing them to 180 degrees (my desired value), for linear bonds. It seems that this is very hard to do.
Fixing the bond angle to 179 degrees (or 179.9 degrees), results in an error saying "Error in internal coordinate system." A lot of people online say this is because of linear/nearly-linear bonds in the molecule being optimized. Some people also say this might be avoided if I try to optimize in Cartesian coordinates, instead of internal coordinates, using the cartesian parameter after opt. However, I do not think I can use ModRedundant and cartesian parameters at the same time.
How can I fix a bond angle to be linear, but still optimize the rest of the structure in Cartesian coordinates? Is it even possible to fix angles to 180 degrees?
No comments:
Post a Comment