I want to simulate the vibrational frequencies of a cluster of molecules (2, 3 or 4 molecules) connected by Hydrogen bond in Gaussian Software. But I am getting an the following error message.
"Error in Internal Co-ordinate System".
Then I tried doing simulations in Cartesian co-ordinates giving the required syntax but it gave an error "Error termination request processed by link 9999."
I would be grateful if anyone can tell me how to simulate such Hydrogen bonded clusters using Gaussian. Thanks in advance.
You should play with initial configuration--see that some dihedral is not near 180, or some angle 0 or 180, or some distance is too short, or..
Link 9999 might mean that the convergence is not attained--in such a case just restart the job with option "restart". It also depends on the method--HF or DFT or something else; also on the set of basis functions.
You should play with initial configuration--see that some dihedral is not near 180, or some angle 0 or 180, or some distance is too short, or..
Link 9999 might mean that the convergence is not attained--in such a case just restart the job with option "restart". It also depends on the method--HF or DFT or something else; also on the set of basis functions.
Dear Shaumik,
If you are getting a "link 9999" error, that signifies that the optimization process is stopped as maximum step limit is exceeded before reaching local minima(As Vladimir already mentioned). You can search for the phrase "Optimization stopped" in your file,you'll find something like this:
Optimization stopped.
-- Number of steps exceeded, NStep= 100
-- Flag reset to prevent archiving.
GradGradGradGradGradGradGradGradGradGradGradGradGradGradGradGradGrad
--------------------------------------------------------------------------------------------------------------------
Which indicates that your structure doesn't reach to the local minima within the 100 optimization steps.(This default step number may be different in your case).There are multiple way to get out of this problem.First one is the restart option(as Vladimir already mentioned) via opt=(restart) option but make sure that .chk file should be alright and both the input file and .chk file should be in your working directory where the job is running otherwise you'll get an error.If any case this option doesn't work,you can always open the output file (.log) via gaussview,which is by default always open the last structure,you can resubmit again by making a new input file from there but it may take a bit more time than "restart" if you calculate the force constant (via calcfc or calcall) again rather than doing "readfc" from the checkpoint.
A third way is using "maxcycle" keyword before submitting the job for the first time(In the initial input file,not while restarting),which will increase maximum number of optimization steps.A sample segment will look like these
opt=(calcfc,cartesian,maxcycle=300) .......
details can be obtained from Gaussian website
http://www.gaussian.com/g_tech/g_ur/k_opt.htm
Hope this helps.Best of Luck.
I agreed with Vladimir comment. Some atoms are crashed in your molecular system. You can resolve this issue using GaussView. Open your input file using GaussView and adjust the dihedral angle (Right click - Bulider -> Dihedral angle) of the crashed atoms or groups. Then save (FIle -> Save) this into new file and recalculate it
Mr. Ray,
The optimization steps are the same like trhose one involving a single molecule computation. You can use also imput coordinates from crystal structure refinment such as the shown example (attachment).
I believe you have the same problem as in the attached question.
https://www.researchgate.net/post/How_can_I_circumvent_Bend_failed_for_angle_errors_in_Gaussian
- Badges
- Science topic