I am new to Oniom calculation available in the Gaussian 09 package. In one of my problem I need to find TS and I have used the command
# oniom(m062x/6-31g(d):m062x/3-21g) Opt=(ts,noeigentest,maxcycle=270,modredundant) freq
in the route section, but still after 270 cycles it doesn't converge. Please help.
Dear Swarnadeep Biswas ,
First of all, if you are using ONIOM, why do you apply a QM:QM scheme in which you use DFT in both cases, just changing the basis? The ONIOM model is useful to deal with large systems, but only if the applied level for the low layer is something like semiempiricals or even molecular mechanics. With the model you are using, the calculation will be even slower than using the M06-2X/6-31G(d) for everything! As the main idea of the low layer is that the relevant chemistry does not take place there, and only exerts a mild influence on the critical region (the high layer), you can reduce the level of theory there. These are two possible alternatives:
1) oniom(m062x/6-31g(d):pm6) A QM:QM approach, but using a less computationally demanding semiempirical for the low layer.
2) oniom(m062x/6-31g(d):uff=qeq)=embedcharge A QM:MM approach, using the UFF force field for the low layer, generating the MM partial charges for all atoms (qeq, so then don't have to be read from anywhere) and using the electronic embedding (embedcharge) to account for the QM-MM interactions.
That being said, to calculate a transition state you don't need the modredundant keyword, so you can delete it. But what you need is to get the force constants at least at the first point, so you (the program) can identify which one is negative and follow that one to reach the transition state. This is achieved with the calcfc keyword. The calcall keyword computes the force constants at every step of the optimization, which is usually more adequate, but for the ONIOM model yields exceedingly large outputs, so just go for the calcfc. Last, the noeigentest option is useful at the beginning if you have more than one negative force constants, i.e. more than one imaginary frequency, otherwise it is not necessary.
Thus, try to use opt=(calcfc,ts) to begin with and, if the problem with the number of imaginary frequencies arises, then switch to opt=(calcfc,ts,noeigentest).
Hope you find it helpful
In addition to Mr. Enrique,I suggest you to add:
1- scf =(qc,maxcyc=1028,novaracc,noinfock).
2- Try first with b3lyp/sto-3g.
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology
- Sectioning