There are lots of options to perform excited state calculations in Gaussian 09

For example, td, cis, and cassscf etc.

Nowadays, the most of excited state calculations have been performed td option.

It shows better performance and time consuming than other options.

You can find information in Gaussian web site

Hope this help

I want to do some geometry optimization in the excited state. Anyone please show how to write the input commands.

This is how you calculate TD-DFT geometry optimization for the 1st excited state using Gaussian

# opt td=(nstates=3,root=1) b3lyp/6-31g(d,p)

I am using Gaussian09 with TD-DFT geometry optimization for the 1st excited state of a system contains 29 atoms. I use the following route section as Muath Nairat answerd:

# opt td=(nstates=3,root=1) b3lyp/6-31g(d,p)

I would like to know what is meaning of (root=1) . I run the program since two days and it is still working. I do not know why it takes a lot of time.

Safiya Amer,

root=1 means that you are using the first excited state gradients to be optimized. in other words, you are looking for the first excited state minima. If you were to use root=2 then you would look for the second excited state minima.

In general, I wouldn't recommend td-dft for finding excited state minima, this is especially for systems where there is a large change in character between the Franck-Condon geometry and the excited state minima. The problem that you are facing is very common, this is when two states are close in energy and the optimization keeps flipping between the states and never converges. You can examine your output and see if the optimization is just switching between two states with small changes in energies within the last steps. Sometimes this can be solved if you increase the number of states that you are solving (nstates=6). If this didn't work then you have to use a multireference method such as casscf with state averaging to make sure that you are actually finding the minima of the state that you are interested in. Please keep in mind that casscf is not a black-box like td-dft and it is way more expensive in terms of computational cost

good luck

Muath Nairat

Thanx a lot for your clear reply.

Hello There, Use this

# Opt td (nstates=n, root=r) level_of_theory/desired_basis_set

and its done.

Regards

Ajay khanna

I want to do some geometry optimization in the ground state. Anyone please show how to write the input commands.

the way you explain things is indeed appreciable. can you please explain me that how to determine exited state of interest for analysis Muath Nairat

Try a quick check with a TDA-DFT then try a TD-SCF-DFT if the energy gap between the states is big enough. If not, a quick check can be to use a CI(S) and compare the results.