try # b3lyp/aug-cc-pvTZ opt freq scf=qc

Hi Alexander

But the program got stuck at the same place....

I think this may due to insufficient memory. What is your operating system? If you use windows, there is problem to allocate RAM. I can't figure out exact problem, without having more details about your calculation.

I believe you are using G09 or higher. Do you realize that augmented-correlation corrected Triple zeta Valence has a very large number of gaussians that may become linearly dependent. Other softwares like GAMESS would tell you the problem before crashing, I don't know G09. This trouble get worse with the size of your molecule.

does it work with # b3lyp/cc-pvTZ opt freq scf=qc

Dear Lakmal

I am using Windows and memory for the calculation is 2GB

%nprocshared=8

%mem=2GB

# opt freq b3lyp/aug-cc-pvTZ scf=qc

The system contains 42 atoms out of which 1 Sulfur and 1 Nitrogen remaining are C,H, O.

Dear Victor

I am Using G09 and my molecule contains 42 atoms.

I got stuck once again at same place even if i use #b3lyp/cc-pvtz

You try with using 256 MB for this. Sometime it may work. I also had these type of problem. I just reduced memory and finally it worked. But I cant sure it for you. Anyhow, try that way. But this need more CPU time. So calculation get slow.

Dear Lakmal

Thank you, i will try and get back to you

I think you must know some stubborn facts about Windows; It can`t handle files larger than 16Gb under Win32-bit, at least that what they claim. In writting G09 instructions you must tell it how much RAM you have reserved (%Mem=2Gb). If you wish to use aug-cc-pvTZ for 42 atoms you have to take into account how much space does this calculations takes, and specialy how much disk is going to use. HUGE AMOUNTS: Get a linux x86_64 machine with 24 cores. Seriously.

Dear Victor. I think that's true. In windows system, there are some problems to reserve RAM. We can overcome this using lower RAM capacity, but it take longer time to complete. However, in Linux system, we can overcome this. I think your point is so important.

I get a 24 core x86_54 with 32 Gb ram and 1.5 Tb disk, however a molecule this size (42 atoms) and a basis set augmented-corelation corrected triple zeta, takes a hell lot of memory and TIME. It would be better to try cc-TZV or cc-DZV before jumping to a diffuse basis set. Chethan, can you try cc-DZV fust? Tell me how ot goes...

Are you using 64 bit system? It is 5 times faster than x86_64.