I'm not sure about a negative oscillator strength, but the only physical way to get a negative wavelength is for there to be a lower-energy configuration possible for the system; then a transition to that state might well show up as a negative wavelength (since it would correspond to an emission rather than an absorption).  I assume that you are not confused between the oscillator strength and the Slater coefficient, since it is quite common for the latter to be negative; if I find anything on negative oscillator strengths, I will reply again.

Thank you for your reply. I will check if it is a TS and let you know

Hi

Dear Rudra

By definition oscillator strength is propotional to square of dipole transition

And its physically imposible, please tell more about your code and computational approach, 

Also part of your output

Bests

i have done hessian calculation and its not showing any negative frequency for ground state. suggest me something   

If you have a negative transition energy from a TD-DFT calculation, then you will also have a negative oscillator strength because the transition energies enter in the conversion from transition dipole moments to oscillator strengths. As Ted Lorance points out, you can only get a negative transition energy if there is a state lower in energy than what you are using as your reference state.

A 'hessian calculation' is a calculation for the vibrational frequencies of a molecule. A TD-DFT calculation is a calculation for the various electronic states of a molecule.

The reason you are getting negative transition energies is (most likely) due to the use of a functional that is not appropriate for excited state calculations. Try B3LYP, PBE0, CAM-B3LYP etc., a functional with some Hartree--Fock exchange in it. You are not telling us what your keywords are and we do not know the molecule either.  If you have a transition metal (high spin) with open shells, then the choice of functional/basis set becomes a real issue. It is quite easy to get negative transition energies on a high spin transition metal compound, and you have a serious computational problem on your hands.

Mine is a simple organ selenium system with no metals. I am using cam-b3lyp with 6-31+G(d,p) basis set. I have used hessian calculation to confirm that the system is a local minima and not a TS. Could you please suggest what I should do

Wow, this is getting weird. The functional is  certainly reasonable for TD-DFT. Try deleting the '+' and 'p' functions from the basis set, i.e. try 6-31G(d).

I presume your molecule is a closed shell singlet in the ground state. The cam-b3lyp calculation apparently finds that an 'excited' open shell singlet state (you wrote that there was an oscillator strength printed in the output, so it must be a singlet and not a triplet state) has an energy lower than what  was assumed in the geometry optimization. You can see from the output (the lines with arrows and orbital numbers, indicating single excitation from orbital XX to orbital YY) how the cam-b3lyp thinks the electron occupancies should be.

Thank you for your reply. I will do it then let you know

Rudra Sahu , Bhai did you find an explanation?

Karsten Krogh-Jespersen Sir, I also found the same from TD-DFT calculation using cam-b3lyp/6-31+g(d)]

Singlet to Triplet (TD-DFT)

Excited State 1: Triplet-A -1.0201 eV -1215.36 nm f=-0.0000 =2.000

286 -> 287 0.69640

This state for optimization and/or second-order correction.

Copying the excited state density for this state as the 1-particle RhoCI density.

However DFT calculation using b3lyp/6-31+g(d) suggest that ground state EE(triplet) - EE(singlet) = +5.7 kcal/mol.

Rudra Sahu Bhai I am facing the same problem?

Excited State 1: Triplet-A -0.7641 eV -1622.59 nm f=-0.0000 =2.000

Did you find any way to solve that problem?

Hello Jhon,

You can take second transition state (the positive one) as your absorption wavelength because as default Gaussian calculates 3 transition states. Neglect the negative transition absorption wavelengths. Why is it coming, this answer has been given already in above comments. You can take the positive absorption wavelength for your molecule. Else you repeat the TD DFT calculation with different functionals.