I have performed absorption calculations for some molecules using TD-DFT in Gaussian 09 program. As I know oscillator strength values should be in the range of 0-1, but when I performed absorption spectra calculations I have observed oscillator strength values more than one. What I am supposed to do to get accurate oscillator strength values? Please suggest me regarding this. Thank you.
Dear Sai Sudhakar,
What basis set was used in your calculation? some times lower basis set in DFT calculations yields poor results. If you don't mind, send the output file and molecular structure.
by
Karthick
Dear Karthick, I have performed our calculations with 6-31G(d) basis set.
While that does sound large, the Thomas–Reiche–Kuhn sum rule is that the sum of all oscillator strengths is equal to the number of electrons, so i think it is not impossible. Note: The functional also makes a difference here too.
Dear Covington, Which functional I supposed to use for thiophene based materials ?? I have used higher basis sets also by adding polarization and diffusion functions. Still i have got the oscillator strengths more than one .
Is this a polymer system? That may be the problem. You have to be careful of semi-large systems in DFT if your functional is not long-range corrected. Look into Spurious Charge-Transfer Excited States. There are several several functionals that are long range corrected and should give better (though not necessarily more correct) results. Cam-b3lyp is a good place to start, though there are better ones.
Yes we have done absorption for oligothiophene molecule. We have used long range corrected functionals like CAM-B3LYP and wB97X-D. Still we observed same oscillator strengths (more than one) as like other functionals B3LYP and M06.
Did you use a better basis set too like 6-311+G** or Aug-cc-pVDZ? If you can find some experimental results that show a similar strong absorption at a similar wavelength, for a similar system, then that would be best. I have looked through some of my old calculations and found oscillator strengths greater than one also, but I always simulate the absorbance spectrum and compare to experiment so i never noticed it.
No I did not use since my absorption values are close to the experimental values when I used 6-31G+(d,p).
I agree with previous post mentioning Thomas–Reiche–Kuhn rule above. See Berkowitz, Photoabsorption, Photoionization and Photoelectron spectroscopy, Academic Press 1979 Appendix 1 p 64 for a proof.
(sorry, I do not have more recent references at hand).
f values > 1 occur especially on highly symmetric molecules (or crystals). We got this recently on the D3d isomer of C4F6 (table 4 in P. Limao-Vieira et al. dx.doi.org/10.1021/jp307599y | J. Phys. Chem. A 2012, 116, 10529 −10538). Since most transitions are forbidde,n the f strengths are concentrated on one main peak. Nothing to worry about... On the experimental side, check work using EELS spectroscopy where f values can be derived experimentally to be compared to calculations.
I have done excitation calculations of supramolecular complexes at ground state and at ist excited state by using B3LYP/631G(d,p) and cam-B3LYP/631(d,p), respectively. I get good results and also in agreement with experimental ones.
See a blog of Felix Plasser for the answer
http://chemical-quantum-images.blogspot.com/2013/12/oscillator-strengths.html
- Badges
- Science method
- Similar topics
- Methodology
- Absorption Spectroscopy