# Can one figure out an optical absorption spectrum from EPR data?

I'm wondering if it's worthwhile for me as an optical spectroscopist to read up a bit on EPR. Do EPR signals reveal where, in terms of wavelength, features in the optical spectrum can be expected to appear? I know it tells me lifetimes of states, but that's about as far as I have dug into it. I have some EPR data on a substance and I wish to figure out where to look for it's absorption with UV/Vis/NIR means.

• To my knowledge, EPR signals cannot generally be used to predict UV-Vis-NIR peak positions. – MarcoB May 31 '15 at 8:14

No, it shouldn't be possible, based on where these spectroscopies lie in the electromagnetic spectrum:

The most common EPR experiments occur at 9.5 GHz (X band), which is in the microwave region. Even the highest-frequency experiments just barely break into the far infrared. Assuming you're doing an X band experiment, and that your system has a transition very close to the free electron $g$-values, 9.5 GHz corresponds to ~0.3 cm$^{-1}$.

For UV-vis, taking the longest wavelength of light in the visible region as 800 nm gives 12,500 cm$^{-1}$. This can be interpreted in two ways:

1. The EPR experiment needs ~42,000x more energy to probe even the lowest possible electronically-excited state.

2. Even the UV-vis transitions with the lowest possible energy will not show any fine structure, like in the sodium lines or benzene vibronic coupling. The energy scales need to be much closer. In the case of the spin-orbit-induced splitting of the sodium spectral lines, the unperturbed gap is 2.1 eV and gap between the split states is 0.0021 eV; this is 3 orders of magnitude between the energies. In the case of the benzene UV spectrum, the gap between peaks is ~10 nm, with the peak center at ~245 nm; this is 1-2 orders of magnitude:

$\hspace{20ex}$

Our EPR/UV-vis example has 5 orders of magnitude difference between the two energy scales. There is (hyper)fine structure in EPR:

which here, for DPPH in toluene is ~0.002 Tesla, or 0.35 GHz, so 1 order of magnitude.