Document Type


Date of Award



Raman effect, Fluorescence spectroscopy

Degree Name

Doctor of Philosophy (PhD)



First Advisor

K. Keith Innes

Second Advisor

Daniel D. Konowalow

Third Advisor

Clifford E. Myers


In 1961, Albrecht predicted that the normal vibration most responsible for vibronic mixing in a weakly allowed electronic transition should show the greatest activity in Raman scattering, and especially so in the pre-resonance Raman region. The pre-resonance effect is an increase of intensity of a Raman line as the exciting frequency approaches that of the weak electronic transition.

The 𝜈10a vibration of pyrazine is known to vibronically mix a weak 1B3u electronic state at 30,876 cm-1 with one (or more) strong 1B2u electronic states while the 𝜈4 vibration is vibronically inactive. The pre-resonance Raman intensity of the 𝜈10a mode relative to the intensity of the 𝜈4 mode was measured for six exciting lines whose frequencies varied from 50% to 80% of the 1B3u transition energy. The pyrazine was in the liquid state at 65°C for this experiment. The areas under the vibrational peaks were measured and corrected for the spectral responses of the Raman instruments. The ratio, I𝜈10a/I𝜈4, increases by a factor of 8 as the frequency increases over this range.

The primary approximation used in testing three pre-resonance Raman theories was that only one 1B2u electronic state was active in the vibronic mixing. Each of the theories, Shorygin’s semi-classical theory, Albrecht’s first-order perturbation theory and Peticolas’s third-order perturbation theory, was reduced to a single term by considering only the largest frequency dependent term and applying the primary approximation. The experimental data agrees within its error with the theory by Peticolas and the identification of the higher electronic state (1B2u) in the vibronic mixing with absorption at 60,700 cm-1. This choice is quite reasonable since the associated transition (1B2u - 1Ag) has the largest oscillator strength in the absorption spectrum.

In 1974, Brus and coworkers obtained a fluorescence spectrum of s-tetrazine from single-mode laser excitation. This spectrum was uniquely simple since the laser mode was so narrow (~0.0002 cm-1) that only 100 transitions in the Q-branch of the 6a10 band of the 18,000 cm-1 1B3u-1Ag transition were absorbing laser photons. The resolved fluorescence emission features were each the result of only about two transitions rather than the hundreds normally seen with large molecules like s-tetrazine.

As a preliminary step in understanding the emission, a rotational analysis of the 6a10 absorption band was performed. Tentative assignments of the sharp features in the P- and R-branches of the experimental spectrum at 300°K [from a photographic plate obtained by Merer and Innes] led to a value for the change in the average of rotational constants A and B. A band contour simulation program for asymmetric top molecules was used to calculate band contours which were compared to the experimental spectrum. The “best” calculated contour reproduced the sharp features well but it did not show the less regular structure in the P- and R-branches properly. The excited state rotational constants of the “best” fit were A′ = 0.22227 cm-1, B′ = 0.21621 cm-1 and C′ = 0.10984 cm-1.

These constants were used as a starting point to analyze several P-branch fluorescence spectra provided by Dr. Brus. Since the s-tetrazine was at ~1 torr the excited states were not collisionally deactivated. Also the experimental conditions indicated that the s-tetrazine absorption was near saturation. For the assignment of the laser mode position within the Q-branch and J values for the P-branch lines, and to reproduce experimental observations of the P-branch intensity pattern, the excited state rotational constants were varied to develop a calculated emission spectrum that approximated part of the experimental fluorescence data. This led to new rotational constants: A′ = 0.22262 cm-1, B′ = 0.21720 cm-1 and C′ = 0.10948 cm-1.The simulated absorption band contour from these constants shows substantial improvement in matching the experimental absorption spectrum as compared to the work described in the preceding paragraph.

There remain discrepancies between simulated and experimental curves. It is concluded that further work on this band will have to consider the possibilities of Fermi resonance or Coriolis type perturbation or that not all absorption transitions were saturated in the fluorescence experiment.