High resolution laser techniques were developed that address the problems listed above. The techniques are collectively called site selective laser spectroscopy. They relied upon the idea that a narrow band laser could be tuned to excite selectively an absorption line of a specific component or site within a sample so the resulting fluorescence spectrum came only from the site or component excited. One could simplify spectral congestion with this approach. One could also eliminate broadening that was caused by inhomogeneities in the samples because excitation within a broadened line would only excite components that had energy states resonant with the laser so the resulting fluorescence spectrum would not reflect the inhomogeneities. These methods were applied to matrix isolation, low temperature organic glasses, Shpol'skii systems, inorganic analysis using precipitates, and supersonic jet spectroscopy to measure a variety of inorganic and organic materials at ultra-trace levels. The methods all relied upon sample fluorescence and they failed if the samples were non-fluorescent. We have recently shown that there is a new family of high resolution laser spectroscopies that have the same capabilities but do not require a fluorescent sample. These spectroscopies are based upon nonlinear mixing where several tunable lasers are focused into a material and new frequencies are formed at all of the sums and differences of the original laser frequencies. This nonlinear mixing is resonantly enhanced when some of the laser combinations match resonances of components in the sample. The nonlinear mixing can be used to perform atomic spectroscopy and molecular spectroscopy. We will concentrate on molecular spectroscopy in this discussion. One can perform component selection by tuning the lasers to match resonances on one specific component in the sample. One would then expect to have that component contribute dominantly to the nonlinear mixing. One can also eliminate inhomogeneous broadening by tuming the lasers to match the resonances of specific sites within the inhomogeneously broadened line. Again, one would expect that those sites would contribute dominantly to the mixing and the nonresonant sites would be discriminated against. We have tested these ideas in several model systems using four wave mixing spectroscopy. The two model systems are pentacene doped into p-terphenyl crystals and pentacene doped into benzoic acid crystals where p-terphenyl was added in small amounts to introduce controlled amounts of inhomogeneous broadening. The experiments were done at 2 K to eliminate thermal effects. There are four schemes that one can use to establish resonances with the pentacene molecules. In all of them, one establishes resonances with the vibrational levels, the excited electronic states, and the vibrational levels of the excited electronic state (which we will call vibronic states). The four schemes differ in which states are involved in the resonance associated with the emitted light. If the emitted light involves transitions between two levels that are not initially populated, the technique is classified as a nonparametric process. If one of the levels were initially populated, the technique is a parametric process. Theories for the ideas predict that each possibility will have a different ability to provide selectivity in the measurement. The pentacene in p-terphenyl system was studied first. Pentacene has four different crystallographic sites in this crystal, some of which differ only slightly from each other. Conventional spectroscopy shows that the transitions from each
Read full abstract