Two-dimensional pattern recognition methods for rapidly recording and interpreting high resolution coherent three-dimensional spectra.

High resolution coherent multidimensional spectroscopy has the ability to reduce congestion and automatically sort peaks by species and quantum numbers, even for simple mixtures and molecules that are extensively perturbed. The two-dimensional version is relatively simple to carry out, and the results are easy to interpret, but its ability to deal with severe spectral congestion is limited. Three-dimensional spectroscopy is considerably more complicated and time-consuming than two-dimensional spectroscopy, but it provides the spectral resolution needed for more challenging systems. This paper describes how to design high resolution coherent 3D spectroscopy experiments so that a small number of strategically positioned 2D scans may be used instead of recording all the data required for a 3D plot. This faster and simpler approach uses new pattern recognition methods to interpret the results. Key factors that affect the resulting patterns include the scanning strategy and the four wave mixing process. Optimum four wave mixing (FWM) processes and scanning strategies have been identified, and methods for identifying the FWM process from the observed patterns have been developed. Experiments based on nonparametric FWM processes provide significant pattern recognition and efficiency advantages over those based on parametric processes. Alternative scanning strategies that use synchronous scanning and asynchronous scanning to create new kinds of patterns have also been identified. Rotating the resulting patterns in 3D space leads to an insight into similarities in the patterns produced by different FWM processes.