Optimizing sparse sampling for 2D electronic spectroscopy

Area of Science:

• Spectroscopy
• Quantum Chemistry
• Data Science

Background:

• Multidimensional electronic spectroscopy is crucial for studying ultrafast chemical dynamics.
• Traditional methods require long acquisition times, limiting their application.
• Optimizing data sampling and reconstruction is key to accelerating spectroscopic techniques.

Purpose of the Study:

• To develop a novel data acquisition strategy for action-based multidimensional electronic spectroscopy.
• To significantly decrease acquisition times while maintaining spectral accuracy.
• To enhance the reconstruction of transient dynamics from spectroscopic data.

Main Methods:

• Implemented optimized non-uniform sampling (NUS) with a genetic algorithm to determine optimal sampling patterns.
• Utilized compressed sensing for reconstructing spectroscopic data from undersampled measurements.
• Transformed two-dimensional (2D) spectra into a 4D time-frequency von Neumann representation for improved sparsity.

Main Results:

• Achieved substantial reduction in data acquisition time for multidimensional electronic spectroscopy.
• Demonstrated successful recovery of transient dynamics in a cresyl violet sample.
• Utilized only 25% of the original data points for accurate spectral reconstruction.

Conclusions:

• The proposed optimized NUS and compressed sensing approach significantly accelerates multidimensional electronic spectroscopy.
• The 4D time-frequency von Neumann representation offers improved data sparsity and reconstruction fidelity.
• This method enables efficient study of ultrafast chemical processes with reduced experimental burden.