Frequency-domain nonlinear regression algorithm for spectral analysis of broadband SFG spectroscopy

Area of Science:

• Spectroscopy
• Nonlinear Optics
• Surface Science

Background:

• Broadband sum frequency generation (BB-SFG) spectra are crucial for surface analysis but often suffer from distortions.
• Nonresonant background and laser pulse lineshapes complicate the interpretation of BB-SFG spectra.
• Accurate retrieval of spectral parameters is essential for understanding surface vibrational properties.

Purpose of the Study:

• To introduce and validate a novel Frequency Domain Nonlinear Regression (FDNLR) algorithm.
• To improve the analysis of BB-SFG spectra by accounting for distortions.
• To retrieve key parameters of resonant vibrational SFG bands without prior assumptions.

Main Methods:

• Development of the Frequency Domain Nonlinear Regression (FDNLR) algorithm.
• Simultaneous fitting of a series of time-resolved BB-SFG spectra.
• Validation using both virtual and experimentally measured SFG spectra.

Main Results:

• The FDNLR algorithm successfully retrieves the first-order polarization induced by infrared pulses.
• Accurate determination of relative phase, dephasing time, and lineshapes of resonant vibrational SFG bands.
• Demonstrated validity and reliability of the FDNLR method on diverse SFG spectra.

Conclusions:

• FDNLR offers a robust method for analyzing complex BB-SFG spectra.
• The algorithm overcomes limitations imposed by nonresonant backgrounds and laser pulse characteristics.
• FDNLR enables precise characterization of surface vibrational modes without restrictive assumptions.