Time-domain fit for improved contrast in quantitative coherent anti-Stokes Raman spectroscopy

Abstract

Among the multiple coherent anti-Stokes Raman scattering (CARS) techniques that provide important quantitative molecular microscopic contrast, Fourier-transform CARS (FT-CARS) stands out with the immunity to nonresonant background and high-speed detection capacity. However, by using FFT for the exponentially decaying signal, FT-CARS faces the dilemma of choosing the delay range of the signal for high SNR or high resolution, the lack of either of which is detrimental to the quantitative contrast of imaging. Here, time-domain fit (TDF) is proposed to fully utilize the time-domain information of FT-CARS, providing optimized SNR and vibrational feature distinguishment. The capacity of noise restriction and feature distinguishment of the traditional FFT and the proposed TDF is analysed with theoretical examination and simulation. Exploiting the matrix pencil extraction of vibrational parameters, TDF is performed for quantitative analysis for simulated FT-CARS signal, and shows more accurate and consistent performance than the FFT method. FT-CARS coupled with TDF intensity evaluation holds the promise to provide micro-spectroscopic contrast with higher SNR and free of spectral overlapping, contributing to a more powerful diagnostic tool.