Analysis of analytic nonresonant background removal algorithm for MCARS spectra

Abstract

Multiplex Coherent Anti Stokes Raman Spectroscopy (MCARS) has been shown to generate a complete Raman spectrum of a material on a millisecond time scale which allows for rapid identification of a wide variety of molecular targets. Along with the desired resonant spectrum due to the vibrational Raman spectroscopy of the analyte, MCARS is known to simultaneously generate a nonresonant spectrum that can obscure the desired Raman spectrum which hinders detection. Extracting the desired resonant Raman signal analytically from the overall MCARS signal has proven difficult without having prior knowledge of the analyte. We have developed an algorithm that utilizes a combination of the maximum entropy method in conjunction with advanced Fourier filtering to analytically remove the nonresonant background from our MCARS spectra without having prior knowledge of the vibrational spectrum of the analyte. In this report, we will report on the theoretical background for this algorithm as well as our experimental work testing this algorithm under various nonresonant spectra conditions for a number of analytes. We will systematically vary the amount of nonresonant background generated in the sample by changing the temporal overlap of the two beams necessary to generate the MCARS signal. Additionally, we place the analyte into increasing concentrations of water to generate increasing amounts of nonresonant background spectra to test the algorithm's effectiveness. Finally, we compare the analyte vibrational spectral output from the algorithm to the Raman spectrum measured with the spontaneous Raman system in the laboratory of the same sample in an effort to ascertain accuracy of the output spectra.