Background Extraction For Fiber Optics Surface-Enhanced Raman Spectroscopy Based On Spectral Matching

Fiber optic surface-enhanced Raman spectroscopy(SERS)combines the unique features of optical fiber in remote sensing,real-time measurements,and the detection of hazardous materials,with the benefit of SERS in significant enhancement of the Raman intensity.It has unique advantages and good application prospect in fields such as vivo detection,water quality monitoring and chemical process control.However,background emission is induced from the core molecules of the fiber in the process of light propagating through the fiber.The strong background emission ought to be considered,for it will bring interferences to the SERS spectrum of the target sample.Focused on the background emission in optical fiber,the meausured optical fiber background is introduced as reference spectrum,two methods of bi-objective optimization and weighted Savitzky-Golay(SG)smoothing have been developed to extract the optical fiber background from the measured SERS spectra of rhodamine 6G(R6G)and crystal violet(CV).The main contents of this article include the following aspects:(1)A bi-objective optimization method for extracting optical fiber background from the measured SERSspectrum of the target sample in the application of fiber optic SERS is proposed.The objective functions are built by using curve fitting to resolve the SERS spectrum into several individual bands,and simultaneously matching some resolved bands with the reference spectrum.The Pearson correlation coefficient(PCor),the first difference PCor,or the first difference Euclidean cosine coefficient(Cos)can be selected as the similarity index in the objective function of spectral matching.According to the background of the sample spectrum under the sample spectrum,a constraint function is built.Based on the objective functions,the constraint function and variable constraints,goal attainment method is applied to build the mathematics model for sloving the problem of bi-objective optimization.Then a detailed algorithm is proposed,implemented and programmed.(2)A novel filter is developed to fit the optical fiber background from the measured SERS spectrum of the target sample.The general model of the filter is built by incorporating a weighted term of matching the similarity between the estimated background spectrum and the reference spectrum into the classic SG smoothing filter model.Through respectively selecting Euclidean cosine coefficient(Cos)and Pearson correlation coefficient(PCor)as the similarity index,two different models of the weighted SG smoothing filter are derived and named as SG-ECos and SG-PCor accordingly.The algorithm compatible with piecewise fitting is presented and implemented.With the developed algorithm,the SG-ECos and SG-ECos filter can be used in optical fiber background removal,while without the help of the reference spectrum they degenerate to SG smoothing filter and can be applied to remove the flourensence background.(3)Based on the change of spectral similary,a method for identifying the sample band is proposed.The method is realized by judging the change of the similarity between the sample spectrum and the reference spectrum within the time before and after the identifying band subtracted from the sample spectrum,to distinguish the identifying band as the background band or the sample band.Based on investigation,Euclidean cosine coefficient is regarded as the best choice to the method,followed by first-difference Euclidean cosine coefficient.It is observed that the method selecting Euclidean cosine coefficient or first-difference Euclidean cosine coefficient as the similarity index can obtain a better effect on the bands’ identification by uplifting these two spectra simultaneously and making their baselines have the identical intercept.(4)On the basis of introducing similarity indices for spectral similarty measurements and focusing on the Pearson correlation coefficient(PCor)which is easily affected by spectral baseline,a new similarity index named Spectral Profile Similarity Coefficient(SPSC)is proposed.Its formula is similar with PCor,but it uses the intensity at the center wavenumber of the peak to correct every element in the spectrum instead of mean intesntiy.When it is used to measure the similary betwee two spectra,it enhances the similarity betwee two spectra located arround the peak and reduces the influence from other peaks and baseline.The above proposed mothods not only can be applied to remove optical fiber background or fluorescence background for SERS spectra,but also can be transferred to conventional Raman spectra recorded using fiber optic instrumentation.It is suggested that the proposed filter may be also applicable for other Raman spectra measurements to remove spectral contaminants originated from sampling substrates.