Surface-enhanced Raman Spectroscopy Data Processing And Analysis Of Biomolecules

Since the discovery of Surface Enhanced Raman spectroscopy（SERS）technology,due to its high sensitivity and the characteristics of reflecting the vibration information of substances,it has become one of the powerful means to explore the molecular structure information at the level of molecular detection,promoting the rapid development of trace amount detection.Silver nanoparticles are a kind of SERS enhanced substrate with high sensitivity,and the preparation by reduction method has the advantages of simple process and low cost.The purpose of this research is to study the widespread abuse of glucocorticoids,A protocol for the detection of the biomolecule glucocorticoid hydrocortisone was designed using SERS as a detection tool and silver nanoparticles as an enhanced substrate.The main content of this thesis covers the following aspects:1)To verify the working principle of Raman spectral acquisition,a Raman spectral acquisition system was designed for validation experiments.Using glucose reduction method and polyvinyl pyrrolidone as protective agent and dispersant,silver nanoparticles reinforced substrate with average diameter of 50nm was prepared at room temperature.Rhodamine 6G was used as probe molecule to investigate the effect of dense distribution or sparse distribution of silver nanoparticles on the enhancement effect.The enhancement reason is simulated by FDTD,and the experimental results agree with the theoretical simulation.2)The Moving Window Median algorithm combined with Ensemble Empirical Mode Decomposition algorithm was proposed to study the denoising optimization method of spike noise,and denoise the spike noise of the Raman spectrum of hydrocortisone solid powder collected in the optical path of confirmatory Raman experiment and SERS of the hydrocortisone.Studied the Raman spectrum and SERS on nano-silver sol substrate of biomolecular glucocorticoid hydrocortisone.The results showed that SERS of hydrocortisone showed selective enhancement of Raman peaks on nano-silver substrate compared with Raman spectrum of hydrocortisone solid powder.The SERS of 0.1 mg·L^-1-10 mg·L^-1 hydrocortisone anhydrous ethanol solution was quantitatively analyzed.Moving Window Averaging algorithm,Moving Window Median algorithm,Savitzky-Golay algorithm smoothing and Wavelet Thresholding Denoise algorithm were used to denoise the spectrum after spike noise removal.Adaptive Iteratively Reweighted Penalized Least Squares algorithm,Asymmetric Least Squares algorithm and BEADS（Baseline Estimation Add Denoising With Sparsity）algorithm were used to perform fluorescence background deduction for SERS of glucocorticoid solution respectively.The Kennard-Stone algorithm was used to divide glucocorticoid solution samples at each concentration into training sets and test sets in a 4:1 ratio.17 Raman feature shifts were optimized by Successive Projections Algorithm.A variety of regression models were used to establish a quantitative model of the characteristic displacement.Finally,the optimal support Vector Regression Model based on linear kernel function was selected,the Root Mean Square Error of test set was only 0.5026mg·L^-1,and the coefficient of determination of test set was as high as 0.9512,providing a scheme strategy for the accurate monitoring of glucocorticoids.