Synthesis Of Nanoparticles,Spectral Data Processing And Optimization For SERS Detection

With the combination of nanotechnology and biomedicine,nanomaterials are widely used in biomedical diagnosis and treatment.Surface-enhanced Raman spectroscopy(SERS)utilizes nanomaterials as the detection substrate,which significantly enhances the strength of the Raman signal,can effectively reduce the detection limit,and has the advantage of being particularly suitable for the detection of liquid substances;therefore,it is widely used in the diagnosis of benign and malignant tumors.Studies have shown that biological samples(especially pathological tissues,body fluids,etc.)exhibit abundant Raman spectrum peaks.However,the similarity of the overall Raman spectrum as well as the major spectral bands that reflect different pathological conditions,is greater than their difference,which is indicated by the weak response of the Raman spectral difference between them.The use of various Raman data analysis methods,for example,multivariate data analysis methods,aims to extract the major and important spectral features,thereby highlighting the distinct,stable and reliable Raman spectral feature differences between normal and abnormal samples,to achieve a final optimized diagnosis and discrimination based on Raman spectroscopic model.During the processing of Raman spectral data,the variations in spectral response caused by subjective and/or objective reasons such as different experimental parameters,instrument response,sample preparation,and individual differences in instrument competency cannot be ignored;therefore,the preprocessing of Raman data is of particular importance.This paper focuses on comparing the effects of four different normalization processing methods on the distinguishing effect of multivariate statistical analysis of SERS data from thyroid patients(both cancer and benign tumors),and our results indicate that best normalization processing is irrelevant to the classification or supervised models.Our results are expected to provide a comprehensive and objective assessment and reference for data normalization before the disease differential diagnosis model established based on Raman spectroscopy.The main research aspects of this paper include:1.Firstly,the research area of nanomaterials and their characteristics are briefly introduced.Secondly,it summarizes the applications of nanomaterials in biomedicine,including molecular diagnosis,biosensing,and nanoimaging,including in vitro and in vivo diagnosis.Finally,it introduces nanomaterials in the treatment of tumor chemotherapy,radiotherapy,photodynamics,photothermotherapy applications.2.The principles of conventional Raman scattering and surface-enhanced Raman spectroscopy(SERS)are briefly introduced,and the application of SERS spectroscopy in the diagnosis of cells,tissues,and body fluids in medicine is summarized.3.Four different SERS substrates(gold nanorods,gold nanostars,silver nanocubes,and silver nanoparticles)were prepared and characterized.By comparison,silver nanoparticles were selected as SERS substrates to collect SERS data of blood plasma from patients with thyroid tumors.4.Data optimization and multivariate statistical analysis were performed on the SERS data of blood plasma from thyroid tumor samples(50 thyroid cancer samples and30 benign thyroid tumor samples).The differential diagnostic results and influence of benign thyroid tumor and thyroid cancer were obtained and compared both in unsupervised methods(principal component analysis)and supervised methods(support vector machine)after the implementation of four normalization methods(vector normalization,area normalization,SNV normalization,and amide band normalization).