Data driven approaches for improving quantification accuracy in surface enhanced Raman spectroscopy sensing
Citation & Export
Hide
Simple citation
Sardar, Sakshi.
Data driven approaches for improving quantification accuracy in surface enhanced Raman spectroscopy sensing. Retrieved from
https://doi.org/doi:10.7282/t3-1nsg-4a10
Export
Description
TitleData driven approaches for improving quantification accuracy in surface enhanced Raman spectroscopy sensing
Date Created2019
Other Date2019-10 (degree)
Extent1 online resource (xviii, 104 pages) : illustrations
DescriptionSurface enhanced Raman spectroscopy (SERS) is one of the most sensitive and selective techniques available. In the past couple of decades numerous applications of SERS for the development of sensors have been reported. Even though it is an excellent qualitative technique, its full quantitative potential has yet to be realized. One of the major categories of SERS based sensors are heterogeneous sensors, based on nanostructured substrates. The performance of these sensors is highly dependent on the distance between the enhancing nanostructure and the analyte, which in turn influences its sensitivity, and the reproducibility of the substrates. These factors play a very important role in controlling the SERS intensity associated with the sensors. We aimed to address these issues through different methodologies to improve the performance of SERS-based sensors. SERS enhancements achieved with the heterogeneous platforms are highly dependent on the surface morphology of the substrates used. For low-cost substrates, which are usually prepared from bottom-up approaches, control over the surface properties is low, resulting in variability among the substrates as well as at different locations within the same substrate. In order to overcome the reproducibility issue, we developed a dual-modality multi-site sensing methodology. In this methodology, we intentionally induced diversity on the substrate to modulate the SERS signal from analyte . Electrochemistry was combined with SERS for dual modality sensing to improve precision by adding redundancy and encoding features, thus increasing measurement robustness and predictability. This technique works by calibrating the SERS response with respect to the active surface area, a parameter known to be proportional to charge, which can be estimated via electrochemical measurements. The dual-modality multi-site measurement demonstrates at least 2.8x improvement in assay precision compared to the traditional single-site Raman measurements. The technique yields overall improved precision of measurement and is not limited to any particular SERS substrate or geometry, and thus can be adapted and incorporated readily in any SERS sensing assay.
Raman spectral variation can be analyzed with another perspective where the pattern obtained for the same analyte for different spot measurements provide information about the spot. In other words, the local environment at the measurement spot have bearing on the Raman spectrum. For instance, the location and orientation of the molecule on the substrate contribute to Raman spectral signature. In addition, the location, orientation, and interaction of the nanoparticles used to prepare the substrates also affect the spectral signature . As a result, the peak intensities and positions are modulated by these factors. Thus, in turn, a Raman spectrum contains all this information, and only by decoding it we can achieve a distinct picture of the local environment, which includes the analyte molecules. In order to understand the contribution of analyte concentration in the spectra, we carried out supervised classification using support vector machines. We found that the classification accuracy can be increased by properly incorporating different features contained in the spectrum.
Selectivity and sensitivity of the sensor are crucial properties for designing a SERS system. The system has to be optimized to achieve acceptable sensitivity and selectivity towards the analyte of interest. In order to improve upon these two properties, we chose phenylalanine (Phe), an biomarker for Phenylketonuria(PKU), an in-born metabolic error that leads to errors in the metabolism of Phe. Patients with PKU can have complications like intellectual disability, microcephaly, severe mental retardation, motor deficits, eczematous rash, autism, difficulty swallowing, seizures/convulsions, developmental problems, aberrant behavior, dystonias, dyskinesias, hyperreflexia, or spasticity, and psychiatric symptom. The treatment usually involves reduce dietary Phe intake and regular monitoring of Phe levels. The sensors were designed to detect Phe using different sensing approaches.
In conclusion, our work addresses important properties and issues that can assist in manufacturing better SERS-based sensors.
NotePh.D.
NoteIncludes bibliographical references
Genretheses, ETD doctoral
LanguageEnglish
CollectionSchool of Graduate Studies Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.