Return to search

Environmental Analysis at the Nanoscale: From Sensor Development to Full Scale Data Processing

Raman spectroscopy is an extremely versatile technique with molecular sensitivity and fingerprint specificity. However, the translation of this tool into a deployable technology has been stymied by irreproducibility in sample preparation and the lack of complex data analysis tools. In this dissertation, a droplet microfluidic platform was prototyped to address both sample-to-sample variation and to introduce a level of quantitation to surface enhanced Raman spectroscopy (SERS). Shifting the SERS workflow from a cell-to-cell mapping routine to the mapping of tens to hundreds of cells demanded the development of an automated processing tool to perform basic SERS analyses such as baseline correction, peak feature selection, and SERS map generation. The analysis tool was subsequently expanded for use with a multitude of diverse SERS applications. Specifically, a two-dimensional SERS assay for the detection of sialic acid residues on the cell membrane was translated into a live cell assay by utilizing a droplet microfluidic device. Combining single-cell encapsulation with a chamber array to hold and immobilize droplets allowed for the interrogation of hundreds of droplets. Our novel application of computer vision algorithms to SERS maps revealed that sialic sugars on cancer cell membranes are found in small clusters, or islands, and that these islands typically occupy less than 30% of the cell surface area. Employing an opportunistic mindset for the application of the data processing platform, a number of smaller projects were pursued. Biodegradable aliphatic-aromatic copolyesters with varying aromatic content were characterized using Raman spectroscopy and principal component analysis (PCA). The six different samples could successfully be distinguished from one another and the tool was able to identify spectral feature changes resulting from an increasing number of aryl esters. Uniquely, PCA was performed on the 3,125 spectra collected from each sample to investigate point-to-point heterogeneities. A third set of projects evaluated the ability of the data processing tool to calculate spectral ratios in an automated fashion and were exploited for use with nano-pH probes and Rayleigh hot-spot normalization. / Ph. D. / How can we understand the dynamic behavior of the cell membrane? Do certain polymeric structures in biodegradable plastic favor bacterial growth and subsequent degradation? To answer these and other intriguing scientific questions, techniques and technologies must be borrowed from a diverse array of fields and combined with fundamental understanding to create innovative solutions. In this dissertation, a two-dimensional surface enhanced Raman spectroscopy (SERS) assay was translated into a live cell assay by utilizing a droplet microfluidic device. Combining single-cell encapsulation with a chamber array to hold and immobilize droplets allowed for the interrogation of hundreds of droplets. Shifting the SERS workflow from a manual cell-to-cell mapping routine to the mapping of tens to hundreds of cells demanded the development of an automated processing tool to perform basic SERS analyses such as baseline correction, peak feature selection, and SERS map generation. Our novel application of computer vision algorithms to SERS maps was able to reveal that sialic sugars on cancer cell membranes are found in small clusters, or islands, and that these islands typically occupy less than 30% of the cell surface area. With an opportunistic mindset, several smaller projects that combine Raman and SERS with extensive data analysis were also pursued. Biodegradable plastics of varying content were studied with Raman spectroscopy. The aliphatic and aromatic polymeric units within these plastics both contain esters, but it is hypothesized that enzymatic hydrolysis occurs at the units asymmetrically. For each of six different samples, five maps were collected, processed using the analysis tool, and then analyzed using a multivariate analysis toolbox. Principal component analysis (PCA) was used to distinguish the polymers and to identify spectral feature changes resulting from an increasing v number of aryl esters. Uniquely, PCA was performed on the 3,125 spectra collected from each sample to investigate point-to-point heterogeneities. A third set of projects evaluated the ability of the data processing tool to calculate spectral ratios in an automated fashion and it was exploited for use with nano-pH probes and Rayleigh hot-spot normalization

Identiferoai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/94644
Date26 April 2018
CreatorsWillner, Marjorie Rose
ContributorsCivil and Environmental Engineering, Vikesland, Peter J., Marr, Linsey C., Zagnoni, Michele, Pruden, Amy
PublisherVirginia Tech
Source SetsVirginia Tech Theses and Dissertation
Detected LanguageEnglish
TypeDissertation
FormatETD, application/pdf
RightsIn Copyright, http://rightsstatements.org/vocab/InC/1.0/

Page generated in 0.0018 seconds