Multivariate hyperspectral Raman and CCD-based Raman spectroscopy applications for pharmaceutical analysis
Abstract
This thesis describes the applications of a newly designed multivariate hyperspectral Raman and CCD-based Raman spectroscopy for pharmaceutical analysis. Raman spectroscopy is a powerful analytical tool providing fingerprints of molecules in a sample. No two molecules give the same Raman spectrum, which makes Raman technique ideal to identify, quantify and map chemical components in pharmaceutical formulations. The first chapter of this thesis introduces newly constructed Multivariate Hyperspectral Imaging (MHI) instrument. MHI is designed to be utilized for a wide range of advanced engineering and smart materials both in industry and academia. Chemical imaging of multi-component materials is required for research and quality control purposes for various fields as well as for pharmaceutical formulations and biological applications. However, a major limitation of Raman, especially for industrial applications, is the relatively long time required (of the order of hours) to generate chemical images. The MHI instrument has been designed and constructed in our lab to achieve increased Raman imaging speeds using a multivariate compressive detection method. More specifically, the MHI is able to speed up the collection by employing a hardware data compression strategy, using spatial light modulator (SLM) to produce programmable optical filters generated implementing various multivariate signal processing or data compression algorithms such as partial least squares (PLS) algorithm. MHI collects Raman images much faster than traditional CCD-based Raman instruments by using a single channel detector instead of an array detector. Single channel detector has a key advantage of producing a better signal to noise ratio compared to CCD. The second chapter describes two applications of MHI for pharmaceutical content analysis and classification of raw materials. MHI is used as a fast imaging instrument to analyze the pharmaceutical ribbons both quantitatively and qualitatively in the first application. For the second application, MHI is used as a fast classification tool to identify various pharmaceutical components by using a set of digital filters that are trained on only two components. The third chapter shifts from multivariate spectral Raman to CCD-based Raman spectroscopy applications. CCD-based Raman spectroscopy is used to examine the structural changes of some active ingredients caused by milling process during the particle size reduction stage. In addition, the normal Raman technique is used for counterfeit detection. Images of a legitimate and a counterfeit tablet are produced to examine the spatial distribution of active ingredients in addition to examining pigments on genuine and counterfeit boxes. Also, spatially offset Raman spectroscopy (SORS) is introduced as a technique to examine the density variations in pharmaceutical ribbons as another application of normal Raman spectroscopy in this chapter. The last application studied is the drop coating deposition Raman (DCDR) application to determine the detection limits for certain small molecule metabolites. The applications presented here demonstrate the viability and also flexibility of both MHI and CCD-based Raman techniques. While the focus of this thesis is on pharmaceutical industry, the ideas can easily be incorporated for other industries, such as chemical, food, biological, cosmetic etc.
Degree
Ph.D.
Advisors
Ben-Amotz, Purdue University.
Subject Area
Analytical chemistry|Pharmaceutical sciences|Organic chemistry
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.