Date of Award

May 2015

Degree Type


Degree Name

Doctor of Philosophy



First Advisor

Carol J. Hirschmugl

Committee Members

John R. Reisel, Hossein Hosseini, Arash Mafi, Jun Zhang


Computed Tomography, FTIR Spectroscopy, Image Analysis, Image Processing, Microscopy, Mixed Polymers


The purpose of this dissertation is to carry out non-destructive 3D imaging by applying Fourier Transform Infrared (FTIR) spectro-microtomographic techniques, and develop corresponding methods of data analysis. This is done by collecting 3D synchrotron-based and lab-based (Thermal) FTIR hyper spectral data at the Synchrotron Radiation Center (SRC) for the first time. Despite other 2D imaging techniques, this does not manipulate the sample, and suppresses the need to microtome 3D biological, material or biomedical samples into slices to study by spectroscopic imaging techniques. Spectro-micro-tomography is applicable for scientific, industrial, energy, biomedical samples such as stem cell characterization and materials such as polymers. Tomographic reconstruction methods are employed to the data to investigate the chemical and morphological localization, and obtain the average spectra of regions of interest as well as spectra for every voxel.

It is assumed that the thermal light has cone geometry, and the data collected needs cone beam reconstruction, whereas the data collected using synchrotron light requires parallel beam reconstruction, since the beam waist created by the focus at IR wavelengths of the synchrotron 12 beams can be approximated well by a parallel beam. While bright synchrotron light provides us with higher SNR data, the capability of doing FTIR spectro-micro-tomographic techniques using thermal light, processing and analyzing it is of a high significance since thermal sources are more readily available. In this study the cone beam reconstruction is implemented and evaluated by applying them to the phantoms such as centered and off-center Polystyrene beads, and samples of mixed-polymers. The results of the cone beam reconstruction show that the cone beam reconstruction does not improve the quality of the reconstruction, and the parallel beam reconstruction is still better. The cone beam is not capable of modelling the optical system of our imaging environment, and the half cone beam angle size is small enough to be considered as parallel beam. Furthermore, the application of the cone beam is limited to the size of the sample.

For further analysis of the 3D reconstructed volumes of the samples, specific signal processing tools are required. The deconvolution algorithm is applied to the 2D projections at all the wavelengths before the reconstruction to increase the image contrast and spectral fidelity, deblur the projections, and finally increase the contrast of the 3D images.

Segmentation methods will be implemented for defining the regions of interest in the 3D structures; this will be used for average spectrum computation as a necessary tool of spectral analysis. The techniques developed here employ thresholding and kmeans clustering are capable of calculating the average spectra of the components found in the data as well as their corresponding renderings.