Polarimetric and hyperspectral imaging methods for characterisation of bio- and nanomaterials

Abstract

Presented here is the development of a time gated Raman spectrometer, a new method for designing polarimeters, testing of an overdetermined Mueller matrix imaging ellipsometer and Mueller matrix imaging of collagen fibre directions and silicon strain. In addition, a new method, spectral correlation, for analysing hyperspectral images, is developed, implemented and tested.

The time gated Raman spectrometer is capable of gating away long lifetime luminescence (> 2 ns), by using a time gated ICCD and a 2 ns pulsed Nd:YAG laser. Furthermore, the system can be used to study UV resonance Raman effects, due to the third (355 nm) and fourth (266 nm) order harmonic excitations available.

For optimising broadband polarimeters, genetic algorithms are employed. The resulting polarimeters are demonstrated to have a lower noise propagation and a broader spectral range, compared to previous commercial and patented designs. One of the designs is realised as a Mueller matrix liquid crystal variable retarder based ellipsometer. Its performance is in good agreement with computer optimised models. For calibrating the instrument, an extended version of the eigenvalue calibration method is employed.

Mueller matrix imaging is used together with state of the art decomposition methods to image collagen fibre orientations in cartilage and tendon. 3D directional imaging of the collagen fibres is achieved by using different angles of incidence upon the sample and a new method for calculating the fibre direction from the linear retardance. The method uses an effective medium approach to justify that the slow axis is along the direction of the fibres. A similar approach is used to study strain in silicon wafers.

Lastly, a spectral correlation method for analysing hyperspectral fluorescence images is developed and demonstrated to be good for analysing double stained amyloid plaques in mice. The resulting analysis allows for calculating the radius dependent mean and standard deviations for a set of plaques as a function of radius. By studying amyloid plaques from mice of different ages, it was found that there were significant changes with age in the structure of the plaques. Furthermore, the method was found to work well for analysing multi-component Raman spectra, where it detected small amounts of tetrahydrofuran in water.