Illumination light manipulation and engineering for sensitive detection
Loading...
Files
Full Text E-thesis
Date
2025-09-29
Authors
Ma, Hui
Journal Title
Journal ISSN
Volume Title
Publisher
University College Cork
Published Version
Abstract
Biophotonic diagnostics provide noninvasive access to structural and biochemical information, yet their performance is often limited by uncontrolled light delivery, depth-dependent signal loss, and a lack of physiological standardisation. This thesis builds on the hypothesis that diagnostic accuracy can be significantly improved by engineering illumination to actively direct photon transport rather than treating it as a static constraint. A series of illumination strategies are therefore developed across multiple platforms, progressing from fundamental beam control to application-specific sensing systems.
The thesis begins by establishing a framework for structured illumination through multimode optical fibres using the transmission matrix (TM). A real-valued intensity transmission matrix (RVITM) approach is implemented to generate controlled circular symmetric output patterns such as rings and combinations of rings without interferometric detection using a 16 x 16 Hadamard calibration basis. The generated patterns, including rings of tunable radius and
composite double ring profiles, remain stable under fibre bending, demonstrating robustness against mode scrambling and providing a compact, axicon free route to programmable beam shaping at the fibre tip.
To address the need for spatially uniform illumination in quantitative imaging, a Kohler integrator based laser microscope incorporating dual microlens arrays (500 um pitch, 13.8 mm focal length) is developed to homogenise excitation across the sample plane, achieving a coefficient of variation (CV) of approximately 1.2 to 1.5% across the field of view. This system is applied to upconversion nanoparticle (UCNP) assisted biomarker imaging for semi-quantitative assessment of HER2 expression (scoring levels 0 to 3+) in breast cancer tissues. The UCNP labelled samples achieved a signal to background ratio of 40, representing a 25 fold improvement over conventional 3,3'-diaminobenzidine (DAB) chromogenic staining (ratio of 1.6), demonstrating substantially enhanced contrast and quantification accuracy in widefield cancer diagnostics.
For subsurface biochemical detection, tailored illumination is applied in Raman spectroscopy. A dual wavelength inverse spatially offset Raman spectroscopy (DWiSORS) system is implemented using axicon based ring illumination with excitation at 730 nm and 830 nm, enabling simultaneous access to the fingerprint (393 to 2195 cm-1) and high wavenumber (2043 to 3846 cm-1) spectral regions. A novel enhancement to noise ratio (ENR) metric is introduced to determine the optimal spatial offset for subsurface probing, yielding patient specific optimal offsets in the range of 5 to 9 mm. Validated through Monte Carlo simulations and two layer tissue mimicking phantom measurements, the system demonstrates sensitive detection of bone mineral and hydration markers at depths of up to 14 mm through overlying soft tissue, with in vivo measurements performed on ten healthy volunteers.
Finally, to enable calibration of optical sensing systems under physiologically relevant conditions, a light guided solid dynamic phantom is constructed. By integrating a polymer optical fibre (3\,mm diameter) with a liquid crystal display modulator, the phantom reproduces wavelength specific pulsatile absorption changes across four wavelengths (455, 530, 660, and 940\,nm). The system accurately simulates photoplethysmography (PPG) signals across a range of heart rates (80 to 120 bpm) and oxygen saturation levels (SpO2: 86 to 100 %), providing a controllable and reproducible platform for pulse oximeter calibration and validation that outperforms conventional liquid blood based phantoms in stability, response time, and spatial controllability.
Across the different systems, analytical modelling frameworks are used to guide illumination strategy and interpret photon transport behaviour. In the Raman studies, these are further supported by Monte Carlo simulations to quantify depth sensitivity and validate experimental trends. Together, this work establishes illumination as a powerful design parameter for enhancing sensitivity, selectivity, and reliability in biophotonic diagnostics, paving the way for future platforms where light delivery is adaptively matched to the sensing target.
Description
Keywords
Beamshaping , Upconversion , Nanoparticles , Raman spectroscopy , Transmission matrix
Citation
Ma, H. 2025. Illumination light manipulation and engineering for sensitive detection. PhD Thesis, University College Cork.