This thesis on luminescent solar concentrators (LSC) presents work carried out as part of the Electronic and Photonic Molecular Materials (EPMM) group of the department of physics and astronomy at the University of Sheffield. The work is presented in five experimental chapters looking at a range of research aspects from film deposition and measurement instrumentation, to exploring LSC optical properties and device performances by spectral based analytical methods.
A Gauge R & R (GRR) study design is used to assess sources of variance in an absolute fluorescence quantum yield measurement system involving an integration sphere. The GRR statistics yield the total variance split into three proportions; equipment, day-to-day and manufacturing variances. The manufacturing variance, describing sample fabrication, was found to exhibit the smallest contribution to measurement uncertainty. The greatest source of variance was found to be from fluctuations in the laser intensity whose uncertainty is carried into the quantum yield determination due to not knowing the exact laser intensity at the time of measurement.
The solvation phenomenon is explored as a potential way to improve LSC device yields; this occurs due to excitation induced changes to a fluorophore's dipole moment which leads to a response by the surrounding host medium resulting in shifts in fluorophore emission energy. This effect is shown to improve self-absorption efficiency by reducing the overlap of absorption and emission for particular organic fluorophores. This is expected to greatly improve energy yields but current dopant materials are too costly to employ according to the cost evaluations of this thesis.
A spray coating deposition tool is considered for the deposition of thin film coatings for bi-layer LSC devices. A screening study design of experiment is constructed to ascertain the level of control and assess the tool's ability to meet thin film requirements. Despite poor control over the roughness of the thin film layer this property was found to lie close to the acceptable roughness limit in most samples. The biggest issue remains the film thickness achieved by the deposition, which was an order of magnitude too small according to Beer-Lambert absorption models. This spray-coating tool is thus unsuitable for the requirements of a bi-layer LSC.
Concentration quenching is explored in the context of LSC device efficiency. Different fluorophores are seen to exhibited varied quenching decay strengths by looking at quantum yield versus fluorophore concentration. For two fluorophores, 4-(Dicyanomethylene)-2-methyl-6-(4-dimethylaminostyryl)-4H-pyran (DCM) and 2,3,6,7-Tetrahydro-9-methyl-1H,5H-quinolizino(9,1-gh)coumarin (C102), the quenching process is explored further using quantum yield and lifetime measurements to extract the quenching rate from rate equations. The form of the quenching rate as a function of molecular separation is shown to be of a monomial power law but distinct from the point-like dipole-dipole coupling of Förster resonant energy transfer (FRET). Additional quenching modes including surface-point and surface-surface interactions are considered to explain the power law form.
Spectral analytical models have been constructed to model performance metrics for square-planar LSC devices. In this model the input solar irradiance is considered to be incident normal to the LSC collection face. Device thickness optimisation is explored to ensure maximisation of the absorption efficiency by the fluorophore using Beer-Lambert absorption modelling. The normalised fluorophore emission spectrum is converted to an equivalent irradiant intensity spectrum based on the amount of energy absorbed. Propagation of this energy through the LSC structure is considered in terms of the mean path length of light rays waveguided by total internal reflection and again Beer-Lambert absorption modelling. Self-absorption and host transport losses are included in some detail. Out-coupling of LSC irradiance at the harvesting edges to connected solar cells is then modelled, using c:Si and GaAs power conversion efficiency spectra, and the resultant power output performance can therefore be estimated. Comparison with real devices from literature show that the model works reasonably well compared to these single device configurations and is somewhat conservative in its estimates. Cost efficiency models based on reasonable assumptions conclude the scope of this work showing that current materials fall short of delivering competitive energy solutions by at least factor of 2 in the case of the best dye modelled here.
