Search

The synthesis of bio-based products, fuels, and materials in large-scale closed-photobioreactors (PBRs) presents a sustainable option for tackling the ever-increasing mass and energy demands of the world's rapidly growing population. With the scale-up of production comes significant cost reductions and increased commercial viability. However, successful PBR scale-up must overcome key hurdles relating to the local environmental conditions, including uneven light distribution caused by cellular absorption and mutual shading, as well as nutrient and biohydrogen partial pressure gradients. Therefore, this thesis confronts these engineering challenges with advanced mathematical modelling techniques by tackling the biotechnology's multi-scale complexities with minimum simulation cost strategies. The proposed models were thoroughly validated using both literature and experimental data collected from cultivating different microbial species in PBRs of different configurations and scales. In a "journal format" style thesis, Chapters 1 to 2 covers the general introduction and comprehensive literature review whereas Chapters 3 to 5 present the published original contributions. More specifically, Chapter 3 proposes the first-ever mechanistic model to directly integrate the effect of PBR mixing-induced light/dark cycles into the biomass growth kinetics. This enables the manipulation of the PBR mixing rate to alleviate light attenuation challenges and maintain higher biomass growth rates. Chapter 4 extends the mechanistic model's capabilities to account for the effects of temperature and PBR biohydrogen partial pressure, which were previously ununified for any microbial species. To evaluate the biotechnological transfer across two types of PBR, namely the Schott bottle-based and vertical tubular-based PBR, two parameters related to the PBR's local environmental conditions were derived: the effective light coefficient and the biohydrogen enhancement coefficient for recalibration. The successful systematic upscaling approach was recommended for other similar biosystems. Building on these achievements, Chapter 5 focuses on the multi-physics coupling within a Computational Fluid Dynamics (CFD) solver to facilitate optimisation and upscaling of biohydrogen production. For this, accelerated growth kinetics and parallel computing were combined to greatly reduce the simulation cost, enabling uncertainty estimation via Monte Carlo simulation for the first time. Finally, Chapter 6 concludes the thesis and presents two future directions: the exploitation of the models developed in Chapters 3 and 5 for (i) model-based optimal control of PBR mixing, (ii) the optimisation of PBR static mixers to enhance biomass growth and biohydrogen productivity, and (iii) application to other scalable PBR configurations.

Electron spin echo envelope modulation (ESEEM) spectroscopy is widely used to study the radical pairs created during the primary steps of photosynthesis. In this thesis the analysis of ESEEM spectra is improved, and some new applications and variations of this experiment suggested. Experimental spectra from species such as P+Q-A, the secondary radical pair formed in the reaction centre of the bacterium Rhodobacter sphaeroides, give information about the exchange and dipolar couplings between the radicals. The model used to analyse the data affects the results; this thesis suggests two improvements. First, the effect of anisotropic hyperfine couplings in the radicals is considered by the addition of a single spin-1/2 nucleus to the model. This approach suggests that previous models neglecting the effect of nuclei may have been slightly in error. Secondly, several model fittings are performed in the time domain. This approach avoids the Fourier transformation to the frequency domain so that experimental dead-time does not corrupt the data. An excellent fit to experimental data is found with a model containing one spin-1/2 nucleus on each radical. The hyperfine coupling parameters resulting from the fit are consistent with independent experimental results. Use is made of the method of Cramér-Rao lower bounds to assess the precision to which experimental parameters are determined from a time domain curve fitting. It is shown that the lower bounds may also be used to determine the optimum sampling strategy for the experiment. An example is given of the novel use of ESEEM to determine the distance between the radicals in the strongly coupled, uncorrelated radical pair Q-AQ-B ESEEM has not yet been used for this purpose, and the simulated spectra produced here indicate that the experiment could be used to evaluate the dipolar coupling and hence the inter-radical distance. This thesis considers the possibility of performing ESEEM at higher frequencies than are usually considered. Calculations show that the increased resolution of the g-tensors allow an experiment performed at the W-band frequency of 95 GHz to make a correlation between the relative orientations of the radicals and the dipolar axis, information which has previously been unavailable from a single experiment.

The energy transfer processes in the LH2 complex from Rps. acidophila 10050 were investigated using a pigment-exchange strategy. In this approach, the Bchla-B800 molecules were selectively removed from their binding pockets and were replaced with modified pigments with altered spectral properties. Spectroscopic analysis of these modified complexes gave valuable insights into the mechanisms of energy transfer within the complex. The Bchla-B800 molecules can be released from their binding pockets by an acid treatment using buffer containing the novel Triton detergent TBGIO. B850-only complexes can then be purified by ion-exchange chromatography. Later, the empty B800 sites can be reconstituted with native Bchk. Maximal occupancy of the B800 sites can be obtained by incubating the B850-only complexes in buffer solution containing the detergent dodecyl-p-D- maltoside (LM) with a 3x fold excess of Bchla for 2 h at pH 8 and at room temperature. The Bchk-B800 reconstituted complex was fully characterised to determine how similar it is to native LH2. The average occupancy of the B800 sites in the Bchk-B800 reconstituted complex is ~ 80%. Resonance Raman spectroscopy showed that the reconstituted pigments are correctly bound within the complex and, as judged by CD spectroscopy, the structure of the LH complex is not affected by the pigment-exchange process. In addition, the reconstituted molecules participate in efficient B800?B850 energy transfer. Abstract The pigment-exchange protocol was then used with a series of modified pigments. The B800 binding sites can be reconstitued with geranylgeraniol Bchla, 132OH Bchla, Zn-Bphe, 31vinyl Bchla, 31OH Bchla, acetyl Chla and Chla. The B800 pockets can not be reconstituted with either pyro-Bchla, Cu- Bphe or Ni-Bphe. The 31vinyl Bchla-, 31OH Bchla-, acetyl Chla- and Chla- B800 reconstituted complexes are of particular interest as they have strongly blue-shifted "B800" Qy absorptions with maxima at 765, 751, 693 and 669 nm. CD spectroscopy again showed that the reconstituted pigments are correctly bound within the respective complexes. The occupancies of the B800 sites in the 31vinyl Bchla-, 31OH Bchla-, acetyl Chla- and Chla-B800 reconstituted complexes have yet to be accurately determined. In all of the reconstituted complexes, those pigment molecules which are correctly bound within the B800 binding pockets participate in efficient B800?B850 energy transfer. The kinetics of B800?B850 energy transfer in native and selected reconstituted complexes were studied by fs transient absorption spectroscopy. In both the native and Bchla-B800 reconstituted complexes, B800?B850 energy transfer takes approximately 0.9 ps. The time constants for B800?B850 energy transfer in the 3Vinyl Bchla-, 31OH Bchla-, acetyl Chla- and Chla-B800 reconstituted complexes are 1.4, 1.8, 4.4 and 8.5 ps respectively. This shows that the rate of B800-.