The operating range and efficiency of a centrifugal compressor is limited by the development of rotating stall and surge at low mass flow rates. To extend the operating range of a compressor, flow control in the compressor can be used to suppress secondary flow structures that lead to rotating stall. The presented work seeks to use the novel idea of placing passive vortex generators (VG) upstream of the impeller to suppress rotating stall, while also developing new concepts and optimization of microvortex generators (MVG). To accomplish this goal, a new SIMPLE-type algorithm for compressible flows was written in Code_Saturne along with a 2nd-order MUSCL scheme for convective terms and an AUSM+-up scheme for mass flux computation. The new algorithm was successfully validated against several widely-used test cases. The new algorithm was used to model the flow of the NASA CC3, a high-speed centrifugal compressor, from choke to rotating stall with a vaneless and vaned diffuser. The new algorithm predicted the performance of the compressor with a vaneless diffuser very well; satisfactory results were obtained for the compressor with a vaned diffuser. The full compressor with a vaned diffuser was used to model rotating stall. A complex stall cycle between the inlet of the impeller and diffuser was observed and studied. The fundament behavior of MVG, i.e. micro (sub-boundary layer) vortex generator, in a turbulent boundary layer was investigated in a channel flow with RANS and LES. Complementary wind tunnel testing was conducted to validate the computational predictions. The configuration of the MVG was studied to determine an optimal configuration and several conclusions were reached on the design of MVG. Most importantly triangle MVG were found to be the most efficient shape followed by NACA0012 and e423-type MVG, and a MVG angle of 18˚ to 20˚ was found to be optimal. Rectangle MVG were observed to suffer flow separation on the vanes which reduced their performance. The circulation and drag of a MVG was found to have a logarithmic relationship with the device's Reynolds number. These findings were incorporated in a LES study to control separated flow on the e387 airfoil and achieved an improvement in lift-to-drag ratio of 11.27%. Additional recommendations for MVG implementation were given. Combining the work on the NASA CC3 with the work on MVG, vortex generators were implemented near the inlet of the impeller. A detailed optimization study was conducted for the implementation vortex generators in the compressor. It was found vortex generators equal to the boundary layer thickness were the most efficient on controlling the downstream flow. The best configuration was implemented into the full compressor with a vaned diffuser to assess the ability of vortex generators to suppress rotating stall. The vortex generators were found to suppress rotating stall and extend the operating range of the compressor.

Stem cells have yet to contribute to their full potential in the field of regenerative medicine and further understanding of the underlying kinetics of cell differentiation could be the step forward. Various methods have been used to characterise stem cell lineage commitment. However, most of these techniques are end-point assays and provide very little information about the changes occurring in the early stages of the differentiation process. This project aims to explore if the structural and geometrical specificity of the cytoskeletal components (actin in particular) encode information regarding cell lineage. Adipogenic and osteogenic differentiation lineages were selected, as they have been extensively studied over the past few decades. We have developed a novel approach to describe cells by defining their cytoskeletal and nuclear morphology in terms of 19 geometric measurements. This set of parameters has a range of complexity, extending from one dimensional (e.g. fibre length, fibre thickness) to compound geometrical readings (e.g. chirality and fibre alignment), while some estimate morphological and mechanical properties of the nucleus i.e. Poisson ratio and chromatin condensation. A proprietary image analysis algorithm is used to analyse fluorescent images of cells biochemically and mechanically stimulated to differentiate for a period of up to 10 days. Our analysis pipeline is currently optimised for images acquired at x20 magnification using epi-fluorescence but can be further adapted for high throughput live cell imaging. Factorial analysis of the measured features showed that some parameters change markedly in the early stages of differentiation. More interestingly we observed these changes to be non-linear and non-monotonic. This analysis, in light with previously published literature on the subject has allowed us to more intricately hypothesise probable mechanisms involved with mechanotransduction which direct the lineage commitments. As our technique quantifies the morphology of individual cells, we used our extracted feature data to characterise each cell using a multivariate predictive model (LDA).

Usability of many chemical substances with significant potential for biomedical applications is limited by their poor solubility in water or limited stability in the physiological environment. One of promising strategies for therapeutic targeted delivery of these types of substances into cells and tissues is their encapsulation inside polyelectrolyte microcapsules (Volodkin et al., 2004b, Sukhorukov et al., 1998b). Successful internalisation of microcapsules loaded with various macromolecules have been observed in several types of living cells (Javier et al., 2008, Kastl et al., 2013), however the mechanisms of the uptake of capsules by living cells are not yet fully understood. Detailed understanding of physico-chemical and mechanical interactions between capsules and living cells is required for specific targeting, effective delivery, and elimination of any potential toxic side effects. This has been largely limited by capabilities of available imaging techniques and lack of specific fluorescent markers for certain types of cellular uptake. The rate of internalisation of microcapsules was primarily studied at the level of cell population using conventional optical/fluorescence microscopy, confocal microscopy, and flow cytometry (Gao et al., 2016, Ai et al., 2005, Sun et al., 2015). These conventional fluorescence methods are known to be prone to overestimating the number of internalised capsules due to their limited capability to exclude capsules which were not fully internalised and remained attached to the cell surface (Javier et al., 2006). Experimental evidence with resolution high enough to resolve the fine membrane processes interacting with microcapsules has been limited to fixed samples imaged by scanning electron microscopy and transmission electron microscopy (Kastl et al., 2013) capturing randomly timed "snapshots" of what is likely to be highly dynamic and complex interaction. Physical force interactions between cellular membrane and capsules during the internalisation were suggested to cause buckling of capsules based on indirect evidence obtained using fluorescence microscopy in live cells 15 (Palankar et al., 2013) and separate measurements of capsule deformation under colloidal probe atomic force microscopy (AFM) outside of the cellular environment (Delcea et al., 2010, Dubreuil et al., 2003). However, our knowledge of the mechanical properties of the fine membrane structures directly involved in the internalisation process or how these structures form during the internalisation is very limited, if non-existent. Here we employ a different approach based on a high-resolution scanning probe technique called scanning ion conductance microscopy (SICM). SICM uses reduction in ionic current through the probe represented by an electrolyte-filled glass nanopipette immersed in saline solution to detect proximity of sample surface (Hansma et al., 1989, Korchev et al., 1997a). The technique has been previously used for high-resolution scanning of biological samples of complexity similar to what can be expected in case of microcapsules interacting with cells (Novak et al., 2014, Novak et al., 2009), and also for mapping mechanical properties at high resolution (Ossola et al., 2015, Rheinlaender and Schaffer, 2013). It has been proved to be able to visualise internalisation process of 200 nm carboxy-modified latex nanoparticles (Novak et al., 2014), however it is not clear whether it would be suitable for visualising internalisation of substantially larger, microscale-sized particles. The aim of this research was to visualize the live internalisation process of microcapsules entering cells by using SICM. The first two chapters of this thesis are introduction and literature review, which summarise the current state of the art. Chapter 3 states the aim and objectives of this study. Chapter 4 introduces the materials and methods we used in our research. Chapter 5, 6, 7 present the main findings of our research. Chapter 5 states the challenges we met in visualising the live internalisation of microcapsule as well as our solution for overcoming those challenges. At the end of that chapter, we describe the detailed procedure we used for recording the live internalisation of microcapsules. The results we got using this procedure are presented in chapter 6 and 7. In chapter 8, we discuss the results we found by comparing them to the results of previous research. In chapter 9, we summarise our study and give some suggestions on future work.

The Cu-Sb-S compounds (Cu12Sb4S13, CuSbS2, Cu3SbS3 and Cu3SbS4) have the advantages of earth-abundance, low-toxicity and low-cost, compared with conventional thermoelectric materials. This work provides a comprehensive study on the synthesis methods, crystal structures and thermoelectric properties of Cu-Sb-S compounds. All of the samples were prepared by mechanical alloying combined with SPS, which had high density, high purity and very fine microstructure. The lone-pair electrons of Sb and the [CuS3] plane play important roles in realizing very low lattice thermal conductivity of these compounds. Except for Cu12Sb4S13, which is known as a good thermoelectric material, the other three compounds showed very poor thermoelectric performance due to their high electrical resistivities. A phase transition at 398 K was found in Cu3SbS3, which makes it unsuitable for applications and attempts to optimize electrical properties of CuSbS2 failed. Different p-type dopants were studied to improve the electrical properties of Cu3SbS4. Both Ge-doping and Sn-doping on Sb sites increased the carrier concentration of Cu3SbS4 significantly. The electrical transport properties were analyzed using SPB model, and a large effective mass of 3.0 me was found for all of the samples. A maximum zT value of 0.69 was obtained at 623 K in 5 mol. % Sn-doped sample which was about 6 times higher than that of undoped sample. The solid-solutions of Cu3SbS4(1-y)Se4y were studied to further improve the thermoelectric properties. The lattice thermal conductivity was reduced in solid-solution due to the local mass contrast and alloying scattering, but there was no further improvement in zT value due to the decrease in Seebeck coefficient. Another solid solution of Cu3Sb1-xBixS4 was studied, but Bi had very low solubility and a second phase was formed instead of forming the solid solution. Future work should focus on reducing the lattice thermal conductivity of Cu3SbS4 without impacting its electrical properties.

Gene delivery can potentially treat acquired and genetic diseases such as cystic fibrosis, haemophilia and cancer. Non-viral gene delivery vectors are attractive candidates over viral vectors such as recombinant viruses, due to their lower cytotoxicity and immunogenicity, despite significantly lower transfection efficiencies. To improve efficiency of non-viral vectors, the investigation of the various parameters influencing DNA transfection is essential. The present study developed a versatile gene delivery system with tailored physicochemical and biological properties. The system used polymer brushes synthesised via atomic transfer radical polymerisation (ATRP), grafted from silica nanoparticles, whose charge density, grafting density, chemistry, length of brush, the size and shape can be altered. The primary focus of the study was poly(2-dimethylaminoethyl methacrylate) (PDMAEMA), known for its positive charge and DNA condensation. The ability of PDMAEMA to interact with DNA was characterised using dynamic light scattering, electrophoretic light scattering methods, surface plasmon resonance and in situ ellipsometry whilst its interaction with cells was studied via cell viability assays. The brush behaviour in response to pH and ionic strength was also studied. The charge density was altered by copolymerising with poly[oligo(ethylene glycol) methyl ether methacrylate](POEGMA) and the effect of such modification on DNA interaction was studied. PDMAEMA-grafted nanoparticles gave the highest transfection efficiency compared to other synthesised polymer brushes, but still displaying almost 2-fold lower transfection efficiency than the commercially available reagent jetPEI®. Different brush chemistries were also investigated. Poly(glycidyl methacrylate) (PGMA) decorated with oligoamines: allylamine, diethylenetriamine and pentaethylene hexamine, and PDMAEMA quaternized with alkyl halides: methyl iodide, allyl iodide and ethyl iodoacetate did not show any significant transfection, despite their performance reported in the literature. The robust system developed is a promising platform for further investigation of parameters influencing cellular uptake and gene expression, and important milestone to develop non-viral gene delivery systems.

Fluorescent carbon dots are a new class of carbon nanomaterials that have emerged recently, and have created a lot of interest as a potential competitor to classical semiconductor quantum dots. Carbon dots possess low toxicity, biocompatibility, easy and low-cost synthesis, and good optical properties. They show huge potential as novel and versatile nanomaterials for a wide range of applications such as bioimaging, drug delivery, chemical sensing, photocatalysis, and as sensitizers for photovoltaic solar cells. The main motivation for this research was the need to produce non-toxic, low-cost nanomaterials with good optical and electrical properties for the use in the fabrication of sustainable, inexpensive nanostructured solar cells with good efficiency. The main aims and objectives of this PhD research were: to synthesize fluorescent carbon dots from biomass-derived precursors by using the hydrothermal synthesis method, to understand and explain structural and optical properties of the as-synthesized carbon dots, and to use the carbon dots as sensitizers for nanostructured solar cells. Carbon dots (CDs) were synthesized using hydrothermal synthesis method from polysaccharides (chitosan and chitin), monosaccharide (D-glucose), amino acids (L-arginine and L-cysteine), and from real food waste in the form of lobster shells. Carbon dots were thoroughly characterized to obtain the information about their structural and optical properties. The as-synthesized carbon dots showed polydispersity and quasi-spherical morphology, with particle sizes ranging from 5-17 nm. Carbon dots showed predominantly amorphous nature, and the functional groups from the starting precursors were successfully incorporated into the as-synthesized carbon dots. Diluted solutions of carbon dots were transparent under daylight and showed blue-green photoluminescence emission under UV excitation. All carbon dots showed excitation-dependent photoluminescence emission which was more pronounced for excitation wavelengths larger than 320 nm. Chitosan CDs, L-cysteine CDs and lobster CDs also showed excitation-independent emission for excitation wavelength in the range of 200 - 320 nm. The highest fluorescence quantum yield of (43.3 ± 2.1) % was calculated for L-arginine CDs. It was concluded that the origin of light emission in carbon dots must be governed by the interplay between the absorption due to the carbon cores and the surface functional groups. Considering the application of the as-synthesized carbon dots, two types of solar cells were fabricated. Carbon dots were used as sensitizers for ZnO-nanorod-based and for TiO2-based nanostructured solar cells. Three types of carbon dots (chitosan CDs, chitin CDs and D-glucose CDs) were used as sensitizers for ZnO-nanorod-based solar cells. ZnO nanorods were successfully coated with carbon dots, and the chitosan-CDs-sensitized solar cells showed the efficiency of 0.061 %. When using layer-by-layer coating method, solar cells with combination of chitosan- and chitin-CDs as sensitizers showed the efficiency of 0.077 %. All six types of carbon dots (chitosan CDs, chitin CDs, D-glucose CDs, L-arginine CDs, L-cysteine CDs, and lobster CDs) were used as sensitizers for TiO2-based nanostructured solar cells. TiO2-based solar cells sensitized with carbon dots showed much higher efficiency compared to the ZnO-nanorod-based solar cells. L-arginine-CDs sensitized TiO2-based solar cells showed the highest efficiency of (0.362 ± 0.007) %, which was the best efficiency of all fabricated solar cells. By surveying a range of biomass-derived carbon dots, and demonstrating a clear link between functionalisation and solar cell performance, this PhD research project provides a guide to direct future development of low-cost, biomass-derived sensitizers for nanostructured solar cells.

Tyres are highly engineered complex rubber composite products. They are constructed from a wide range of different materials in addition to the rubber. In different parts of the tyre's construction, the rubber elements are expected to perform different functions and as a consequence many different types of rubber are used, each of which will have its own specific detailed compound formulation. These different regions of a tyre's construction are joined together by different types of molecular bonding. This variety of materials introduces potential sources of failure both in the homogenous regions within the tyre's construction but also at the interfaces between them. This thesis investigates the crack growth resistance of the rubber materials used in different regions of a tyre's construction as well as the interfaces that are found between the different parts of the tyre. A fracture mechanics framework was used to investigate the fatigue behaviour of bulk rubber and some of the interfaces. The loading of a tyre is periodic in nature as a consequence of the wheel's rotation therefore the materials were characterised over a range of loading conditions. The effect of cyclical loading frequency on the fatigue behavior of the bulk rubber was also investigated. This work discovered that the amount of crack growth per cycle was comprised from two different crack growth contributions. The first is related to the steady tear which is related to the length of time the load is applied. The second resulted from additional damage caused by the repeated loading and unloading of the material. Potential reasons for this additional crack growth contribution are discussed. The interfacial fatigue properties between adjoining and potentially dissimilar rubber compounds were examined using a fatigue peeling experiment. A novel test piece geometry was developed to evaluate the fatigue properties of interfaces in tyres and it was also used to investigate how different processing parameters such as the pressure at the interface during vulcanisation alter the interfacial strength. A significant effect was observed and this was related to the different phenomena occurring when two rubbery polymers are brought into contact. Finally, a fracture mechanics approach was also used to derive the value of the tearing energy, the variable governing crack growth propagation in the rubber materials found in tyres, using submodelling technique in finite element analysis. The tearing energy values at different locations within a tyre were calculated and are shown not to exceed the minimum energy criteria for crack propagation under normal service conditions.

Energy storage is currently gaining considerable attention due to the current energy crisis and severe air pollution. The development of new and clean forms of energy and related storing devices is in high demanded. Dielectric capacitors, exhibiting high power density, long life and cycling life, are potential candidates for portable devices, transport vehicles and stationary energy resources applications. However, the energy density of dielectric capacitors is relatively low compared to that of traditional batteries, which inhibits their future development. In the current work, three types of dielectrics, namely antiferroelectric samarium-doped BiFeO3 (Bi1-xSmxFeO3), linear dielectric (potential antiferroelectric) BiNbO4 and incipient ferroelectric TiO2, have been investigated to develop their potential as energy storage capacitors. For the samarium-doped BiFeO3 (Bi1-xSmxFeO3) system, the effect of samarium content in the A-site (x=0.15, 0.16, 0.165 and 0.18) on the structural phase transitions and electrical properties across the Morphotropic Phase Boundary (MPB) were studied. A complex coexistence of rhombohedral R3c, orthorhombic Pbam and orthorhombic Pnma was found in the selected compositions. The R3c phase is the structure of pure BiFeO3, the Pbam phase has a PbZrO3-like antiferroelectric structure and the Pnma phase has a SmFeO3-like paraelectric structure. The presence of the PbZrO3-like antiferroelectric structure was confirmed by the observation of the 14{110}, 14{001}, 12{011} and 12{111} superlattice reflections in the transmission electron microscopy diffraction patterns. The weight fractions of the three phases varied with different calcination conditions and Sm substitution level. By increasing the calcination temperature, the weight fractions of the Pbam increased, while that of the R3c decreased. The fraction of the Pnma phase is mainly derived by the Sm concentration and is barely affected by the calcination temperature. The increase of Sm concentration, determined an increase of the weight fraction of the Pnma phase and a decrease of the Pbam and the R3c phases. Temperature dependent dielectric measurements and high temperature XRD of Bi0.85Sm0.15FeO3 revealed several phase transitions. The drastic weight fraction change between the Pbam and the Pnma phase around 200 °C is assumed as the Curie transition of the antiferroelectric Pbam phase. The transition at 575 °C is related to the diminishing of the R3c phase and is suggested as the Curie transition of the ferroelectric R3c phase. The Curie point of the antiferroelectric Pbam phase and the ferroelectric R3c phase in the Bi1-xSmxFeO3 ceramics shifted towards lower temperature with an increase of the Sm concentration. Current peaks were obtained in current-electric field loops in Bi0.85Sm0.15FeO3, which are correlated to domain switching in the R3c phase. The ferroelectric behavior was suppressed in Bi1-xSmxFeO3 (x=0.16, 0.165, 0.18), which is due to the gradually diminished contribution from the R3c phase. The system Bi0.82Sm0.18FeO3 showed the highest energy density of 0.64 J cm-3 (error bar ±0.02). For the BiNbO4 system, single phase α-BiNbO4 (space group Pnna) and β-BiNbO4 (space group P-1) powder and ceramics were produced. The longstanding issue related to the sequence of the temperature-induced phase transitions has been clarified. It is demonstrated that the β phase powder could be converted back to the  phase when annealed in the temperature range 800 °C -1000 °C with certain incubation time. The β to  phase transition is a slow kinetic process because sufficient temperature and time are required for the transition. In bulk ceramics with β phase, this transformation is impeded by inner stress, while it is favored by graphite-induced reducing atmosphere. A high temperature  phase has been revealed and the structure has been resolved. The structure of the  phase is monoclinic with a space group of P21/c. The lattice parameters are: a = 7.7951(1) Å, b = 5.64993(9) Å, c = 7.9048(1) Å,  = 104.691(2) Z=4. The volume is 336.76 (2) Å3. The calculated density is 7.217 g cm-3. The phase relationships among ,  and  phases have been clarified. It was found that the  phase (for both powder and ceramic) transforms into the  phase at 1040 °C on heating, and that the  phase always transforms into the  phase at 1000 °C on cooling. Meanwhile, a reversible first-order  to  phase transition is observed at ca. 1000 °C for both powder and ceramic if no incubation is processed on heating. The electric properties of both α- and - BiNbO4 have been investigated. The breakdown field of both ceramics were too low to observe any possible field-induced transition. As a result, linear P-E loops were obtained in each phase. The energy densities of α- and - BiNbO4 ceramics are 0.03 and 0.04 J cm-3 (error bar ±0.001), respectively. For the TiO2 system, ceramics were produced by conventional sintering and spark plasma sintering (SPS). Compared to conventional sintering, SPS technique produced dense ceramics without using sintering aids and avoided abnormal grain growth. Relaxation behavior related to the oxygen hopping among vacant sites is observed in the temperature range of 200 to 600 °C. TiO2 exhibits ultra-low loss at terahertz frequencies due to the reduced contribution of oxygen vacancies relaxation. TiO2 has a high breakdown field, but still has low polarization. The highest energy density obtained inTiO2 ceramics is 0.3 J cm-3 (error bar ±0.01).

Dielectric elastomers are electroactive polymers, which change size and shape in response to an electrical field. Dielectric elastomer actuators (DEAs) are highly promising new technologies in optical applications such as tuneable optical lenses, diffraction gratings and active camouflage. This thesis aims to develop a new approach to create a strain actuated compliant colour changing device that is controlled using DEAs as they offer stretchability, low weight, high efficiency, low cost and the possibility for miniaturisation. Conventional DEAs use transparent elastomeric materials with no significant colour change with strain. Conversely, liquid crystal materials are known to display dynamic colour changing behaviour, thereby making them good candidate materials. The thesis examines both the potential for colour changing soft actuators and the upcoming challenges in this field as well as the key concepts around liquid crystals that exhibit colour change. An initial approach was aimed at creating colour changes using dielectric elastomer actuators that drove a masked positioner. This method showed colour change since the mask changes the colour visualisation. The second approach used polymer dispersed liquid crystals, such as a nematic liquid crystal within a reactive silicone resin. The immiscibility of these compounds resulted in a dispersion of the liquid crystal droplets in the silicone matrix. However, the optical properties could not be controlled through mechanical deformation alone and the alignment of resulting LC droplets in the PDLC films was sensitive to the substrate used to perform the actuation. The next approach used reactive cholesteric liquid crystals (CLC) instead. A thin film coating process was preferred to carefully control the film's thickness by stretching. In free standing films a planar cholesteric alignment was obtained with mesogens aligned parallel to the substrate and colour was achieved based on the selective reflection of light. A transfer print technique was introduced to combine CLC coatings with elastomeric substrates that can be stretched. However, no colour change was achieved in response to mechanical deformation primarily due to the modulus and strength mismatch between the thin film and the elastomeric susbstrate material. Finally, lightly crosslinked liquid crystal elastomers using a combination of reactive and non-reactive liquid crystals were produced that were compatible with elastomer substrate materials. In free standing films planar cholesteric alignment was obtained with mesogens aligned parallel to the substrate. Successfully a reversible colour change based on selective reflection of light was achieved in response to a mechanical deformation.

This thesis investigates stabilisation of the SIMPLE-family discretisations for incompressible flow and their discrete adjoint counterparts. The SIMPLE method is presented from typical \prediction-correction" point of view, but also using a pressure Schur complement approach, which leads to a wider class of schemes. A novel semicoupled implicit solver with velocity coupling is proposed to improve stability. Skewness correction methods are applied to enhance solver accuracy on non-orthogonal grids. An algebraic multi grid linear solver from the HYPRE library is linked to flow and discrete adjoint solvers to further stabilise the computation and improve the convergence rate. With the improved implementation, both of flow and discrete adjoint solvers can be applied to a wide range of 2D and 3D test cases. Results show that the semi-coupled implicit solver is more robust compared to the standard SIMPLE solver. A shape optimisation of a S-bend air flow duct from a VW Golf vehicle is studied using a CAD-based parametrisation for two Reynolds numbers. The optimised shapes and their flows are analysed to con rm the physical nature of the improvement. A first application of the new stabilised discrete adjoint method to a reverse osmosis (RO) membrane channel flow is presented. A CFD model of the RO membrane process with a membrane boundary condition is added. Two objective functions, pressure drop and permeate flux, are evaluated for various spacer geometries such as open channel, cavity, submerged and zigzag spacer arrangements. The flow and the surface sensitivity of these two objective functions is computed and analysed for these geometries. An optimisation with a node-base parametrisation approach is carried out for the zigzag con guration channel flow in order to reduce the pressure drop. Results indicate that the pressure loss can be reduced by 24% with a slight reduction in permeate flux by 0.43%.