In this study, ultrathin CuIn_xGa_1-xSe₂ (CIGS) films with two distinct thicknesses were grown on n-Si substrates by pulsed laser deposition (PLD), with thickness controlled by varying the number of laser pulses. Au plasmonic nanoparticles were subsequently incorporated into the CIGS layers using the same PLD technique for photodetection applications. Owing to the localized surface plasmon resonance (LSPR) induced by the embedded Au nanoparticles, photon absorption within the CIGS layers was significantly enhanced across the visible and NIR spectral regions. Increasing the film thickness in the presence of Au nanoparticles promoted the formation of larger grains and yielded notable improvements in crystallinity. The dark electrical behavior of plasmonic and non-plasmonic p-CIGS/n-Si heterojunctions was analyzed using the conventional J–V method, as well as the Cheung–Cheung and Norde methods. Key diode parameters, including barrier height, ideality factor, and series resistance, were extracted and comparatively evaluated. Among the studied devices, the Au-CIGS2A heterojunction (based on Au-embedded CIGS thin film produced with 86,400 laser pulses) exhibited the most ideal diode characteristics, whereas the CIGS1A device (based CIGS thin film produced with 43,200 laser pulses) demonstrated the least favorable electrical performance. Under illumination, the combined effect of increased CIGS thickness and the LSPR-driven optical enhancement provided by the Au nanoparticles resulted in higher photovoltaic conversion efficiency in the corresponding heterojunction diodes.

This dissertation investigates how interaction strategies shape occupant responses to automated façades in office buildings, contributing to reducing the gap between predicted and actual performance. Automated façades, such as roller shades, venetian blinds, and switchable glazing, offer potential to reduce energy demand and improve indoor environmental quality. However, their effectiveness is often undermined by misalignment between automated control logic and occupant comfort requirements, leading to dissatisfaction and frequent manual overrides.
Using a mixed-methods approach, the research combines systematic literature reviews, controlled laboratory experiments, and large-scale surveys. The findings show that while automated façades can substantially reduce lighting energy use, their overall performance depends strongly on occupant interaction. Five key factors influence acceptance: personal preferences, environmental conditions, context, façade technology, and control logic. In particular, interaction design, such as how quickly and in what way façade systems adjust, control usability, and information availability, plays a critical role in shaping satisfaction and behavior.
Experimental results demonstrate that disruptive automation increases override actions, whereas systems that are understandable, predictable, and easy to control foster trust and acceptance. Survey data further show that occupants prefer differentiated levels of automation depending on context, building service, and time of day, with a general tendency toward “mixed-control” strategies.
The dissertation frames an “interaction gap,” arguing that energy efficiency and occupant comfort depend on occupant-centered design. It concludes that effective façade strategies must integrate clear communication, low-effort override options, and context-sensitive control.

To create an X-ray Interferometry test bed a high precision adjustable slit mechanism is needed. Com pliant mechanisms are ideal solutions for such high precision mechanisms. This research presents the design, manufacturing, and testing of such a compliant mechanism. Furthermore, topology optimization is explored and evaluated as an alternative design method to find a novel mechanism that performs better than a traditionally synthesized counterpart. A traditionally designed mechanism was first designed and characterized using a 405 nm laser source and Fraunhofer diffraction analysis. Although optical verification was limited to a minimum slit width of 3 μm, experimental results demonstrate that the traditionally developed design has the potential to achieve dimensional requirements, with demonstrated step sizes of 0.2 μm. Simultaneously, a topology optimization model was developed, implementing penalized strain energy, parasitic displacement, and decoupling constraints, along with a robust formulation, to generate an alternative multi-degree-of-freedom mechanism. Although topology optimization proved a tool capable of producing an alternative compliant mechanism, there is still work to be done to fully mature this synthesis method. The research concludes that the developed mechanism meets the requirements set for the X-ray Interferometry test bed, whilst the topology optimization proves a viable alternative, albeit complex, method for future high-performance iterations.

Quantum computing has seen remarkable progress over the past decade, with superconducting qubits emerging as one of the leading hardware platforms for its physical realisation. Among the various superconducting qubit architectures, the transmon has been one of the most widely adopted candidates, owing to its relatively simple design and compatibility. However, the transmon suffers from key limitations, most notably its weak anharmonicity and multi-qubit gates using transmon being difficult to execute faster without errors.
In this context, alternative qubit architectures have attracted interest. The fluxonium qubit, with its large anharmonicity and long coherence times, has emerged as a promising candidate for overcoming shortcomings of the transmon. Similarly, the cos(2φ) qubit offers an intriguing design with inherent noise protection, making it a compelling subject of investigation for next-generation superconducting hardware.
Beyond the question of which qubit to build, a central challenge in quantum computing is the realisation of high-fidelity two-qubit gates. Higher gate fidelity directly translates to better quantum operations and is essential for the implementation of error correction protocols. Equally important is the question of how to perform these gates in a fast and practical manner—ideally using simple pulse techniques that minimise the time and effort required for calibration.
This thesis addresses these themes across two broad directions. First, we explore the fabrication of fluxonium and cos(2φ) qubits, developing improved processes and cleanroom techniques to achieve more reliable and reproducible devices. Second, we investigate two-qubit gate operations using analytic pulse techniques, with the goal of realising faster and simpler gates that reduce the overhead associated with pulse calibration. Together, these contributions advance the development of superconducting qubit platforms towards more practical and scalable quantum computing.

The vibrations of electrically and symmetrically actuated micro-rectangular plates are considered. Coulomb and geometric nonlinearities are taken into account, as well as small linear damping. The Berger model and the Kantorovich procedure are used, which allows one to reduce the original PDE to a nonlinear ODE. Based on this ODE, the dynamic pull-in phenomenon is investigated, i.e., the transition from the oscillatory regime to an attracting one is studied. An effective numerical algorithm to determine the voltage for which the system collapses is proposed. Furthermore, the bounds for the regions of parametric instability can be computed. Applications with only an AC voltage and various variants with combined AC and DC voltages were considered. The study was carried out for a wide range of AC voltage frequency variations, that is, from zero to twice the natural frequency of the linear plate. In all cases, the instabilities that arise are parametric in nature. Along with the study of the original nonlinear equation, its simplified versions are also considered. Linearization of the Coulomb nonlinearity leads to a Mathieu–Duffing equation with double parametric forcing. Neglecting the geometric nonlinearity, a Mathieu equation with double parametric forcing is obtained. An interesting conclusion of the numerical experiments is that linearization of the original nonlinear equation and its replacement by the Mathieu equation allow one to determine quite accurately the dynamic pull-in of a microplate upon which a symmetric electrostatic actuation is applied. Since the theory for Mathieu equation is well developed, it is possible to analytically determine the frequency of the parametric resonances of the system. This is important, since in the vicinity of the corresponding AC voltage frequencies, sharp decreases in the dynamic pull-in values occur.