UWSpace

UWSpace is the University of Waterloo’s institutional repository for the free, secure, and long-term home of research produced by faculty, students, and staff.

Depositing Theses/Dissertations or Research to UWSpace

Are you a Graduate Student depositing your thesis to UWSpace? See our Thesis Deposit Help and UWSpace Thesis FAQ pages to learn more.

Are you a Faculty or Staff member depositing research to UWSpace? See our Waterloo Research Deposit Help and Self-Archiving pages to learn more.

Photo by Waterloo staff

Communities in UWSpace

Select a community to browse its collections.

Recent Submissions

  • Thumbnail Image
    Item type: Item ,
    Finite-Size Orientational Crossover in Water–Argon Chains with Optional Carbon Nanotube Confinement
    (University of Waterloo, 2026-08-25) Dass, Ajay
    Nanoscale confinement can strongly alter the orientational behaviour of water. This thesis investigates whether local axial orientation persists when neighbouring water molecules are separated by fixed argon atoms, and how optional confinement within a (6, 5) carbon nanotube modifies that response. Finite water–argon chains were modelled as fixed-centre asymmetric-top water rotors with stationary argon spacers and nearest-neighbour water–water interactions. Ground states were calculated using the density-matrix renormalization group in a matrix product state representation as the water–water separation, R, was varied. The analysis combined energy, central von Neumann entropy, signed and absolute axial orientation, site-resolved profiles, angular distributions, and numerical convergence diagnostics. The principal result is a finite-size orientational crossover obtained for the ordinary open-boundary finite-chain Hamiltonian. The sampled central-entropy maximum moves to larger R with increasing chain length and approaches the 9–10 Å region for the larger chains. In the same main crossover window, the signed mean axial orientation is strongly reduced while the mean absolute local orientation remains appreciable. Site-resolved results show that differently biased parts of the chain make cancelling contributions while local axial orientation remains. The ground-state energy per water varies smoothly and has no corresponding feature. CNT confinement shifts and modestly suppresses the sampled entropy response, while the matched orientation curves remain similar over most of the main crossover window. CNT OFF and CNT ON are therefore consistent with the same general response, although an identical microscopic mechanism is not established. The separate short-R boundary-sensitive feature near R ≈ 7 Å was examined over R = 6.80–7.30 Å using sine-square deformation. Its entropy and branch-selection patterns change under SSD, showing sensitivity to boundary weighting without isolating an individual boundary contribution. This dataset does not test the boundary robustness of the finite-size orientational crossover. The conclusions apply to ground states of the finite, fixed-centre, nearest-neighbour Hamiltonian and do not establish a thermodynamic phase transition or physical excitation gaps.
  • Thumbnail Image
    Item type: Item ,
    Re-examining Contribution Fairness in Federated Learning
    (University of Waterloo, 2026-08-25) Costa Campos, Guilherme
    Federated Learning (FL) allows multiple data owners to train a shared model collaboratively without exposing their local datasets. Because participation imposes real computation, communication, and data-collection costs, sustaining long-term collaboration requires reward mechanisms that satisfy contribution fairness, the principle that each client should be rewarded commensurately with its contribution to the training. A substantial body of research pursues this goal through the distribution of contribution-based Top-K sparsified gradient rewards, such that higher-contributing clients receive denser, and, therefore, more useful, gradients than lower-contributing ones. Top-K sparsification, however, was originally devised as a gradient-compression technique intended to preserve convergence, which raises the question of whether it can actually produce the client-level model differentiation that contribution fairness demands. Furthermore, existing frameworks often combine distinct contribution estimation algorithms, reward generation rules, and client model update methods, making it unclear which component is responsible for the observed fairness behavior. This thesis investigates these questions by implementing two representative state-of-the-art frameworks, namely ACGSV and CFFL, and evaluating them across four benchmark datasets under distinct training-data partition settings. Building on this comparative evaluation, a series of controlled ablation studies is conducted to isolate the effects of three key components: the gradient used for reward generation, Top-K sparsification, and the retention of each client's locally accumulated gradient. The analysis yields three main findings. First, contribution fairness is commonly evaluated using Pearson’s r between clients’ standalone and federated test accuracies, with standalone accuracy serving as the reference contribution value and forming a set of contribution ranks. However, Pearson's r can remain high even when most clients converge to nearly identical federated accuracies and the reference contribution ranking is not preserved. Thus, Pearson’s r alone may provide an incomplete fairness assessment. To address this limitation, this thesis proposes a novel protocol that uses Kendall’s τ to measure the reference contribution ranks preservation and the Gini Mean Difference to quantify differentiation among the final client models. Second, Top-K sparsification induces significant model differentiation primarily when reward sparsity is high, corresponding to at least 70% in the studied settings. At lower sparsity levels, clients tend to converge to models with nearly identical performance. These findings indicate that sparsifying the aggregated gradient through Top-K is not, by itself, an effective and calibrated mechanism for estimated-contribution-based reward allocation. Third, local gradient retention, whereby clients retain their locally accumulated gradients generated during model training alongside the received estimated-contribution-based reward, is commonly treated as a secondary detail in framework design because it is not governed by the FL framework's contribution estimation algorithm and the contribution-score-to-reward mapping pipeline. However, it is an under-acknowledged confounding factor that drives much of the model differentiation, preservation of the reference contribution ranking, and, consequently, the high fairness scores reported by these frameworks.
  • Thumbnail Image
    Item type: Item ,
    Polymer-Integrated Halide Perovskite Solar Cells and Photoelectrochemical Cells for Solar Energy Conversion
    (University of Waterloo, 2026-08-25) Khamgaonkar, Saikiran
    Energy demand is skyrocketing, which has led to excessive use of non-renewable resources such as fossil fuels, which are major contributors to climate change and global warming. Therefore, to meet the increased energy demand while reducing carbon footprints, the implementation of renewable technologies for the production of clean energy is needed. In this context, the utilization of solar energy is one of the most appealing strategies for clean energy production. In recent years, halide perovskite solar cells and photoelectrochemical cells have emerged as potential candidates for harvesting solar energy into electricity or storing it directly as chemical fuels. This is mainly due to their excellent optoelectronic properties, along with simple, low-cost solution processing. Although high efficiencies have been achieved for both devices, their long-term stability remains one of the major issues that impedes their commercial deployment. In this thesis, the stability challenge of halide perovskite devices are addressed by implementing polymeric passivators to improve both extrinsic and intrinsic stability. Additionally, novel electrocatalysts have been designed and successfully integrated with these passivated perovskite devices to develop efficient and stable photoelectrochemical cells for green hydrogen production. In Part A, Chapter 2, we focus on simultaneously passivating both bulk and interfacial defects present in perovskite thin films using polystyrene as a bulk additive along with 4-fluorophenethylammonium iodide as the interfacial passivating agent. The addition of polystyrene modulates the perovskite crystallization kinetics, leading to the formation of larger grains with fewer grain boundaries. Meanwhile, 4-fluorophenethylammonium iodide passivates surface defects such as undercoordinated Pb²⁺ sites and vacancies. This combined passivation leads to perovskite solar cells with a high efficiency of 22.4%, along with significant improvement in stability, retaining 92% and 99% of their initial efficiency after 1008 h and 560 h under ISOS-D-1 and ISOS-D-2 conditions, respectively. Part A, Chapter 3 focuses on addressing thermal instability challenges in perovskite solar cells. Most high-efficiency halide perovskite compositions consist of either MA as an A-site cation or as an additive to stabilize the alpha phase of the perovskite. The addition of MA⁺ cations, even in small amounts, can lead to thermal instability due to their volatile nature. In this work, a dipolar polymeric passivator is developed by systematically modulating the electronic structure of polystyrene. A novel poly(pentafluoropolystyrene) passivator is identified, which consists of a distinct electronic distribution with highly diffused electron-rich and electron-deficient regions. These regions not only passivate both cationic and anionic defects in the bulk and at the surface of the perovskite but also exhibit strong interactions with MA⁺ species. This results in perovskite solar cells with efficiencies as high as 24%, along with significant improvement in thermal stability, retaining 95% of the initial efficiency after 2000 h under ISOS-D-2 conditions. Additionally, the storage and operational stability of the devices are greatly improved, retaining 97% and 95% of their initial efficiency after 3500 h and 500 h under ISOS-D-1 and ISOS-L-1 testing conditions, respectively. In Part B, Chapter 4, polystyrene bulk-passivated perovskite solar cells are integrated with bimetallic catalysts to fabricate perovskite photocathodes for green hydrogen production. Here, polystyrene is incorporated into the bulk of the perovskite to improve the intrinsic stability of the solar cells against ion migration. Additionally, using a facile self-assembly process of Au nanoparticles, a bimetallic Au–Pt–Ni catalyst is designed, which shows low overpotential and fast kinetics for the hydrogen evolution reaction in both acidic and basic conditions. The polymer-incorporated photocathodes show excellent performance with a half-cell solar-to-hydrogen (HC-STH) efficiency of 10.11%, along with significant improvement in stability, with T₈₀ values of 70 h (in H₂SO₄) and 78 h (in KOH), indicating improved device lifetime. In Part B, Chapter 5, highly efficient and stable bias-free water splitting is achieved by developing a polymer integrated perovskite photoelectrochemical cell. In this work, two critical challenges of PEC cells are addressed: (a) intrinsic instability of perovskite solar cells under external bias and (b) sluggish catalyst performance. Improved intrinsic stability against ion migration is achieved by implementing 4-fluoropolystyrene as both bulk and interfacial passivating agent. Meanwhile, an Au–Pt–Ni–PBA bifunctional catalyst is developed which shows low overpotential and fast kinetics for both hydrogen and oxygen evolution reactions. These combined effects lead to the fabrication of polymer-incorporated perovskite-based photocathodes with half-cell STH efficiencies of 18.6% and photoanodes with applied bias photon-to-current efficiency of 12.7%, along with excellent stability. Most impressively, the polymer-incorporated photoanodes maintain 95% of their initial performance after 270 h of operation under external bias. This improvement is primarily attributed to the incorporation of the polymer, which significantly reduces ion migration under applied bias conditions. Finally, the combined photoelectrodes enable unassisted water splitting with an STH efficiency of 13.56% and a T₉₀ of 190 h. Therefore, this work highlights the importance of improving the intrinsic stability of perovskites under external bias, along with efficient catalyst engineering.
  • Thumbnail Image
    Item type: Item ,
    Mid-Infrared Photoacoustic Spectroscopy for Non‑Invasive Glucose Monitoring: Design, Skin Phantom Development, and Numerical Modelling of PA-Tissue Interactions
    (University of Waterloo, 2026-08-25) Chu, Wing Yan
    Non-invasive glucose monitoring research and development has gained increasing attention over the past two decades, driven by the rising diabetic population and the limitations of conventional glucose monitoring approaches. Photoacoustic spectroscopy has emerged as a promising technique for painless glucose sensing. This thesis explores the use of dual-wavelength (9.25 μm and 10.3 μm) mid-infrared quantum cascade lasers in a photoacoustic spectroscopy system for glucose monitoring. Through improvements to the developed system and the design of a more systematic evaluation using controlled phantom fabrication and measurement conditions, the proposed approach demonstrated more consistent results, with a 75% reduction in frequency shift compared with previous studies. The fabricated skin phantoms exhibited close similarity to human skin in both mechanical and acoustic properties. With the implementation of a machine learning regression model, promising performance was achieved, with 100% of predictions falling within Zone A and B of the Clarke Error Grid. The agreement between the experimental measurements and the COMSOL photoacoustic simulation further supports the validity of both the fabricated skin phantoms and the developed mid-infrared photoacoustic spectroscopy system. Overall, this research demonstrates the potential of a non-invasive photoacoustic glucose monitoring system with an improved evaluation methodology, providing a strong foundation for future in vivo studies.
  • Thumbnail Image
    Item type: Item ,
    Machine Learning-Driven Decision Support Framework to Improve Musculoskeletal Health in Electrical Workers Using Wearable Devices
    (University of Waterloo, 2026-08-25) Attalla, Ahmed
    The construction industry is currently facing a significant shortage of skilled labour, limiting its ability to meet societal demands and sustain economic contributions. A key contributor is the high prevalence of musculoskeletal disorders (MSDs), which are often caused by repetitive motions, awkward postures, and sustained physical exertion inherent to various construction trade activities. This research focuses on electrical workers, a subgroup within the construction workforce that has been shown to experience high rates of MSDs. Recently, wearable support devices have emerged as a promising intervention for reducing fatigue and MSD risk. Although the construction sector has historically been slow to adopt new technologies due to fiscal and practical constraints, the growing availability of cost-effective, lightweight, and task-specific wearables has made their application increasingly feasible. Such innovations hold significant potential to enhance worker well-being, decrease the risk of work-related musculoskeletal injuries, and ultimately extend career longevity and earning potential within the industry. This research examines the effectiveness of wearable support devices in reducing the risk of MSD injuries among electrical construction workers. To facilitate evidence-based implementation, a decision-support framework is developed to recommend cost-effective wearable devices tailored to individual workers’ needs while accounting for the financial constraints faced by both employees and organizations. The framework comprises three principal components: (1) analysis of injury databases and task-specific ergonomic assessments to identify high-risk body regions and exposure patterns; (2) evaluation and classification of commercially available wearable devices according to their provided biomechanical support levels; and (3) development of a decision support system (DSS) that integrates machine learning and optimization techniques to generate cost-effective, personalized recommendations. First, publicly available injury data for electrical construction workers from WorkSafeBC and the Workplace Safety and Insurance Board (WSIB) were analyzed to identify the most frequently injured body parts, highlighting the three main areas most in need of support, namely the back, leg, and shoulder. These findings were validated through detailed task analysis and ergonomic assessment of common electrical construction activities. Next, a representative subset of commercially available wearable support devices targeting these body parts was experimentally evaluated. A diverse group of participants, including professional electricians and novice users, performed experimental tasks while performance metrics were recorded. Clustering analysis classified the devices into three support categories (Low, Medium, and High). Finally, a DSS was developed, integrating: (1) a machine learning model to predict support requirements, and (2) an optimization model that recommends specific wearable configurations aligned with worker needs and budget constraints. The DSS was validated through field testing with electrical workers who provided positive feedback on the wearables’ real-world effectiveness and usability. The proposed framework provides a novel decision-support approach for recommending cost-effective wearable devices from a repository of low-cost, commercially available options. It integrates objective and subjective inputs, performance metrics, machine learning tools, and optimized decision logic to address a critical occupational health and safety concern prevalent in construction. Although this study focused on electrical workers and a specific subset of support wearables, the framework is inherently scalable. It can be adapted to other trades with different ergonomic demands, as well as to a broader and continuously expanding range of wearable technologies entering the market. By supporting the selection of effective and affordable devices, the framework has the potential to reduce MSD occurrences, enhance worker well-being, and improve overall productivity within the construction sector.