AUS Repository

Cholesterol detection is essential for early diagnosis and monitoring of cholesterol-related diseases, such as atherosclerosis, hypercholesterolemia, and liver diseases. A variety of nanomaterials have been designed and synthesized for cholesterol detection via electrochemical and spectrophotometric techniques. Metal organic frameworks (MOFs) have emerged as promising detector materials for cholesterol sensing. Recent research has explored MOFs as spectrophotometric cholesterol sensors with remarkable performance in terms of limit of detection (LOD) and selectivity. Given the growing interest in cholesterol sensing and limitations of existing biosensors, this thesis aims to develop a novel MOF-based spectrophotometric sensing platform for cholesterol detection. First, this thesis reviews recent advances in MOF-based spectrophotometric cholesterol sensors, outlining the different mechanistic roles of MOFs in cholesterol detection, current challenges, and potential applications of MOF-based sensors for cholesterol detection in point-of-care devices and medical diagnostics. Then, iron-based MOF (Fe-BTC) is introduced as a novel, peroxidase mimic nanozyme for cholesterol detection, marking its first reported use in this application. Characterization studies, including Fourier transform infrared spectroscopy, X-ray diffraction, and zeta potential, revealed stable, amorphous nature of Fe-BTC and potential peroxidase activity. Parametric studies including pH, time, temperature, and reagent concentrations were performed to determine optimal conditions for H2O2 and cholesterol detection. Mechanistic studies demonstrated biosensor operation via OH· radical formation by Fe-BTC. The present biosensor achieved a cholesterol limit of detection (LOD) of 2.91 μM and 2.88 μM at 25 ⁰C and 37 ⁰C cholesterol incubation, respectively, with a linear detection range of 6.56–78.75 μM. The biosensor had good selectivity to cholesterol in the presence of interfering analytes, including glycine, uric acid, glucose, and NaCl. Overall, our novel Fe-BTC-based biosensor demonstrated comparable performance to nanomaterial-based cholesterol sensors reported in the literature and shows great promise for cholesterol detection via spectrophotometric methods.

Audit quality plays a key role in shaping corporate financial stability, especially in heavily regulated sectors such as pharmaceuticals. While past research has examined the general effects of audit quality on companies, its specific impact on the financial stability of pharmaceutical firms remains underexplored. This study addresses this gap by comparing the United States and European markets, focusing on firms audited by the largest global audit networks (Big Four) versus those audited by smaller audit firms. Using a comprehensive dataset, the analysis reveals that firms audited by the largest global audit networks exhibit lower financial volatility and more stable returns, likely due to enhanced transparency and reduced information asymmetry. In contrast, firms audited by smaller audit firms demonstrate higher average returns but greater instability, suggesting a risk-return trade-off that may appeal to risk-tolerant investors. Notably, the largest global audit networks deliver consistent financial stability in both the United States and Europe, underscoring the robustness of their standardized practices. However, smaller audit firms in Europe show significantly higher volatility than their United States counterparts, implying that regulatory compliance alone does not ensure audit quality. The study also challenges the assumption that frequent auditor rotation improves outcomes; instead, longer auditor relationships, coupled with rigorous standards, may enhance financial oversight, particularly in complex industries like pharmaceuticals. These findings advocate for reconsidering mandatory auditor rotation policies and strengthening oversight of smaller audit firms to improve reporting quality and investor confidence.

RF energy harvesting has emerged as a viable energy source for low-powered devices in wireless sensor networks. It also acts as a replacement for conventional power sources such as batteries. RF energy harvest uses an unlimited source and makes efficient use of the existing energy in the surrounding environment. The use of machine learning techniques to predict the suitability of RF energy harvest under specific conditions further enhances the performance of energy harvesters. Such a prediction depends on several parameters, such as the time of the day, the temperature, the distance from source, the water density in the air, etc. These have a direct effect on the quality of the received signal at the harvesting node and thus, the harvested energy. In this thesis, a simulation of an RF energy harvesting network using MATLAB to collect relevant data is proposed. This data is used to train different machine learning models: Logistic Regression, Classification Trees, Support Vector Machines and Naïve Bayes in RStudio. The outcomes of the machine learning models are used to enhance the energy harvesting modules’ performance by scheduling them to be on or off with a given set of parametric values. The most suitable model for the dataset being used is chosen based on accuracy, F-Measure and Area Under the Curve. All the models evaluated in this thesis show a performance of 95% and above when tested.