The oxygen evolution reaction (OER) is a crucial step in water electrolysis for sustainable hydrogen production. However, the dependence on noble-metal catalysts like IrO2 and RuO2 limits scalability due to their high cost and scarcity. This study aims to develop a cost-effective and sustainable OER electrocatalyst by synthesizing a bimetallic Bi-Fe/N-doped carbon (Bi-Fe/NC) catalyst derived from zeolitic imidazolate framework-8 (ZIF8). By leveraging Fe-Bi interactions and nitrogen-doped carbon support, this research explores an economically viable approach for improving OER efficiency in non-noble electrocatalyst.
The catalyst was synthesized via a one-pot solvothermal method at 120 °C for 4 h, followed by pyrolysis. The study examined the effects of pyrolysis temperature, Bi to Fe ratio, and metal-to-ZIF8 ratio on catalytic performance. Higher pyrolysis temperatures enhanced conductivity and electrochemically active surface area (ECSA), while an optimal 2:8 Bi-to-Fe ratio improved graphitization and active oxide formation. Additionally, in an effort to remove oxides without acid treatment, lower metal loading (0.025) resulted in reduced OER activity due to fewer active sites. The optimized Bi2Fe8NC 0.05 catalyst exhibited an overpotential of 397.48 mV at 10 mA cm-2 and a Tafel slope of 80.61 mV dec-1 in 1.0 M KOH, highlighting electronic conductivity and ECSA as the dominant factors influencing OER performance. Stability tests showed a 21.20% increase in overpotential after 2000 cycles and stable chronoamperometric performance for 7.5 h before noticeable degradation. These findings highlight Bi-Fe/NC catalysts as promising alternatives to noble-metal-based OER catalysts. However, long-term stability remains a challenge, likely due to metal oxide leaching and structural degradation. Future studies should focus on enhancing catalyst durability and employing computational methods to further investigate
reaction mechanisms.

Nonlinearity in communications systems negatively impacts its performance. In radio over fiber systems, lasers are one of the primary contributors to link nonlinearities. There are existing major techniques in link linearization, namely, hardware compensation, and digital signal processing (DSP). DSP offers promising link performance at a lower cost. Improvements in the computational time and complexity of the system are needed for the system to adapt with the conditions of the link close to real-time.

The dissertation aims to characterize and linearize two different Intensity Modulation - Direct Detection RoF systems and analyze the degree of nonlinearity caused by the laser diode. Models are used to create a DSP-aided RoF link linearization, exploring the different techniques: predistortion, postdistortion, and dual compensation. The links are electronically linearized by estimating the predistortion through Memory Polynomial Method (MP) and Generalized Memory Polynomial Method (GMP), improved through the novel implementation of the Improved Cann Model as postdistortion, and a combination of the two techniques through dual compensation.

In digital predistortion, ACPR improved by 5 dB, and EVM improved by up to 11%. Memory effects were observed to be mitigated as well in the AM-AM and AM-PM plots. This technique requires initial training and high memory depth with increased computation time and complexity. Performance improvement using the novel digital postdistortion is
seen in the EVM of the link, with as much as a 12% difference from the original received OFDM signal, improving the scatter and correcting memory effects. Improvements observed in the system are limited to in-band signals. Dual compensation proposed improved both in-band and out-of-band signals, yielding the best ACPR values, despite having a lower memory depth with approximately 7 dB improvement. Application of the postdistortion reduces computational complexity in the predistorter, lowering the memory depth required to have a better performance of the link.

Autoimmune diseases arise from immune-mediated attacks on healthy tissues, driven by autoantibodies that target self-antigens and amplify inflammation through sustained activation of the Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathway. Pro-inflammatory cytokines, often elevated in autoimmunity, hyperactivate JAK/STAT signaling, which in turn promotes autoreactive B and T cell responses, further increasing autoantibody production. This self-reinforcing cycle disrupts immune tolerance, perpetuates tissue damage, and accelerates disease progression, highlighting JAK/STAT as a critical node linking autoantibody-driven inflammation to autoimmune pathogenesis. While JAK inhibitors have proven effective in controlling such diseases, their clinical use has been limited by adverse effects, including increased risks of infections, cardiovascular events, and certain cancers.
This study aimed to identify potential inhibitors of Janus kinases (JAK1, JAK2, JAK3, and TYK2), key components of the JAK/STAT pathway, by screening 2,607 phytochemicals from the MMRL virtual library. ADMETLab 3.0 was used to filter compounds based on drug-likeness and safety properties, and 24 candidates met all criteria for further analysis. Molecular docking analyses using AutoDock 4.2 and AutoDock VINA revealed that five promising candidates—cycloeicosane, ambrettolide, (10Z)- oxacyclononadec-10-en-2-one, (Z)-oxacyclopentadec-6-en-2-one, and myricetin 3- rutinoside—have competitive docking scores across all JAK kinases, comparable to or exceeding those of reference drugs such as peficitinib, upadacitinib, and tofacitinib. Molecular dynamics simulations indicated that these five phytochemicals are stable when in complex with the four Janus kinases within the 100-ns simulation timeframe. Hydrogen bond heatmaps highlighted the unique ability of myricetin 3-rutinoside to form consistent hydrogen bonds, while other candidates primarily relied on hydrophobic interactions. The MM/PBSA results revealed that cycloeicosane exhibited the strongest binding affinity across all JAK kinases—JAK1 (-40.31 kcal/mol), JAK2 (-38.70 kcal/mol) JAK3 (-38.69 kcal/mol), and TYK2 (-35.54 kcal/mol). Myricetin 3-rutinoside, ambrettolide, (Z)- oxacyclopentadec-6-en-2-one, and (10Z)-oxacyclononadec-10-en-2-one also showed competitive binding affinities. Based on these findings, cycloeicosane, ambrettolide, (10Z)-oxacyclononadec-10-en-2-one, (Z)-oxacyclopentadec-6-en-2-one, and myricetin 3- rutinoside are recommended for further in vitro and in vivo evaluations to validate their inhibitory potential against JAK kinases, paving the way for the development of novel therapeutics for autoimmune diseases.

Anti-cancer drugs non-selectively interact with healthy cells apart from the target cells, which causes severe side effects suffered by cancer patients. Two of the existing nanocarriers that are widely studied are liposome-based and carbon-based. However, liposomal formulations, which are already available in the market, have only limitedly improved the overall survival of the treated patients (Gabizon et al., 2016; Moosavian et al., 2021). While carbon nanotubes reportedly persist in the organs for several weeks, which causes long-term toxicity (Liang et al., 2020;W. Chen et al., 2023 ). In this study, Rosette Nanotubes (RNTs) are introduced as the drug vehicles for anti-cancer drugs. RNTs such as K1, xK1, and iEt-xK1 are potential as anti-cancer drug vehicles with their self-assembling, biocompatible, amphiphilic, and low toxicity properties. Chlorambucil (CBL), Camptothecin (CPT), Doxorubicin (DOX), Flutamide (FLU), and Paclitaxel (PTX) molecules are selected for this study and are widely known anti-cancer drug molecules used to treat various kinds of cancers, including lung, breast, prostrate, leukemia, and lymphomas. This study comprises three phases: 1) molecular and thermodynamic stability of each drug with each RNT, 2) self-assembly pathways of each RNT for encapsulation, and 3) release mechanisms of each drug from each RNT. Molecular dynamics (MD) simulations and energetic analysis using the three-dimensional reference interaction site model (3DRISM) are used to investigate the molecular and thermodynamic stability of each RNT as drug carriers for the selected drugs. This investigation involved detailed stability, interactions, structural, and energetics analyses of the complexes. The potential for encapsulation of these drugs by RNTs was also explored by simulating and calculating the energetic analysis of the proposed self-assembly pathways of RNT motifs with the drug molecules. The mechanisms involved in the release of drug molecules from RNTs were also investigated by reflecting the acidic environment of the cancer cells at low pH (pH = 5.1).
CPT and DOX demonstrated strong π–π stacking and stable intercalation within K1 (-24.41 and -21.18 kcal/mol), and they also showed favorable binding (-27.29 to -4.5 kcal/mol) at the ends of each RNT, attributed to their planar aromatic ring structures. On the other hand, CBL and FLU consistently favored end-localization (-15.47 to -4.83 kcal/mol) for all RNTs. PTX, due to its bulky structure, exhibited repulsive interactions (1.31 to 790.60 kcal/mol) within the inner channels of all RNTs, indicating poor encapsulation potential. Energetic profiling identified K1 as the potential drug vehicle candidate for internal encapsulation through self-assembly, with the presence of post-assembly potential energy barriers contributing to drug retention. In comparison, xK1 and iEt-xK1 showed minimal energy differences between their unloaded and drug-loaded forms, indicating limited internal barriers and a tendency toward end-localization. Based on energetic analysis, aggregate-based stacking emerged as a viable self-assembly mechanism, favorably for xK1 and iEt-xK1, offering an alternative to ring-stacking. Protonation simulations confirmed successful drug release following structural disassembly, validating the responsiveness of these systems under acidic conditions. Structural, visual, and energetic analyses collectively highlight K1 as the potential carrier for stable drug encapsulation among the types of RNTs studied here. The findings here offered insights on the viability of RNTs for targeted drug delivery, which warrants further experimental studies.

Microbial fuel cells (MFCs) represent a promising technology for sustainable
electricity generation and wastewater treatment. However, their long-term efficiency is
hindered by limited substrate availability for electroactive bacteria such as Shewanella
oneidensis MR-1. This study introduces a three-layered MFC design incorporating MR-1
with Pseudomonas strains and ammonium-oxidizing bacteria (AOB) to enhance
bioelectricity production. Extracellular polymeric substances (EPS) produced by AOB
strains—particularly Nitrosomonas europaea and Nitrosococcus watsonii—were found to
be rich in proteins and polysaccharides, serving as degradable organic substrates.
Introducing Pseudomonas strains improved EPS breakdown, promoting MR-1 biofilm
development and electron transfer efficiency. Peak power densities of 0.2794 W/cm² and
0.2045 W/cm² were achieved with S. oneidensis MR-1 + Pseudomonas putida + N.
europaea and N. watsonii, respectively. In contrast, P. aeruginosa–based systems produced
lower peaks (0.0819 W/cm² and 0.05903 W/cm²) but maintained power generation for
longer durations—up to 69.5 hours. These results highlight a trade-off between power
density and operational longevity, offering insights into tailoring microbial consortia for
specific energy and treatment goals. The structured integration of MR-1, Pseudomonas, and
AOB provides a scalable strategy for enhancing MFC performance in ammonium-rich
wastewater environments.