70040 research outputs found
Sort by
Promoted osteogenesis on calcium modified surface of stainless-steel lattice produced by selective laser melting
No full textStainless steel has been widely used as an implant material for its good biocompatibility, suitable mechanical strength, and high corrosion resistance in vivo. However, its biomedical applications suffer from delayed healing due to its high density and stiffness. Here we proposed body-centered cubic lattice structures with various unit sizes to adjust the density and stiffness of 17-4 PH stainless steel implants to simulate the bone structure and mechanical performance. The mechanical properties satisfy the requirement to be used with the human body with a yielding strength over 60 MPa and Young’s modulus over 1.7 GPa. Corrosion resistance characterization indicates that the implants have negligible changes in microstructures and mechanical properties in simulated body fluid for 6 months. The implants were modified by inserting calcium sulphate-based bone cement into the voids of the lattice to improve their biocompatibility. Cytotoxicity results showed that both the implants and modification have no toxicity to human bone marrow mesenchymal stem cells. In vivo safety and osseointegration testing of the implants were conducted by implantation in rabbit distal femur, showing an improved recovery and bone integration of the implants. The presence of calcium sulphate and tailored lattice structure synergistically promotes osteogenesis through controlled calcium ions release and matching the mechanical properties of the bone
Eco-pilgrimages: Linking humans, heritage, and hydrology
No full textOver the last century, the health of aquatic ecosystems around the world has reached critical levels. In the UK, waterways are severely polluted, and yet many wells and springs are still venerated as ‘sacred’. This article presents ‘eco-pilgrimages’ as a sustainability strategy to connect key heritage sites through ecological corridors. This aims, simultaneously, to strengthen biodiversity; to enable immersive historical and ecological education; to contribute to human well-being; and to provide more effective flood amelioration in river catchment areas
Hydrogen Sulfide Mitigation in Pulp and Paper Mill Water Systems
No full textThe presence of hydrogen sulfide (H2S) in pulp and paper mills poses substantial work health safety risks due to its toxicity, while also introducing technical challenges related to corrosion and odour. The existence of recalcitrant sulfate and organic compounds under anaerobic conditions exacerbates this issue as it promotes the biogenic production of dissolved sulfide via sulfate-reducing bacteria (SRB). This thesis aims to systematically evaluate a range of biological, chemical, and physical treatment methods to limit this biogenic production, and the subsequent presence of dissolved sulfide in a paper machine process water system through four distinct approaches: mitigation of SRB activity (sodium nitrate and sodium nitrite), chemical oxidation (hydrogen peroxide and potassium permanganate), metal sulfide precipitation (ferrous, ferric, zinc, and cupric salts), and aeration (fine bubble membrane diffusion).
Nitrate and nitrite, each dosed at 30 mg-N/L, inhibited the rate of sulfate reduction by 49 and 96% respectively after 48 hr. Both studied chemical oxidants achieved dissolved sulfide removals greater than 85%, however overall reaction efficiency was low, requiring 10.6 g of hydrogen peroxide or 30.6 g of potassium permanganate per g of dissolved sulfide. This low performance is attributed to an upstream reductive bleaching process. Metal sulfide precipitation required between 3.9 to 5.3 g of each metal ion per g of dissolved sulfide to achieve reductions between 87 to 96%, though all caused negative discolouration of the process water apart from zinc. Pilot scale aeration provided an attractive chemical-free approach, leveraging both degassing and oxidation to achieve a maximum of reduction of 65% via waterfall aeration, and up to 85% when combined with diffused aeration. In the dynamic pilot system, it was found that 5.3 g of zinc ions were required per g of dissolved sulfide to achieve 88% reduction. Zinc sulfide was also shown to be successfully retained in kraft handsheets without impacting chromatic properties, highlighting a potential treatment method for full scale implementation.
Future work should address the compatibility of sodium nitrite within the paper manufacturing process, downstream effects of zinc sulfide formation, the impact of diffused aeration mediated foaming and iron precipitation, and alternative techniques for the separation of insoluble metal sulfide precipitates from valuable fibres
SolarShift Customer Hot Water Roadmap Report
No full textPublic facing online tool to help households in deciding the most suitable water heating technology in Australi
Iridium Nanocrystals Enriched with Defects and Atomic Steps to Enhance Oxygen Evolution Reaction Performance
No full textThe presence of defects can significantly improve catalytic activity and stability, as they influence the binding of the reactants, intermediates, and products to the catalyst. Controlling defects in the structures of nanocrystal catalysts is synthetically challenging. In this study, we demonstrate the ability to control the growth of Ir nanocrystals, enabling the tuning of both structural and surface defects. The Ir nanocrystals have unique structures that range from single crystals of a few nanometers to twinned nanoparticles and multiply twinned crystallites with a high density of atomic steps. This approach of defect engineering enables us to understand their roles in enhancing the performance of the OER and producing an Ir catalyst with both high activity and stability. Our results show the importance of the concept of using synthetic control of structural and surface defects in metal nanoparticles as a strategy to improve catalytic performance
Understanding regional occurrence patterns of pyroCb over temperate southeast Australia
No full textPyrocumulonimbus clouds (pyroCb) are extreme fire-driven convective storms that pose significant threats to ecosystems, human safety, and global atmospheric systems. In recent years, the frequency of pyroCb events in Australia has risen markedly, driven by escalating wildfire activity and changing climatic conditions. Particularly in the spring and summer of the 2019–2020 (Black Summer) fire season, multiple pyroCb events occurred simultaneously at a regional scale, challenging the conventional understanding of pyroCbs as spatiotemporally isolated events. Furthermore, with global warming, extreme wildfires are expected to become more frequent and intense, suggesting that pyroCb events have the potential to become a more common large-scale hazard. In this context, understanding the regional-scale occurrence patterns of pyroCbs is critical for informing future wildfire management strategies. This thesis employs statistical modelling to provide insights into pyroCb drivers and their geographical variability at a regional scale and leverages these insights to project how pyroCb occurrence may change in response to climate change.
This thesis first investigated the effect of atmospheric instability (represented by the Continuous Haines Index, C-Haines) and fuel moisture (represented by the Fuel Moisture Index, FMI) on pyroCb occurrence. The variables were related to the probability of pyroCb occurrence by employing logistic regression. The analysis revealed that high atmospheric instability and low fuel moisture are representative of favourable conditions for pyroCb occurrence. However, the model using only atmospheric instability and fuel moisture demonstrated limited explanatory power for pyroCb occurrence in Victoria, particularly when discerning pyroCb events from large wildfires, highlighting the necessity of incorporate additional variable for more comprehensive analyses.
Subsequently, this thesis employed logistic and random forest modelling techniques to further analyse the geographical drivers of pyroCb occurrence, encompassing four categories: atmosphere, surface weather, topography, and vegetation. These analyses quantified the importance of each variable on pyroCb occurrence. While the variable importance rankings differ slightly between the random forest and logistic models, atmospheric variables, particularly C-Haines, serve as the most robust index of pyroCb occurrence across both modelling approaches. Furthermore, significant regional differences in pyroCb drivers were identified.
To gain deeper insights into these regional differences, Geographically Weighted Regression (GWR) was integrated with the logistic model to reveal the regional differences in the influence power of each variable. Atmospheric and surface weather variables were found to have significant influence in the northeastern part of the study area, while topographic and vegetation variables are more influential in the southwestern area. Notably, the GWR model can improve large-scale risk assessment by considering local variations in the influence of different factors. However, the analyses also revealed the need for sufficient data to support robust analysis.
Finally, using the two most representative variables, C-Haines and FMI, and the results of the GWR model, this thesis evaluated the response of pyroCb risk to climate change under different future climate scenarios. The findings indicate that in the near future (2025–2059), pyroCb risk is projected to increase across most of temperate southeast Australia under both high and low greenhouse gas emission scenarios. In the far future (2065–2099), pyroCb risk is projected to decrease compared to near-future in some inland areas under a low-emission scenario, while a decrease is projected in some coastal regions under a high-emission scenario.
Overall, the findings of this thesis provide valuable insights into the occurrence patterns of pyroCb over southeast Australia, which enhance extreme wildfire risk management and improve risk assessments in the context of climate change
Catalytic Mechanism of Zirconium Metal-Organic Frameworks for the Selective Capture and Degradation of Organophosphates
No full textZirconium-based metal-organic frameworks (Zr-MOFs) are studied as versatile materials for both sensing and catalytically degrading hazardous organophosphorus (OP) compounds such as pesticides and nerve agents. This thesis explores the catalytic performance and sensing capabilities of multiple Zr-MOFs in degrading environmentally persistent OPs, specifically glyphosate (GPh), malathion, and nerve agents such as sarin (GD) and VX. Degradation studies were completed comparing the performance of different crystal sizes of structures such as MOF-808, UiO66 and UiO-67-NH2. The studies focused on solid spectroscopy analysis of MOFs through synchrotron techniques that allow for the identification of coordinated reactive species, which, combined with nuclear magnetic resonance (NMR) and chromatographic techniques, permit the elucidation of the degradation routes. The studies showed that GPh can be degraded at room temperature using MOF-808, with better performance obtained by nanosized crystals named (nMOF-808, 65 nm). The reaction allows to obtain benign products such as N-formyl glycine and hydroxymethyl-phosphonate that remains attached to the MOF. The very reactive performance is attributed to higher hydroxyl and water coordination on the secondary building units (SBUs), as revealed by X-ray absorption spectroscopy (XAS), enabling fast ligand exchange at unsaturated Zr sites. Similarly, malathion degradation was studied across MOF-808, UiO-66, and UiO-67-NH₂ frameworks at two different crystal sizes. NMR and synchrotron techniques confirmed effective breakdown, yielding significantly less toxic products—ethyl succinate and O,O-dimethyl hydrogen phosphorothioite—at rates up to 0.3 g of malathion per gram of MOF. These results support the use of Zr-MOFs as biomimetic phosphatase-like catalysts for pesticide decontamination. To enhance Zr-MOFs for real-world applications, MOF-808 was also functionalized with a Rhodamine-derived sensor (RDS) for fluorescence-based detection of OP. Crystal size influenced sensor coordination and, consequently, the optical response observed. The RDS-modified MOFs demonstrated strong fluorescence and visible changes upon exposure to diethyl chlorophosphate (DCP), while also showing degradation activity against real chemical warfare agents (CWAs) such as GD (sarin), HD (mustard gas), and VX. Together, these findings highlight the tunable catalytic and sensing potential of Zr-MOFs for environmental remediation and chemical defense, driven by crystal engineering and ligand-field control strategies
Unveiling the evolutionary code of NOTCH3: mammalian bioinformatics sheds light on human pathogenicity
No full textNOTCH3 is a highly conserved transmembrane receptor implicated in CADASIL, a hereditary small vessel disease driven by mutations in its extracellular EGF-like repeats. The mechanism by which these mutations cause pathology remains unclear. We present the first large-scale comparative bioinformatic analysis of NOTCH3 across 113 mammalian species, uncovering three novel insights: i) a remarkable evolutionary conservation of all 204 cysteines, with the only exception being eight naturally occurring cysteine mutations in jaguar (EGFr13-15); ii) a unique deletion in Brandt’s bat regulatory region, which may expose it to proteases, potentially altering signaling; iii) a rare human NOTCH3-X1 isoform, absent in most mammals but shared with select primates, a bat, and elephants, involving a cysteine-depleting deletion spanning EGFr20-22. These features provide novel evolutionary insights into human pathogenicity and suggest testable targets for in vivo experiments. Our study highlights the potential of comparative bioinformatics to identify previously hidden functional elements in disease-associated mammalian proteins
Explainable Bias Mitigation in Transformer-Based Medical Image Analysis
No full textFairness is one of the major concerns when utilising deep learning models for medical image analysis in clinical practices. In
this context, fairness is often defined as ensuring minimal performance gaps across various cohorts of the input data based
on participants’ demographics, such as skin tone, age, gender, etc, which are often referred to as sensitive attributes.
Vision transformers (ViT) have emerged as powerful deep learning models in medical image analysis, demonstrating
outstanding performance across a wide range of tasks. These models are designed to learn very complex patterns in the
given data, making them susceptible to unreasonably associating sensitive attributes with the outcomes. That is why
their adoption in clinical settings raises critical concerns about their ethical deployment and trustworthiness. Despite
recent advances in bias mitigation techniques, existing approaches primarily focus on convolutional neural networks
(CNNs) and rely heavily on retraining or global modifications, which are computationally expensive and lack precision
in addressing the root causes of bias. Moreover, the unique challenges posed by ViTs, particularly their global attention
mechanisms that could amplify biases, remain unexplored.
To fill these gaps, this thesis proposes two novel post-processing methods to mitigate bias in ViTs. Our initial work
introduced XTranPrune, an explainability-aware pruning method that leverages explainability methods to identify and
remove discriminatory nodes within ViT architectures. By utilising the derived attribution score from the explainability
method, we pinpoint nodes responsible for biased predictions while preserving performance-critical components. Building
upon this foundation, in our subsequent work, we redesign the approach for identifying the biased nodes, enhancing the
efficiency of the pruning mask generation module. Additionally, we incorporate uncertainty quantification into the pruning
process. This enhanced method employs a contrastive framework to refine node attribution scores and iteratively generate
robust pruning masks, further improving fairness.
We evaluated these methods on three diverse medical imaging datasets: Fitzpatrick17k, PAD-UFES-20, and Harvard-
GF3300. We use a comprehensive set of fairness metrics alongside traditional classification measures. Comprehensive
ablation studies demonstrated that both methods significantly reduced bias across multiple sensitive attributes while
preserving classification performance. Notably, our advanced method outperformed state-of-the-art bias mitigation
techniques with greater precision and robustness. These contributions establish a new standard for fairness enhancement
in transformer-based medical image analysis, paving the way for more equitable and reliable machine-assisted diagnostics
Teacher collaboration: conceptualisation and practice
No full textTeacher collaboration has been recognised as one of the fundamental approaches to enhancing learning and teaching. However, despite its prominence as a professional development approach, it is conceptually under-theorised, with competing definitions, understandings and practices with gaps remaining in our understanding of teacher collaboration practice. Using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline, we accessed and reviewed 66 studies to investigate how teacher collaboration is conceptualised and practised in the literature. Findings indicate that the construct is predominantly shaped by Western research, lacks a universally accepted definition, and is often used interchangeably with related terms. The analysis further reveals that practices of teacher collaboration exist along a continuum from informal to highly structured approaches, highlighting the need for a comprehensive framework that balances structure with contextual flexibility. Our analysis shows geographical, conceptual and practical gaps which have implications for the understanding and practice of teacher collaboration