http://repositorio.ufc.br/handle/riufc/492 2026-11-15T03:13:50Z 2026-11-15T03:13:50Z Pimentel, Helano Wilson http://repositorio.ufc.br/handle/riufc/88388 2026-10-28T16:25:48Z 2026-01-01T00:00:00Z

Título: Desenvolvimento de concreto permeável utilizando escória de aciaria BSSF: avaliação do desempenho técnico e ambiental Autor(es): Pimentel, Helano Wilson Abstract: The increasing imperviousness of urban areas and the high consumption of mineral resources by the construction industry have driven the search for solutions that integrate sustainable drainage, circular economy, and environmental impact reduction. This thesis aimed to develop and evaluate lower-environmental-impact pervious concretes through the complete replacement of natural aggregates with Baosteel Slag Short Flow (BSSF) steel slag and the exploratory use of alkali- activated binders. The methodology comprised the development and characterization of different mixtures, including physical and morphological analyses using the Aggregate Imaging Measurement System (AIMS), mechanical and hydraulic tests according to ABNT NBR 16416:2015, and evaluations of clogging, ultrasonic pulse velocity, electrical resistivity, thermal behavior, and Life Cycle Assessment (LCA). Mixtures containing alkali-activated binders based on fly ash and BOF slag, combined with BSSF aggregate, were also evaluated. The results demonstrated the technical feasibility of completely replacing natural aggregates with BSSF, culminating in the granting of a patent by the National Institute of Industrial Property (INPI). The P1 mixture exhibited mechanical and hydraulic performance compatible with the normative requirements, allowing the slab thickness to be reduced from 8 to 6 cm. The thermal results indicated a potential contribution to urban heat island mitigation, while drainage capacity was recovered after cleaning procedures. The LCA indicated reductions of 23.8% in CO₂-e emissions compared with conventional commercial pervious concrete and 48.4% compared with an academic reference mixture, even when the impacts associated with slag processing and transportation were considered. The alkali- activated systems showed additional potential for reducing environmental impacts. It is concluded that the complete replacement of natural aggregates with BSSF steel slag represents a technically feasible and environmentally promising alternative for sustainable pervious pavements. Tipo: Tese

2026-01-01T00:00:00Z Maia, Luana Samara Paulino http://repositorio.ufc.br/handle/riufc/88341 2026-10-21T00:30:33Z 2026-01-01T00:00:00Z

Título: Design direto e inverso de nanoestruturas plasmônicas metálicas por meio de simulações FEM e modelos de regressão Autor(es): Maia, Luana Samara Paulino Abstract: The growing interest in surface plasmonics is driven by the ability of metallic nanostructures to confine and enhance electromagnetic fields, enabling applications in biosensing, optical com- munications, data storage, and energy conversion. However, research in this field is constrained by the high costs associated with fabrication and characterization, the need for specialized infrastructure, and the significant computational cost of numerical simulations. This study inves- tigated the plasmonic properties of cylindrical metallic nanostructures and developed regression models for the direct prediction of their optical responses and the inverse prediction of their geometric parameters. Gold, silver, and copper nanocylinders were modeled using COMSOL Multiphysics 6.2, considering heights ranging from 100 to 300 nm, radii from 50 to 150 nm, and wavelengths between 400 and 1200 nm. A total of 693 simulations were performed, re- sulting in a database comprising approximately 139,293 samples, which was used to train and evaluate the CatBoost, Random Forest, and Extra Trees models. The simulations revealed a redshift of the plasmonic resonances with increasing nanostructure dimensions, along with the emergence of different resonant modes. In the direct design approach, the models predicted the optical scattering response with coefficients of determination close to 0.999, whereas in the inverse design approach, the CatBoost model achieved an absolute error of only 0.25 nm when predicting a radius that was excluded from the training dataset. In addition to their high predictive accuracy, the proposed models substantially reduced the time required to obtain new predictions: while a single finite element simulation could require approximately two hours of computation, depending on the available computational resources, model training and prediction were completed in approximately two minutes. The analysis of experimental absorbance data demonstrated the initial feasibility of applying the proposed methodology to real experimental data, although its performance was limited by the quantity and variability of the available sam- ples. The results demonstrate that the proposed machine learning models provide a promising approach for the direct prediction of optical responses and the inverse estimation of the geometric parameters of plasmonic nanostructures, reducing the need for additional numerical simulations and accelerating their analysis and design process. Tipo: Tese

2026-01-01T00:00:00Z Lima, Pedro Henrique Pinheiro http://repositorio.ufc.br/handle/riufc/88215 2026-10-06T14:56:52Z 2026-01-01T00:00:00Z

Título: Enhancement of mechanical properties of near-eutectoid steel: phase transformations and crystallographic texture Autor(es): Lima, Pedro Henrique Pinheiro Abstract: This study investigates the processing–microstructure–property relationships in a near- eutectoid pearlitic steel (~Fe–0.7 wt.% C) subjected to controlled thermal and thermomechanical treatments, with emphasis on martensitic tempering, carbon redistribution, crystallographic evolution, and cross-rolling processing with 90° strain-path rotation relative to the rolling direction (RD). Particular attention is given to the mechanisms governing carbide precipitation, defect relaxation, texture evolution, mechanical performance, and corrosion behavior. Initially, the tempering behavior of quenched martensitic structures was investigated through dilatometry, differential scanning calorimetry (DSC), X-ray diffraction (XRD), electron backscatter diffraction (EBSD), and transmission electron microscopy (TEM). The results demonstrated that tempering at 250 °C promotes the precipitation of coherent or semi-coherent Fe2C transition carbides associated with high lattice distortion and elevated dislocation density. Increasing the tempering temperature to 350 °C promotes carbon redistribution, cementite (Fe3C) formation, and recovery processes, resulting in reduced lattice distortion, lower defect density, and greater crystallographic homogeneity. This condition provided the best balance between strength and toughness, reaching an ultimate tensile strength increase of approximately 15.56% compared with the as-received condition. The second stage focused on thermomechanical controlled processing (TMCP), particularly the effects of cross rolling with 90° rotation between passes compared with conventional rolling. The results revealed that the deformation route strongly influences colony fragmentation, crystallographic texture, and anisotropy. Conventional rolling promoted elongated pearlitic colonies and strong γ-fiber intensification, whereas cross rolling redistributed deformation across multiple crystallographic variants, reducing texture sharpness and promoting more homogeneous lamellar fragmentation. Processing at 700 °C enabled partial recovery and improved deformability, while severe deformation at 450 °C resulted in intense strain hardening and premature cracking at high reductions. Among all investigated conditions, the cross-rolled sample processed at 700 °C with 75% reduction exhibited the most favorable combination of microstructural refinement, reduced anisotropy, moderated hardening, and mechanical stability, demonstrating the effectiveness of strain-path rotation as a strategy for microstructural optimization in eutectoid steels. Finally, corrosion behavior was evaluated through immersion tests in synthetic seawater combined with grazing incidence X-ray diffraction (GIXRD) and optical microscopy. The results showed that corrosion morphology is directly controlled by the underlying microstructure. Microstructures with high defect density and strong phase heterogeneity promoted localized electrochemical activity and heterogeneous oxide formation, whereas more stable and homogeneous conditions produced comparatively more uniform corrosion layers. Magnetite was identified as the dominant corrosion product in most conditions, with partial transformation to hematite in the outer regions of the oxide scale. Overall, the results demonstrate that controlled thermal and thermomechanical processing can significantly modify the crystallographic configuration, defect structure, and phase distribution of near-eutectoid steels. In particular, the application of cross rolling with 90° strain-path rotation proved highly effective in reducing anisotropy and improving microstructural homogeneity, establishing a direct relationship between deformation route, texture evolution, mechanical performance, and corrosion behavior. Tipo: Tese

2026-01-01T00:00:00Z Bampoky, Nayuca Alberto http://repositorio.ufc.br/handle/riufc/87940 2026-09-11T19:33:52Z 2026-01-01T00:00:00Z

Título: Green synthesis of CoFe2O4 nanoparticles and their application as bifunctional electrocatalysts in water electrolysis Autor(es): Bampoky, Nayuca Alberto Abstract: The development of efficient and sustainable electrocatalysts based on earth-abundant elements is essential for advancing green hydrogen production through water electrolysis. This doctoral thesis investigates cobalt ferrite (CoFe2O4) nanoparticles synthesized by environmentally friendly routes and evaluates their performance as bifunctional electrocatalysts for the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER) in alkaline media. The study was conducted in two stages: (i) optimization of the sonochemical synthesis by varying the ultrasonic power and (ii) comparison of the optimized sonochemical route with hydrothermal and protein-assisted sol–gel methods. The results demonstrate that increasing the ultrasonic power promotes structural disorder and a higher density of surface defects, as evidenced by Mössbauer spectroscopy, leading to a significant increase in the electrochemically active surface area and improved electrocatalytic kinetics. The optimized sample, synthesized at 60% ultrasonic power (designated as CoFe2O4-60), exhibited the best overall performance for both HER and OER. Comparative analyses revealed that the protein-assisted sol–gel route produced smaller nanoparticles with intrinsic lattice strain, whereas the hydrothermal route resulted in incomplete phase formation with hematite impurities. In contrast, the sonochemical method yielded highly pure spinel CoFe2O4 nanoparticles without requiring post-synthesis calcination. Electrochemical measurements confirmed the superior catalytic activity and long-term stability of the sonochemically synthesized material compared with the hydrothermal and protein-assisted sol–gel counterparts. These findings demonstrate that sonochemical synthesis is an effective route for tailoring the structural properties of CoFe2O4 and producing highly active bifunctional electrocatalysts for alkaline water electrolysis. The work provides a sustainable and scalable approach to developing low-cost alternatives to noble-metal-based catalysts for green hydrogen production. Tipo: Tese

2026-01-01T00:00:00Z