The Nelson Mandela African Institution of Science and Technology
Institutional Repository
Research Collection
Select to browse Research collections.
Recent Submissions
Enhancement of eggshell-kaolin composite for fluoride removal in water
(NM-AIST, 2026-07) Kitemangu, Aisha
Fluoride (𝐹⁻) in groundwater is a major public health concern when concentrations exceed the WHO guideline of 1.5 mg/L, particularly in fluoride-prone regions such as Manyara. This study investigated the spatial distribution of 𝐹⁻, its relationship with hydrogeochemical and physicochemical parameters, and the associated non-carcinogenic health risks in the Manyara region. To address this, a novel ternary eggshell-kaolin-MgO composite was synthesized via wet impregnation and co-precipitation methods; its surface and physicochemical properties were characterized, and then assessed for defluoridation in batch and fixed-bed column systems. Results show that the highest percentage of groundwater samples that exceeded the WHO limit was found in Simanjiro (85.71%) and Hanang (79.25%). The dominant hydrogeochemical facies was Ca-Mg-HCO3, while groundwater with high fluoride concentrations (>10 mg/L) was predominantly associated with the Na-HCO3 hydrogeochemical facies. The ionic relationships suggest that fluoride mobility and removal may be influenced not only by pH, alkalinity, and the (Ca2+ + Mg2+)/(Na+ + K+) ratio, but also by evaporation, water-rock interaction, and ion-exchange processes. Batch adsorption studies showed that equilibrium data followed the Freundlich isotherm, indicating multilayer adsorption on heterogeneous surfaces, while kinetics were best described by the pseudo second-order model (R2 = 0.996). The process was spontaneous and endothermic (ΔH° = 52.28 kJ/mol, ΔS° = 180.07 J/mol·K, ΔG° = –1.41 to –5.01 kJ/mol). Combined characterization and adsorption modelling suggested that fluoride removal by the CEKM composite involved both physisorption and chemisorption via electrostatic attraction, hydrogen bonding, and surface complexation. Maximum 𝐹⁻ removal (95%) occurred at 180 min, 0.6 g, pH 7, and 5 mg/L initial 𝐹⁻ concentration. Column experiments demonstrated high adsorption capacity (6.35 mg/g), good model fit (R2 ≥ 0.94), and effective regeneration over four cycles. Co-existing anions inhibited 𝐹⁻ uptake in the order PO₄³⁻ > SO₄²⁻ > NO₃⁻ > Cl⁻. The synthesized composite exhibited good adsorption performance, which may be attributed to its high specific surface area (158.5 m²/g) and the presence of active Ca2+ and Mg2+ sites, suggesting its potential as an efficient and sustainable adsorbent for fluoride remediation in groundwater.
Sulfated and non-sulfated titanium-niobium mixed oxide catalysts for the selective dehydration of biomass-derived xylose to furfural: catalyst designe and process optimization
(NM-AIST, 2026-06) Saidi, Sophia
Lignocellulosic biomass is increasingly recognized as a sustainable alternative source of carbon-based chemicals and fuels. It includes agricultural residues, wood, and forestry waste, which are abundant, low-cost, and widely available. One important product derived from this biomass is furfural (FUR), a versatile platform chemical produced from the dehydration of hemicellulose-derived sugars such as xylose. Furfural is widely used in the production of fuels, solvents, and other industrial chemicals. Conventional furfural production employs strong mineral acids (e.g., H₂SO₄, HCl), which present significant environmental and operational challenges, including waste disposal and equipment corrosion. To address these limitations, solid acid catalysts, particularly mixed metal oxides, have been explored due to their recyclability and lower environmental impact. However, many reported catalysts still suffer from low selectivity and limited stability. In this study, titanium (Ti) and niobium (Nb) mixed oxide catalysts were prepared and tested for xylose dehydration to furfural. Seven catalysts (A–G) with different Ti:Nb ratios were prepared and characterized using FTIR, SEM, N₂ adsorption–desorption, XRD, and XPS. Results confirmed mesoporous structures with type IV isotherms, a predominantly amorphous TiNb₂O₇ phase, and Ti–O, Nb–O, and SO₄²⁻ functional groups. XPS revealed Nb⁵⁺, Ti⁴⁺, and sulfate species, confirming oxidation states essential for acidity. Catalytic screening in aqueous medium showed catalyst F (TiNb₂O₇) exhibited the highest performance, attributed to synergistic TiO₂–Nb₂O₅ interaction providing balanced Lewis and Brønsted acid sites for xylose isomerization and dehydration. This catalyst was further evaluated in a toluene/water biphasic system and optimized using central composite design within response surface methodology. Under optimized conditions, furfural yield, selectivity, and xylose conversion reached 52%, 57%, and 91%, respectively. To further enhance performance, a sulfated catalyst, SO₄²⁻/TiO₂ Nb₂O₅ (STNO), was synthesized via a modified sol–gel method. Sulfate groups increased Brønsted acidity, improving dehydration efficiency. Under optimized conditions, STNO achieved a furfural yield of 63%, selectivity of 74%, and xylose conversion of 98%. This study demonstrates that catalyst design and acidity tuning significantly enhance furfural yield and selectivity, establishing a clear relationship between catalyst composition, structure, and performance, and providing a practical, sustainable approach for biomass conversion into value-added chemicals.
Computational and experimental performance evaluation of a solar air heater integrated with a photovoltaic system for tobacco curing
(NM-AIST, 2026-07) Kilakoi, Seuri
Conventional solar air heater (SAH) drying systems often use opaque baffle materials such as wood, aluminium and mild steel, which limit solar radiation penetration into the airflow channel and reduce thermal performance. This study developed and optimised a solar-assisted tobacco curing system that integrates semi-transparent glass baffles with automated drying control to enhance thermal performance, energy efficiency, and operational reliability. A three dimensional computational fluid dynamics (CFD) model simulated airflow, heat transfer, radiation and moisture transport using Reynolds-Averaged Navier-Stokes (RANS), Renormalisation Group k-epsilon (RNG k-ε), Shear Stress Transport k-omega (SST k-ω) and Discrete Ordinates (DO) radiation models. Response surface methodology (RSM) was used to optimise baffle length, height and spacing, and the optimised system was validated experimentally. The optimised SAH configuration achieved a thermal efficiency of approximately 89.3% at a mass flow rate of 0.02 kg s⁻¹ because of enhanced radiative penetration and convective heat transfer. The optimisation model showed strong predictive capability (R² > 0.98), while validation showed close agreement between simulations and experiments, with a mean absolute percentage error (MAPE) of approximately 3.9%, a root mean square error (RMSE) of approximately 1.7°C, and R² of approximately 0.99. Sensitivity analysis confirmed robustness, with ±5% variations in input parameters producing less than ±3% performance deviation. Quality evaluation indicated that nicotine content (≈2.1-3.03%) and reducing sugar levels (≈17.6-19.1%) remained within acceptable ranges for flue-cured Virginia tobacco. The results show that semi-transparent glass baffles combined with hybrid photovoltaic-thermal (PV-T) technology can improve SAH thermal performance while maintaining desirable tobacco quality. The system is recommended for field-scale testing, economic assessment, and adaptation to other crop drying applications in regions with high solar potential.
The influence of engine mileage on oil performance and emissions: comparing gasoline and compressed natural gas in aged retrofitted spark-ignition engine
(NM-AIST, 2026-06) Kyando, Michael
This study investigates the influence of fuel type and progressive operating duration on lubricant degradation and emission characteristics in an aged, retrofitted spark-ignition (SI) engine. The research is motivated by the increasing adoption of compressed natural gas (CNG) in developing countries, where high-mileage engines are commonly retrofitted without redesign. Experiments were conducted on a Toyota 1NZ-FE SI engine with over 160 000 km service history under controlled laboratory conditions. The engine was operated for 100 hours per fuel (gasoline and CNG) at 2000 rpm, with oil samples collected at 20 hours intervals. Lubricant analysis was performed using standardized methods, including viscosity, total base number (TBN), total acid number (TAN), wear metals, additive elements and fuel dilution. Results indicate that fuel type significantly influences lubricant degradation behavior under identical conditions. Fuel dilution increased from 0 to 1.5% under gasoline compared to 0 to 0.5% under CNG, highlighting the reduced tendency of gaseous fuel to contaminate the lubricant through blow-by. The study identified wear- and viscosity-related parameters as significant indicators of lubricant degradation, particularly iron (Fe), aluminum (Al) and viscosity at 100°C (V100). Iron increased from 0 to 21 ppm under gasoline and 0 to 11 ppm under CNG, while aluminum increased from 0 to 4 ppm and 0 to 2 ppm, respectively. In contrast, viscosity at 100°C decreased slightly under gasoline (9.8 to 9.7 mm²/s) but increased marginally under CNG (9.8 to 9.9 mm²/s), reflecting differing degradation mechanisms and confirming reduced mechanical wear under CNG operation. The study observed that CO and HC emissions complied with EURO IV standards for both fuels, whereas NOₓ emissions exceeded the standard under certain operating conditions. A multiple linear regression model was developed to predict oil change intervals (OCI), achieving strong predictive performance (R² = 0.98). The model identified Fe as the dominant predictor, while viscosity at 100°C and Al were significant negative predictors of oil service life. This study evaluates progressive degradation within a single aged engine, rather than comparative mileage effects. Overall, the findings demonstrate that CNG operation reduces fuel dilution, wear, and lubricant degradation, thereby supporting extended oil service intervals.
Mathematical modelling and performance analysis of a novel hybrid solar-biogas dryer for drying agricultural products
(NM-AIST, 2026-06) Yawe, John
Post-harvest losses in agricultural products, particularly fruits, can reach up to 45% due to inadequate preservation technologies, posing a significant threat to food security and sustainability in developing countries. Hybrid solar-biogas drying systems (HSBD) present a promising solution by combining renewable energy sources for efficient, continuous operation. However, a comprehensive mathematical model that accurately incorporates solar collector thermodynamics, biogas energy supplementation, and drying kinetics under dynamic atmospheric conditions is lacking. This gap hinders the design, performance prediction and optimization of these hybrid dryers, resulting in non-uniform drying, inefficient moisture removal and high operational costs. This study developed and validated a mathematical model to analyze the thermodynamic and kinetic behavior of an innovative HSBD for drying agricultural products. The design includes dual U-grooved solar collectors, uniform hot-air distribution, independent moist-air exhaust outlets, a wind-driven ventilator and locally made biogas burners. The model was based on energy balance equations and experimentally validated using Cavendish banana slices. The drying and exergy performance were assessed using mango slices through energy analysis, with experimentally determined moisture ratios fitted to six thin-layer drying models and run by a computer Programme in the Matrix Laboratory (MATLAB) Version R2023b software to identify the optimal moisture removal kinetic model. Key quality parameters such as color, Vitamin C content, and microbial load were measured using a colorimeter, redox titration, and serial dilution techniques, respectively. Results indicated that the developed model effectively predicts system behavior (R² > 0.94). The HSBD significantly outperformed traditional drying systems, achieving a drying efficiency of 19.88%, compared to 19% in previous studies. The Midilli Model was identified as the best fit for drying mango slices, with R² = 0.998, RMSE = 0.0185 and SSE = 0.0021. Dried slices exhibited superior color retention, with Lightness (L*) increasing from 66.37 in fresh mangoes to 72.22 in dried slices, and Vitamin C content rising to 294.46 mg/100 g, compared to 126.07 mg/100 g in fresh samples. Additionally, yeast count in HSBD mango was significantly lower (3.95 × 10² CFU/g) than in fresh samples (2.63 × 10³ CFU/g). These findings demonstrate that the HSBD effectively addresses the limitations of conventional dryers by ensuring uniform drying, reducing energy loss, and enhancing product quality. This scalable, sustainable solution is particularly beneficial for smallholder farmers in areas with high solar irradiance and abundant biomass.