Martinson Addo Nartey

Interface passivation is one of the few design levers that directly affects non-radiative recombination in n-i-p perovskite solar cells. Yet, buried-interface treatment, surface treatment, absorber thickness, and trap populations do not act independently. One-factor-at-a-time simulation explores this space slowly and unreliably. SCAPS-1D was therefore coupled with a machine-learning surrogate for an ITO/SnO2/BCH/PVSK/PEAI/Spiro/MoO3/Ag architecture, treating BCH as a model descriptor for buried-interface passivation and phenethylammonium iodide (PEAI) as a surface-passivation modifier. Cleaning left 11,142 valid simulations described by 21 leakage-safe numeric features. Five regressors were benchmarked on a fixed 80:20 held-out split, with Bayesian tuning confined to the training partition. CatBoost was the strongest power-conversion-efficiency (PCE) model (R2 = 0.9981; root-mean-square error 0.105 percentage points; mean absolute error 0.073 percentage points), while a shuffled-target control collapsed to R2 = −0.0064. Rank correlation, binned response curves and SHAP attribution independently identified the same governing variables: absorber defect density, absorber thickness and the PVSK/PEAI and PEAI/Spiro interface-defect descriptors. An ensemble then ranked 100,000 virtual devices by a score penalized by inter-model disagreement, and the 200 best were returned to the simulator. Of 198 candidates that produced complete outputs, 170 exceeded the previous simulated maximum of 22.53%, and the leading design reached 23.32% with VOC = 1.253 V, JSC = 26.41 mA cm−2 and FF = 70.45%. The novel contribution is a closed-loop, leakage-audited search that concentrates expensive SCAPS evaluations inside a dual-interface, passivation-dominated region while leaving the physical simulator as the sole arbiter of performance.

Operational stability remains a central barrier to perovskite-photovoltaic deployment, yet literature-derived lifetime data are heterogeneous and frequently right-censored because many tests end before a device reaches T80. We developed a reproducible censoring-aware machine-learning workflow that combines publication-grouped validation, separate transport-layer and absorber-adjacent-interface representations, an illuminated maximum-power-point-tracking (MPPT) primary cohort, and applicability-domain-gated candidate screening. A frozen extract of 51,539 records from the NOMAD Perovskite Solar Cell Database yielded 381 devices from 203 publications, with 134 observed T80 events and 247 right-censored observations. Kaplan–Meier, ridge-penalized Cox, and random survival forest models were evaluated with publications kept intact across five-fold cross-validation. Mean publication-grouped concordance was 0.601–0.642, whereas refitting under a random device-level split inflated concordance by 0.02–0.06 because 64.9% of validation devices shared a publication with training. Cox coefficients and forest permutation importance independently emphasized encapsulation, hole-transport-layer family, and absorber-adjacent interfaces. Screening literature-supported absorber–stack recombinations identified 44 previously unreported combinations, represented by 20 model-distinguishable signatures; all lay within the encoded applicability domain, and materials review retained 11 for experimental triage. These results show that censoring and source-group dependence materially affect apparent machine-learning performance in perovskite stability modelling. The workflow supports evidence-constrained data-driven prioritization, not causal optimization or guaranteed lifetime prediction.

Green synthesis of biomass precursor into GO and rGO was achieved using hydrothermal carbonization. The synthesized material exhibited an enhanced specific capacitance of 552.48 F g ⁻¹ at 1 A g ⁻¹ with high redox activity.

This study explores the use of synthesized calcium silicate hydrate (CSH), produced with the sol‐gel method, to enhance the performance of calcined clay pozzolan cement, a lower‐carbon alternative to traditional cement. The synthesized CSH was characterized by x‐ray diffraction (XRD) and Fourier‐transform infrared (FTIR) spectroscopy. The CSH was used as a minor admixture between 1 and 3 wt.% of calcined clay Portland cement. In this study, 30% of the Portland cement was substituted with calcined clay. The test methods that were investigated included compressive strength, absorption, and porosity. Environmental impact assessment was performed on the optimum mix. The results of the study indicated that the CSH was found to improve compressive strength, especially at 3% dosage, which showed the highest strength at 28 days. This optimum mix achieved a 12.6% reduction in CO ₂ emissions compared with reference samples. Although increased CSH reduced flowability, it significantly improved durability by lowering water absorption by up to 34% at 3 days and enhancing pore structure, as evidenced by reduced sorptivity. Overall, the synthesized CSH enhanced early strength development and durability, making calcined clay cement a more viable and sustainable option.

With rising global energy demands and increasing environmental concerns, sustainable energy technologies are crucial for ensuring a cleaner future. Water splitting has emerged as a key method for producing clean hydrogen fuel. However, developing efficient bifunctional electrocatalysts for this process remains a significant challenge. Traditional noble metals like platinum (Pt) and ruthenium (Ru) offer high performance but are costly and scarce, highlighting the need for cost-effective alternatives. MoS₂-based electrocatalysts have attracted attention due to their promising performance and cost-effectiveness compared to noble metals. Despite their potential, challenges such as optimizing efficiency, reducing overpotential, and enhancing stability must be addressed. This review explores the MoS₂ as an alternative to noble metals, focusing on its performance metrics through the lens of materials engineering. A detailed analysis of the structural and electronic characteristics of MoS2 is provided, examining how these factors influence its bifunctional electrocatalytic performance. Additionally, the review delves into the reaction mechanisms, critically evaluating key performance parameters such as overpotential, Tafel slope, exchange current density, stability, and electrochemically active surface area. Furthermore, the impact of synthesis and processing methods on MoS₂’s performance is discussed, including phase and defect engineering, morphology optimization, and heterostructure construction. By integrating current research and offering insights into the structure-property-performance relationships, this review aims to guide future developments in sustainable energy technologies.

Biomass and biowaste resources can be used to create self‐doped carbon with a distinctive microstructure. Using an economical and environmentally friendly method to create heteroatom‐doped carbon electrode materials with excellent electrochemical performance has attracted much attention in the energy storage industry. A novel facile two‐step, low‐cost, and eco‐friendly synthesis method for Colocasia esculenta peels has been developed to manufacture activated carbon (CEPAC) and used as an electrode material for supercapacitor application. The CEPAC 1:1 displayed a high specific surface area of 910 m²/g with oxygen‐heteroatom polar sites in the carbon network. A specific capacitance of 525.3 F/g was recorded in the three‐electrode system using a 3 M KOH solution. The assembled symmetric cell delivered an impressive specific capacitance of 98.7 F/g at 1 A/g while maintaining 98.4% of the initially recorded capacitance after 10 000 charge–discharge cycles. These results present a promising low‐cost and simple processing route for synthesizing electrode materials with superior surface properties for high‐performance supercapacitors.

Carbon black, a nano-porous material usually derived from the pyrolysis of waste tyres possesses varied particle sizes and morphology making it a viable material for several engineering applications. However, the high tendency for CB to agglomerate remains a challenge. To address this, bio-templating has been employed to produce a nanostructured porous carbon electrode material for supercapacitor applications using diatomite as a template. Diatomite-synthesized activated carbon (DSAC) was fabricated through a three-step process involving acid treatment of diatomite, thermal activation of carbon black, and bio-template synthesis. The resulting material was thoroughly characterized using XRD, Raman spectroscopy, BET analysis, and SEM imaging. Its electrochemical properties were assessed through cyclic voltammetry, galvanostatic charge–discharge, and electrochemical impedance spectroscopy. The DSAC material exhibited a high specific surface area of 266.867 m² g⁻¹, pore volume of 0.6606 cm³ g⁻¹, and mean pore radius of 1.8943 nm. The electrochemical evaluation revealed that DSAC demonstrates excellent electrochemical performance, achieving a high specific capacitance of 630.18 F g⁻¹ and retaining 94.29% capacitance after 5000 cycles at 1 A g⁻¹. The DSAC electrode is eco-friendly and a promising candidate for supercapacitor applications.