Sebelas Maret University

This study investigates the effect of Fe metal loading and support type on the characteristics and catalytic behavior of Fe-based catalysts to produce gasoline-like hydrocarbons (GHCs) via the hydrodeoxygenation (HDO) of oleic acid. The catalysts were prepared by the wet impregnation method using HZSM-5, mordenite (MOR), and amorphous silica-alumina (SiAl) supports, with Fe loadings of 3, 6, and 9 wt%. The HDO reaction was conducted under mild operating conditions at atmospheric pressure. Characterization results indicate that Fe impregnation does not significantly alter the crystalline structure of the supports, as confirmed by XRD analysis. However, SEM observations reveal an increase in surface roughness after metal loading. An increase in Fe content leads to a decrease in surface area and total acidity, likely due to partial pore blockage and reduced accessibility of acid sites. Catalytic performance evaluation shows that both catalytic and non-catalytic experiments achieve high oleic acid conversion (>95%), indicating a significant contribution from thermal decomposition under the applied reaction conditions. Among all catalysts, Fe6/HZSM-5 exhibits the highest GHCs selectivity of 73%, outperforming Fe6/MOR (46.60%) and Fe9/SiAl (56.20%). The superior performance of Fe6/HZSM-5 is attributed not solely to acidity, but rather to the combined effects of pore structure, metal-support interaction, and the presence of Fe species that facilitate deoxygenation pathways and subsequent hydrocarbon formation. Overall, this study provides valuable insights into optimizing support selection and Fe metal loading for designing efficient catalysts for fatty acid-based gasoline-like hydrocarbons production under mild operating conditions.

Science process skills (SPS) are essential for elementary students because they enable active engagement in scientific practices such as observing, questioning, predicting, experimenting, collecting data, interpreting data, and concluding. However, science instruction still offers limited opportunities for students to engage directly in these scientific activities, tending to emphasize conceptual understanding over process. This study analyzed the effect of board game-assisted Game-Based Learning on elementary students' SPS in energy-transformation content and identified how these skills developed following the intervention. A quantitative, quasi-experimental pretest-posttest control group design was used with 52 fourth-grade students: 24 in the experimental group and 28 in the control group. SPS were measured using a performance assessment covering seven indicators. Data were analyzed using normality and homogeneity tests followed by an Independent Sample t-Test. The experimental group's mean score increased from 66.88 to 84.17, while the control group increased from 65.18 to 77.68; the Independent Sample t-Test indicated a significant difference between groups after the intervention (t(50) = -3.593, p < .001), with a large effect size (Cohen's d ≈ 1.00). These findings suggest that board game-assisted Game-Based Learning (JoyR1D3) is a promising interactive alternative for developing elementary students' SPS.

Nano-enhanced bio-based phase change materials have emerged as promising thermal management materials for solar cooling applications because of their ability to improve heat transfer and thermal energy storage performance. This study presents a systematic review and quantitative synthesis of 50 peer-reviewed studies published between 2019 and 2025, focusing on thermodynamic mechanisms, thermophysical characterization, and cooling applications in photovoltaic and photovoltaic/thermal systems. The results demonstrate that nanoparticle incorporation significantly enhances thermal conductivity by approximately 15–400%, particularly with carbon-based nanofillers such as graphene and multi-walled carbon nanotubes, whereas latent heat retention generally decreases by 5–15% due to nanoparticle volume displacement and molecular interactions. The synthesized datasets indicate that an optimal nanoparticle loading of 2–4 weight percent provides a balanced enhancement of thermal conductivity and latent heat storage capacity. The correlation analysis reveals a minimal linear relationship between dispersion quality and latent heat retention, with a coefficient of r = 0.0935. This indicates that dispersion quality alone does not adequately account for the variations in latent heat retention observed across the reviewed studies. Instead, it suggests that factors such as nanoparticle morphology and interfacial compatibility are likely to have a more significant impact on thermodynamic behavior. The synthesis further shows that photovoltaic operating temperature can be reduced by up to 11°C under peak solar irradiance, resulting in an approximately 7.2% improvement in electrical efficiency. Exergy evaluation indicates that the latent phase contributes the highest useful exergy, approximately 42%, while demonstrating lower thermal irreversibility. Despite substantial material-level improvements, the aggregated system-level performance remains constrained by thermal resistance, methodological inconsistencies, and integration challenges in practical PV and PV/T configurations.

This study explores immersive virtual reality media to support environmental literacy and analytical skills among elementary school students within the context of zero-emission education. Employing an exploratory design integrating qualitative analysis, descriptive statistics, and the Analytic Hierarchy Process (AHP), data were collected through interviews, observations, expert evaluations, written feedback, and questionnaires involving three experts and 100 fifth-grade students across eight elementary schools. The results indicate a media feasibility score of 90.34% and a student response score of 81.99%. Specific ratings were 94.05% for content, 91.07% for pedagogy, and 85.71% for media quality. Performance metrics included 85.38% for analytical skills, 83.11% for environmental literacy, 86.89% for media presentation, and 70.81% for practical application. Student experiences were shaped through contextual exploration, environmental change simulations, decision-making, and visual feedback. The media assisted students in identifying problems, tracing cause-and-effect relationships, determining solutions, and drawing conclusions. The assessment prioritized responsible behavior, causal analysis, and solution determination. Successful implementation requires device orientation, simple navigation, clear instructions, and teacher guidance. The study establishes a concept for decision-based immersive environmental learning that emphasizes student agency, environmental consequence evaluation, and the integration of environmental literacy with analytical skills, providing both a theoretical foundation and practical implementation guidelines for schools.

Fused deposition modeling (FDM) is governed by process–structure–property (P–S–P) relationships, in which nominal process settings shape the realized as-built structure, and that structure, in turn, governs the final properties and manufacturing quality. This systematic literature review examines how artificial intelligence (AI) and machine learning (ML) have been used to model, optimize, and operationalize these relationships in FDM. Using Scopus and Web of Science with a PRISMA-guided screening procedure, we reviewed English-language journal articles published between 2021 and 2025 and identified 65 studies for final synthesis. The evidence converges around four linked themes: direct process–property prediction, structure-aware data and feature representation, surrogate-assisted optimization and inverse design, and robustness-oriented deployment for quality assurance and real-time monitoring. The most important advances arise when nominal process variables are interpreted as latent drivers of the realized infill, porosity, bonding quality, anisotropy, and geometric deviation, rather than as flat predictors of performance. Tree-based ensembles, boosting methods, NNs, Gaussian process-based models, and hybrid deep learning architectures dominate predictive benchmarks, whereas explainable learning, multimodal sensing, and measured or inferred structural descriptors improve both interpretability and process insight. A recurring methodological pattern is the sequence design of experiments, ML surrogate construction, and downstream optimization, compensation, or decision support. Despite this progress, the literature remains constrained by mismatches between nominal and realized structures, limited cross-printer transferability, underdeveloped treatment of uncertainty, small and heterogeneous datasets, and inconsistent external validation. Overall, the review shows that AI contributes most in FDM when it makes the structure-forming role of process variables explicit and turns learned P–S–P mappings into more reliable manufacturing decisions.