This study explores the impact of pore volume distribution on the structural, thermal, and mechanical properties of spinodal phase-separated silica gels synthesized with poly(ethylene oxide) as a phase-separating agent. By systematically varying gelation temperatures between 20 and 60 °C, we investigate how reaction kinetics influence the resulting pore architecture, thermal conductivity, and elasticity. Nitrogen sorption, mercury intrusion porosimetry, and SEM analysis reveal a transformation from a bimodal pore structure at low temperatures, featuring interconnected macropores, to a predominantly mesoporous network with loss of bimodality. This shift in the diameter of the macropores significantly impacts the thermal insulation properties of the gels as thermal conductivity decreases from 68 to 27 mW (m·K)⁻¹ due to reduced macroporosity, enhanced mesoporosity, and the Knudsen effect. Mechanical testing revealed a substantial decline in Young’s modulus with increasing gelation temperature. These changes are attributed to the interplay of mesoscale structural differences and density variations, driven by increasing gelation temperatures. While higher temperatures lead to reduced strut thickness and the loss of interconnected macropores, the substantial decline in Young’s modulus highlights the critical role of mesoscale structural integrity in maintaining mechanical stability. The findings underscore the importance of an optimized pore volume distribution in tailoring pore structure and performance characteristics, providing a pathway for optimizing silica gels for applications in thermal insulation, filtration, and catalysis.

In the context of environmental protection, the construction industry plays a key role with significant CO₂ emissions from mineral-based construction materials. Recycling these materials is crucial, but the presence of hazardous substances, i.e., in older building materials, complicates this effort. To be able to legally introduce substances into a circular economy, reliable predictions within minimal possible time are necessary. This work introduces a machine learning approach for detecting trace quantities (≥0.06 wt%) of minerals, exemplified by siderite in calcium carbonate mixtures. The model, trained on 1680 X-ray powder diffraction datasets, provides dependable and fast predictions, eliminating the need for specialized expertise. While limitations exist in transferability to other mineral traces, the approach offers automation without expertise and a potential for real-world applications with minimal prediction time.