The presented multimodal transformer networks quantitatively reproduce experimental proton conductivity and the underlying conduction mechanism and provide predictive uncertainty estimates.

Powders with phase composition including quasi-amorphous phases and calcium carbonate CaCO₃ in the form of calcite or aragonite and sodium halite NaCl as a reaction by-product were synthesized from 0.5M aqua solutions of sodium silicate and 0.5M aqua solutions of calcium and/or magnesium chlorides. Starting solutions were taken in quantities which could provide precipitation of hydrated calcium and/or magnesium silicates with molar ratios Ca/Si = 1 (CaSi), Mg/Si = 1 (MgSi) or (Ca+Mg)/Si = 1 (CaMgSi). Hydrated calcium and/or magnesium silicates, hydrated silica, magnesium carbonate, hydrated magnesium carbonate or hydrated magnesium silicate containing carbonate ions are suspected as components of quasi-amorphous phases presented in synthesized powders. Heat treatment of synthesized powders at 400, 600, 800 °C and pressed preceramic samples at 900, 1000, 1100 and 1200 °C were used for investigation of thermal evolution of the phase composition and microstructure of powders and ceramic samples. Mass loss of powder samples under investigation during heat treatment was provided due to evacuation of H₂O (m/z = 18), CO₂ (m/z = 44) and NaCl at temperatures above its melting point. After sintering at 1100 °C, the phase composition of ceramic samples included wollastonite CaSiO₃ (CaSi_1100); enstatite MgSiO₃, clinoenstatite MgSiO₃ and forsterite Mg₂SiO₄ (MgSi_1100); and diopside CaMgSi₂O₆ (CaMgSi_1100). After sintering at 1200 °C, the phase composition of ceramics CaSi_1200 included pseudo-wollastonite CaSiO₃. After heat treatment at 1300 °C, the phase composition of MgSi_1300 powder included preferably protoenstatite MgSiO₃. The phase composition of all samples after heat treatment belongs to the oxide system CaO–MgO–SiO₂. Ceramic materials in this system are of interest for use in different areas, including refractories, construction materials and biomaterials. Powders prepared in the present investigation, both via precipitation and via heat treatment, can be used for the creation of materials with specific properties and in model experiments as lunar regolith simulants.

The tunable structure of metal–organic frameworks (MOFs) is an ideal platform to meet contradictory requirements for proton exchange membranes: a key component of fuel cells. Nonetheless, rational design of proton-conducting MOFs remains a challenge owing to the intricate structure–property relationships that govern the target performance. In the present study, the modeling of quantities available for hundreds of MOFs was scaled up to many thousands of entities using supervised machine learning. The experimental dataset was curated to train multimodal transformer-based networks, which integrated crystal-graph, energy grid, and global-state embeddings. Uncertainty-aware models revealed superprotonic conductors among synthesized MOFs that have not been previously investigated for the application in question, thus highlighting magnesium-containing frameworks with aliphatic linkers as high-confidence candidates for experimental validation. Furthermore, classifiers trained on the activation energy threshold effectively discriminated between well-known proton conduction mechanisms, thereby providing physical insights beyond the black-box routine. Thus, our findings prove high potential of data-driven materials design, which is becoming a valuable addition to experimental studies on proton-conducting MOFs.