Modeling and machine learning elucidate morphology evolution and size control in spray pyrolysis

A comprehensive understanding of the mechanisms governing particle size and morphology in spray pyrolysis of basic magnesium carbonate microspheres is essential for advancing reactor design and process optimization. A multicomponent mathematical model was developed and experimentally validated to simulate the processes of droplet evaporation and pyrolysis. The effects of the initial droplet size, precursor concentration, and ambient temperature on the process were investigated. Morphology formation was investigated through the Péclet number, while a support vector machine model was employed to predict morphological outcomes in a three-dimensional parameter space, achieving an accuracy of 94%. The results indicated that evaporation time increased with droplet size but decreased at elevated temperatures. In contrast, the decomposition process was significantly faster, suggesting the dominance of solvent evaporation. Particle size exhibited a positive correlation with droplet diameter and a negative correlation with temperature, whereas the influence of precursor concentration was minimal. Hollow structures were formed at high Péclet numbers, which were induced by either large droplets or high temperatures. Critical transition boundaries were identified, corresponding to a diameter of 15.5 μm, a temperature of 405.3 K, and a mass fraction of 4.4 wt. %. This study integrates physics-based modeling with machine learning to elucidate the physicochemical processes governing single-droplet evaporation and pyrolysis. The developed approach provides novel insights at the droplet scale and serves as a predictive tool for spray pyrolysis reactor design and optimization. Moreover, the methodology can be applied to other material systems as well.