Purpose – The purpose of the study was to investigate the thermo-reactive behaviour of the MnO–Al system and to develop a scalable reaction vessel capable of sustaining the thermal conditionsrequired for complete aluminothermic reduction of manganese oxide and effective metal–slag sep-aration, while enabling circular utilisation of the resulting slag.

Methodology – This quasi-experimental study combined three complementary approaches. Three laboratory-scale reaction trials were conducted in different vessel configurations with a fixed re-actant charge, and thermophysical properties of the vessel material and slag were characterised by DSC and LFA. A finite-element model in COMSOL Multiphysics, coupling Heat Transfer in Solids and Fluids under the Boussinesq approximation, was validated against the experimental temperature–time profiles and applied to a scaled-up vessel size. An analytical circularity model linked vessel geometry to a slag-derived material balance.

Findings – The thermal response of the MnO–Al reaction depended strongly on vessel configuration rather than on reactant chemistry alone. Peak inner temperatures ranged from 1849◦C in the thin-walled quartz vessel to above 2315 ◦C in the custom alumina vessel, with the time above the 1500 ◦C separation threshold ranging from 22 to 38 s. The scale-up simulation predicted a separation window of approximately 23 min above 1500 ◦C at the 100 kg scale, and the circularity model showed that a self-sustaining slag-to-vessel loop is governed primarily by vessel geometry and fractional vessel loss per run.

Implications – The study extends the engineering understanding of metallothermic reactor design by integrating thermo-reactive, geometric, and material-circularity perspectives, and shows how surface-to-volume ratio governs the trade-off between peak temperature and separation duration in aluminothermic systems. The findings provide structured guidance for the design of reaction vessels in aluminothermic manganese production, offering a foundation that can be used to balance thermal stability, operational safety and slag reuse when scaling such processes toward low-carbon manganese production.

Limitations – Only one trial was performed per vessel configuration, so statistical uncertainty could not be quantified. Thermocouple range and instrumentation issues limited validation of the FEM-predicted gradients and DSC/LFA data covered only restricted temperature ranges, high- temperature properties were taken from Thermo-Calc. The FEM model captures the cooling phase with a prescribed initial condition and the circularity model relied on an assumed fractional loss parameter. Pilot- and production-scale validation remain pending.