Hydrogen Reduction of Bauxite Residue-Calcite Pellets and Mechanisms of Iron Particles Growth: A Comparative Study at Lab and Pilot Scales

Abstract Bauxite residue, an alkaline by-product, poses significant challenges for the alumina industry due to the complexity of its utilization. This study investigates the hydrogen reduction of self-hardened bauxite residue-calcite pellets examined in lab-scale stationary bed reactor and in a pilot-scale rotary reactor. The reduced pellets from two pilot campaigns were compared to two lab-reduced pellets with similar reduction conditions, focusing on iron particle growth and the presence of leachable adjacent mayenite (Ca 12 Al 14 O 33 ) phases. Notably, the mayenite (Ca 12 Al 14 O 33 ) supergroup of minerals relative peak intensity of XRD patterns was higher in the pilot-reduced pellets than in the lab-reduced ones. Various analytical techniques were employed, including X-ray fluorescence (XRF) for compositional analysis and LECO combustion analysis for oxygen quantification. Additionally, X-ray diffraction (XRD) was used for phase identification, and electron probe microanalysis (EPMA) with wavelength-dispersive spectroscopy (WDS) was used for microstructural and elemental characterization. The results revealed that average metallic iron particle size in the pilot-reduced pellets was approximately 15–20 µm, significantly larger than the 1–2 µm particles observed in the lab-reduced pellets. Based on the experimental observations and thermodynamic calculations, two mechanisms for iron particles growth were proposed: particles accretion in a deformable oxide matrix, and a nucleation–ionic diffusion growth mechanism. In the particle’s accretion in a deformable oxide matrix mechanism, the rotation of furnace and movement of pellets provides vibrational, rotational and impact energy to deform and move the phases in the pellets. In the nucleation–ionic diffusion mechanism, diffusion of Fe 2+ ions within the oxide matrix leads to particle growth, unlike the short diffusion lengths to many nucleated iron particles in lab-scale trials, longer diffusion lengths to fewer nucleated iron particles in pilot trials lead to larger particles.