Abstract:
In aluminium reduction cell, anode cover is added onto the anodes and the bulk bath to maintain the cell heat balance, as well as to reduce the anode air burn and fluoride loss. Alumina and crushed bath are the most common raw materials for the anode cover. Investigation and understanding of the thermochemical behaviour and stability of the crust is important for maintaining the cell heat balance. In a reduction cell, the anode cover at room temperature is heated by high heat flux from the hot bulk bath. Liquid bath appears in the anode cover due to the penetration of liquid bath from the bulk bath and melting of the crushed bath in the anode cover. The liquid bath penetrating upwards is cooled down by the cold upper part of the anode cover. As a result, the solid phases of cryolite and alumina crystallize out of the liquid bath, which reduces the cryolite ratio (CR) of the remaining liquid bath. The liquid bath and the precipitated crystals contribute to the formation of sintered crust at the bottom part of the anode cover. However, too much liquid phase in the crust can lead to the melting and weakening of the sintered crust. Based on the heat and mass transfer theory and phase equilibrium diagrams, a theoretical thermochemical model was developed to quantitatively simulate the thermochemical evolution process during sintered crust formation. An effective specific heat capacity was deduced to couple the temperature and chemical composition variables together. The thermochemical model was solved by a designed computational code applying finite element method (FEM). The simulation results were compared with experimental data. Seven crust pieces were taken from industrial reduction cells. According to their appearance, shape and crystal texture, three types of crystalline crust were distinguished from the sintered crust, and accounted for by crystallization from the bulk bath under different cooling rates. Crust samples were taken from different vertical positions in these crust pieces. XRD analysis, Rietveld refinements, and LECO oxygen analysis were carried out to analyze the chemical compositions of the crust samples. A non-commercial and calibrated DTA system was used to measure the liquidus and melting temperatures of the crust samples. The measured results show that cryolite content and CR are 60~70 wt% and 2.58~2.64 in the lower crust respectively, and decrease significantly to 20~30 wt% and 1.75~2.10 in the upper crust. The liquidus temperature of the crust sample is depressed by the decreasing of CR and by other fluoride additives, and is consistent with the measured chemical composition by XRD. The melting behaviour of the crust was studied by the shape analysis of the melting peak of the DTA heating curve. Crust samples with high cryolite content and CR have sharp peak shape indicating narrow melting temperature range. In chiolite enriched crust, high liquid bath content that incongruently melts from chiolite can melt solid cryolite over a wide temperature range with low energy intensity.