The Optimization of Salt Additions during Processing of Aluminium-Magnesium Alloys

Abstract

Sodium is well known to be a cause for edge cracking during rolling of aluminium-magnesium alloys (AA5xxx). Due to this, requirements for the sodium contents in Al-Mg alloys are often less than 1,0 ppm. The electrolytic produced aluminium have a sodium content in the range of 30 to 200 ppm, where the highest levels are present in the most modern pot lines.To reduce sodium and other impurities like oxides and carbides several techniques are available for casthouses today. Furnace fluxing and degassing with chlorine gas has been in use for a long time. Due to issues related to health, safety and the environmental, the industry is trying to reduce or replace its use of chlorine gas.Salt mixtures with magnesium chloride or aluminium fluoride have the possibility to remove alkaline and alkaline earth metals in liquid aluminium. Aluminium fluoride can, however, only be used when magnesium free aluminium are processed. Magnesium chloride can on the other hand be used when magnesium is present, and can thermodynamically reduce the sodium concentration to less than 1 ppm.In the present work the salt ability to remove sodium and calcium was tested both laboratory and full-scale (industrial scale). Laboratory scale batch experiments were performed with 2 kg?s of an Al-Mg alloy, and salt was added to the top of the melt. The use of salt doubled the rate constant for sodium removal from the melt compared too without salt. Calcium was not removed without the presence of salt. The lowest Na and Ca values achieved were 0,2 and 0,1 ppm respectively.In the full-scale test conducted at Alcoa Mosjøen, the effect of stirring time and salt amounts was studied. The removal of Na and Ca were shown to follow first order reaction kinetics. The removal constants were calculated and compared to salt amounts, temperature and magnesium concentration in the melt. Temperature and Mg concentration did not show to have an impact on the removal rates in the intervals tested. The rate constant for Na removal showed to be linear to the salt amounts in the region tested. The overall amounts of Ca in the melt was low, which resulted in the small amount removed gave large outcomes on the removal constants.By estimating the need for removal of Na, the amount of salt and the stirring time needed can be calculated. By increasing the stirring time the additions of salt needed is lowered, and salt charges will with time be less.