Sources of Potroom Dust Emissions From Aluminium Smelters

Abstract

'Potroom dust', the air-suspended particulates in the potrooms of aluminium smelters, is a major source of particulate emissions from a smelter to the environment. With regulatory limits for total particulates continually tightening, there is an increasing need for smelters to understand the sources and pathways by which potroom dust is generated. Only armed with this understanding can smelters develop targeted strategies to reduce these emissions. Research has been undertaken to determine: firstly, the major sources of dust, by examining the compositional, morphological and particle size nature of dust; secondly, the mechanisms and operations by which this dust is generated; and thirdly, strategies that smelters can employ to reduce potroom dust. The study examined dust generation across four modern pre-bake, point-fed smelters, with varying raw materials properties and transport systems, operating practices and cell technologies. A robust sampling strategy was developed and adapted for each smelter, targeting airborne and settled dust across multiple potroom locations and elevations, particularly the operating floor and roof levels. Suspected sources of dust (pot fume, cover material, alumina) were also collected, 'fingerprinted' using XRD, SEM-EDS and particle sizing techniques, and compared with collected dust samples. Real-time studies also characterised dust generation in response to operations in the potroom, over entire cycles of operation. The study found a remarkable level of consistency between the four smelters, in the composition and key operational contributors to dust. Airborne dust differed significantly in composition from settled dust. Airborne dust from all smelters were high in bath or electrolyte content, being dominated by ultrafine particles of pot fume and condensed bath vapour; these originate from open cells, pot fume leakages from closed cells and from emission sources external to the pot. In contrast, settled dust varied from being predominantly cover material to 50/50 mixtures of cover and alumina; these originate from the loading of raw materials to cells and were dependent on the granulometry of cover (more cover-based dust where cover material was 'fine' as opposed to 'coarse') and the delivery method for feed alumina (more alumina-based dust where alumina was loaded to cells by crane as opposed to being conveyed). Despite an initial industry hypothesis, alumina fines were not found to be a major factor to dusting; however their role in potroom dust could not be discounted due to the presence of gibbsite (uncalcined alumina) in both settled and airborne dust. The study confirmed that potroom dust is largely generated by process-related sources, rather than those external to the potroom. Differences in airborne and settled dust point to differences in the particle size and hence mobility (aeration vs. settling behaviour) of their contributing sources. The major contributors to dust at a general smelter, as found by real-time studies, differed at the potroom roof and operating floor elevations. At the roof level, anode change was by far the most significant contributor to total dust (up to 25-60% at different smelters), followed by metal tapping, cover application, bath transfers and background emissions (not attributed to any identifiable activity, from pot fume leakage and recirculating dust). In contrast, at the operating floor level, background emissions represented the greatest contributor (up to 50-80% at different smelters) to total dust, followed by anode change, cover application, bath transfers, metal tapping, pot tending and heavy vehicle movements. Contributors were classified under two pathways for dust generation: newly generated dust (primary generation) from open-cells, pot external sources and pot fume leakage, and old recirculating dust (secondary generation) from re-aeration of settled dust by vehicles, wind/air movement and housekeeping activities. Pot fume was a major source of newly generated dust, being generated from operations with open-cells and hot, fuming materials (cooling anodes, bath cleanings), and was well associated with HF. Being very fine in particle size, pot fume is highly mobile and tends to escape through the roof. While dusting from cover and alumina also formed newly generated dust, they are coarser in size and less mobile than pot fume; hence these tend to form settled dust, which ultimately becomes a source of old recirculating dust. The significant role of anode change, cover application and metal tapping led to a test of high pot draft vs. normal draft at two smelters during these open-cell operations. In general, emission reductions during the three operations can be achieved using high draft; however the reductions achieved are impacted by the actual rates of draft increase (from normal to high draft) and can be masked by other significant pot-external emission sources, particularly cooling spent anodes/bath cleanings and recirculating dust from external winds. Minimising the number and duration of hoods removed for open-cell operations was also confirmed as an important strategy for reducing dust. Progressive removal of hoods on two test cells resulted in a steady increase in both dust and HF. The use of high draft significantly reduced open-cell emissions (50-80% reductions in dust for 1-3 hoods, compared to normal draft), however its effectiveness rapidly deteriorates with more hoods removed. Emissions measured with even 1 hood removed demonstrate that no 'safe level' of removed hoods can be assumed (even with high draft), and highlights the importance of hooding quality and pot sealing to minimise background dust from leakages of pot fume. While not tested, reducing the role of fines in cover, e.g. by using coarser crushed bath, was also recommended as an important strategy to reduce the settled dust in potrooms. Other suggested dust reduction strategies included: (i) redesigning cover and alumina loading systems to reduce dusting; (iii) reducing emissions of bath fume from spent anodes, cavity cleanings and liquid bath moulds using passive lids/covers; (iv) diversion of crucible exhausts during tapping back to the cell; and (v) increased focus on housekeeping to reduce settled dust inventories in the potroom and hence reduce recirculating dust.