Recycling of aluminium dross under industrial conditions

What benefits does sampling aluminium dross provide?

Aluminium dross, often referred to as salt dross, oxide dross or simply "dross", is an unavoidable by-product of aluminium smelting and recycling processes. During the smelting process, liquid aluminium reacts with oxygen. This reaction leads to the formation of aluminium oxide (Al₂O₃), which settles on the surface as dross [1].

The composition of aluminium dross is heterogeneous and depends heavily on the type of smelting process, the raw materials used (scrap, primary metal or alloying elements) and the furnace technology. In addition to metallic aluminium and aluminium oxide, aluminium dross contains salts (NaCl, KCl) and heavy metal oxides [2]. Depending on the source, different distribution ranges are given, and the residual aluminium content in the dross varies between 15 % and 80 % [1, 2, 3, 4].

For aluminium dross recyclers, including AMAG casting GmbH, dross is not considered waste but rather a secondary raw material containing valuable resources. In rotary and tilting drum furnaces (Figure 1), dross can be reprocessed under salt and the metallic aluminium recovered. Therefore aluminum dross is heated to temperatures of approximately 650–800 °C so that the metallic aluminum melts, while the oxide components remain predominantly solid. The rotating motion and the addition of molten salt as a flux break up break up oxide crusts, improve wettability, and cause aluminum droplets to coalesce. The liquid aluminum then separates from the dross matrix by gravity and is skimmed off or tipped out, while an oxidic, salt-containing residue remains. The economic value of the dross is measured by the respective proportion of metallic aluminium and expressed as a percentage metal content. In order to determine the metal content value of individual dross deliveries, a representative sampling procedure is required. AMAG casting GmbH is intensively engaged in optimising these processes. On the one hand, fair competitive conditions for suppliers and recyclers are essential; on the other hand, the process control in reprocessing is also highly dependent on correct sampling values. In the case of negative metal content deviations, the recovery potential cannot be fully exploited, resulting in additional manufacturing costs for recyclers.

The recycling of aluminium dross is of significant energy-economic relevance and has considerable economic potential. The recovery of aluminium and aluminium oxide produces valuable secondary raw materials that contribute significantly to the competitiveness of the aluminium industry [5]. Professional and representative sampling is therefore a basic prerequisite for the proper and efficient use of aluminium dross as a secondary raw material with a recycling rate of almost 100 % based on the amount used. The melting process recovers the proportion of metallic aluminum. The remaining salt slag fraction undergoes separate processing, during which the salts it contains are recovered and can be fed back into the process cycle. The remaining oxide components can be reused as a secondary mineral raw material. The distribution is directly dependent on the aluminum yield of the dross used.

Challenges in the sampling process

Sampling aluminium dross is a challenge due to the highly heterogeneous nature of the material. Not only does the composition vary between different batches, but it also varies within individual samples. This is due to the way dross is formed during the smelting process, in which oxide phases, metal inclusions, salts and mineral components mix in different size distributions [6].

If the dross is mechanically treated during the process, for example by pressing or draining, this results in further inhomogeneity, as the metal-containing fraction in particular is largely recovered, while an increasingly oxide-rich fine fraction remains [7]. As a result, pressed or drained dross differs significantly from untreated material, making the entire sampling process considerably more difficult. Pressed aluminium dross is usually easily recognisable by a press stamp and drainage holes (Figure 2).

Another key factor to consider when sampling is the proportion of fine fractions. Fine-grained material has a large specific surface area and is therefore particularly susceptible to oxidation and moisture absorption. For this reason, fine fractions tend to have a significantly lower metal content than coarse fractions.

The combination of these factors means that representative sampling, and thus reliable determination of the metal content is no trivial task. Minimal discrepancies in particle distribution or residual metal content can significantly influence the results.

Fractionation for homogenisation

A fundamental step in ensuring representative sampling is to divide the heterogeneous delivery into fractions that are as homogeneous as possible. On an industrial scale, this is achieved at AMAG casting GmbH by physically separating the material into coarse and fine fractions. The separation is conducted using wheel loaders and excavator grabs. The fine fraction is defined as the portion of material that can still be "shovelled" manually by an employee using a standard 10-litre bucket (Figure 2).

The coarse fraction consists of the remaining residues, which are either pressed and drained, as shown in Figure 2, or simply skimmed off in tubs without any mechanical post-treatment.

After separating the dross delivery into coarse and fine fractions, both fractions are weighed using suitable weighing systems. This step is essential, as it is the only way to ensure reliable determination of the mass distribution. Due to the higher bulk density of the fine fraction, its proportion is often underestimated in practice, even though it has a significant influence on the overall metal content. In addition, there are considerable differences in the distribution of the fractions depending on the respective supplier and the point of disposal. The results from the sampling department at AMAG casting GmbH show that the mass fraction of the fine fraction can be between 10 % and 70 % of the total delivery.

Correct sampling

To determine the total metal content of a delivery, melting trials are carried out in a 100-kg rotary drum furnace. A representative sample of the coarse and fine fractions is taken in order to determine the corresponding aluminium metal content. Considering the mass distribution previously determined by separation and weighing, the weighted total metal content of the delivery can be calculated (Table 1). The evaluation of the sampling data over a longer period of time shows that the aluminium metal content of the coarse fraction can vary between 32 % and 89 %, depending on the supplier and material quality. Metal contents in the range of 18 % to 72 % were determined for the fine fraction.

The central and most demanding step in the entire sampling process is representative sampling, as it forms the basis for a valid determination of the aluminium metal content.

Sampling the fine fraction is less complicated, as the smaller particle sizes mean that the distribution is generally more homogeneous. Nevertheless, it is necessary to further homogenise the total fraction before sampling by mixing it mechanically. During the subsequent sampling, care should also be taken to ensure that the particle size distribution of the sample taken corresponds to that of the total fraction. There are methodological limitations as fine particles in particular tend to segregate, which can compromise the representativeness of the sample. Careful handling and standardised procedures are therefore essential.

Additional treatment steps are required to produce a representative sample of the coarse fraction. On the one hand, the material must be prepared in such a way that it can be charged into the sampling melting furnace, and on the, the coarse fraction is not sufficiently homogeneous due to its nature. This can be exacerbated by the fact that individual deliveries may contain different types and qualities of aluminium dross. This heterogeneity is particularly evident in the coarse fraction, as it can include pressed and drained as well as untreated dross lumps. The metal content of the individual lumps can also vary considerably. Independent melting tests carried out by AMAG casting GmbH have shown that the metal content fluctuations within the coarse fraction of a delivery can be as high as 20 %. For this purpose, two lumps from the same delivery were crushed using an excavator chisel and then melted down individually. The causes of these deviations lie primarily in the different pre-treatment of the lumps (pressed, drained or untreated), in varying oxide and salt contents, and in residual moisture or adhesions. These factors have a significant influence on both the metal content and the homogeneity of the fraction.

In addition, it was found that increasingly compressed lumps exhibit pronounced inhomogeneity. During the pressing process, the metallic aluminium is pushed outwards to be recovered, resulting in the formation of thin metal layers, particularly on the surfaces of the lumps. This metallic skin often gives the impression of a higher metal content, while the interior consists mainly of previously pressed material with a correspondingly low metal content (Figure 3).

For representative sampling, it is therefore essential to prepare the lumps by mechanical crushing. In this process, approximately 20 % of the material per type of lump is broken up, from which the sample for the coarse fraction is then created. This procedure ensures sufficiently meaningful sampling, as the partial breakdown covers both the outer and inner areas of the material. The 20 % proportion represents a practical compromise between statistical representativeness and operational efficiency. In this way, the significant heterogeneity of the coarse fraction can be represented while minimising potential sources of error in the sampling process. Figure 3 shows a lump before and after processing.

Results

The investigations show that aluminium dross is a highly heterogeneous material, with the metal contents of individual batches within a delivery varying considerably. Compressed and drained drosses are particularly problematic because their inhomogeneous structure gives the appearance of a high aluminium content on the outside due to a metallic layer, while the inside often contains compressed, low-metal material. Added to this is the uneven distribution of coarse and fine particles, which makes representative sampling even more difficult. Since many deliveries are also characterised by fluctuating quality, careful mechanical processing and the preparation of meaningful samples are essential to minimise sources of error and obtain reliable results.

In summary, only a clearly structured and reproducible sampling process can provide reliable information about the quality and usability of aluminium dross. Figure 4 illustrates the results presented and shows the complete sampling process in a structured form. This not only ensures methodological traceability and reproducibility but also creates a reliable basis for the practical evaluation and usability of aluminium dross.