**4. Discussion**

The slag structure is an important factor for the potential recycling of various valuable metal contents. By adjustment of the process parameters, the structure can be tailor-made for optimizing the recovery of di fferent valuable materials [29].

Firstly, it has to be stated that the results from XCT and MLA are in good agreement. In both cases, clear di fferences between sample H on the one hand and sample M and S on the other hand were visible. There were two main occurrences of Cu-metal within the slag structure. The first represents the presence of single Cu droplets within the amorphous matrix in the outer region of the TBRC. These droplets appeared detached from the surrounding amorphous slag structure (see Figure 13). In this region, the cooling rate was rather fast, whereby the droplets could not sink back into the metal phase and were entrapped within the glass-like slag structure. This area is of less importance, as this zone 1 (see Figure 4) was determined to sum up to <1 wt. % of the total mass in the process. Nevertheless, if possible, a further reduction of this zone could be beneficial.

The main slag material showed a structure as identified in samples M and S. Although these two samples originated from di fferent zones of the furnace (zones 2 and 3, see Figure 4), no significant di fferences between their phase structures were identified. There were two main patterns present in these samples: an Al-rich slag phase and a Cu-rich phase. These main areas showed significant di fferences regarding their structure and elemental contents. With respect to the solidification order of these main phases, the following assumptions can be made: the Al-rich phase (a) represents the material that solidifies first, followed by the Cu-metal phase. The metal droplets then accumulate as loos agglomerates within the Cu-rich matrix (b and c). Meanwhile the Cu-rich dendritic structures are formed, and phase b and c finally solidify.

The distribution of various valuable metals (e.g., Au, Ag, Ge, In) between the di fferent phases of the slag, their deportment, and in which phases they accumulate have to be investigated further [30–32]. Special focus should also be on the e ffect of the applied slag system [33] and the adjustment of the components to improve the breakage properties with respect to a downstream comminution process [18,34].

By adjusting the cooling rate, the occurrence of the described Cu droplets has to be minimized. The slag should be kept in a certain temperature window to allow these droplets to settle back onto the metal face at the bottom of the furnace. As these isolated droplets predominantly appear in the outer regions of the furnace and the Cu content is already lower compared to other regions, the handling and adjustment of this structure is of less importance. By lowering the cooling rate at the same time, the Cu structure in the inner regions of the slag can also be altered. The target is a better recyclability of the Cu contents of these materials. A detailed understanding of the occurrence of the main phases and their formation process is necessary for a better understanding of potential parameters to optimize the processability. The structures have to be adjusted to the requirements of the following processes. A potential downstream grinding and separation cascade possibly implies increasing the size of the Cu-metal agglomerates.

In the end, there are two main aspects to improve the recyclability of potential valuable metals within the slag structure. The first point, which is only indirectly linked with the improvement of recyclability, is to reduce these metal contents in the slag by allowing the metal droplets to sink back into the metal phase on the bottom of the furnace. The second point is the adjustment of the slag structure to improve the performance of potential downstream processes [35]. Both require a detailed understanding of the process–parameter–slag–structure relationship and the proposed investigation procedure is able to reveal and quantify the critical e ffects. Additionally, other observation techniques should complement the presented approach to ge<sup>t</sup> the full picture [36]. Finally, it has to be stated

that the found structures within the slag analyzed via XCT and MLA strongly depend on the applied reactor technology and the resulting flow patterns. These dependencies have to be investigated further.
