*3.3. Sorting of Fragmented Material*

In the next step, the fragmented and screened materials were sorted according to optical criteria in order to obtain the exposed regenerates in the mono-fraction. Due to the significant color differences, the desired regenerates of the sample material RefMat-1 (Figure 7), the zirconia-alumina fused aggregate (ZAC) and the bauxite could be separated from the remaining material of the sample with high accuracy for all grain size fractions. The optical sorting was performed in several runs. In the first run, the yellow or flesh colored ZAC was sorted out with as little residual material as possible. Afterwards, the greyish brown to black colored bauxite was separated from the residual material. The remaining residual material consists mainly of coarse grains of light grey colored binder matrix and aggregates with residual material on its surface (ZAC and bauxite). The largest proportion of regenerates could be obtained from the fraction >3 mm. A total of approximately 5.3 kg (26 wt.%) of ZAC and approximately 2.7 kg (13 wt.%) of bauxite were exposed from 20 kg of the refractory material (RefMat-1). However, the results also showed that there was still a high proportion of residual material in the 2–3 mm fraction.

**Figure 7.** Sample RefMat-1 before and after optical sorting.

In case of the sample material RefMat-2, it was possible to separate the desired regenerate white corundum (glassy transparent grains) for each grain size fraction in just one pass (Figure 8). However, the output in all three fractions was very low, and the proportion of residual material, i.e., white colored binder matrix and not cleanly exposed white corundum, was relatively high. In all three fractions, only approximately 3.3 kg (13.8 wt.%) could be separated cleanly from the binder matrix, whereby approximately 23.6 kg of sample material (RefMat-2) was fragmented. The largest proportion of white corundum could be found in the 1–2 mm fraction.

**Figure 8.** Sample RefMat-2 before and after optical sorting.

From the RefMat-3 sample, two different regenerates (tabular alumina and white corundum) needed to be sorted out, as for the RefMat-1 sample. However, it was discovered that due to the optical properties, the glassy transparent white corundum could hardly be distinguished from the grayish colored binder grains by the laboratory sorting system and could therefore not be sorted out. The optical filters could not be adjusted accordingly on the laboratory equipment, so a small amount of white corundum from the 2–3 mm fraction was sorted out by hand. Nevertheless, it was possible to sort out the white colored tabular alumina from all fractions from the sample material RefMat-3. This regenerate could be easily distinguished from the remaining material by optical sorting and could be sorted out for all three fractions (Figure 9). In addition to the non-sortable white corundum, the remaining material consists mainly of binder matrix and not cleanly exposed tabular alumina. With this sample material, the yield of regenerates is not very high. From the total fragmented sample material (RefMat-3) of approximately 31.5 kg, only approximately 4.8 kg (15.3 wt.%) of regenerates distributed over three grain size fractions could be obtained.

**Figure 9.** Sample RefMat-3 before and after optical sorting.

It is obvious that the desired regenerates from the three different samples could be cleanly separated out of each sample material. However, it can also be seen that the yield of regenerates is not very high. Regenerates can still be found in the remaining materials that have been sorted out, but these have not been sufficiently liberated from the binder matrix by the fragmentation.

### *3.4. Characterization of the Fragmented Material*

On the basis of the microscopic examinations, the extent of binder matrix adhesions to the regenerates obtained was investigated. In the recovered material from the sample material RefMat-1, namely the sorted out ZAC (Figure 10a) and bauxite (Figure 10b), binder adhesion could be detected on rare occasions. Generally, the individual grains were very effectively separated from the binder matrix using electrodynamic fragmentation.

**Figure 10.** Micrographs of the fragmented sample RefMat-1, sorted out ZAC (**a**) and sorted out bauxite (**b**).

The sorted out white corundum regenerates, which were obtained from the sample material RefMat-2, also show hardly any binder residues on the grain surfaces (Figure 11a). Only sporadic white binder residues are still visible (Figure 11b). The degree of detachment for these regenerates is thus evaluated as very high in purely optical terms.

**Figure 11.** Micrographs of the fragmented sample RefMat-2. Sorted white corundum (**a**), single grain of white corundum (**b**).

The microscopic images of the fragmented and unsorted RefMat-3 sample (Figure 12a) clearly show why the white corundum could not be clearly detected by optical sorting. Due to the different blue coloration of the binder, the glassy transparent white corundum cannot be clearly distinguished optically. On the other hand, the exposed white colored tabular alumina can be recognized and can therefore be sorted out without difficulty by optical sorting. The sorted out tabular alumina (Figure 12b) shows only slight binder adhesion.

**Figure 12.** Micrographs of the fragmented sample RefMat-3. Unsorted material (**a**) and sorted out tabular alumina (**b**).

For a better evaluation of the chemical purity of the recovered regenerates, they were investigated in terms of their chemistry and mineralogical phase composition. Table 7 shows the results of the XRF measurement of the ZAC regenerates and bauxite regenerates from the sample material RefMat-1. For comparison, primary materials (ZAC and bauxite) were also analyzed and then cross-checked with the results of the recovered materials.


**Table 7.** Comparison of the chemical composition (XRF) of ZAC regenerates and bauxite regenerates with primary material (b.d.l.—below determination limit).

It becomes clear that the chemical composition of the regenerates does not differ greatly from the primary material. There are no significant differences between the regenerate and the primary bauxite, so the chemical purity of the obtained bauxite regenerate can be confirmed. The recovered ZAC has a slightly higher Al2O3 content and slightly lower ZrO2 and SiO2 contents compared to the primary material. As these are the three main elements in ZAC, and no unusually high contents of CaO, Na2O or Fe2O3 or other impurities were measured, the deviating element contents are due to the inhomogeneity of the material. The material can therefore be regarded as a regenerate with a high chemical purity.

Additionally, the mineralogical phase composition does not show high impurities (Table 8). Only the grossite (CaAl4O7) detected in both regenerates, which is a relic of the hydrated alumina cement, indicates that there is a small amount of binder residue.


SiC 2.1 1.4

**Table 8.** Mineralogical phase composition of the regenerates ZAC and bauxite without X-ray amorphous content.

The determined chemical composition of both regenerates, white corundum from sample RefMat-2 and tabular alumina from sample RefMat-3, show no impurities (Table 9). The chemical analysis of the tabular alumina obtained could be compared to an analysis of primary material. This confirms the very high chemical purity of the tabular alumina regenerates. The phase composition (Table 10) also shows no major impurities. As with the recovered material from the RefMat-1 sample, minimal contents of grossite and spinel are detectable, which indicate binder residues.


**Table 9.** Chemical composition of regenerates white corundum from sample RefMat-2 and tabular alumina from sample RefMat-3 compared to a primary sample of tabular alumina (AlfaTab 30) (b.d.l.—below determination limit).

> **Table 10.** Mineralogical phase composition of the regenerates white corundum (RefMat-2) and tabular alumina (RefMat-3).


In summary, based on these results, it can be seen that the regenerates obtained by fragmentation have a high chemical purity, which is a basic prerequisite for a high-quality recycling.

### *3.5. Reuse of the Regenerates*

Only after it was proved by analysis that the regenerates obtained possess a high chemical purity were the regenerates used in new refractory ceramics. The aim was to evaluate whether the recovered materials can substitute for the primary raw materials without disadvantages in terms of rheological, mechanical and refractory properties. A refractory concrete (mixture RC) and a refractory tamped concrete (mixture TC) were selected for the recycling tests, in which the primary raw materials were substituted by the regenerates obtained in this study. In the refractory concrete, the regenerates ZAC and bauxite were used, while in the refractory tamped concrete, the tabular alumina was used. In parallel, both mixtures were produced using primary raw materials, acting as reference samples for later performance comparisons. During the production of the different mixtures (RC and TC), it was found that the rheology, and thus the workability of the fresh mixtures, did not change and was absolutely comparable to the reference mixtures. The important factor here was to reach a comparable flow behavior and thus a comparable workability without changing the water demand. Regenerates of lower quality might require more water in order to achieve satisfactory workability. A higher water demand is known to have negative effects on the strength development of refractory concretes as well as tamped concretes. This is also known from concrete technology [15].

Several test specimen prisms were produced with the mixed concretes. All prisms were stripped after 24 h, stored for another 24 h at room temperature and then dried for another 24 h in a drying oven at 120 ◦C. It was found that the setting and drying behavior of the regenerate masses and reference masses is close to identical. After drying all prisms, some of the samples were burned at different temperatures in a high temperature furnace. This high-temperature treatment (sintering) can be used to check the refractoriness of the samples produced. The strength development by sintering at different temperatures is also an important characteristic value for refractory materials. The specimens of mixture RC were sintered at 1000 and 1450 ◦C, and the specimens of mixture TC were sintered at 1000 and 1500 ◦C.

As a result of this sintering, no negative changes were found in the regenerate samples. Thus, it could be proved that using the regenerates provided by the fragmentation results

in a comparable refractoriness compared to the refractory material made of primary raw materials.

For a final evaluation of the quality and usability of the regenerates obtained, a test of cold compressive strength and cold bending tensile strength was carried out. By previously treating the samples at different temperatures, three different strength characteristics could be determined for each mixture: the test specimens that were dried in the drying oven at 120 ◦C only (mixture RC and TC) and the test specimens that were sintered at 1000 ◦C (mixture RC and TC) as well as at 1450 ◦C (mixture RC) and 1500 ◦C (mixture TC).

The test results for the samples of mixture RC are shown in Figure 13. The direct comparison of the achieved strengths of the regenerate samples and the reference samples shows that there are no significant differences in strength. The achieved strengths of the regenerate samples are at least as high as the strengths of the reference samples. Some of the regenerate samples even show slightly higher strengths.

**Figure 13.** Comparison of the achieved cold bending tensile strength (**a**) and cold compressive strengths (**b**) of the refractory concrete (mixture RC) with regenerates (ZAC and bauxite) and the corresponding reference samples.

After testing the cold bending tensile strength, the fracture surfaces of the individual test specimens could also be checked (Figure 14). It is clear that the microstructure of the refractory concrete with regenerates is completely comparable to the microstructure of the refractory concrete with primary raw materials. Thus it is shown that the regenerates used have no negative influence on the quality of the refractory concrete and can therefore substitute for primary raw materials equally.

**Figure 14.** Test specimen of mixture RC after the cold bending tensile strength test.

A similar result was found for the samples of mixture TC with the tabular alumina. The determined strengths of the regenerate samples and the reference samples are absolutely comparable (Figure 15). The strengths achieved in the regenerated sample are even slightly higher than in the reference samples. When examining the fracture surfaces of the test specimens (Figure 16), no differences can be found. Thus, the suitability of the recovered tabular alumina used could also be proven on the basis of these results.

**Figure 16.** Test specimen of mixture TC after the cold bending tensile strength test.

### **4. Discussion**

The described results of this study prove that the technology of electrodynamic fragmentation is suitable for breaking down the composite material refractory into its individual components. While a conventional jaw crusher can only crush composite materials to a small size but still maintain the bond, electrodynamic fragmentation allows the valuable aggregates or regenerates respectively to be separated from the composite or binder matrix selectively. The advantage of the technology used is primarily that the electrical impulses run along the grain boundaries of the material, and the composite is torn from the inside to the outside by an expanding plasma channel. Another advantage of the fragmentation technology is that it can be considered as a dust-free and, above all, contamination-free process. Metallic abrasion, as is the case with jaw crushers, cannot take place.

After fragmentation, the various regenerates could be sorted by optical sorting, which is a necessary requirement for a high-quality reuse. Only the white corundum from sample material RefMat-3 could not be sorted out because the color differences compared to the matrix material are too small. In order to optically separate these materials with slight color differences, more powerful sorting approaches must be developed and used.

It could also be shown that the separated and sorted regenerates, namely tabular alumina, bauxite, ZAC and white corundum, have a very clean surface and are almost completely free of binder residues and other adhesions. By means of chemical–mineralogical analyses, a very high chemical purity of the recovered regenerates could be achieved. A comparison of the chemical composition with original raw materials clearly showed that there are almost no differences observable.

Finally, the recovered regenerates could be reused in refractory concrete or refractory tamped concrete without sophisticated post-treatment, thus replacing primary raw materials without affecting the properties of the refractory products. The processing properties of the fresh masses as well as the mechanical test results of the sintered samples using regenerates do not show any adverse effects, as is usually the case when using conventionally crushed material as aggregate material. However, no testing of the newly developed refractory material under working conditions (e.g., lining in an aluminum melting furnace) has been taken place so far. Thus, the material must be checked for further aspects such as corrosion resistance, thermal shock resistance and abrasion resistance.

The technology of electrodynamic fragmentation can be a promising alternative to existing recycling technologies. It is possible to recover high-quality secondary raw materials for the refractory industry and thus realize a high recycling rate for the material itself. However, further research and development work is needed to make a recycling process for refractory materials economically realizable on an industrial scale. This is especially true for the fragmentation process. For this purpose, the plant must be designed in such a way that a continuous throughput of material for an industrial viable throughput is possible.

### **5. Patents**

Based on this study, a patent was applied for with the title "Method for recycling ceramics, regenerated materials obtained thereby, and use of the regenerated materials for manufacturing ceramics" (DE 10 2017 217 611).

**Author Contributions:** Conceptualization, S.S., S.D. and J.B.; methodology, S.S. and S.D.; validation, S.S., S.D. and J.B.; investigation, S.S. and S.D.; resources, J.B.; data curation, S.S. and S.D.; writing—original draft preparation, S.S.; writing—review and editing, S.D. and J.B.; visualization, S.S.; supervision, S.S.; project administration, S.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data available on request due to restrictions e.g., privacy or ethical. The data presented in this study are available on request from the corresponding author. The data are not publicly available due to ongoing patent process.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**

