**1. Introduction**

Today, refractory materials are used in a wide variety of industrial sectors. These ceramic materials are products that are generally used under a high-temperature load of >1500 ◦C [1] for lining and delivery of thermotechnical facilities, such as blast furnaces or converters, as well as for transport devices in iron and steel, aluminum, cement and ceramic industries, incineration plants and refineries. The material separates the reaction zone from the outer parts of the process devices and are in contact with solid and liquid but also gaseous, partly very aggressive reaction components and reaction products [2]. Without these refractory materials there would be no technical thermal processes, which are fundamental to the production of steel, iron, aluminum, cement and glass. For example, in 2016 the German steel industry produced a total of 40.2 million t of steel products, requiring 0.5 million t of refractory materials [3]. This means approximately 10 kg of refractory materials were needed to produce 1 t of steel.

The necessary refractability as well as other important properties such as zero shrinkage, high thermal shock resistance, chemical resistance and mechanical or temperaturedependent strength are given to refractory materials by their non-metallic, inorganic composition. The main components consist of the six basic oxides SiO2, Al2O3, MgO, CaO, Cr2O3 and ZrO2, often in combination with carbon (e.g., SiC) [4]. These oxides form refractory compounds such as bauxite, corundum or white corundum, tabular alumina, zirconium (zirconium silicate), fireclay or silicon carbide through complex and high-emission thermal processes (dehydrogenation, sintering reactions or melt flow processes). These

**Citation:** Seifert, S.; Dittrich, S.; Bach, J. Recovery of Raw Materials from Ceramic Waste Materials for the Refractory Industry. *Processes* **2021**, *9*, 228. https://doi.org/10.3390/ pr9020228

Academic Editor: Daniel Vollprecht

Received: 23 November 2020 Accepted: 21 January 2021 Published: 26 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

raw materials or semi-finished products are then used to produce the refractory materials. The largest raw material resources are located mainly in China, Russia, South Africa, South America and Australia. Accordingly, the German and European refractory industry is highly dependent on these imports. The refractory materials are exposed to high thermal, physical and chemical loads and serve as wear material during the use in industrial processes.

The average lifetime of refractory materials depends on the application and ranges from few days (e.g., purging plugs) to a few weeks (e.g., steel-casting ladles and converters) to long-term applications from one to several years (e.g., converters and heatexchangers/preheaters). After the utilization phase, the refractory lining must be repaired or at least relined. The majority of the lining wears out completely and is no longer usable or is chemically contaminated such that it requires an extensive replacement. A smaller amount of the material (approximately 5–10% of the waste material) is currently recycled after stripping and used as so-called regenerates for the production of new refractory materials. For this purpose, the stripped material is pre-sorted and afterwards crushed in a jaw crusher, for example. The subsequently sieved material can partly be used as aggregate in new refractory products. However, a higher chemical purity of these regenerates required for the refractory industry cannot be achieved. For these reasons, little recycled material is used in the refractory industry today. Predominantly waste material and thus valuable secondary raw material for the refractory industry is deposited in special landfills for final disposal. A comprehensive industrial recycling process for refractory materials is a major challenge that offers the chance to reduce the dependency on primary raw material imports and the global CO2 emissions.

A suitable method for processing refractory waste material in order to enable a high rate of recycling could be electrodynamic fragmentation. This innovative method uses pulsed high voltage discharges to separate bulky multi-phase material selectively along grain boundaries. The discharge has to take place underwater to enable the solid material to be penetrated. Thus, the whole approach is a "wet process". The potential of this pulsed power processing approach was demonstrated for the recycling of waste concrete [5] or the separation of municipal waste incineration ash [6,7]. Like refractory material, the investigated ash is a mineral-based compound material. Reusable components like iron metal, glass or ceramics are enclosed by a siliceous matrix. It was proved that after processing the ash with electrodynamic fragmentation the individual components of the ash could be recovered selectively.

The aim of the study presented here was to use the innovative process of electrodynamic fragmentation as a recycling strategy for refractory waste products in order to obtain regenerates with a high chemical purity. With these high-quality regenerates, new sources of raw materials for the refractory industry could be made accessible, and thus primary resources are conserved. The increased use of such high-purity refractory regenerates would lower the cost of refractory materials and could reduce dependencies on world markets.

### **2. Materials and Methods**

### *2.1. Electrodynamic Fragmentation*

Electrodynamic fragmentation is a technique using pulsed power discharges to separate compound materials selectively. The method itself was first investigated and described at the University of Tomsk in the late 1940s [8]. The efficiency of the method is the higher the more electrical discharges are generated over time. Usually, a so-called Marx generator is used to enable a high rate of high voltage discharges [9]. Besides a high voltage, a high slew rate is of special interest. The slew rate describes how fast the maximum voltage of a discharge can be achieved and thus determines the possible pulse length. With a Marx generator a slew rate that enables pulse lengths of a few nanoseconds can be realized. This is very important in order to force the electrical discharge into a solid material. Whether an electrical discharge can penetrate a material or not depends on the dielectric strength

of the given material. The dielectric strength of a material is not a constant but varies in dependency with the slew rate and therefore the length of a discharge [10]. At a pulse rise time lower than 500 nanoseconds, the dielectric strength of water is higher than that of solid material (see Figure 1). Electrodynamic fragmentation utilizes this physical principle.

**Figure 1.** Dielectric strength of different materials as a function of the pulse rise time (modified from [10]).

In the process vessel of the fragmentation unit, a solid material sample is placed underwater. The bottom of the process vessel serves as an electrode. A second electrode is placed on top of the sample. The distance between sample and upper electrode can be varied. After voltage application, a polarization of the sample takes place. This leads to electrostriction, meaning micro stress within the sample as the charge carrier (e.g., electrons) of the solid material are not freely movable [11]. As different components in a compound material, such as refractory waste material, differ in their dielectric strength, the grain boundaries represent the regions with the biggest contrast in polarization. The grain boundaries therefore are the path of least resistance for the discharge. Along the grain boundaries, so-called streamers infiltrate the material during a first stage of the pulse discharging. When the first streamer reaches the electrode at the bottom of the process vessel, the complete energy of the discharge runs along the corresponding path. A short-lived plasma channel is generated reaching temperatures of up to 104 K. During the collapse of the plasma channel, a shockwave of up to 10 GPa shatters the material having the strongest influence for the separation. The shockwave is reflected at the process vessel and penetrates the material a second time. As the described procedure occurs with each discharge, a fast and efficient separation of a solid material is achieved (Figure 2).

**Figure 2.** Principle of electrodynamic fragmentation of a composite material.

In this study a laboratory plant was used, which works in batch mode with a five liter process vessel [12]. The general setup of the lab plant allows a processing of samples with a diameter of around 40 mm. The maximum volume of sample material for a single fragmentation step depends on the density of the material and is about several hundred grams of material generally. The default operating parameters for the lab plant are a pulse rate of 5 Hz and a voltage of 180 kV per pulse, though these parameters are adjusted to achieve an optimum in separation for each sample processed.

### *2.2. Sample Material*

In this study, different refractory waste or break-out material was investigated and chosen from different industrial fields of application that were as varied as possible. It was also important that the selected materials were exposed to different temperatures during the manufacturing process or during use. In refractory materials, higher temperatures lead to a stronger sintering of the individual components, which makes it considerably more difficult to process or cleanly separate the individual components with conventional processing methods.

For this study, three different refractory waste materials were collected (Figure 3), from different fields of application and containing potential regenerates, such as bauxite or zircon. A brief description of the sample material with details of the potential regenerates is given in Table 1. All materials used in this study originate from shaped bricks.

**Figure 3.** Refractory waste materials from various applications. (**a**) Highly sintered brick for sintering or rotary kilns; (**b**) corundum stone for melting tanks; (**c**) functional refractory ceramic from the steel industry.



The sample material RefMat-1 (Figure 3a) consists of preformed bricks, which are fired at high temperatures (1250 ◦C) before being used in a sintering furnace. Only after the firing is the stone used for the kiln lining. These stones are mainly composed of bauxite, zirconia-alumina fused grain (ZAC), SiC and microsilica as well as a matrix of aluminacement-based binder. Of this refractory material, there is interest above all in recovering bauxite and ZAC as regenerates.

The sample material RefMat-2 (Figure 3b) is a pure corundum stone, which is used at very high temperatures and has a high chemical load. The main components are white corundum and alumina-cement-based binder. These highly sintered bricks are used as lining for melting furnaces, e.g., for aluminum melts. The recovery of the white corundum as regenerate would considerably save primary resources.

The sample material RefMat-3 (Figure 3c) consists of functional refractory ceramics from the steel industry. These functional refractory ceramics have fine channels and are used in the furnace linings or at the bottom of the blast furnaces and ladle linings of steelworks to inject various gases into the molten steel via the channels. The main components of these functional refractory ceramics are tabular alumina and white corundum in a binder matrix of alumina cement. Potential regenerates with this sample material would be tabular alumina and the white corundum.

### *2.3. Fragmentation of the Sample Material*

In order to obtain coarse sample fragments, prior to the fragmentation experiments, the three different blocks were crushed down using a sledgehammer. The resulting size of sample material was about 4 to 5 cm each. At the beginning of the fragmentation experiments (Figure 4), several small samples of each material were processed, whereby the parameters voltage, electrode-sample gap, number of pulses and pulse frequency were varied. The goal was to identify the optimal fragmentation parameters for each sample material in order to subsequently process a larger amount of sample material with the optimized parameters.

**Figure 4.** Fragmentation of the RefMat-2 sample. Placing the sample in the process vessel (**a**) and underwater sample in the process vessel (**b**).

After the optimal fragmentation parameters for the individual sample materials had been determined in the preliminary tests, larger quantities of the individual materials could be processed in the next step in order to obtain sufficient material or regenerates for the subsequent recycling tests. Finally, between 20 and 32 kg of each of the three samples was fragmented in batches. After fragmentation, the separated material was dried and classified into different grain size fractions by means of sieve classification. In addition, the process water from the individual tests was filtered and the filter residue dried so that a fine fraction could be obtained.

In the next step, the fragmented and sieve-classified material was sorted in order to keep the separated aggregates, or regenerates, sorted by type. Therefore, the fragmented material was sorted by optical sorting, meaning by differences in color and translucency. The sorting was carried out with a laboratory system for optical bulk material sorting (TableSort), which was developed at the Fraunhofer IOSB [13,14]. Equipped with a RGB filter camera and a filigree blow-out device, this system is suitable for small amounts of material. The electro-optical sorting was carried out in several steps or passes for each individual grain size fraction. In each pass, the optical filter was adjusted in such a way that the desired material, meaning the regenerate in question, was blown out from the bulk mass flow.

### *2.4. Methods of Investigation*

To characterize the collected material, all samples were analyzed before the fragmentation process by X-ray phase analysis (XRD) to determine the mineralogical phase composition and by X-ray fluorescence analysis (XRF) to investigate the chemical composition. For the XRD analysis, a D2 Phaser (Bruker) was employed, and for the XRF analysis, an Epsilon 3 XL (Panalytical) was used. The XRD analysis was performed using powder samples. For the XRF analysis, powder compacts were used. Before sample preparation, all

materials were crushed and milled down. For the analysis before the fragmentation experiments, the fragments obtained from crushing the refractory blocks with a sledgehammer were milled in a two-step process. Firstly, a vibratory disc mill was used to achieve a grain size below 1 mm. Secondly, the material was milled down using a McCrone micronizing mill. With this wet-milling process a powder fineness was achieved suitable for XRD analysis. After the material was fragmented and sorted, the obtained regenerates were examined by optical microscopy to evaluate the degree of the achieved detachment. To determine the chemical purity of the recovered material, the regenerates were analyzed by XRD and XRF. Subsequently, the analytical results were compared with the analysis of primary aggregates (e.g., ZAC or tabular alumina).

The sample material obtained from the fragmentation experiments was milled down in a two-step process as well. The aforementioned McCrone mill was used after the material was pretreated in a ball mill to obtain the particle size required in the wet-milling process.

Furthermore, the recovered material was used to produce the refractory test specimen (prisms with a dimension 4 × 4 × 16 cm) in accordance with industrial requirements. The cold compressive strength as well as the cold bending tensile strength of all prisms was determined using a Z100 testing machine from ZwickRoell and an Alpha 3-3000 S testing machine from Form + Test, respectively.

### **3. Results**

### *3.1. Characterization of the Refractory Waste Material*

As expected, all samples show a very high Al2O3 content, which is mostly distributed between the mineralogical phases α-alumina (corundum) and β-alumina. Therefore, all samples are high alumina refractories.

In addition to the high Al2O3 content, the sample material RefMat-1 shows high contents of SiO2 and ZrO2 (Table 2). This chemism is reflected in the mineralogical phases corundum (Al2O3), baddeleyite (ZrO2) and mullite (3Al2O3-2SiO2). Minor phases are potassic and alkali feldspars, SiC and grossite (CaAl4O7) and some zirconium (ZrSiO4). The two mineral phases corundum and baddeleyite are components of the zirconia-alumina fused grain (ZAC), which is one of the main components of these refractory materials. Furthermore, the two highly refractory mineral phases corundum and mullite are components of calcined bauxite, which is also used as an important raw material in many shaped and unshaped refractory products.


**Table 2.** Chemical and mineralogical composition of the sample material RefMat-1.

The sample material RefMat-2 is an almost pure Al2O3 product (Table 3). Besides the very high Al2O3 content, only very low contents of Fe2O3 and CaO could be detected. The Al2O3 is mainly found in the mineral phase corundum or in white corundum. White corundum, a chemically pure alumina, is extracted from the melt and is used both in unshaped and shaped refractory products. Figure 5 shows an example of the X-ray diffractogram of sample RefMat-2.


**Table 3.** Chemical and mineralogical composition of the sample material RefMat-2.

**Figure 5.** X-ray diffractogram of sample RefMat-2.

Sample material RefMat-3 has a slightly lower Al2O3 content than sample RefMat-2. In addition, contents of MgO and traces of SiO2, Fe2O3, CaO and Cr2O3 were analyzed (Table 4). The two mineral phases α-alumina (corundum) and spinel (MgAl2O4) could be detected by phase analysis. The α-alumina can be attributed to the tabular alumina contained in the sample, but also to the white corundum. The two phases can hardly be distinguished by X-ray diffraction. Tabular alumina is chemically equivalent to the white corundum described above. The only difference is that white corundum is obtained from the melt and tabular alumina is recrystallized or sintered α-alumina with a high density, the morphology of which consists of large (50–200 μm), tabular corundum crystals. The two mineral phases catoite (Ca3Al2(OH)12) and grossite (CaAl4O7) are residual components of the alumina-cement-based binder.


**Table 4.** Chemical and mineralogical composition of the sample material Ref-Mat-3.

### *3.2. Fragmentation of the Refractory Waste Material*

Based on the fragmentation pre-tests, the individual parameters could be determined for each sample material, allowing a larger amount of material to be processed in the next step. The resulting material was then dried and subsequently screened into four grain size fractions (<3, 2–3, 1–2 and >1 mm; see Figure 6). In addition, a fine fraction was obtained from the process water by filtration.

**Figure 6.** Sample material RefMat-3 before (central) and after fragmentation and subsequent classification by grain size.

Table 5 shows the mass balance of the individual grain size fractions. It can be seen from this table that the proportion of the fine fraction (filter residue) from the process water for sample MatRef-1 and sample MatRef-3 is almost 45 wt.% and for sample MatRef-2 even 60 wt.%. This high proportion is due to the relatively high number of pulses per material input (Table 6). In order to separate the individual components from the binder matrix, a high number of pulses with a high voltage had to be applied. This resulted in a lot of fine material. After drying, it was also observed that the fine material from the process water partially solidified. One reason for this is that a large amount of binder accumulates in the filter residue, and the binder also has a residual hydraulic activity.


**Table 5.** Weight percentages of the individual fractions after fragmentation.

**Table 6.** Process data.


The various aggregates of the fragmented refractory ceramics, which are to be recycled as regenerates, have mostly accumulated in the coarser fractions (1 to >3 mm). In addition to the very cleanly exposed aggregates, binder residues can also be detected in the coarser fractions.
