*2.3. Experimentation*

The leaching of the required weight of ESP dust in distilled water was carried out at L:S = 10:1 at 95 ◦C for 60 min in a 0.5 dm<sup>3</sup> Lenz Minni thermostatic reactor (Lenz Laborglas GmbH & Co. KG, Wertheim, Germany) fitted with an overhead stirrer and ports for a refrigerator and samplers.

Autoclave leaching of the ESP dust sample at L:S = 3:1 was carried out at 160, 200, and 240 ◦C in an aluminate mother liquor for 15–90 min in a 1 dm<sup>3</sup> Parr autoclave (Parr Instrument, Moline, IL, USA). Leaching in an aluminate liquor at 90 ◦C also used a thermostatic Lenz Minni reactor.

After water and alkaline leaching, the resulting red mud was separated from the aluminate liquor by filtration on a Buchner funnel. After washing and drying of the bauxite residue for 8 h at 110 ◦C, the content of rare-earth elements and other components was measured. The solid phase obtained after water leaching of ESP dust is designated as ESPDW, and that obtained by autoclave leaching with an alkaline aluminate liquor at 240 ◦C for 90 min is designated as ESPDA.

#### **3. Results and Discussion**

It can be seen from Table 1 that the chemical composition of ESP dust and bauxite sinter are different. The high LOI in the dust is associated with the incomplete decomposition of soda and calcite; however, there is also a significant decrease in the content of silica, magnesium, and calcium in dust relative to alumina and an increase in sodium content. A decrease in silica content in dust leads to a significant increase in the silicon modulus (μSi). The results of X-ray diffraction analysis (Figure 2) confirm the presence of sodium aluminate in the ESP dust; however, most of the soda remains unreacted.

<sup>1</sup> LOI—loss on ignition at 1000 ◦C. 2 μSi (silicon modulus)—mass ratio of Al2O3 to SiO2.

According to the previous research [19,21,43,44], the REE and Sc are possibly adsorbed on goethite, hematite and in the channels of aluminosilicates in the typical Bayer bauxite residue. The primary container of these elements could be cancrinite [45]. However, during the sintering process iron minerals are transformed into sodium ferrite and aluminosilicate—into sodium silicate and sodium aluminate (Figure 2). Therefore, REE and Sc could be liberated through the sintering process, which could lead to an increase in their concentration in the by-products and increased leachability than in typical Bayer bauxite residue.

Figure 3 shows the content of rare-earth elements in ESP dust and sinter. The obtained data show that the ESP dust from the sintering process contains more REE than the sinter itself. Moreover, if we detract LOI (consisting of water and CО2 carbonates, which are removed from sinter at 1000 ◦C and which remain in ESP dust, according to Table 1), the REE content in the calcined ESP dust will be approximately 70–80% higher than in the sinter. The obtained data is in good agreemen<sup>t</sup> with the results presented previously in the patent [42], where it was shown that ESP dust contains about 50 ppm of scandium. This may be due to the repeated circulation of fine dust in the process and mineralogical transformation. Consequently, ESP dust can be a source of REE, and we can consider the option of removing it from the cycle in order to extract valuable components and increase the efficiency of the sintering kiln.

**Figure 3.** Content of rare earth elements in electrostatic precipitator dust and bauxite sinter, ppm (LOI—loss on ignition at 1000 ◦C).

#### *3.1. Water Leaching of ESP Dust*

The XRD analysis of ESP dust (Figure 2) shows a high content of soluble minerals (such as sodium silicate and sodium aluminate). Tests were carried out to leach ESP dust with water for 1 h at 95 ◦C, which is necessary for the complete conversion of sodium salts into liquor. The output of the solid phase (ESPDW) after leaching was 40% of the initial weight of ESP dust. Table 2 shows the chemical composition of the resulting red mud, and the degree of recovery of the main components into the liquor is shown in Figure 4. Part of the aluminum in electrostatic precipitator dust was converted to sodium aluminate; it is therefore easily leached with water. Notably, aqueous leachate has a higher silicon modulus than the original dust, as silicon has a degree of recovery higher than aluminum. This indicates that the degree of transformation of silica to sodium silicate is higher than alumina to sodium aluminate. Na2O is almost completely leached, which indicates that with water leaching for less than 1 h the components in the resulting liquor do not have time to form a disilication product according to the following Equation (1):

$$6\text{Na}\_2\text{SiO}\_3 + 6\text{NaAl(OH)}\_4 + \text{Na}\_2\text{X} \rightarrow \text{Na}\_6\text{IAl}\_6\text{Si}\_6\text{O}\_{24} \text{I} \cdot \text{Na}\_2\text{X} + 12\text{NaOH} + 6\text{H}\_2\text{O},\tag{1}$$

where X represents various inorganic anions, most often sulfate, carbonate, chloride, aluminate, etc. [46].

**Table 2.** The chemical composition of electrostatic precipitator dust after water leaching, wt. %.

**Figure 4.** The extraction degree of the main components into the liquor by water leaching of electrostatic precipitator dust.

It can also be seen in Figure 4 that REE (Sc, Y, La, Ce, Nd) are slightly extracted during water leaching, because of their interaction with alkali and soda [47]. As shown in some studies [39,40], the degree of scandium recovery from red mud to soda liquor usually does not exceed 20%.

As it was stated in the introduction section, scandium is the most valuable REE in red mud. It was previously shown that scandium in red mud is mainly associated with iron minerals [48]. However, some researches show [45] that cancrinite could accumulate Sc as well. Therefore, to study the association of rare-earth metals with different phases, we performed ESPD and ESPDW surface mapping using EPMA (Figure 5).

It can be seen from Figure 5 that scandium in ESP dust is mainly associated with iron minerals, and to a lesser extent—with silicon compounds. It could be associated with the fact that iron minerals (hematite) are not fully transformed into sodium ferrite during sintering because small particles of dust pass very quickly through the hot zones of the furnace and a disilication product is not formed yet. Almost the same picture is observed in ESPDW; however, the distribution of scandium is more uniform than in ESP dust. This may be due to the formation of a disilication product (sodalite or cancrinite), which can adsorb the REE. This suggests that destruction of the hematite matrix is required at first for the complete recovery of REE from ESPD, as is the case with the typical Bayer process red mud. This also explains the low degree of REE recovery at the water leaching stage, since hematite cannot be dissolved by sodium carbonate liquor.

#### *3.2. Kinetics of ESP Dust Leaching by Alkaline Aluminate Liquor*

To study the mechanism and effect of temperature on the leaching of initial ESPD and ESPDW with the alkaline aluminate mother liquor, experiments were carried out to measure the aluminum extraction degree from the liquor with variable duration and temperature of the process. Aluminum was chosen as an indicator of leaching efficiency as it was the only element extracted from the liquor during alkali leaching of ESPDW. The temperature range was selected, taking into account that, after the extraction of readily soluble sodium salts, the alumina in the material is mainly represented by boehmite, the leaching of which requires a temperature of more than 160 ◦C [49]. The experimental results are shown in Figure 6.

(b) 

**Figure 5.** Mapping of electrostatic precipitator dust surface (**a**) and water-leached electrostatic precipitator dust surface (**b**) using the electron probe micro-analyzer.

It can be seen from Figure 6 that the kinetics of aluminum extraction into the liquor is quite high at all temperatures for the first 30 min for ESP dust. Even at 90 ◦C, more than 40% of aluminum is extracted from ESP dust after 30 min, apparently because of a certain degree of conversion of boehmite to sodium aluminate in the sintering kiln. However, after pre-leaching in water (ESPDW), sodium aluminate had already been extracted into the first-stage liquor; therefore, the rate of leaching of aluminum from ESPDW at 90 ◦C is significantly lower. Although at higher temperatures, due to a lower silica content, the efficiency of alumina recovery from ESPDW increases, reaching 90% after 90 min of leaching at 240 ◦C.

To study the leaching mechanism, the obtained kinetic curves were processed using the shrinking core model [50]. We studied six kinetic equations [51] describing the process in various modes, from kinetic to intra-diffusion; however, the models shown below (Equations (2) and (3)) proved to be most promising for the process description:

$$1 - \Im(1 - X)^{2/3} + \Im(1 - X) = k\_1 t,\tag{2}$$

$$1/3\ln(1-X) + ((1-X)^{-1/3}-1) = k\_2t,\tag{3}$$

where *X* is the degree of aluminum recovery into the liquor at a time *t*, *ki* is the apparent rate constant. Equation (2) describes the process in the intra-diffusion area, while Equation (3) describes the process limited by interfacial transfer and diffusion through the product layer. The plot of 1/3ln(1 − *X*) + ((1 − *<sup>X</sup>*)−1/<sup>3</sup> − 1) versus t for ESP dust leaching gives a straight line (Figure 7a) with the determination coefficient R<sup>2</sup> the highest among all models used (more than 0.98 for all temperatures except 90 ◦C), which indicates that leaching, in this case, is most likely limited by interfacial transfer and diffusion through the product layer. It can be concluded that during leaching, a disilication product is formed around the core of the boehmite (Figure 8), which slows down the leaching process. Also, a film of sodium titanate can form on the surface of boehmite, which is known [52] to reduce the rate of dissolution of aluminum hydroxides significantly. For water-leached dust, the highest determination coefficient is observed, when using Equation (2) (Figure 7c), which implies that the process is limited by diffusion through the layer of the product or unreacted matter.

**Figure 6.** Effect of leaching time and temperature for Al extraction from raw ESP dust (**a**), from water-leached ESP dust (**b**).

Using the obtained values of the apparent rate constant (*ki*) in Figure 7b,d and the Arrhenius equation (Equation (4)), we determined the values of the apparent activation energy for leaching ESPD and ESPDW (Figure 6) to be 24.98 kJ/mol and 33.19 kJ/mol, respectively.

$$k\_i = \text{Aexp}(-E\_a/\mathbb{R}T),\tag{4}$$

where A is the Arrhenius constant, R is the universal rate constant (8.314 J/mol·K), *T* is the temperature (K), *Ea* is the apparent activation energy (J/mol).

The obtained values of the activation energy also confirm diffusion limitation. However, for an intradiffusion stage, the activation energy should be in the range of 8–22 kJ/mol. The higher values, in this case, maybe due to the fact that higher activation energy is required for the dissolution of the boehmite since a temperature of more than 160 ◦C is required for its extraction into alkali liquor. In the second case, the activation energy is higher, since there is no easily soluble phase of aluminum left after water leaching and a low degree of leaching efficiency of boehmite is observed at 90 ◦C.

Figure 8 shows XRD patterns of ESPDW and ESPDA, from which it can be seen that alkaline leaching results in disappearing of boehmite peaks and appearing instead of the peaks corresponding to cancrinite; iron, in contrast to the Bayer red mud, is represented by both hematite and hydroxide phases.

The yield of solid residue (red mud) from the leaching of ESPD in alkaline aluminate liquor at 240 ◦C for 90 min was 29.8% of the initial weight of dust. At the same time, the yield of red mud (ESPDA), after ESPDW leaching with alkali liquor at 240 ◦C for 90 min was 21.0%. As a result, the degree of REE enrichment of red mud in the second case was higher. Figure 9 shows the content of rare earth metals in these products.

**Figure 7.** Results of substituting the data for ESPD leaching to Equation (3) (**a**); substituting the data for ESPDW leaching to Equation (2) (**c**); dependence ln*k* − *T*−<sup>1</sup> for leaching the original ESP dust (**b**) and water-treated ESP dust (**d**).

**Figure 8.** XRD pattern of water-leached electrostatic precipitator dust (**green**) and alkali-leached electrostatic precipitator dust (**red**).

The data in Figure 9 show that the total amount of REE in the red mud after ESP dust leaching was more than 1700 ppm in the first process and over 3200 ppm in the second process. Thus, we demonstrated that, in principle, it is possible to concentrate rare-earth elements in red mud by leaching ESP dust with water and alkaline aluminate liquor. The REE content in the red mud obtained thereby is three times higher than in conventional red mud of the alumina refinery, which can significantly reduce the cost of obtaining the REE concentrate in the future. Moreover, the recovery of additional components from natural raw materials can improve the efficiency of processing of bauxite raw materials in general. Based on these findings, we propose the following ESPD processing scheme (Figure 10), which enables efficient recovery of both alumina and rare earth elements. The stage of leaching REE from ESPDA will be discussed in the next paper.

**Figure 9.** Content of rare earth elements in water-leached electrostatic precipitator dust and alkali-leached electrostatic precipitator dust.

**Figure 10.** A flowsheet of rare-earth elements (REE) concentration from electrostatic precipitator dust by water/alkali leaching.
