Acid and Acid-Alkali Treatment Methods of Al-Chloride Solution Obtained by the Leaching of Coal Fly Ash to Produce Sandy Grade Alumina

: Sandy grade alumina is a valuable intermediate material that is mainly produced by the Bayer process and used for manufacturing primary metallic aluminum. Coal ﬂy ash is generated in coal-ﬁred power plants as a by-product of coal combustion that consists of submicron ash particles and is considered to be a potentially hazardous technogenic waste. The present paper demonstrates that the Al-chloride solution obtained by leaching coal ﬂy ash can be further processed to obtain sandy grade alumina, which is essentially suitable for metallic aluminum production. The novel process developed in the present study involves the production of amorphous alumina via the calcination of aluminium chloride hexahydrate obtained by salting-out from acid Al-Cl liquor. Following this, alkaline treatment with further Al 2 O 3 dissolution and recrystallization as Al(OH) 3 particles is applied, and a ﬁnal calcination step is employed to obtain sandy grade alumina with minimum impurities. The process does not require high-pressure equipment and reutilizes the alkaline liquor and gibbsite particles from the Bayer process, which allows the sandy grade alumina production costs to be to signiﬁcantly reduced. The present article also discusses the main technological parameters of the acid treatment and the amounts of major impurities in the sandy grade alumina obtained by the di ﬀ erent (acid and acid-alkali) methods. Polypropylene, which was stable in hydrochloric acid, was used as a ﬁlter cloth, and acetone was used as a washing liquid. After washing, the ACH crystals obtained were dried in a drying oven at 105 ◦ C for 2 h, and the impurities in the crystals were then analyzed.


Introduction
Sandy grade alumina (SGA) is the aluminum oxide that is used for metallic aluminum production and is mainly produced by the Bayer alkaline method. In this method, bauxite is used as a raw material, which may differ in quality and composition. For bauxite with a high silica content, the sintering process is used [1,2]. A decrease in the bauxite quality and orientation of the main aluminum producers (Russia and China) with regards to their own resources has led to investigations of alternative raw materials for SGA production [3]. In fact, it can be obtained from both natural (kaolin clay [4,5] or pyrophyllite ore [6]) and technogenic (red mud [7], coal fly ash (CFA) [8], and secondary aluminum dross [9,10]) materials. The CFA waste generated from coal-fired power plants is one of the most promising raw materials for SGA production [11], since it does not require preliminary grinding, as in Figure 1. A flowsheet of sandy grade alumina (SGA) production from the Al-chloride liquor by the conventional acid and novel acid-alkali treatment methods (red arrows show the traditional acid method used for SGA production, and green arrows show the new acid-alkali method).

Materials
Raw CFA was collected from the Omsk TPP-4 ash residue storage (Google Earth coordinates: 55.129231, 73.167497). Magnetite concentrate and unburned carbon were separated from CFA by wet magnetic separation and by foam flotation with kerosene as a collector, respectively [30]. The nonmagnetic fraction, after the enrichment process, was subjected to high-pressure HCl leaching to maximize the aluminum extraction from the liquor (T = 210 °C, CHCl = 30%, τ = 3 h, and the solid liquid ratio = 1:5) [29]. The chemical composition of the liquor is presented in Table 1. Table 1. The chemical composition (g/l), measured by inductively coupled plasma optical emission spectrometry (ICP-OES), of the aluminum chloride liquor obtained by high-pressure HCl leaching of coal fly ash (CFA).

Analytical Methods
The metal concentrations in the liquor and solid samples were measured by inductively coupled

Materials
Raw CFA was collected from the Omsk TPP-4 ash residue storage (Google Earth coordinates: 55.129231, 73.167497). Magnetite concentrate and unburned carbon were separated from CFA by wet magnetic separation and by foam flotation with kerosene as a collector, respectively [30]. The non-magnetic fraction, after the enrichment process, was subjected to high-pressure HCl leaching to maximize the aluminum extraction from the liquor (T = 210 • C, C HCl = 30%, τ = 3 h, and the solid liquid ratio = 1:5) [29]. The chemical composition of the liquor is presented in Table 1. Table 1. The chemical composition (g/l), measured by inductively coupled plasma optical emission spectrometry (ICP-OES), of the aluminum chloride liquor obtained by high-pressure HCl leaching of coal fly ash (CFA).

Analytical Methods
The metal concentrations in the liquor and solid samples were measured by inductively coupled plasma optical emission spectrometry (ICP-OES) using an atomic absorption spectrometer AA-240FS Metals 2020, 10, 585 4 of 18 (Varian, Belrose, Australia). Solid samples were dissolved beforehand by distilled water (ACH) or hydrochloric acid (alumina).
An X-ray diffraction analysis (XRD) of samples was carried out using an Ultima IV diffractometer (Rigaku, Tokyo, Japan) with Cu-Kα; the radiation was in a range of 2θ angles of 10-100 • with a 0.02 • step. The quantitative and qualitative structure of samples were investigated using the PDXL (Rigaku) program.
The change of mass and temperature of phase transformations of the ACH sample were determined via differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA), using an STA 409 Luxx synchronous thermal analyzer (Netzsch, Selb, Germany) combined with a QMS 403 Aeolos quadrupole mass spectrometer with a capillary connection (Netzsch, Selb, Germany). The heating rate was 5 • C/min in the temperature range of 100-1200 • C.
The morphology of the samples was examined by scanning electron microscopy (SEM) using EVO LS 10 (Carl Zeiss, Oberkochen, Germany) and Vega 3 (Tescan, Brno, Czech Republic) microscopes.
The particle size distribution and specific surface area of the samples were determined by the laser diffraction method (LD) using an Analysette 22 NanoTec analyzer (Fritsch, Idar-Oberstein Germany) and by the Brunauer-Emmett-Teller method (BET) using a NOVA 1200e apparatus (Quantachrome Instruments, Hook, UK).

Experiments
The salting-out of ACH from the aluminum chloride liquor was carried out by the flushing of gaseous HCl mixed with Ar at a rate of 1.5 g/min ( Figure 2) through a Drexel flask with 10 mL of the liquor, which was thermostated at temperatures of 0, 25, 50, 75, and 95 • C for 10 min. The ACH crystals obtained were separated from the liquor in a Schott funnel. Polypropylene, which was stable in hydrochloric acid, was used as a filter cloth, and acetone was used as a washing liquid. After washing, the ACH crystals obtained were dried in a drying oven at 105 • C for 2 h, and the impurities in the crystals were then analyzed. 2020, 10, x FOR PEER REVIEW 4 s quadrupole mass spectrometer with a capillary connection (Netzsch, Selb, Germany). g rate was 5 °C/min in the temperature range of 100-1200 °C. he morphology of the samples was examined by scanning electron microscopy (SEM) u LS 10 (Carl Zeiss, Oberkochen, Germany) and Vega 3 (Tescan, Brno, Czech Repu scopes. he particle size distribution and specific surface area of the samples were determined b diffraction method (LD) using an Analysette 22 NanoTec analyzer (Fritsch, Idar-Ober any) and by the Brunauer-Emmett-Teller method (BET) using a NOVA 1200e appa tachrome Instruments, Hook, UK).
xperiments he salting-out of ACH from the aluminum chloride liquor was carried out by the flushin us HCl mixed with Ar at a rate of 1.5 g/min ( Figure 2) through a Drexel flask with 10 mL o r, which was thermostated at temperatures of 0, 25, 50, 75, and 95 °C for 10 min. The A ls obtained were separated from the liquor in a Schott funnel. Polypropylene, which was s drochloric acid, was used as a filter cloth, and acetone was used as a washing liquid. A ing, the ACH crystals obtained were dried in a drying oven at 105 °C for 2 h, and the impu crystals were then analyzed. morphous Al2O3 (ACH400) was obtained by the calcination of ACH crystals in an electric ce (SNOL, Moscow, Russia) in a corundum boat. The heating time to T = 400 °C was 15 min t of the ACH sample was 15 g, and the calcination time was 1 h. Exhaust gases (HCl) d through a Drexel flask with distilled water to regenerate HCl. Amorphous Al 2 O 3 (ACH400) was obtained by the calcination of ACH crystals in an electric tube furnace (SNOL, Moscow, Russia) in a corundum boat. The heating time to T = 400 • C was 15 min, the weight of the ACH sample was 15 g, and the calcination time was 1 h. Exhaust gases (HCl) were passed through a Drexel flask with distilled water to regenerate HCl.
The leaching of ACH400 was carried out in a thermostatically controlled stainless-steel reactor equipped with an overhead stirrer. The alkaline aluminate liquor (500 mL) with a concentration of Na 2 O k = 150 g/L and α k = 3.4 was loaded into the reactor. After reaching 60 • C, 200 g/L of a 2:1 mixture of aluminum hydroxide from the Ural Aluminum Smelter (coarse Al(OH) 3 in Figure 1) and ACH400, respectively, was added to the reactor. Then, the pulp was maintained for 36 h with constant-speed stirring of 200 rpm. The precipitate obtained was separated from the liquor by filtration, washed with hot water, and dried at 105 • C for 8 h.
The conventional SGA obtained from ACH crystals (SGA-C) and the novel SGA obtained from Al(OH) 3 (SGA-A) were then annealed in a muffle furnace HTC 03/18/3N/PE (Nabertherm, Lilienthal, Germany) in open corundum crucibles. The heating time to T = 1000 • C was 1 h, the weight of the sample was 5 g, and the calcination time was 1 h.

Equations
The ACH crystal size distribution was determined by an analysis of the SEM images. It was impossible to use LD analysis, since the crystals dissolve in water. The Sturges equation [31] was employed to find the number of intervals of the particle size distribution histogram (n): where N is the number of measured ACH particles. The data obtained were analyzed by the Origin Pro 8 Software, and 1120, 790, and 813 ACH particles were determined for each ACH sample (0, 50, and 95 • C, respectively).
The ACH recovery level (R) from the aluminum chloride liquor was calculated by the following equation: where Al 1 and Al 2 represent the aluminum content (g/l) in the aluminum chloride liquor after and before salting-out experiments.

ACH Precipitation
The temperature of the Al chloride liquor and the HCl content are the two most important parameters of the salting-out process, which affect both the precipitation of the main impurities (Fe, Ca, Na, and K) and the yield of the ACH crystals [32,33]. It was found that the temperature of the liquor does not significantly affect the ACH solubility. At 125 • C, this value was 4.012 mol/kg, and at 0 • C, it was 3.283 mol/kg [34]. The HCl concentration imposes a much greater effect: with water liquor (0 mol/kg HCl), the ACH solubility was 31.20%, while with 10 mol/kg HCl, this value decreased to 0.13%. The solubility of iron chloride hexahydrate was practically unchanged and stayed at a~50% level [35,36]. The effect of temperature on the ACH yield is illustrated in Figure 3. Decreasing the temperature from 95 to 0 • C enhances the ACH yield from 58% to 99%, respectively. 0 °C, it was 3.283 mol/kg [34]. The HCl concentration imposes a much greater effect: with water liquor (0 mol/kg HCl), the ACH solubility was 31.20%, while with 10 mol/kg HCl, this value decreased to 0.13%. The solubility of iron chloride hexahydrate was practically unchanged and stayed at a ~50% level [35,36]. The effect of temperature on the ACH yield is illustrated in Figure 3. Decreasing the temperature from 95 to 0 °C enhances the ACH yield from 58% to 99%, respectively. The dependence obtained correlates with the ACH crystal size distribution shown in Figure 4. At 95 °C , the number of small crystals ≤100 µm is significantly lower than that at 50 and 0 °C. The average crystal size for 0, 50, and 95 °C is 85, 127, and 135 µ m, respectively. A large number of small crystals at 0 °C can simultaneously work as precursors for the formation of new ones. The dependence obtained correlates with the ACH crystal size distribution shown in Figure 4. At 95 • C, the number of small crystals ≤100 µm is significantly lower than that at 50 and 0 • C. The average crystal size for 0, 50, and 95 • C is 85, 127, and 135 µm, respectively. A large number of small crystals at 0 • C can simultaneously work as precursors for the formation of new ones.  Figure 5a-c provide the SEM images of ACH crystals obtained at different temperatures. It can be seen that the temperature greatly affects the ACH crystal size, which can be associated with a change in the supersaturation and density of the liquor caused by an increasing temperature [22,34,37]. An increase of the density, apparently, leads to a decrease in the supersaturation of the liquor at T > 50 °C and results in a lower degree of the ACH yield. With high supersaturation, the rate of nucleation often exceeds the crystal growth rate [38,39], which accounts for the formation of smaller crystals at 0 °C ( Figure 5a). However, at the surface of large crystals (>100 µm), the offset of secondary nucleation and crystallization can be observed (see Figure 5b, where yellow arrows indicate secondary crystals). This is most clearly seen in the sample obtained at 95 °C, where the sizes Figure 5a-c provide the SEM images of ACH crystals obtained at different temperatures. It can be seen that the temperature greatly affects the ACH crystal size, which can be associated with a change in the supersaturation and density of the liquor caused by an increasing temperature [22,34,37]. An increase of the density, apparently, leads to a decrease in the supersaturation of the liquor at T > 50 • C and results in a lower degree of the ACH yield. With high supersaturation, the rate of nucleation often exceeds the crystal growth rate [38,39], which accounts for the formation of smaller crystals at 0 • C ( Figure 5a). However, at the surface of large crystals (>100 µm), the offset of secondary nucleation and crystallization can be observed (see Figure 5b, where yellow arrows indicate secondary crystals). This is most clearly seen in the sample obtained at 95 • C, where the sizes of the secondary crystals exceed 100 µm (Figure 5c). The secondary nucleation of gibbsite was also studied by Li et al. [40]. The chemical composition of the ACH crystals obtained at 0-95 °C is presented in Table 2. It can be seen that the total content of impurities in the crystals stays within a range of 0.05-0.1 wt.%. It can also be concluded from the results presented in Table 2 that a decrease of temperature of the aluminum chloride solution reduces the solubility of K, Na, and Si impurities, which was also confirmed by Gao et al. for CaCl2 and KCl [41].
Despite the fact that a purer ACH can be obtained at 95 °C , the results of the ACH yield and crystal size advocate the use of a low temperature (e.g., 0 °C ) as optimal for the salting-out process.  The chemical composition of the ACH crystals obtained at 0-95 • C is presented in Table 2. It can be seen that the total content of impurities in the crystals stays within a range of 0.05-0.1 wt.%. It can also be concluded from the results presented in Table 2 that a decrease of temperature of the aluminum chloride solution reduces the solubility of K, Na, and Si impurities, which was also confirmed by Gao et al. for CaCl 2 and KCl [41]. Despite the fact that a purer ACH can be obtained at 95 • C, the results of the ACH yield and crystal size advocate the use of a low temperature (e.g., 0 • C) as optimal for the salting-out process.

ACH Calcination for SGA Production
To study ACH decomposition by calcination, a large number of ACH crystals were produced in an additional experiment using 200 mL of the liquor, T = 0 • C, and a duration of 1 h. Figure 6a shows the phase composition of the crystals obtained. It can be seen that the impurities did not form individual phases and were incorporated into the crystalline phase. Figure 6b shows the ACH crystal size distribution. The average ACH crystal size was 156.57 µm. Increasing the salting-out time leads to a significant increase in the crystal size. It can be seen from Figure 6c-d that surfaces of large particles >200 µm became nucleation sites for smaller ones. Some crystals were greater than 800 µm in size (Figure 6d).

ACH Calcination for SGA Production
To study ACH decomposition by calcination, a large number of ACH crystals were produced in an additional experiment using 200 mL of the liquor, T = 0 °C , and a duration of 1 h. Figure 6a shows the phase composition of the crystals obtained. It can be seen that the impurities did not form individual phases and were incorporated into the crystalline phase. Figure 6b shows the ACH crystal size distribution. The average ACH crystal size was 156.57 µ m. Increasing the salting-out time leads to a significant increase in the crystal size. It can be seen from Figure 6c-d that surfaces of large particles >200 µm became nucleation sites for smaller ones. Some crystals were greater than 800 µm in size (Figure 6d). The TGA/DSC analysis was carried out in air atmosphere to determine the mass loss and main thermal effects present upon the heating of ACH. There are two endothermic peaks and one exothermic peak shown in Figure 7. The first peak (before 100 °C) indicates the evaporation of physical water with a loss of ~1 wt.%. The second peak at 180 °C indicates ACH decomposition, which is accompanied by a loss of more than 78 wt.% from the total mass. Ivanov et al. [42] showed that this peak corresponds to the decomposition of ACH to Al2O3, H2O, and HCl by the following reaction: The TGA/DSC analysis was carried out in air atmosphere to determine the mass loss and main thermal effects present upon the heating of ACH. There are two endothermic peaks and one exothermic peak shown in Figure 7. The first peak (before 100 • C) indicates the evaporation of physical water with a loss of~1 wt.%. The second peak at 180 • C indicates ACH decomposition, which is accompanied by a loss of more than 78 wt.% from the total mass. Ivanov et al. [42] showed that this peak corresponds to the decomposition of ACH to Al 2 O 3 , H 2 O, and HCl by the following reaction: A small exothermic peak at 850 • C indicates the phase transformation of amorphous alumina into γ-Al 2 O 3 ; a small peak at 950 • C can indicate the transformation of γ-Al 2 O 3 into α-Al 2 O 3 [43]. The process of ACH decomposition could only be completed above 250 • C, when the ACH phase became undetectable by XRD, while the Al-product could be transformed into amorphous alumina until the temperature reached~850 • C, when the γ-Al 2 O 3 phase appeared.  The direct production of sandy grade alumina (SGA-C) from ACH crystals was carried out by calcination at 1000 °C for 1 h. Using these technological parameters, it is possible to obtain SGA with the smallest chlorine content (less of 0.2 wt.%) [44] and a high specific surface (>60 m 2 /g). A further increase of the temperature will result in a high content of α-Al2O3 and a decrease in the specific surface to less than 60 m 2 /g [45].

ACH Calcination to Produce Amorphous ACH400
Further study will focus on producing amorphous ACH400 and obtaining sandy grade alumina from it. The idea is to remove chlorine and water from ACH, but to avoid the crystallization of alumina, which can occur at temperatures higher than 800 °C. It can be seen from Section 3.2 that the decomposition of ACH crystals into aluminum oxide via reaction (3) starts at temperatures of around 100-150 °C. However, the decomposition of ACH usually occurs in two steps: first, it decomposes into an aluminum oxychloride compound at temperatures up to 350 °C , which can be presented by the following reaction: 2AlCl3·6H2O(s) = Al2(OH)4Cl2·H 2O(s) + 4HCl(g)↑ + 7H2O(g)↑, and second, the aluminum oxychloride decomposes into Al2O3 at higher temperatures. However, if the heating rate is high enough, reactions (3) and (4) occur almost simultaneously (see, for example, [42]), and this results in one broad peak from 100 to 300 °C, as can be seen in Figure 7. In the present study, it was found that in order to maximize the removal of chlorine ions and thereby ACH400 formation, it is necessary to anneal ACH at temperatures above ~400 °C. Taking into account the above, ACH400 was obtained by the calcination of ACH crystals at T = 400 °C for 1 h. This was further used to obtain sandy grade alumina (SGA-A) by a novel alkali The direct production of sandy grade alumina (SGA-C) from ACH crystals was carried out by calcination at 1000 • C for 1 h. Using these technological parameters, it is possible to obtain SGA with the smallest chlorine content (less of 0.2 wt.%) [44] and a high specific surface (>60 m 2 /g). A further increase of the temperature will result in a high content of α-Al 2 O 3 and a decrease in the specific surface to less than 60 m 2 /g [45].

ACH Calcination to Produce Amorphous ACH400
Further study will focus on producing amorphous ACH400 and obtaining sandy grade alumina from it. The idea is to remove chlorine and water from ACH, but to avoid the crystallization of alumina, which can occur at temperatures higher than 800 • C. It can be seen from Section 3.2 that the decomposition of ACH crystals into aluminum oxide via reaction (3) starts at temperatures of around 100-150 • C. However, the decomposition of ACH usually occurs in two steps: first, it decomposes into an aluminum oxychloride compound at temperatures up to 350 • C, which can be presented by the following reaction: and second, the aluminum oxychloride decomposes into Al 2 O 3 at higher temperatures. However, if the heating rate is high enough, reactions (3) and (4) occur almost simultaneously (see, for example, [42]), and this results in one broad peak from 100 to 300 • C, as can be seen in Figure 7. In the present study, it was found that in order to maximize the removal of chlorine ions and thereby ACH400 formation, it is necessary to anneal ACH at temperatures above~400 • C. Taking into account the above, ACH400 was obtained by the calcination of ACH crystals at T = 400 • C for 1 h. This was further used to obtain sandy grade alumina (SGA-A) by a novel alkali treatment method described in Section 3.5.

Comparison of SGA-C and ACH400
This section provides a comparison of the sandy grade alumina (SGA-C) produced by high-temperature calcination and amorphous alumina (ACH400) produced by the low-temperature calcination of ACH crystals. The samples of Al 2 O 3 obtained have different phase compositions, particle sizes, and forms ( Figure 8). Figure 8a presents the results of XRD analysis of powders. It can be seen that the sample calcined at 400 • C is fully amorphous, while the sample calcined at 1000 • C contains crystalline phases, which are stable and metastable polymorphs of Al 2 O 3 . The main phase is found to be metastable γ-Al 2 O 3 , while the most stable form, corundum (α-Al 2 O 3 ), does not exceed 10 wt.%. After low-temperature ACH calcination with the subsequent removal of H 2 O and HCl, the particle size of the initial ACH crystals increases. As a result, ACH400 has an average particle size of 95.19 µm, while that of SGA-C calcined at 1000 • C is 43.67 µm (Figure 8b). Amorphous alumina powder retains an initial hexagonal shape of the particles similar to that of crystalline SGA-C (Figure 8c,d). The ACH400 particle surface partly remains smooth and is partly covered with long pores that are~5 µm in length, which were formed when water and HCl were evaporated (Figure 8e). The SGA-C powder becomes brittle, and the entire surface of the SGA-C particles contains micro-and nanopores (Figure 8f).
The BET analysis of samples showed that the flat surface of the initial ACH crystals provides a low specific surface area of about 0.7 m 2 /g. During ACH decomposition, pores appeared on the particle surface, which led to an increase of the specific surface area for ACH400 to 36.9 m 2 /g. Further calcination at 1000 • C resulted in the entire surface of the particles consisting of pores, and the specific surface area for SGA-C increased to 71.2 m 2 /g. Due to such a high specific surface area of the SGA-C obtained by the acid treatment, its dissolution rate in cryolite (Na 3 AlF 6 ) was faster than the SGA obtained by the Bayer alkaline method [46].
The chemical composition of the initial ACH, ACH400, and SGA-C samples is shown in Table 3. It can be seen that the amount of impurities increases and the chlorine content decreases significantly when increasing the calcination temperature. (e) (f) Figure 8. Data of ACH400 (T = 400 °C; τ = 1 h) and SGA-C (T = 1000 °C; τ = 1 h) obtained from ACH calcination: (a) XRD analysis; (b) distribution of the particle size; (c) and (d) SEM images of the particle shape for ACH400 and SGA-C powders, respectively; (e) and (f) SEM images of the particle surface for ACH400 and SGA-C powders, respectively.
The BET analysis of samples showed that the flat surface of the initial ACH crystals provides a low specific surface area of about 0.7 m 2 /g. During ACH decomposition, pores appeared on the particle surface, which led to an increase of the specific surface area for ACH400 to 36.9 m 2 /g. Further Figure 8. Data of ACH400 (T = 400 • C; τ = 1 h) and SGA-C (T = 1000 • C; τ = 1 h) obtained from ACH calcination: (a) XRD analysis; (b) distribution of the particle size; (c) and (d) SEM images of the particle shape for ACH400 and SGA-C powders, respectively; (e) and (f) SEM images of the particle surface for ACH400 and SGA-C powders, respectively.

SGA Production from Amorphous Alumina
The SGA-C obtained by the direct acid treatment has a high specific surface area; however, the particle shape and the content of impurities do not meet the technical standards accepted at RUSAL plants [44]. It is possible to reduce the impurities content, if recrystallization or solvent extraction methods are utilized [47].
Another approach involves alkaline treatment of the Al 2 O 3 by the Bayer process [48]. In this case, Al 2 O 3 from the acid treatment stage is first leached in an alkaline aluminate solution to obtain a liquor supersaturated by alumina. For this process, high-pressure leaching equipment should be used. It is also necessary to obtain the alkaline aluminate liquor with a caustic module (the molar ratio of Na 2 O to Al 2 O 3 in the solution) below 1.5 [49]. Then, the supersaturated alkaline aluminate liquor can enter a precipitation stage, where gibbsite (Al(OH) 3 ) precipitates out in the presence of a large number of seed crystals (to achieve that, the ratio of Al 2 O 3 in the liquor to that in the seed crystals must be greater than 4). To obtain coarse particles of Al(OH) 3 (only 10% of particles should have a size below 45 µm) and a precipitation rate of more than 50% required for SGA production, the duration of the whole process should be 48-72 h [50]. Therefore, in order to produce SGA that suits the required standards, the use of high-pressure equipment and large decomposers, as well as additional energy and raw materials, is required, which significantly increases the cost of the final product.
It is also possible to apply the above alkaline recrystallization process at lower material/energy costs using an enlargement method [51] based on the difference in solubility of coarse particles of gibbsite or boehmite (AlOOH) and amorphous alumina (or highly dispersed hydroxide) [52,53]. Previous studies on the leaching of amorphous alumina in hot alkaline liquor have shown that, in this case, the process can occur at atmospheric pressure [54,55] and no high-pressure leaching equipment is necessary.
In this paper, we have proposed, for the first time, the use of the intermediate product of the acid treatment stage, the amorphous ACH400, in the alkaline recrystallization process. In the present research, the amorphous alumina obtained after the acid treatment stage was subjected to further alkaline treatment and subsequent recrystallization by the mother liquor and coarse Al(OH) 3 particles, respectively, formed during a stage of the Bayer process on the Ural Aluminum Plant (UAP). As a result, the amorphous alumina dissolved, increasing the supersaturation of the aluminate solution, which in turn led to the precipitation of aluminum hydroxide on the coarse Al(OH) 3 particles. Finally, the sandy grade alumina could be obtained after filtration and calcination. The proposed process flowsheet is shown in Figure 9. The results of the proposed treatment are discussed in the rest of the section. The particle size distribution of the initial Al(OH)3 particles produced on the UAP and of Al(OH)3 particles after the recrystallization stage is shown in Figure 10. The initial powder contains about 25% of a fraction of fewer than 50 µ m, which prevents the production of SGA from this powder on the UAP. To increase the size of the coarse Al(OH)3 particles, ACH400 powder was used. When mixed with an alkaline liquor, due to the high specific surface area (36.9 m 2 /g for ACH400 and 0.1 m 2 /g for coarse Al(OH)3), The results of the proposed treatment are discussed in the rest of the section. The particle size distribution of the initial Al(OH) 3 particles produced on the UAP and of Al(OH) 3 particles after the recrystallization stage is shown in Figure 10. The initial powder contains about 25% of a fraction of fewer than 50 µm, which prevents the production of SGA from this powder on the UAP. To increase the size of the coarse Al(OH) 3 particles, ACH400 powder was used. When mixed with an alkaline liquor, due to the high specific surface area (36.9 m 2 /g for ACH400 and 0.1 m 2 /g for coarse Al(OH) 3 ), it will be dissolved first, and then precipitated on the Al(OH) 3 surface. This process can be described by the following reactions: NaAl(OH) 4(l) + Al(OH) 3(s) = 2Al(OH) 3(s) + NaOH (l) .
According to reaction (5), amorphous alumina is dissolved in the liquor to form hydrated sodium aluminate, thereby increasing the supersaturation of the liquor relative to the coarse Al(OH) 3 particles [52]. Next, dissolved alumina in the form of hydrated sodium aluminate precipitates on the surface of Al(OH) 3 via reaction (6). Li et al. showed that Al(OH) 3 agglomeration also proceeds during the late stages of the precipitation process [56]. It can be seen ( Figure 10) that the particle size distribution after recrystallization shifted towards a greater average size and a higher volume fraction of coarse particles.   Figure 11c); a large number of small particles were not found, as was the case before ( Figure 11a). An analysis of the particle shape of the coarse Al(OH)3 revealed voids and surface inhomogeneity (Figure 11b), whereas, after recrystallization, the Al(OH)3 particles had an almost uniform structure of a dense pack of small particles, consisting of larger agglomerates (Figure 11d).   (Figure 11a). The Al(OH) 3 sample after recrystallization mainly consisted of large particles with a size of 100-125 µm ( Figure 11c); a large number of small particles were not found, as was the case before ( Figure 11a). An analysis of the particle shape of the coarse Al(OH) 3 revealed voids and surface inhomogeneity (Figure 11b), whereas, after recrystallization, the Al(OH) 3 particles had an almost uniform structure of a dense pack of small particles, consisting of larger agglomerates (Figure 11d). The proposed process proceeds at atmospheric pressure and the liquor temperature of 60 °C . The Al2O3 precipitation degree after 36 h was only 4.5%. The XRD analysis shows that the powder obtained only consists of the gibbsite phase ( Figure 12). The low Al2O3 precipitation degree and the absence of amorphous alumina in the recrystallized sample ( Figure 11c) confirm that the growth of Al(OH)3 occurs due to the leaching and recrystallization of ACH400 powder. After filtration, the mother liquor was reused to leach a new portion of amorphous alumina. For SGA production, it is necessary to calcinate Al(OH)3 at T = 1000 °C and τ = 1 h. The removal of external moisture and crystalline water in Al(OH)3 occurs according to the following reaction: It can be seen from Figure 10 that the number of particles smaller than 45 µ m becomes lower than 5%. Since calcination in this research was carried out in a muffle furnace, the attrition index (AI) was not studied. However, Wind et al. showed that decreasing the fraction of particles greater than 45 µm increases the AI in the calcination furnace [57]. Therefore, if the Al(OH)3 after recrystallization contains 5% of particles smaller than 45 µm, the AI is 5%. This means that, even after abrasion, the obtained SGA-A must comply with the SGA requirements.
The XRD pattern of the SGA-C obtained after Al(OH)3 calcination is shown in Figure 12. SGA-C consists of two phases: 95% γ-Al2O3 and 5% α-Al2O3. The proposed process proceeds at atmospheric pressure and the liquor temperature of 60 • C. The Al 2 O 3 precipitation degree after 36 h was only 4.5%. The XRD analysis shows that the powder obtained only consists of the gibbsite phase ( Figure 12). The low Al 2 O 3 precipitation degree and the absence of amorphous alumina in the recrystallized sample ( Figure 11c) confirm that the growth of Al(OH) 3 occurs due to the leaching and recrystallization of ACH400 powder. After filtration, the mother liquor was reused to leach a new portion of amorphous alumina. For SGA production, it is necessary to calcinate Al(OH) 3 at T = 1000 • C and τ = 1 h. The removal of external moisture and crystalline water in Al(OH) 3 occurs according to the following reaction: It can be seen from Figure 10 that the number of particles smaller than 45 µm becomes lower than 5%. Since calcination in this research was carried out in a muffle furnace, the attrition index (AI) was not studied. However, Wind et al. showed that decreasing the fraction of particles greater than 45 µm increases the AI in the calcination furnace [57]. Therefore, if the Al(OH) 3 after recrystallization contains 5% of particles smaller than 45 µm, the AI is 5%. This means that, even after abrasion, the obtained SGA-A must comply with the SGA requirements.
The XRD pattern of the SGA-C obtained after Al(OH) 3 calcination is shown in Figure 12. SGA-C consists of two phases: 95% γ-Al 2 O 3 and 5% α-Al 2 O 3 . Metals 2020, 10, x FOR PEER REVIEW 16 of 19 The chemical composition of SGA-C is presented in Table 4. Since ACH400 was completely dissolved in the alkali liquor, the maximum dilution of impurities also occurred. Therefore, the chlorine content in the SGA-A was quite low. Senyuta et al. [48] showed that, in a real metallurgical process, chlorine can be accumulated in circulating liquors at the amount of 40-90 g/L, which may lead to an increase of the chlorine content in the SGA to about 0.013 wt.%. Using the proposed novel process, it is possible to reduce the cost of SGA production, which is necessary for modern electrolysis processes. Acid/alkali treatment enables the leaching stage of the amorphous alumina at high pressure to be eliminated; the long precipitation time to be significantly shortened; and the sandy grade alumina, which is appropriate with regards to the SGA standards, to be produced.

Conclusions
In this work, a novel method has been developed for sandy grade alumina production from the Al-chloride liquor obtained by the HCl leaching of coal fly ash. In this method, the amorphous alumina obtained after traditional acid treatment of the Al liquor is subjected to alkaline treatment by the mother liquor from a Ural Aluminum Plant and to subsequent precipitation using the coarse Al(OH)3 particles from UAP, with further calcination to obtain SGA suitable for metallic aluminum production. The following points concerning the technological parameters of various stages of the novel process can be made: 1. The temperature of the salting-out process affects the ACH crystal size and impurities content. At T = 0 and 95 °C , the total amount of impurities is 0.1 and 0.06 wt.%, and the average particle size is 85 and 135 µm, respectively; 2. ACH calcination at T = 400 °C produces amorphous alumina with the specific surface of 36.9 m 2 /g and content of chlorine ions of 5.27 wt.%; 3. ACH calcination at T = 1000 °C allows SGA production with the specific surface area of 71.2 m 2 /g and chlorine ion content of 0.15 wt.%; 4. Alkaline recrystallization of amorphous alumina obtained after the calcination of ACH at 400 °C at T = 60 °C and Na2O = 150 g/L, allows the average particle size of coarse Al(OH)3 to be increased from 76 to 99 microns; The chemical composition of SGA-C is presented in Table 4. Since ACH400 was completely dissolved in the alkali liquor, the maximum dilution of impurities also occurred. Therefore, the chlorine content in the SGA-A was quite low. Senyuta et al. [48] showed that, in a real metallurgical process, chlorine can be accumulated in circulating liquors at the amount of 40-90 g/L, which may lead to an increase of the chlorine content in the SGA to about 0.013 wt.%. Using the proposed novel process, it is possible to reduce the cost of SGA production, which is necessary for modern electrolysis processes. Acid/alkali treatment enables the leaching stage of the amorphous alumina at high pressure to be eliminated; the long precipitation time to be significantly shortened; and the sandy grade alumina, which is appropriate with regards to the SGA standards, to be produced.

Conclusions
In this work, a novel method has been developed for sandy grade alumina production from the Al-chloride liquor obtained by the HCl leaching of coal fly ash. In this method, the amorphous alumina obtained after traditional acid treatment of the Al liquor is subjected to alkaline treatment by the mother liquor from a Ural Aluminum Plant and to subsequent precipitation using the coarse Al(OH) 3 particles from UAP, with further calcination to obtain SGA suitable for metallic aluminum production. The following points concerning the technological parameters of various stages of the novel process can be made: 1.
The temperature of the salting-out process affects the ACH crystal size and impurities content. At T = 0 and 95 • C, the total amount of impurities is 0.1 and 0.06 wt.%, and the average particle size is 85 and 135 µm, respectively; 2.
ACH calcination at T = 400 • C produces amorphous alumina with the specific surface of 36.9 m 2 /g and content of chlorine ions of 5.27 wt.%; 3.
ACH calcination at T = 1000 • C allows SGA production with the specific surface area of 71.2 m 2 /g and chlorine ion content of 0.15 wt.%;

4.
Alkaline recrystallization of amorphous alumina obtained after the calcination of ACH at 400 • C at T = 60 • C and Na 2 O = 150 g/L, allows the average particle size of coarse Al(OH) 3 to be increased from 76 to 99 microns; 5.
The calcination of recrystallized Al(OH) 3 at T = 1000 • C and maintenance for 1 h allows SGA with a content of chlorine ions of 0.004 wt.% to be obtained.