Extraction of scandium from bauxite residue by high-pressure leaching in sulfuric acid solution

Abstract

Rare earth elements, such as yttrium, scandium, neodymium, and praseodymium, have been reported to be associated with minerals in bauxite and transferred to the refining residue when bauxite is refined to alumina (Al₂O₃) by Bayer Process. In terms of price, scandium is the most valuable rare-earth element in bauxite residue. This research discusses the effectiveness of scandium extraction from the bauxite residue through pressure leaching in sulfuric acid solution. The method was selected to obtain high scandium recovery and leaching selectivity to iron and aluminium. Series of leaching experiments were conducted under variations of H₂SO₄ concentration (0.5–1.5 M), leaching duration (1–4 h), leaching temperature (200–240 °C), and slurry density (10–30% (w/w)). Taguchi method with L9:34 orthogonal array was adopted to design the experiments. Analysis of Variance (ANOVA) was performed to determine the most influential variables of the extracted scandium. The experimental result and statistical analysis revealed that the optimum condition for scandium extraction was at 1.5 M H₂SO₄, a leaching duration of 1 h, a temperature of 200 °C, and a slurry density of 30% (w/w). The leaching experiment carried out at this optimum condition resulted in scandium extraction of 90.97% and co-extracted iron and aluminium of 32.44% and 75.23%, respectively. Analysis of Variance showed that solid/liquid ratio was the most influential variable with a contribution of 62%, followed by acid concentration (21.2%), temperature (16.4%), and leaching duration (0.3%).

1. Introduction

The refining process of bauxite into alumina using the Bayer Process produces a processing residue known as red mud because of its color. The bauxite residue contains iron, aluminium, titanium, silicon, calcium, and sodium compounds that are not dissolved during digestion using a hot sodium hydroxide solution. Typically, 1–2 tons of red mud are generated per ton of alumina produced. It was reported that approximately 4 billion tons of red mud were stockpiled worldwide in 2010, with an annual growth rate of around 160 million tons [1]. Stockpiling of materials with such a large volume poses environmental risks, especially in areas with high rainfall and seismic activity. On the other hand, it was reported that red mud contains valuable metals such as gallium and rare earth elements (REE), with a total REE of about 0.1% [2]. The REEs found in red mud are cerium (Ce), scandium (Sc), lanthanum (La), neodymium (nd), yttrium (Y), and several others with lower concentrations [2]. Scandium is the most valuable element of REE in terms of price, representing ±95% of the REE value in the bauxite residue [3]. In China, ores with scandium content in the range of 20–50 mg/kg are considered as a resource that is suitable for exploitation [4]. Because of its abundance, bauxite residue is a promising source of scandium and other valuable elements, such as gallium. To reduce the bauxite residue dumping problem, extracting more voluminous components in the bauxite residue, such as iron, must also be carried out. If feasible, then utilize the remaining mass as a construction material.

Scandium is needed in strategic applications such as fabricating high-strength aluminium alloys used in fighter aircraft and preparing solid electrolytes used in solid oxide fuel cells (SOFC). In addition, the demand for scandium is increasing for applications in laser glass, catalysis, electronic components, nuclear technology, alloys, high-intensity metal halide lamps, and solid oxide fuel cells [5].

Several previous researchers have investigated the extraction processes of scandium from the bauxite residue [[3], [4], [5], [6], [7], [8], [9], [10]]. The selected process routes include the hydrometallurgical and combined pyro-hydrometallurgical processes (where scandium is enriched first in the smelting slag or sulfation-roasting of the bauxite residue, followed by water leaching [6,7]). Leaching reagents that have been studied include sulfuric acid, hydrochloric acid, nitric acid, and organic acids such as methane sulfonic acid, citric acid, and acetic acid [3,8]. Leaching efficiency depends on the bauxite residue composition and process parameters, such as reagent type and concentration, retention time, temperature, and solid/liquid ratio [2,3,5,6,9,10].

Ochsenkühn-Petropulu et al. [9] showed that 29–96% of REEs in bauxite residue could be extracted by direct leaching using 0.5 M nitric acid, at the liquid to solid ratio of 50, the retention time of 24h, and at room temperature. Under this condition, a scandium extraction of 83.8% was obtained, with co-extracted iron of 2.2%. Meanwhile, Borra et al. [3] have examined the effects of various acid leaching agents, acid concentration, liquid to solid ratio, retention time, and process temperature on the extracted REE from the bauxite residue. The experimental results showed that the highest level of REEs extraction was obtained when HCl was used as the leaching agent. The extracted Sc of 70–80% was achieved by leaching experiment using 6 N HCl at a liquid to solid ratio of 50:1 for 24h at a temperature of 25 °C. Extraction of the REEs increased with acid concentration, leaching duration, and liquid to solid ratio. Although it can produce a high extraction value of REEs, using HCl as a leaching reagent resulted in the high co-dissolution of iron (∼60%). The low selectivity of REE leaching to iron using HCl solution would complicate the downstream process of the pregnant leach solution.

Using acidic reagents to treat bauxite residue produces a large volume of acid wastewater, requiring intensive neutralization. The bauxite residue leaching process to extract REEs can also be performed using alkaline reagents [10]. Yatsenko and Pyagai [10] conducted a leaching experiment of bauxite residue in an alkaline solution using NaHCO₃ as a leaching reagent to extract scandium. Scandium leaching occurs by forming complex anionic compounds with carbonate ions. Alkaline leaching was performed at 50 °C for 2h with a solid to liquid ratio of 1:2.5. The experimental results showed that the solubility of scandium oxide was 16.7 g/l in the initial NaHCO₃ concentration of 100 g/l. The dissolved scandium was precipitated by adding sodium aluminate at 80 °C for 2h. The precipitate was rinsed with NaOH (10–15%) while heated until the solution boiled. The remaining precipitate was rinsed with NaOH solution (1–5%). Before being filtered, the precipitate from the previous stage was dissolved in HCl (1–5%). Finally, ammonia (10–25%) was added to the filtrate to precipitate scandium [10].

One of the critical issues in extracting scandium from the bauxite residue by acid leaching is the selectivity of the leaching to iron. To overcome the problem of iron co-dissolution in the leaching of scandium in an acidic solution, Zhou et al. add EDTA (ethylene diamine tetra acetic acid) during scandium leaching of the bauxite residue in hydrochloric acid [5]. EDTA is a chelating agent that can form complex bonds with certain metals, such as iron. Zhou et al. [5] conducted an experiment on bauxite residue leaching in HCl solution added with EDTA with the liquid to solid ratio of 4:1 using 10 g of the bauxite residue sample. The bauxite residue, a certain amount of EDTA, and concentrated HCl were mixed in the triangular beaker and agitated by a magnetic stirrer at various temperatures. The optimum parameters obtained were at leaching reagent: bauxite residue: EDTA ratios of 40 ml: 10 g: 2 g; at a temperature of 70 °C, and retention time of 4h. Under the optimum condition, the extracted scandium and iron were 79.6% and 6.12%, respectively. This method reduced the acid consumption and enhanced the Sc/Fe mass ratio in the pregnant leach solution, making the downstream process of separating Sc from Fe easier.

Another method to reduce the dissolved iron in the scandium leaching process from the bauxite residue is combining the smelting of the bauxite residue and the leaching of the scandium-bearing slag. The iron oxide in the bauxite residue will be reduced to iron metal in the smelting stage, while scandium is concentrated in the slag. Borra et al. [3] showed that more than 95% of the iron in the bauxite residue could be separated at the smelting stage. The bauxite residue sample was mixed with graphite and flux to make a pellet and then smelted in a furnace at 1500–1600 °C for 1h, while argon gas constantly flowed to the reactor to produce a non-oxidative atmosphere. After the slag was separated from the molten iron, the slag was crushed and ground to produce a particle size of <80 μm before passing through a magnetic separator to remove the remaining iron particles. The remaining slag was then leached in a solution of HCl, HNO₃, dan H₂SO₄ with a liquid to solid ratio of 50/1, respectively. The acid concentration was varied from 0.25 to 6 N. The low-temperature leaching was carried out at 25 °C for 24h, while the high temperature leaching was done at 90 °C for 1h. Low-temperature of the slag leaching gave low extraction yields, whereas high-temperature leaching improved the extraction yields. The scandium and most other REEs could be well extracted at 90 °C in 3 N acid concentration. At the conditions that give the highest scandium extraction, the concentrations of scandium and iron in the pregnant leach solution were 3 and 250 mg/l, respectively. The advantage of slag leaching compared to direct leaching of the bauxite residue is that most REEs can be extracted with a minimum co-dissolution of iron. However, this method requires high energy and flux consumption in the smelting stage due to the high alumina content in the bauxite residue.

Although several leaching processes can partially or entirely recover REEs from the bauxite residue, these procedures are typically accompanied by complications, notably due to the co-dissolution of iron and silicon, which makes subsequent processing problematic. The co-dissolution of iron is undesirable since it is difficult to separate it from the REEs, especially scandium, due to their comparable physicochemical properties, necessitating a considerable amount of reagents during downstream processing. Scandium(III) ions are present in the iron(III)-rich oxide lattice, which prevents its complete dissolution [11,12]. In order to enhance the extraction and leaching selectivity of REEs, iron must be previously extracted or dissolved and then re-precipitated. Meanwhile, a high silicon concentration in the leach solution might result in the formation of silica gel, which considerably reduces the leaching efficiency because the gel solution can no longer be filtered. As they are partially adsorbed on the surface of bauxite residue minerals, it is hypothesized that these aggregates could be a barrier to the acid leaching of the bauxite residue [13]. Under acidic conditions, below the isoelectric point for silica in the solution (i.e., pH between 1.7 and 2.2) [14], silica hydrolysis occurs rapidly to produce silica monomers, which tend to form cyclic oligomers through Ostwald ripening until a gel network is formed [[15], [16], [17]]. During typical acid leaching of bauxite residue at room temperature, it has been observed that a substantial decomposition of silicate compounds occurs, which favors silicon dissolution. As a result, the leach solution cannot be filtered due to the polymerization of silica monomers.

This study investigated the extraction of scandium from bauxite residue obtained from an alumina refinery in West Kalimantan, Indonesia, through High-Pressure Acid Leaching (HPAL) in a sulfuric acid solution. HPAL process offers a high extraction level of scandium and a quick process. In addition, it can reduce the dissolution of silica, which can lead to the formation of silica gel during the leaching process. Besides, the solubility of main impurities, such as iron, aluminium, and titanium, can also be reduced due to the hydrolysis of their ions at high temperatures (200–240 °C). The bauxite residue characteristics and the effects of leaching variables, namely acid concentration, temperature, leaching duration, and slurry density on the extracted scandium were experimentally investigated. The experiment was designed by the Taguchi Method with an orthogonal array under various variables to determine the optimum condition and degree of significance of the experimental parameters to influence the experiment's output, namely, scandium extraction and the leaching selectivity to iron.

2. Materials dan methods

2.1. Sample preparation and Characterisation

As aforementioned, the bauxite residue sample used in this study was received from an alumina refinery in the West Kalimantan Province of Indonesia. The refinery treats gibbsite type of washed bauxite that typically contains 41–50% Al₂O₃, 20–29% SiO₂, 9–18% Fe₂O₃, 1–2% TiO₂, and several other constituents with lower concentrations. The homogenized bauxite residue sample was dried in an electric oven at 110 °C for 24h, ground in a ball mill, and then dry-sieved in a sieve shaker through a 200 mesh sieve before being used in the leaching experiments. Measurements of the moisture content showed that the bauxite residue sample contained 25.1% moisture. The chemical composition of the bauxite residue sample for the major components determined by X-Ray Fluorescence (XRF) is shown in Table 1. The main components of the bauxite residue sample are Fe, Si, Al, Na, Ti, and Ca, with Fe₂O₃, SiO₂, and Al₂O₃ contents of 30.9%, 27%, and 24.2%, respectively. Scandium content was analyzed by Inductively Coupled Plasma Mass Spectrometry (ICP-MS), and the head content of the Sc was 63 ppm. The chemical composition analysis result with XRF agreed with the X-Ray Diffraction (XRD) analysis (Fig. 1) which identified hematite (Fe₂O₃), sodalite ((Na₈(Al₆Si₆O₂₄)Cl₂), and quartz (SiO₂) as the most dominant minerals.

Table 1.

Chemical composition of the bauxite residue sample used in the investigation.

Component	Content (%)
Fe2O3	30.9
SiO2	27.0
Al2O3	24.2
Na2O	11.2
TiO2	2.55
CaO	1.63
Cl	1.03
SO3	0.563
P2O5	0.249
Rh2O3	0.212
K2O	0.204
MnO	0.0775
Cr2O3	0.0449
ZrO2	0.037
MgO	0.0348
PbO	0.0277
CuO	0.0204
Sc	0.0063

Fig. 1.

XRD difractogram of the bauxite residue sample.

2.2. Leaching experiment

A series of high-pressure leaching experiments were done in an autoclave using a sulfuric acid solution to selectively dissolve scandium from the bauxite residue. The leaching experimental design was made using the Taguchi Method. There were 4 (four) variables varied in the leaching experiments, namely concentration of sulfuric acid (H₂SO₄), temperature, duration of leaching, and solid to liquid ratio (S/L). The experimental design with the Taguchi method used was the Orthogonal Array from L9: 3⁴, where nine experiments were performed with four variables varied at three different levels of variation. The experimental variables and their levels of variation based on the Taguchi L9: 3⁴ design are shown in Table 2. Furthermore, detailed experimental variations were arranged based on the experimental design in Table 2. The detail of the experimental variations is listed in Table 3. All experiments were carried out in duplicate to determine the reproducibility of the experimental results under the same conditions so that deviations from the experimental results could be evaluated.

Table 2.

Experimental variable and its level of variation designed by the Taguchi L9: 3⁴ orthogonal array.

Parameters	Level
	1	2	3
H2SO4 concentration (M)	0.5	1	1.5
Leaching duration (hours)	1	2	4
Temperature (°C)	200	220	240
Slurry density (wt%)	10	20	30

Table 3.

The detail of the experimental variations designed by the Taguchi L9: 3.⁴.

Exp. number	Variables and their levels of variation
	[H2SO4] (M)	T (°C)	Slurry density (wt%)	leaching duration (h)
1	0.5	200	10	1
2	0.5	220	20	2
3	0.5	240	30	4
4	1	220	30	1
5	1	240	10	2
6	1	200	20	4
7	1.5	240	20	1
8	1.5	200	30	2
9	1.5	220	10	4

Fig. 2 shows the experimental set-up of scandium leaching from the bauxite refining residue. The slurry agitation was performed using a mechanical agitator, and the temperature of the solution during the leaching was measured by a thermocouple immersed in the solution. The solution temperature was maintained by controlling the temperature of the reactor heater; meanwhile, the speed controller adjusted the agitation speed. After measuring the weight of the bauxite residue sample and the sulfuric acid solution according to the desired slurry density, the bauxite residue and the sulfuric acid were loaded in a Teflon container and put into the reactor. During pre-heating, the stirring speeds were kept constant at 50 rpm and increased to 100 rpm when the desired temperature was reached.

Fig. 2.

Schematic picture of the autoclave used in the leaching experiment.

After the leaching experiment was completed, the stirring was stopped, and the autoclave was left until the reactor temperature reached room temperature. Next, solid-liquid separation was conducted to separate the pregnant leach solution from the leaching residue using filter paper. The concentrations of scandium, iron, and aluminium in the pregnant leach solutions were analyzed by inductively coupled plasma mass spectroscopy (ICP-MS). Meanwhile, the leaching residue was cleaned from the remaining solution that was still attached by stirring with distilled water at 70 °C using a magnetic stirrer with a stirring speed of 300 rpm. After washing the leaching residue was accomplished, the residue was dried in an electric oven at 110 °C for 24 h. The dry residue was then dissolved in aqua regia solution. Analysis with ICP-MS was carried out on the solution obtained from the digestion of leaching residues to determine the concentrations of dissolved scandium, iron, and aluminum. Finally, the percentage of metal extracted was calculated from the ratio of the mass of metal dissolved in the pregnant leach solution (PLS) and the total mass of metal dissolved in PLS with the mass of metal remaining in the leaching residue as formulated by the following equation:

$$extractedmetal=\frac{m_1}{m_1+m_2}x100\%$$ (1)

in which m1 is the mass of metal dissolved in PLS and m2 is the mass of metal remaining in the leaching residue. As previously noted, the leaching objective is to maximize scandium dissolution while minimizing the co-dissolution of iron and aluminium.

3. Result and discussion

3.1. Effect of leaching variables on the extracted scandium, iron, and aluminium analyzed by ANOVA

Table 4 presents the data of extracted Sc and co-extracted Fe and Al from the leaching experiment. The results of the leaching experiments showed that the extracted scandium was in the range of 0.54%–84.91%, while the co-extracted iron and aluminum were in the range of 0.00%–33.43% and 0.00%–47.10%, respectively. From the data of the extracted metals, the Analysis of Variance (ANOVA) was carried out to determine the significance of the effect of each leaching variable on the extraction percentages of scandium, iron, and aluminium. This analysis determines the sum of squares (SS), mean square, and F distribution values. In addition, the significance of the influence of the leaching variable was obtained by comparing the value of the standard F distribution with the value of the F distribution of each variable. Furthermore, the levels of influence of each variable on the metal extraction percentages were obtained by comparing the value of the sum of the squares of each variable with the value of the total sum of squares. The standard F-distribution value was obtained from the F-distribution Table using α = 0.05, df1 = 2, and df2 = 9, which is 4.26. The ANOVA results for the extracted Sc, Fe, and Al are presented in Table 5, Table 6, Table 7, respectively. Using the standard F distribution value of 4.26, it was found that H₂SO₄ concentration, temperature, leaching duration, and slurry density significantly affected the scandium extraction and co-extraction of iron and aluminium. Based on the analysis results, slurry density was known to have the most decisive influence on the extraction percentage of scandium and co-extraction percentages of iron and aluminium, with respective contribution levels of 62.0%, 53.9%, and 47.5%. These results agree with the investigation result of Wang et al. [18], who stated that the most influential parameter in the leaching of bauxite refining residue in the acid solution is the ratio of solid to liquid. An Optimum slurry density can provide good mass transfer and contact between minerals and the leaching reagent. The slurry density can affect the saturation of metal ions in the solution, preventing further dissolution of the metals. In addition, the slurry density also affects silica gel formation, which can be an obstacle in the bauxite residue leaching using an acid solution.

Table 4.

Results of the leaching experiment.

No.	Extracted Sc (%)			Fe co-dissolution (%)			Al co-dissolution (%)
	Duplicate 1	Duplicate 2	Mean	Duplicate 1	Duplicate 2	Mean	Duplicate 1	Duplicate 2	Mean
1	32.04	35.07	33.56	2.97	3.63	3.30	1.65	1.73	1.69
2	0.95	1.03	0.99	0.01	0.03	0.02	0.00	0.04	0.02
3	0.54	0.55	0.54	0.00	0.00	0.00	0.00	0.01	0.00
4	0.58	0.59	0.58	0.14	0.24	0.19	0.37	0.39	0.38
5	81.83	84.91	83.37	20.89	20.77	20.83	17.57	21.25	19.41
6	21.99	24.49	23.24	3.40	3.35	3.37	1.69	1.52	1.61
7	54.89	55.31	55.10	9.32	9.28	9.30	2.78	1.92	2.35
8	10.17	12.87	11.52	2.56	1.60	2.08	3.49	2.66	3.07
9	61.00	60.40	60.70	29.46	33.43	31.45	47.19	41.06	44.12

Table 5.

Results of the analysis of variance for the extracted scandium.

Parameters	Degree of Freedom	Sum of squares	Mean square	F distribution	Contribution (%)
H2SO4 concentration (M)	2	3136.08	1568.04	862.72	21.18
Leaching duration (hour)	2	43.64	21.82	12.00	0.29
Temperature (°C)	2	2427.34	1213.67	667.75	16.39
Slurry density (%)	2	9184.27	4592.13	2526.55	62.02
Error	9	16.36	1.82	1.00	0.11
Total	17	14807.68			100

Table 6.

Results of the analysis of variance for the co-extracted iron.

Parameters	Degree of Freedom	Sum of squares	Mean square	F distribution	Contribution (%)
H2SO4 concentration (M)	2	521.10	260.55	273.66	26.39
Leaching duration (hour)	2	162.21	81.10	85.18	8.21
Temperature (°C)	2	218.77	109.39	114.89	11.08
Slurry density (%)	2	1064.18	532.09	558.86	53.89
Error	9	8.57	0.95	1.00	0.43
Total	17	1974.82			100

Table 7.

Results of the analysis of variance for the co-extracted aluminium.

Parameters	Degree of Freedom	Sum of squares	Mean square	F distribution	Contribution (%)
H2SO4 concentration (M)	2	770.60	385.30	132.01	21.76
Leaching duration (hour)	2	571.84	285.92	97.96	16.15
Temperature (°C)	2	491.16	245.58	84.14	13.87
Slurry density (%)	2	1681.99	840.99	288.14	47.49
Error	9	26.27	2.92	1.00	0.74
Total	17	3541.85			100

The leaching variable that has the second most significant influence on the extracted scandium is sulfuric acid concentration. The levels of contribution of the effect of the sulfuric acid concentration on the extracted scandium, iron, and aluminium are 21.2%, 26.4%, and 21.8%, respectively. The sulfuric concentration determines the leaching agent's availability to react with the minerals and the stability of the dissolved metal ions. A sufficient acid concentration is required to ensure the reaction takes place entirely and to keep it soluble. Furthermore, the leaching temperature and retention time also affected the extractions of scandium, iron, and aluminium. The effects of temperature and the leaching time are related to the leaching kinetics. Generally, the dissolution rate of metals in an aqueous solution is enhanced when the leaching temperature is higher; however, certain metals undergo a hydrolysis reaction at a high temperature which can re-precipitate the dissolved metals, such as iron and aluminium [19].

3.2. Optimum condition for scandium, iron, and aluminium extractions

A signal-to-noise (S/N) ratio parameter is used to measure the experimental variable's effect on the experiment's results by minimizing the effect of noise factors. The value of the S/N ratio was obtained from processing the experimental data for several combinations of variable levels. Each combination will have its value. This value can indicate two possibilities of the resulting output response. The first is when the variable significantly influences a process so that it can be called a "signal." Second, when these variables only have a minor effect on the resulting output response, they are considered as a "noise" [20]. In this investigation, the S/N ratios were calculated for each variable level to determine the optimum condition. As a maximum dissolution of scandium in the pregnant leach solution was desired, the calculation of S/N ratios for scandium followed the larger-the-better S/N formula. On the other hand, co-dissolutions of iron and aluminium as the main impurities in the PLS were desired as minimum as possible so that the S/N smaller-the-better formula was used. The followings are the formula for S/N larger the better and S/N smaller the better [20]:

$$\mathrm{L}\mathrm{a}\mathrm{r}\mathrm{g}\mathrm{e}\mathrm{r}-\mathrm{t}\mathrm{h}\mathrm{e}-\mathrm{b}\mathrm{e}\mathrm{t}\mathrm{t}\mathrm{e}\mathrm{r}:\mathrm{S}/\mathrm{N}=-10\log \left(\frac{1}{\mathrm{n}}\sum _{\mathrm{i}=1}^{\mathrm{n}}\frac{1}{{{\mathrm{y}}_{\mathrm{i}}}^2}\right)$$ (2)

$$\mathrm{S}\mathrm{m}\mathrm{a}\mathrm{l}\mathrm{l}\mathrm{e}\mathrm{r}-\mathrm{t}\mathrm{h}\mathrm{e}-\mathrm{b}\mathrm{e}\mathrm{t}\mathrm{t}\mathrm{e}\mathrm{r}:\mathrm{S}/\mathrm{N}=-10\log \left(\frac{1}{\mathrm{n}}\sum _{\mathrm{i}=1}^{\mathrm{n}}{y_i}^2\right)$$ (3)

where y_i are responses (the extracted metals) for a given level of experimental variable variation, and n = number of replication in each variable variation (in this experiment = 2).

The profiles of S/N ratios for Sc, Fe, and Al extraction versus leaching variables are shown in Fig. 3, Fig. 4, Fig. 5. The desired condition for maximizing Sc dissolution and minimizing co-dissolution of Fe and Al were indicated by the highest S/N value in each figure. Moreover, the line gradient's magnitude indicates the significance of the change in the variable level in affecting the extracted metals. Based on the profiles of the S/N ratio for Sc extraction, the optimum condition for maximizing Sc dissolution was found at H₂SO₄ concentration of 1.5 M, a temperature of 200 °C, a leaching duration of 1 h, and a slurry density of 10% (w/w). The optimum condition for minimizing Fe extraction was obtained at H₂SO₄ concentration of 0.5 M, a temperature of 220 °C, a leaching duration of 4 h, and a slurry density of 30% (w/w), while the optimum condition for minimizing Al extraction is at H₂SO₄ concentration of 0.5 M, temperature 240 °C, leaching duration 4 h, and slurry density 30% (w/w). Since the goal of the leaching process is to get as much scandium as possible in the pregnant solution, the condition with H₂SO₄ concentration of 1.5 M, a temperature of 200 °C, leaching duration of 1 h, and slurry density of 10% (w/w) was considered as an optimum condition for the leaching process.

Fig. 3.

Profile of S/N ratio at various levels of the experimental variables which respect to the scandium extraction percentage data.

Fig. 4.

Profile of S/N ratio at various levels of the experimental variables which respect to the iron extraction percentage data.

Fig. 5.

Profile of S/N ratio at various levels of the experimental variables which respect to the aluminium extraction percentage data.

3.3. Effect of H₂SO₄ concentration, leaching duration, and temperature on the extracted scandium, iron, and aluminium

Based on the optimum condition for the leaching experiment determined by the Taguchi method, further leaching experiments were carried out under single variations in H₂SO₄ concentration (0.5–2 M), temperature (180–240 °C), and leaching duration (30–240 min) at 10% (w/w) slurry density and stirring speed of 100 rpm. These experiments were conducted to validate the optimum condition and determine the influence of each leaching variable on the dissolutions of Sc, Fe, and Al under optimum conditions of the other variables. The pregnant slurry samples were filtered through filter paper, washed with distilled water, and dried at 110 °C for 24 h. The ICP analyses were subsequently done on the filtrates and the solution produced by the digestion of the leaching residues to determine the concentrations of Sc, Fe, and Al and the percentages of metal extractions according to Equation (1). The results of the leaching experiments under variation of a single variable are shown in Fig. 6, Fig. 7, Fig. 8. The experimental results show that the highest scandium extraction of 90.97% was obtained at an H₂SO₄ concentration of 1.5 M, a temperature of 200 °C, and a leaching duration of 60 min.

Fig. 6.

Effect of acid concentration on the extracted Sc, Fe, and Al from the bauxite residue (T: 200 °C, S/L: 30% (w/w), t: 60 min).

Fig. 7.

Effect of leaching duration on leaching on the extracted Sc, Fe, and Al from the bauxite residue ([H₂SO₄]: 1.5 M, T: 200 °C, S/L: 30% (w/w)).

Fig. 8.

Effect of leaching temperature on the extracted Sc, Fe, and Al from the bauxite residue ([H₂SO₄]: 1.5 M, S/L: 30% (w/w), t: 60 min).

The results of the experiments with the variation of sulfuric acid concentration (Fig. 6) show that the increase of acid concentration from 0.5 to 2 M consistently enhances the extracted Fe and Al during 1h of leaching. Meanwhile, the extracted Sc increased significantly from 42.22% to 90.97% as the acid concentration increased from 0.5 M to 1.5 M. A further increase of the acid concentration from 1.5 M to 2 M resulted in the decrease of the extracted Sc to 80.32%. Iron minerals in bauxite (i.e., goethite and hematite) are reported as host minerals for scandium [11]. Scandium is found in the bauxite residue, mainly in the iron-rich oxide lattice with an isomorphic structure that is not uniformly distributed [12]. The presence of scandium in the iron oxide lattice is because scandium has similar ion radii to iron; thus, scandium can substitute iron in iron oxide minerals. Moreover, scandium is expected to be present in the outer layer of iron oxide particles or attached to the iron oxide phase's surface [21]. As a result, scandium dissolution is limited in the absence of iron mineral dissolution. The decrease of the extracted Sc by the increase in acid concentration from 1.5 M to 2 M might be due to the co-precipitation of Sc with iron. At acid concentrations above a certain threshold level, iron precipitation in the forms of basic ferric sulfate or jarosite tends to occur in the high pressure acid leaching processes [22], [23, [24]. Previous mineralogical studies also indicated that Sc³⁺ substitutes Fe³⁺ in the jarosite [25].

The effect of leaching duration on the extracted Sc, Fe, and Al is shown in Fig. 7. The scandium extraction increased significantly from 77.85% to 90.97% as the leaching duration was extended from 0.5h to 1h. Further extension of the leaching duration from 1h to 4h negatively affected the extracted Sc. The extracted Sc was reduced from 90.97% to 85.66% by extending the leaching duration from 1h to 4h. The results are consistent with the S/N profile for scandium extraction presented in Fig. 3, where the highest S/N ratio was obtained at a leaching duration of 1h. The profile of the extracted scandium is comparable to that of iron, indicating a close correlation between scandium dissolution and the dissolution and precipitation behaviour of Fe. Dissolution of iron decreased after 2 h, indicating precipitation of the dissolved iron, which can simultaneously adsorb scandium in the precipitate. Dissolved iron(III) tends to be hydrolyzed at high-temperature in the acidic solution to form hematite or hydronium jarosite according to the following reactions [23,24]:

Fe₂(SO₄)_3(aq) + 3H₂O_(l) → Fe₂O_3(s) + 3H₂SO_4(l) (4)

3Fe₂(SO₄)_3(aq) + 14H₂O_(l) → 2[H₃OFe₃(SO₄)₂(OH)₆]_(s) + 5H₂SO_4(aq) (5)

Meanwhile, aluminium dissolution increased in the first hour of leaching. However, it decreased after 1h, indicating that aluminium in the bauxite residue is initially dissolved and then precipitated through the following hydrolysis reaction [26]:

3Al₂(SO₄)_3(aq) + 14H₂O_(l) → 2[H₃OAl₃(SO₄)₂(OH)₆]_(s) + 5H₂SO_4(aq) (6)

The effect of temperature on the extracted Sc, Fe, and Al is shown in Fig. 8. The extracted scandium increased from 84.14% to 90.97% when the temperature increased from 180 °C to 200 °C. Additionally, the extracted scandium decreased and remained relatively constant as the temperature increased to 220 °C and 240 °C. The decrease of Sc extraction by raising the temperature from 200 °C to 240 °C can be associated with the entrapment or adsorption of the scandium ion onto the iron precipitates. In the case of iron, the extraction percentage decreased from 40.22% at 180 °C to 30.61% at 220 °C. The hydrolysis of ferrous sulfate might cause a decrease in iron extraction to hematite at temperatures greater than 200 °C. However, the iron extraction raised to 41.45% at 240 °C, which is postulated due to the generation of more sulfuric acid by hydrolysis reaction at this temperature. The XRD analysis of the leaching residues from the leaching experiment at 240 °C revealed the formation of jarosite (Fig. 9), which was more stable at higher acidity levels. Aluminium extraction decreased as the leaching temperature increased due to the precipitation of aluminium in the form of alunite, which was confirmed by the presence of alunite in the leaching residue. The previous investigators also reported the precipitation of alunite from the sulfuric acid media at high temperatures [[27], [28], [29]].

Fig. 9.

XRD diffractograms of the leaching residues obtained from various temperatures ([H₂SO₄]: 1.5 M, S/L: 30% (w/w), t: 60 min).

The XRD diffractograms in Fig. 9 illustrate the mineral composition of the leaching residue from the leaching experiments at temperatures of 180, 200, 220, and 240 °C. Quartz, hematite, calcite, alunite, magnetite, and jarosite are the main constituents detected in the leaching residue. Quartz exhibited the highest peak intensity in the XRD diffractogram of the leaching residue resulting from the entire temperature variations, confirming that quartz is not dissolved during leaching. Alunite is the mineral detected in the leaching residue with the second-highest peak after quartz. Alunite was not present in the initial bauxite residue XRD spectra but detected in the residue due to the conversion of the dissolved aluminum sulfate to alunite at elevated temperatures according to the reaction in Equation (4). Calcite deposits are generally formed due to the low solubility of calcium sulfate in water [30]. Magnetite in the leaching residue originates from the magnetite in the initial bauxite residue that was not dissolved. Jarosite is not present in the XRD spectra of either the bauxite residue sample or the leaching residue at temperatures ranging from 180 to 220 °C. As previously stated, jarosite is more likely to form at higher free acid concentrations, generated at higher temperatures through hydrolysis reactions. Kaya and Topkaya [31] reported that the increase in the sulfur content of the leach residue by increasing temperature indicated the formation of a sulfur-containing hydrolysis product, such as basic ferric sulfate or jarosite.

Under optimum conditions, the percentage of Sc extraction was 91.97%, whereas the co-dissolution of Fe was less than 33%. These results indicate that high-pressure acid leaching gives a high recovery of scandium with higher selectivity to iron and lower acid requirements than direct acid leaching at atmospheric pressure that yields extracted scandium of 80% with an iron co-extraction of about 60% [32]. In addition, the leaching with inorganic acid at atmospheric pressure faces filtering issues due to silica gel formation. It also requires a greater quantity of reagent in the downstream process since it is less selective to iron. Meanwhile, when organic acids such as acetic or citric acid were used as the leaching agents, scandium recoveries from the bauxite residue leaching were relatively low [33,34]. Other methods that are typically considered more ecologically acceptable, such as alkaline leaching and bioleaching, result in significantly lower scandium recoveries (<45%) [35,36]. The high-pressure acid leaching of the bauxite residue in sulfuric acid resulted in scandium recovery rates comparable to or even higher than the energy-intensive pyrometallurgical pre-treatment followed by leaching methods [37,38].

4. Conclusion

This study discusses the extraction of scandium from bauxite residue as a by-product from an alumina refinery through High-Pressure Acid Leaching (HPAL) in sulfuric acid solution. The high-pressure leaching in sulfuric acid could effectively dissolve scandium, with minimum silica gel formation and relatively lower co-extracted iron compared to direct atmospheric acid leaching but still had a high co-extracted aluminium. Based on the data analysis by ANOVA, the contributions of acid concentration, leaching duration, temperature, and slurry density in influencing the extraction of scandium were 21.2%, 0.3%, 16.4%, and 62%, respectively. The optimum condition for maximizing scandium extraction was obtained at an acid concentration of 1.5 M, a temperature of 200 °C, a slurry density of 10% (w/w), and a leaching duration of 1 h. The validation experiment carried out under these conditions resulted in scandium extraction of 90.97%, with iron and aluminium co-extractions of 32.34% and 75.23%, respectively. The extracted scandium increased as the acid concentration increased to 1.5 M and decreased by a further increase of the acid concentration to 2 M. Meanwhile, iron and aluminium dissolution increased consistently with increasing acid concentration. The extracted scandium, iron, and aluminium increased significantly in the first 1 h but did not change significantly with a further extension of the leaching duration. Increasing the temperature from 180 °C to 200 °C increased the extracted scandium, but a further increase of the leaching temperature to 220 °C and 240 °C decreased the extracted scandium. The XRD analysis revealed that quartz, calcite, magnetite, alunite, hematite, and jarosite are the main constituent of the leaching residue. The investigation results show the utilization prospect of bauxite residue, a by-product from an alumina refinery, by conducting a high-temperature acid leaching in sulfuric acid.

Author contribution statement

Zaki Mubarok: Conceived and designed the experiments; Contributed reagents, materials, analysis tools or data; Wrote the paper. Mikhael Reynaldo Kanekaputra: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Funding statement

This work was supported by MIND ID.

5. Data availability statement

Data will be made available on request.

6. Declaration of interest's statement

This research work is university-based and totally independent, and financially supported by MIND ID (i.e., a state holding mining company in Indonesia). The bauxite residue sample was provided by the company which has given us permission to publish the research results. The authors declare no conflict of interest.