Producing High Purity Metallic Iron from Bauxite Residue through Hydrogen Reduction Followed by Flux Smelting

Abstract

The effective management and utilization of bauxite residue poses a significant challenge for the alumina industry, especially given the escalating demand for aluminum. This investigation focuses on utilization of bauxite residue to produce high-purity iron and the creation of leachable calcium aluminate slag through processes involving hydrogen reduction and smelting. Bauxite residue was pelletized and then underwent reduction at 1000°C in the presence of hydrogen. The resulting reduced bauxite residue was subsequently milled and blended with varying percentages of CaO, followed by smelting at 1500°C to recover iron and a leachable calcium aluminate slag. Analytical techniques such as X-ray diffraction, Electron Probe Microanalysis, scanning electron microscopy, and X-ray fluorescence were employed to assess the phases, microstructure, and chemical compositions. The hydrogen reduction process successfully transformed iron oxide in the bauxite residue into metallic iron. The increasing amounts of CaO in smelting led to the dominance of the calcium aluminate (CaO·Al₂O₃) phase in the slag products, with the phase composition remaining relatively stable after reaching a 40% CaO content. Key phases identified in the smelted slag included CaO·Al₂O₃, Ca₂Al₂SiO₇, and CaTiO₃. Notably, the iron produced during smelting exhibited a purity exceeding 99.5%, comparable to electrolytic iron. Experimental analysis revealed a positive correlation between the purity of iron and the concentration of CaO in the slag melt. Only a minimal amount of iron, primarily in the form of FeO, was observed in the slag, a phenomenon corroborated by FactSage analysis.

1. Introduction

With sustainable development, a primary goal for industry is to minimize waste generation with a focus on preventing industrial pollution. With the progress of time, growing environmental concern have made waste generation and storage increasingly problematic for the alumina industry. During the extraction of alumina from bauxite ore through the Bayer Process a byproduct is generated known as red mud.^1,2) The red mud comprises of a slurry of unleached components of bauxite ore suspended in a highly alkali solution. In the past, industries used to dispose of the slurry directly either to the sea or to the soil via storing in ponds. However, due to more recent environmental regulations this slurry is classified hazardous, and hence for better handling in some alumina refineries it is now subjected to a filtering to remove the excess water and alkali content and enable its safer discarding or its valorization.^3,4) This dewatered form of red mud is known as bauxite residue (BR). This residue is highly alkaline, pH level exceeding 11, and contains a significant amount of heavy metals.⁵⁾ The particles of residue are very fine in size with D₅₀ around 2.2 μm.⁶⁾ Due to the leachability of the alkali and heavy metal to the ground water and atmosphere, residue cannot be safely dumped in open soil, therefore it stored with specific methods i.e. in the cemented floor.⁷⁾ Bauxite residue contains varieties of oxides such as Fe, Al, Si, Ca, Ti, and certain rare earth elements.⁸⁾ Due to this composition, it can be used as a source for several applications such as for the recovery of specific elements, use as a major component in the manufacturing and construction industry, soil amelioration or landfill capping.⁹⁾ Nevertheless, recovery of metallic value from the residue will be more economical as compared to other applications. A major portion of bauxite residue consists of iron and alumina ranging from around 50 to 70 wt.% and this percentage may vary depending upon the Bayer process parameters and also the origin of the bauxite ore.^2,5,10) The utilization of bauxite residue for iron and alumina recovery has the potential to significantly decrease the residue volume.

Both iron and alumina can be recovered from the bauxite residue through processes such as smelting or reduction roasting employing reductant. The previous research relied on carbon as reductant, leading to environmental consequences.^11,12,13) As the concern about environment increases, particularly concerning global warming primarily to greenhouse gas emission, industries are actively transitioning towards the use of hydrogen as a reductant, replacing traditional carbon-based methods. This transition to hydrogen represents a significant step towards sustainability and offers a promising reducing emission in the future. Hydrogen holds to serve as a sustainable reductant, heading a cleaner and more environmentally friendly era for various industrial processes and its use in bauxite residue valorization has positive environmental impacts.

Iron production through carbothermic reduction from bauxite residue has been studied in recent years. Lazou et al. (2021) studied the iron production from the bauxite residue and bauxite ore by smelting-reduction using carbon as a reductant.^1,14) This research has identified that maintaining a CaO: Al₂O₃ mass ratio within 1.3 to 1.4 in the slag composition can yield more alumina containing leachable phases such as CaO.Al₂O₃ and 12CaO.7Al₂O₃. They also emphasis that the cooling rate of the slag has the significant effect on the characteristics and phase formation of the slag. A slower cooling rate leads to an enhanced formation of calcium aluminate leachable phase in slag. Majority of minor elements such as V, Ni, Cr and P are concentered in the molten pig iron, while elements like Ca, Al, Mg, and S are concentrated in the slag, and Si and Ti are distributed between the two phases.

Borra et al. (2015) studied the smelting reduction of bauxite residue using CaSiO₃ as a flux and carbon as reductant to recover iron and rare earth elements from the slag.²⁾ The optimized condition for iron recovery (above 85%) were determined to be heating to 1500°C with adding 29 wt.% flux and 5 wt.% of reductant. Following the removal of iron, the slag undergoes treatment with different acids resulting the recovery of Sc above 95%, rare earths and titanium above 70 wt.%. In another study by Borra et al. (2017), a two-step process investigated, which includes alkali roasting followed by smelting and slag leaching with acid to recover alumina, iron and rare earths, respectively.¹⁵⁾ The bauxite residue is blended with Na₂CO₃ and roasted at 950°C to facilitate the recovery of alumina by water leaching. After alumina recovery, the leaching residue was mixed with graphite to reduce iron oxide and the generated slag is utilized for rare earths recovery by acid leaching. In this process more than 90% of iron and above 80% of scandium and titanium were recovered. The iron and alumina recovery of bauxite residue by carbothermic reduction were previously summarized.⁵⁾ More recently, iron and alumina recovery from bauxite residue by hydrogen reduction has been studied.^16,17,18,19) In certain studies, physical separation techniques have been employed to recover iron, however, those methods show low efficiency/recovery.

There are some literatures related to iron production through hydrogen reduction from iron ore^{20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37)} however, here we mostly discussed about the iron production from bauxite residue.

Iron produced by carbothermic smelting has several drawbacks. Firstly, it contributes to the emission of greenhouse gases which is the major cause of global warming. Furthermore, the produced iron contains significant number of impurities majorly includes carbon, silicon (Si), titanium (Ti) and the cumulative impurity percentage goes above 5 wt.%. These impurities can adversely affect the quality and usability of produced iron as a pig iron source or additive. The chemical composition of the produced iron from smelting-reduction of bauxite residue from various studies are presented in the Table 1.

Table 1. Composition of iron produced by carbothermic reduction-smelting of bauxite residue (wt.%).

Fe	C	Si	P	S	Ni	Cr	Ti	Mn	Cu	V	Ref
94.256	5.1	0.19	0.12	0.004	–	–	–	–	–	–	2)
99.1	–	0.26	–	–	0.32	0.19	0.05	–	–	–	12)
92.74	5.22	0.22	0.08	0.001	0.32	0.03	0.75	0.04	0.03	0.15	1)
93	5.1	0.1	0.2	0.006	–	–	0.3	–	–	0.2	11)

In this work, the reduction of bauxite residue by hydrogen followed by the flux-melting of the reduced pellets is investigated. The aim is to produce high purity iron, and a highly leachable calcium-aluminate slag via lime use as flux. This research outlines a process to make value from the bauxite residue waste regarding environmental sustainability.

2. Materials and Methods

2.1. Materials

Bauxite residue was provided from Mytilineos, Greece, the residue filter cake was received as lumpy form. The lumps were dried in an oven at a temperature 80±5°C overnight to remove excess moisture. This dried bauxite residue was then deagglomerated in a ball mill and sieved to achieve particle size below 500 μm. The calcium oxide used for this experimental work are lab grade with purity above 99.9% (Sigma-Aldrich).

2.2. Pelletization

The deagglomerated bauxite residue was pelletized in a drum pelletizer via the addition of approximately 10 wt.% water. During pelletization, as the drum undergoes rotation, bauxite residue was charged at the bottom of the drum. Upon contact with the water, bauxite residue particles join to form micro-pellets. Rolling action combined with the water addition to bauxite residue, promotes the agglomeration and increases in pellet size. The combined addition of bauxite residue and water spraying continue until the pellet size increased to above 6 mm. At this stage, pellets were sieved and sorted within the size of 6–10 mm. These green pellets were dried overnight in an oven at 80±5°C.

2.3. Hydrogen Reduction

Hydrogen reduction was carried out in a thermogravimetry furnace. The details and operational procedure of the furnace was described elsewhere.¹⁰⁾ The pellets were heated to the targeted temperature of 1000±10°C at a heating rate of 10°C/min in the presence of argon, with a flow rate of 1 NL/min. Upon reaching the targeted temperature, the gas was switched to hydrogen with a flow rate of 4 NL/min. Hydrogen gas was purged once the weight loss during heating stabilized. The hydrogen reduction was carried out for duration of 90 minutes at the targeted temperature. After the reduction cycle was completed, the reduced pellets were cooled down to the room temperature in the presence of argon to avoid oxidation of the metallic iron. The reduced pellets underwent milling in ring mill for 60 sec at a speed of 900 revolution per minute (rpm). The reason of milling was to have proper mixing with the added lime flux powder in the next smelting step.

2.4. Flux-smelting

The smelting experiments were carried out in a sealed stainless steel vertical reactor. The schematic drawing of the furnace and the operational procedure were descried previously.³⁸⁾ Five set of smelting experiments were conducted with varying calcium oxide (CaO) as flux and with the aim of forming calcium-aluminate slags. For each smelting experiment, 15 g of reduced bauxite residue mixed with selected amount of CaO powder (0, 20, 30, 40 and 50 wt.%) and the mixture was added to a high purity alumina crucible (External diameter of 3.5 cm and height of 4 cm). The alumina crucible was placed in a graphite substrate inside a well-calibrated heating zone of the furnace. To monitor the furnace and crucible temperature two thermocouples were used, one on the furnace wall, and another was inserted from the top of the furnace inside the crucible, slightly above the sample. All smelting experiments were conducted at 1500°C in the presence of argon atmosphere (1 Nl/min) and the holding time at the targeted temperature were maintained 60 mins. The heating and cooling to this temperature were with the rate of 25°C/min. The Experimental overview for the entire process is shown in Fig. 1.

Fig. 1. Experimental overview and the materials flow. (Online version in color.)

2.5. Characterization

Phase analysis of the raw materials, reduced pellets and slag were carried out by X-ray diffraction (XRD) (Bruker AXS GmbH, Karlsruhe, Germany) with Cu Kα radiation (wavelength λ=1.54 Å). The diffractometer scans in the range of 15° to 75° 2θ with a step size of 0.03° were used. The qualitative phase analysis of the raw data was analyzed by using DIFFRAC.EVA software with the data base of PDF-4+ (2014, ICDD, Philadelphia, Pennsylvania, USA). To prepare the standard sample for XRD, the materials were milled in WC vibratory disk mill (RS 200, RETSCH GmbH, Haan, Germany) for 45 sec at an 800 revolution per minute(rpm). The microstructural analysis and elemental mapping of the materials were done by an electron probe microanalyzer (EPMA) (JXA-8500F, JEOL Ltd., Akishima, Japan), supported with wavelength dispersive spectroscopy (WDS) for the quantitative measurements.

3. Results

The chemical composition and phase analysis of the bauxite residue are shown in Table 2 and Fig. 2, respectively.

Table 2. XRF elemental analysis of BR and calcite.

Materials/oxides	Al2O3	CaO	Fe2O3	K2O	MnO	MgO	Na2O	P2O5	SO3	SiO2	TiO2	LOI
BR	22	8.8	40.71	0.09	0.08	0.23	3.1	0.11	0.95	7.1	5	12.4

Fig. 2. Phase analysis of bauxite residue.

The XRD analysis of the bauxite residue reveals that the major phases present are hematite, diaspore, katoite, cancrinite, and sodalite. The majority of the iron and alumina in the residue are primarily found in the form of hematite and diaspore.

3.1. Phase Analysis of Reduced Pellet

The phase analysis of reduced bauxite residue pellets is shown in Fig. 3. The XRD of raw bauxite residue was presented in our previous work³⁹⁾ and indicated the hematite (Fe₂O₃), goethite (FeHO₂), anatase (TiO₂), perovskite (CaTiO₃), calcite (CaCO₃), diaspore α-AlO(OH), sodalite (Al₃Cl₁K₁Na₄O₁₂Si₃) and cancrinite (C_0.76H_1.7Al₃Ca_0.75Na₃O_14.4Si₃) are the major phases found. In the reduced pellet the major peaks are for iron (Fe), gehlenite (Ca₂Al₂SiO₇), perovskite (CaTiO₃), corundum (Al₂O₃) and nepheline (Al₄Ca_0.285K_0.1Na_3.33O₁₆Si₄). The absence of iron oxide is indicating that the complete reduction of iron oxide occurred in the hydrogen reduction.

Fig. 3. XRD of reduced bauxite residue.

3.2. Phase Analysis of Produced Slags

The phase analysis of different slags resulting from various mixtures of reduced bauxite residue and calcium oxide are presented in the Fig. 4. In the reduced bauxite residue without calcium oxide, the generated slag primarily composed of nepheline, perovskite and larnite with the presence of minor fraction of metallic iron. Furthermore, in this slag some peaks corresponding to corundum phase were also observed. As the percentage of calcium oxide increases the krotite (CaAl₂O₄) phase becoming more pronounced. At a calcium oxide 50 wt.%, alumina predominantly formed krotite with a minor fraction present in gehlenite (Ca₂Al₂SiO₇). As indicated by the XRD result, the intensity of krotite phase increased more intensely after 30 wt.% added calcium oxide. Presence of perovskite (CaTiO₃) phase is observed in all slag samples. Slag with 40 wt.% and 50 wt.% of calcium oxide, the phases are identical and exhibiting almost similar intensities. Figure 3 shows that the intensity of gehlenite decreases with increasing in the percentage of calcium oxide in the charge. Gehlenite phase has the most intense intensity at 30 wt.% calcium oxide. In slag with a higher percentage of calcium oxide, a portion of alumina exists as gehlenite and mullite, while krotite is the primary phase. Alumina predominantly forms three phases in the smelted slag gehlenite, mullite, and krotite. Of these, only krotite is alkali-leachable, whereas both gehlenite and mullite are not alkaline leachable. Form previous work in which the leachability of the calcium aluminate phase has this order: 12CaO.7Al₂O₃ > 3CaO.Al₂O₃ > CaO.Al₂O₃ > CaO.2Al₂O₃.⁴⁰⁾

Fig. 4. XRD spectra of different slags obtained with varying CaO content of smelting feed. (Online version in color.)

As shown in Fig. 5 of the phase diagram, increasing the calcium oxide content pushes the slag into the krotite (CaAl₂O₄) phase region, which is the desired leachable phase. Phase analysis (Fig. 4) reveals that when the calcium oxide concentration exceeds 30 wt.%, krotite becomes the dominant phase. This transition is also evident in the phase diagram, where the melilite phase is observed at 30% CaO and shifts towards the krotite region as the concentration exceeds 30 wt% CaO.

Fig. 5. Different slag composition with varying CaO wt.% presented in CaO–Al₂O₃–SiO₂ phase diagram. (Online version in color.)

3.3. Microstructural Analysis

The elemental mapping of the smelted reduce bauxite residue is presented in the Fig. 6. For elemental mapping major elements Al, Ca, Fe, Na, Si and O were considered. Right side of the color bar represents the intensity of presence of elements. Bright areas indicate the presence of metallic iron phase and these bright areas do not over lapping with any other element, which is indicating that iron is in pure form. The overlapping of Si, O, Al, Na and Ca in certain areas indicates the presence of nepheline phase. The coexistence of Ca, Ti and O elements in some area suggests the presence of perovskite phase.

Fig. 6. Elemental mapping of smelted reduce bauxite residue. (Online version in color.)

The elemental analysis results of the smelted reduce bauxite residue are presented in Fig. 7. The given elemental analysis was obtained by averaging three measurements for each specific areas at higher magnification aiming to reduced signals from surrounding phases. The bright area is corresponding to metallic iron with minor fraction of Al, Na. The dark area is comprising elements such as Al, Ca, Si, Na and O indicating the phase of nephelite. These results are correlated with both with elemental mapping and XRD results. The light dark may be corresponding to calcium titanate phases. The overall of slag composition is also presented the Fig. 7.

Fig. 7. EDS point analysis of the smelted reduce bauxite residue. (Online version in color.)

Figure 8 shows the elemental mapping of the smelted sample of BR+30 wt.% CaO mixture. Obviously, the majority of analyzed area is covered with metallic iron. Similar to smelted reduce BR pellet, iron is presented separately without overlapping with other elements.

Fig. 8. Elemental mapping of melted BR+30 cal. (Online version in color.)

The elemental analysis of different areas in smelted reduce BR+30 wt.% CaO sample are presented in Fig. 9. The bright area is the metallic iron along with a minor fraction of Na, Ti, and Al. The dark grey phase is composed of Al, Ca and O suggesting the presence of calcium aluminate phase with a minor Na present. In this elemental analysis the presence of iron may be due to dispersed white particles (iron fine particles) inside the dark color phase. The light dark areas are likely to be the calcium titanate phase, perovskite. The elemental analysis of overall slag composition is presented in upper left corner of the Fig. 9.

Fig. 9. EDS point analysis of the smelted reduce BR+30 wt.% CaO sample. (Online version in color.)

The elemental mapping of smelted reduce BR+50 wt.% CaO pellets are shown in Fig. 10. Major bright areas are indicative metallic iron evident the fact that they do not overlap with oxygen and other elements. Ca, Al and O are over lapping in some areas which may be the calcium aluminate phase.

Fig. 10. Elemental mapping of smelted reduce BR+50 wt.% CaO sample. (Online version in color.)

The elemental analysis of various regions in the smelted reduce BR+50 wt.% CaO mixture is presented in Fig. 11. Metallic iron is composed of majority of Fe with minor fraction of Na and Si. From the elemental analysis it was found that the dark phase is mainly calcium aluminate (CaAl₂O₄) with minor fraction of silicon and iron. The light-grey area is calcium titanate phase. In the overall composition of slag, majority of phases are dominated by the elements with Ca, Al, and O. The elemental analysis of the different phases is calculated in higher magnification to overcome the signals from the surrounding phases.

Fig. 11. EDS point analysis of the smelted reduce BR+50 wt.% CaO sample. (Online version in color.)

For these three samples, the phases identified with elemental analysis align with the phases detected in XRD.

4. Discussion

4.1. Phase Formation during Reduction

During the reduction of bauxite residue, the reaction between calcium oxide (CaO), alumina (Al₂O₃) and silica (SiO₂) results in formation of gehlenite (Ca₂Al₂SiO₇) phase, which exhibits thermodynamically stability at the given reduction temperature (1000°C). In the specified reduction temperature along with the formation of gehlenite, perovskite (CaTiO₃) phase is also produced. The alumina present in the bauxite residue reacts with sodium oxide and silica to form nepheline (Al₄Ca_0.285K_0.1Na_3.33O₁₆Si₄) phase. This phase is also stable at the specified given temperature as per Gibbs free energy calculation shown in Eqs. (1), (2), (3). As per XRD phase analysis nepheline (Al₄Ca_0.285K_0.1Na_3.33Si₄O₁₆) phase has minor fraction of Ca and K in its structure, which suggest the possibility of high temperature diffusion of Ca and K into the lattice of nepheline. The free energy formation of nepheline phase without Ca and K is presented below as this stochiometric is not present in thermodynamics databases such as FactSage.

  

$$\begin{array}{l}2CaO+A{l}_2{O}_3+Si{O}_2\to C{a}_{2}A{l}_{2}Si{O}_7,\\ \mathrm{\Delta }°G_{1\mathrm{\hspace{0.17em} }000°\mathrm{C}}=-163.84\mathrm{\hspace{0.17em} }kJ/\text{mol}\end{array}$$(1)

  

$$CaO+Ti{O}_2\to CaTi{O}_3,\mathrm{\hspace{0.17em} }\mathrm{\Delta }°G_{1\mathrm{\hspace{0.17em} }000°\mathrm{C}}=-78.60\mathrm{\hspace{0.17em} }kJ/mol$$(2)

  

$$\begin{array}{l}2A{l}_2{O}_3+2N{a}_{2}O+4Si{O}_2\to A{l}_{4}N{a}_4{O}_{16}S{i}_4,\\ \mathrm{\Delta }°G_{1\mathrm{\hspace{0.17em} }000°\mathrm{C}}=-178.18\mathrm{\hspace{0.17em} }kJ/mol\end{array}$$(3)

The Gibbs free energy of formation for the primary calcium aluminate and calcium silicate phases at the specified melting temperature are presented in Eqs. (4), (5), (6), (7). Alumina remains in the slag in form of Al₄CaO₇, CaAl₂O₄, Ca₂Al₂SiO₇ and Al₆SiO₁₁ with CaAl₂O₄ phase constituting the majority of alumina in the slag.

  

$$2CaO+Si{O}_2\to C{a}_{2}Si{O}_4,\mathrm{\hspace{0.17em} }\mathrm{\Delta }°G_{1\mathrm{\hspace{0.17em} }500°\mathrm{C}}=-146.54\mathrm{\hspace{0.17em} }\text{kJ}/\text{mol}$$(4)

  

$$CaO+2A{l}_2{O}_3\to A{l}_{4}Ca{O}_7,\mathrm{\hspace{0.17em} }\mathrm{\Delta }°G_{1\mathrm{\hspace{0.17em} }500°\mathrm{C}}=-71.72\mathrm{\hspace{0.17em} }kJ/mol$$(5)

  

$$CaO+Si{O}_2\to CaSi{O}_3,\mathrm{\hspace{0.17em} }\mathrm{\Delta }°G_{1\mathrm{\hspace{0.17em} }500°\mathrm{C}}=-87.99\mathrm{\hspace{0.17em} }kJ/mol$$(6)

  

$$CaO+A{l}_2{O}_3\to CaA{l}_2{O}_4,\mathrm{\Delta }°G_{1\mathrm{\hspace{0.17em} }500°\mathrm{C}}=-52.82\mathrm{\hspace{0.17em} }kJ/mol$$(7)

4.2. Thermodyamices Calculation

The thermodynamics calculation of slag phase was conducted by using FactSage thermodynamics software, version 8.1. In this calculation the major oxides (CaO, Al₂O₃, TiO₂ and SiO₂) were used as input oxides for calculation. Na₂O was excluded from the thermodynamics calculation due to the lower vapour pressure at the operating temperature. This exclusion also aligns with our previous experimental findings, which indicates the loss of sodium during the high temperature process.¹⁰⁾ Thermodynamics calculations are performed under the assumption of a closed system. However, in actual experimental condition often involve open system. Due to the reason mentioned earlier, Na₂O is mostly lost during experiments. Additionally at higher CaO additions, sodium-containing phases were not detected in the XRD, and as Na is light element, we cannot rely on EDS analysis. The phases formed with varying CaO wt.% in the slag are presented in Fig. 12 and show that calcium aluminate exists in various forms including CaAl₂O₄, CaAl₄O₇, CaAl₁₂O₁₉ and Ca₃Al₂O₆. Additionally, some alumina is present in the form of CaAl₂Si₂O₈ and melilite phase. Melilite is a solution phase of Ca, Al, Si and O. In slag with 50 wt.% CaO, the primarily phases include CaAl₂O₄, Ca₃Al₂O₆, Ca₂SiO₄ and Ca₃Ti₂O₇ and Ca₅Ti₄O₁₃ phase. In thermodynamics calculation of calcium titanate solid solution various phases are observed, including CaTiO₃, Ca₃Ti₂O₇, Ca₅Ti₄O₁₃. When comparing these findings to the XRD analysis of a slag with the addition of 50 wt.% CaO, the identified phases include CaAl₂O₄, Ca₂Al₂SiO₇, and CaTiO₃. In 40 wt.% CaO slag calcium aluminate (CaAl₂O₄, Ca₃Al₂O₆) and calcium silicate (Ca₂SiO₄) phases are present similar to 50 wt.% slag. Additionally, both the 40 wt.% and 50 wt.% CaO added slags contain Ca₃Ti₂O₇ and Ca₅Ti₄O₁₃ as calcium titanate phase. However, in XRD of all slag indicates presence of CaTiO₃ phase. With increase in CaO percentage the stability of CaAl₂O₄ is observed and after 40% CaO addition, it is becoming constant. As per thermodynamics calculation above 30 wt.% CaO added slag, calcium aluminate (CaAl₂O₄) phase becoming stable whereas below 30 wt.% CaO, CaAl₄O₇ and CaAl₁₂O₁₉ calcium aluminate phase are stable. In smelted reduce BR pellets CaAl₁₂O₁₉, CaAl₂Si₂O₈ and CaTiO₃ major phases found as per thermodynamics calculation, however, in actual calculation it was found to be present in nepheline phase, which is solid solution of Ca, Al, Si, O, and Na. Additionally perovskite, larnite phases are present.

Fig. 12. Equilibrium calculation of reduced BR and varying CaO at 1500°C at 1 atm using FactSage version 8.1. (Online version in color.)

4.3. Viscosity Model for Different CaO Addition

In the above microstructural analysis, it was found that there are some small metal particles in the produced slags as observed in Figs. 6, 7, 10 and 11. Moreover, more metallic particles were found in the less added lime experiments. This may indicate that the viscosity is different for the produced slags, and it may be the reason of not complete metal droplets separation and settlement. In principle, viscosity plays a major role in effective slag metal separation. Hence, the viscosity of slag produced by the addition of CaO with reduced bauxite residue pellet was calculated by FactSage 8.1. This involved utilization of thermodynamics models and data to estimate the viscosity of the slag at specified condition. In these thermodynamics viscosity calculation components considered includes CaO, TiO₂, Al₂O₃ and SiO₂. Viscosities of slags calculated was plotted against wt.% CaO added in the Fig. 13. As shown, with increase in CaO percentage in the smelting feed, the viscosity of the produced slag decreases. However, there is a lower rate in decrease of viscosity at higher calcium oxide additions than 30%. The viscosity of the industrial blast furnace (BF) slag typically falls within the range of 0.3 to 0.5 Pa.s at 1500°C with the major oxides are CaO, MgO, Al₂O₃ and SiO₂.⁴¹⁾ As compared to BF slag, reduced bauxite residue calcite slag viscosity is similar. The viscosity of reduced bauxite residue slag is around 0.75 Pa.s which may be due to the higher silica and alumina present as compared calcium oxide. In the thermodynamics viscosity calculation, the assumption was made that all oxides are in liquid state, however, in actual experiments some solid oxides such as CaTiO₃ may present in the liquid melt leading to increase in viscosity.

Fig. 13. Viscosity of slag with varying CaO content of molten slag at 1500°C. (Online version in color.)

5. Purity of Produced Iron

Figure 14 shows the scanning electron micrographs of powdered electrolytic metallic iron used as a reference material. The particles are irregular in shape accompanied by variation in the particle size. The EDS analysis of particles is shown in Fig. 14 and elemental mapping in Fig. 15. It is shown that, spectrum 1 reveals pure iron and spectrum 2 indicates the presence of minor amount of oxygen along with metallic iron. The presence of oxygen may be due to the oxidation of the metallic iron in the production step or further material handling.

Fig. 14. Scanning electron micrograph and EDS point analysis of a commercial electrolytic iron. (Online version in color.)

Fig. 15. Elemental mapping of pure metallic iron. (Online version in color.)

In comparison with the metallic iron produced from bauxite residue had a purity above 99 wt.% like the pure metallic iron. The iron produced from bauxite residue holds potential as a source for high purity metallic iron for some applications. As presented above the produced iron has minor elements of Na, Si, Ti and Al, which are totally less than 1.0 wt.%. The hydrogen reduction followed by flux-smelting in this study is advantageous compared to a similar carbothermic process that yields a pig iron (saturated of carbon) and a similar calcium aluminate slag. In that carbothermic process a pig iron with considerable amounts of C, Si, Ti, V, Cr is produced as presented in Table 1.^1,2,11) This may indicate the effect of carbon in molten iron on the reduction of metal oxides from the slag.

The above chemical analysis results indicated that the purity of produced Fe is higher when more CaO added in the smelting of the reduced BR pellet. Figure 16(a) illustrates the FactSage results pertaining to the presence of FeO in the slag relative to the addition of CaO and Fig. 14(b) shown the Al, Si and Ti in the metal and slag. As shown in the Fig. 16(b), with increases in the CaO, Al, Si and Ti in the metal decreases. This may be due to the decrease of the chemical activities of Al₂O₃, SiO₂ and TiO₂ in the slag via more CaO addition that causes more Al, Si and Ti transfer from the metal to the slag, yielding a higher purity molten iron. At lower weight percentages of CaO, the metallic iron content was lower compared to the calculated quantity, and as CaO increased, a lower proportion of iron entered the slag phase. Notably, beyond a 30 wt.% CaO addition, the FeO content in the slag remained constant as 0.1 g. Hence, the use of more CaO in the flux-smelting of the hydrogen reduced BR pellets will yield a higher iron purity.

Fig. 16. Thermodynamics calculation of FeO in slag with respect to CaO, calculated by using FactSage version 8.1 (a), Al, Si and Ti in oxide slag(S) and metal(M) (b) (NOTE: Fe input for calculation=Fe in metal+Fe in slag, Total mass of metal=Fe in metal+ Al, Si and Ti in metal. (Online version in color.)

6. Conclusion

In this study, bauxite residue pellets underwent a reduction process using hydrogen, followed by smelting of the reduced pellets with different amounts of CaO flux to attain high-purity metallic iron and leachable calcium aluminate slag. The primary conclusions drawn from this investigation can be summarized as follows:

(1) The complete reduction of iron oxide within bauxite residue into metallic iron was successfully achieved during the reduction process at 1000°C. Following the reduction process, the resulting major phases comprise metallic iron, perovskite, larnite, and corundum.

(2) Beyond the threshold of 40% CaO, a significant portion of the alumina present in the bauxite residue undergoes conversion into mono calcium aluminate phase (krotite), which is expected for further alkali leaching.

(3) The viscosity of the produced slag is decreased with increasing CaO in the slag, and it improves metal-slag separation.

(4) The iron obtained through the smelting process exhibits a purity of approximately 99.5 wt.%, with a minor/trace fraction of Na, Si, Ti, Al. This level of purity positions the produced iron comparable with electrolytic iron.

(5) The experimental analysis revealed a correlation between the augmentation of CaO and increase in the produced iron purity which aligns with the analysis conducted using FactSage.

(6) The addition of more lime in the flux- smelting of pre-reduced pellet shows higher iron product purity due to the more distribution of main impurities such as Si, Al, and Ti into the slag phase.

Statement for Conflict of Interest

Authors have declared that there is no conflict of interest related to the conduct of this research.

Acknowledgement

This project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No. 958307. This publication represents only the authors’ views, exempting the Community from any liability.

Funding

This project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No. 958307.

References

• 1)   A.  Lazou,  C.  Van Der Eijk,  K.  Tang,  E.  Balomenos,  L.  Kolbeinsen and  J.  Safarian: Metall. Mater. Trans. B, 52 (2021), 1255. https://doi.org/10.1007/s11663-021-02086-w
• 2)   C. R.  Borra,  B.  Blanpain,  Y.  Pontikes,  K.  Binnemans and  T.  Van Gerven: J. Sustain. Metall., 2 (2016), 28. https://doi.org/10.1007/s40831-015-0026-4
• 3)   K.  Evans: J. Sustain. Metall., 2 (2016), 316. https://doi.org/10.1007/s40831-016-0060-x
• 4)   K.  Evans,  E.  Nordheim and  K.  Tsesmelis: Light Met. 2012, 63. https://doi.org/10.1007/978-3-319-48179-1_11
• 5)   M. K.  Kar,  M. A. R.  Ӧnal and  C. R.  Borra: Resour. Conserv. Recycl., 198 (2023), 107158. https://doi.org/10.1016/j.resconrec.2023.107158
• 6)   C.  Cardenia,  E.  Balomenos and  D.  Panias: J. Sustain. Metall., 5 (2019), 9. https://doi.org/10.1007/s40831-018-0181-5
• 7)   G.  Power,  M.  Gräfe and  C.  Klauber: Hydrometallurgy, 108 (2011), 33. https://doi.org/10.1016/j.hydromet.2011.02.006
• 8)   A.  Lazou,  C.  Van der Eijk,  E.  Balomenos,  L.  Kolbeinsen and  J.  Safarian: In Proceedings of the 3rd International Bauxite Residue Valorisation and Best Practices Conference 2020, BR 2020, Virtual, (2020), 77.
• 9)   C.  Klauber,  M.  Gräfe and  G.  Power: Hydrometallurgy, 108 (2011), 11. https://doi.org/10.1016/j.hydromet.2011.02.007
• 10)   M. K.  Kar,  C.  Van Der Eijk and  J.  Safarian: Metals, 13 (2023), 644. https://doi.org/10.3390/met13040644
• 11)   P. T.  Yin,  B. F.  Buhle Xakalashe,  D.  Panias and  V.  Vassiliadou: In Proceedings of 35th International Committee for Study of Bauxite, Alumina & Aluminium conference 2017, ICSOBA 2017, Germany, (2017), 603.
• 12)   K. E.  Ekstrøm,  A. V.  Bugten,  C.  Van Der Eijk,  A.  Lazou,  E.  Balomenos and  G.  Tranell: J. Sustain. Metall., 7 (2021), 1314. https://doi.org/10.1007/s40831-021-00422-7
• 13)   H. A.  Cheema,  S.  Ilyas,  M.  Farhan,  J.  Yang and  H.  Kim: J. Ind. Eng. Chem., 136 (2024), 201. https://doi.org/10.1016/j.jiec.2024.02.007
• 14)   A.  Lazou,  C.  Van der Eijk,  E.  Balomenos and  J.  Safarian: In Proceedings of the Proceedings of European Metallurgical Conference 2019, EMC 2019, Düsseldorf, Germany, (2019), 17.
• 15)   C. R.  Borra,  B.  Blanpain,  Y.  Pontikes,  K.  Binnemans and  T.  Van Gerven: J. Sustain. Metall., 3 (2017), 393. https://doi.org/10.1007/s40831-016-0103-3
• 16)   G.  Pilla,  S. V.  Kapelari,  T.  Hertel,  B.  Blanpain and  Y.  Pontikes: Mater. Today Proc., 57 (2022), 705. https://doi.org/10.1016/j.matpr.2022.02.152
• 17)   G.  Pilla,  T.  Hertel,  B.  Blanpain and  Y.  Pontikes: In Proceedings of the 8th International Slag Valorisation Symposium 2023, SVS 2023, Belgium, (2023), 146.
• 18)   A.  Hassanzadeh,  M. K.  Kar,  J.  Safarian and  P. B.  Kowalczuk: Metals, 13 (2023), 946. https://doi.org/10.3390/met13050946
• 19)   M. K.  Kar,  A.  Hassanzadeh,  C.  Van der Eijk,  P. B.  Kowalczuk,  K.  Aasly and  J.  Safarian: Int. J. Hydrog. Energy, 48 (2023), 38976. https://doi.org/10.1016/j.ijhydene.2023.09.212
• 20)   Y.  Kashiwaya and  M.  Hasegawa: ISIJ Int., 52 (2012), 1513. https://doi.org/10.2355/isijinternational.52.1513
• 21)   S. K.  Dutta and  A.  Ghosh: ISIJ Int., 33 (1993), 1168. https://doi.org/10.2355/isijinternational.33.1168
• 22)   M.  Sugata,  T.  Aida,  Y.  Hara and  S.  Kondo: Jpn. Inst. Met., 31 (1967), 574. https://doi.org/10.2320/jinstmet1952.31.4_574
• 23)   S.  Taniguchi,  M.  Ohmi and  M.  Yamada: Trans. Iron Steel Inst. Jpn., 14 (1974), 241. https://doi.org/10.2355/isijinternational1966.14.241
• 24)   N.  Takeuchi,  Y.  Iwami,  T.  Higuchi,  K.  Nushiro,  N.  Oyama and  M.  Sato: ISIJ Int., 54 (2014), 791. https://doi.org/10.2355/isijinternational.54.791
• 25)   O.  Hessling,  J. B.  Fogelstrom,  K.  Kojola and  J.  Martinsson: ISIJ Int., 64 (2024), 1493. https://doi.org/10.2355/isijinternational.ISIJINT-2023-443
• 26)   T.  Murakami,  H.  Wakabayashi,  D.  Maruoka and  E.  Kasai: ISIJ Int., 60 (2020), 2678. https://doi.org/10.2355/isijinternational.ISIJINT-2020-180
• 27)   A.  Abdelrahim, M. lljana, M. Omran, T. Vuolio, H. Bartusch and T. Fabritius: ISIJ Int., 60 (2020), 2206. https://doi.org/10.2355/isijinternational.ISIJINT-2019-734
• 28)   K.  Kato and  H.  Ono: ISIJ Int., 64 (2024), 881. https://doi.org/10.2355/isijinternational.ISIJINT-2023-398
• 29)   A.  Bonalde,  A.  Henriquez and  M.  Manrique: ISIJ Int., 45 (2005), 1255. https://doi.org/10.2355/isijinternational.45.1255
• 30)   M.  Bai,  H.  Long,  S.  Ren,  D.  Liu and  C.  Zhao: ISIJ Int., 58 (2018), 1034. https://doi.org/10.2355/isijinternational.ISIJINT-2017-739
• 31)   J.  Kim,  S.  Kubota,  S.  Natsui,  T.  Iwanaga,  Y.  Miki and  H.  Nogami: ISIJ Int., 63 (2023), 1595. https://doi.org/10.2355/isijinternational.ISIJINT-2023-152
• 32)   I.  Iwamoto,  A.  Kurniawan,  H.  Hasegawa,  Y.  Kashiwaya,  T.  Nomura and  T.  Akiyama: ISIJ Int., 62 (2022), 2483. https://doi.org/10.2355/isijinternational.ISIJINT-2022-155
• 33)   O.  Kovtun,  M.  Levcheko,  M.  Ilatovskaia,  C.  Aneziris and  O.  Volkova: Steel Res. Int., 95 (2024), 2300707. https://doi.org/10.1002/srin.202300707
• 34)   Z.  Shen,  S.  Sun,  J.  Xu,  Q.  Liang and  H.  Liu: ISIJ Int., 63 (2023), 42. https://doi.org/10.2355/isijinternational.ISIJINT-2022-280
• 35)   J.  Zhang,  S.  Li and  L.  Wang: Int. J. Hydrogen Energy, 49 (2024), 1318. https://doi.org/10.1016/j.ijhydene.2023.09.243
• 36)   Y.  Kashiwatani and  M.  Hasegawa: Tetsu-to-Hagané, 100 (2014), 302 (in Japanese). https://doi.org/10.2355/tetsutohagane.100.302
• 37)   A.  Ohkura,  M.  Tokuda and  Y.  Matsushita: Tetsu-to-Hagané, 48 (1962), 1039 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.48.9_1039
• 38)   A.  Rasouli,  K. E.  Herstad,  J.  Safarian and  G.  Tranell: Metall. Mater. Trans. B, 53 (2022), 2132. https://doi.org/10.1007/s11663-022-02513-6
• 39)   M. K.  Kar and  J.  Safarian: Processes, 11 (2023), 137. https://doi.org/10.3390/pr11010137
• 40)   F.  Azof,  M.  Vafeias,  D.  Panias and  J.  Safarian: Hydrometallurgy, 191 (2020), 105185. https://doi.org/10.1016/j.hydromet.2019.105184
• 41)   M.  Chen,  D.  Zhang,  M.  Kou and  B.  Zhao: ISIJ Int., 54 (2014), 2025. https://doi.org/10.2355/isijinternational.54.2025