**1. Introduction**

The iron and steelmaking processes have been always related to the generation of dust/sludge, slag and other wastes or byproducts. Dust and sludge from the gas cleaning systems of the blast furnace (BF) and the basic oxygen furnace (BOF) contains significant amounts of iron, carbon and other elements. Such materials cannot be recycled because of its impurities (nonferrous metals, mainly zinc) and represent a loss of 2% of the iron and coal contained in the raw material. In 2018, the world mine production was of 2.5 billion tons of iron ore; then, dust and sludge can be representing a loss of 15 to 20 million tons of Fe units annually) [1,2]. The composition of the blast furnace and basic oxygen furnace dusts varies significantly depending on the process conditions. Blast furnace sludge (BFS) containing 0.77–5 wt.% Zn has been reported [3,4]. A high content of zinc in the raw material might damage the blast furnace refractory materials and, consequently, shorten its campaign life [5,6]. In the case of alkalis (Na and K), they increase coke consumption due to the transfer of heat to higher levels in the blast furnace, causing catastrophic swelling and disintegration of ores, pellets and coke. It also causes deterioration of the furnace lining. In brief, zinc and alkalis cause irregularities in the

operation process and decrease the production, so only it is recycled up to 20% of the dust and sludge generated [7–9].

As mentioned, not recycling of BFS represents an economical lost, due to the relatively high contents of iron (up to 50 wt.% Fe). This also represents an environmental issue, because this sludge contains leachable compounds of lead, cadmium and other elements that are classified as hazardous wastes [10]. This "waste" material is pile stocked on the steelmaking sites, which is an environmental hazard and occupies valuable land. To recycle BFS to the process, it is necessary to reduce their zinc and alkalis content to at least 0.03 wt.% and 0.12wt.%, respectively [11,12].

Pyrometallurgical and hydrometallurgical processes have been developed to recycle dust and/or to recover zinc from it. Pyrometallurgical routes are considered the primary choice because of its high potential metal recovery and among the most important processes are: Waelz, rotary heart furnace, PRIMUS, OXYCUP, coke-packed bed, Ausmelt, electric smelting reduction furnace, Plasamadust, plasma-arc, Elkem multi-purpose furnace, submerged plasma, etc. [13,14]. It is worth noting that in the metallurgical industry, the energy cost and pollution problems produced by pyrometallurgical processes (e.g., combustion gas emissions) have led to a more intensive search of hydrometallurgical alternatives [3].

The hydrometallurgical processes developed for dust treatment might be classified as leaching in acid or alkaline media [15,16]. The efforts have been focused on conventional technology, based on H2SO4, as is shown in Table 1. As can be observed, almost all of studies have been realized for dust from the electric arc furnace steelmaking process (EAF). It is interesting to observe the relativity high H2SO4 concentration used (>1 M) and temperature (>50 ◦C). In the case of the treatment time, it varies from 15 min (microwave heating leaching assisted) to 1.5 h [17–19].


**Table 1.** Leaching studies for zinc removal base in H2SO4 media.

Table 1 shows that Zn removal is >70% but the iron loss has a great variability (4% to 60%), which could be due to the different composition between each material. Kukurugya et al. [5] studied the behavior of various elements (i.e., zinc, iron and calcium) during the leaching of electric arc furnace dust in sulfuric acid solutions. The authors proposed that the rate limiting step for zinc and calcium was diffusion, while for iron it was the chemical reaction. This confirms the high solubility of some iron oxide compounds, but the iron loss will depend of the type of iron oxides contained in the material.

Other acid media has been studied to steelmaking dust. Shawabkeh [19] studied the zinc extraction from a Jordanian electric arc furnace dust by testing different acids (i.e., nitric, hydrochloric and sulphuric) at different concentrations. The highest zinc extraction (i.e., 72%) was obtained with H2SO4 and 50 ◦C. Although 10% nitric acid concentration can dissolve approximately 33% of zinc [19], this leaching agent must be discarded due to the large amount of iron dissolved (it is preferable that the iron remain in the dust to recycle it as raw material).

Novel leaching agents has been used for the treatment of EAF, BF and BOF dust as is shown in Table 2. Organic acid as butyric, citric or iminodiacetic have been studied to extract zinc from basic oxygen sludges or dust and electric arc furnace dust by using the coordination reaction between the organic ligand and zinc ions [20–23]. According to results shown in the table, in some case high Zn removal are obtained (up to 100%) with low Fe loss. The table also shows the alkaline leaching (hydroxide and carbonate solutions) that has been studied due to the insolubility of the ferric in these medium [24,25]. However, pre-treatment or high concentration of lixiviant is necessary to obtain a good Zn extraction (>70%).


**Table 2.** Leaching studies for Zn removal using different aqueous medium.

It is worth mentioning that some processes based on chlorides were developed in the past by the industry. In the Hoogovens Ijmuiden and Delft process [26,27] the EAF dust is leached under oxidant conditions at approximately 140 ◦C with HCl to dissolve zinc and a small quantity of iron. This process only considers the Zn dissolution from zinc oxides, but not the presence of zinc ferrite. The limited amount of iron dissolved can be achieved controlling the pH of the solution to limit the attack of the ferric oxides and leaching with O2 to get enough oxidation conditions. A similar process was developed by Terra de Gaia, which is based on the dissolution of ZnO and ZnFe2O4 by FeCl3–HCl at 175 ◦C [28]. The dust is mixed with the solution of chloride, reacting with the dust of iron and producing gas chloride, then it is injected into the autoclave at 175 ◦C. The other one was the Ezinex process de Engite Impianti [29] that uses ammonium chloride as aqueous media at 80 ◦C. The zinc oxide in the dust was dissolved but the zinc ferrite was not perceptibly attacked.

Most of the studies mentioned so far, have been focused on the removal and recovery of zinc, mainly in the electric arc furnace. Fewer studies have been done to treat basic oxygen furnace and blast furnace dust/sludge. For the case of the alkalis, the problem has been addressed since the operation conditions of the blast furnace perspective. Practically no study addresses the simultaneous removal and recovery of zinc and alkalis, respectively.

Considering the previous developments and the most recent studies, it could be concluded that the issue of the reduction of zinc and alkalis in iron and steelmaking dusts is still relevant for the industry due to the iron-containing values and the environment aspects. The aim of the present investigation was to evaluate the elimination of zinc and alkali in a BFS using H2SO4, HCl and NH4Cl solutions and employing different oxidizing agents (i.e., ferric chloride, ferric sulfate, oxygen and ozone) at 27 and 80 ◦C. Eh–pH diagrams were added to support the experimental results obtained.

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

#### *2.1. Material*

Samples of blast furnace sludge (BFS) were obtained from the wet-off gas cleaning system of an iron blast furnace (the collection system of the blast furnace dusts consists of three devices: Cyclone, baghouse and wet-off gas cleaning or wet scrubber; the samples were obtained from the latter). The sample was dried (110 ◦C by 24 h) and crumbled manually. The particle size was determined by wet sieving, using a Tyler sieve.

The BFS sample was analyzed by atomic absorption spectroscopy (AAS) using a Thermo Electron Solaar S4 spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA, USA); Ca, K and Na (reported as CaO, K2O and Na2O, respectively) (Sigma-Aldrich Inc., San Luis, MO, USA), a 0.2% *w*/*v* Cl3La and 0.1% *w*/*v* CsCl (Sigma-Aldrich Inc., San Luis, MO, USA)were added to prevent reduction of the signal by aluminum and as an ionization buffer, respectively. Mg and Al (reported as MgO and Al2O3) were measured under more oxidative conditions using N2O/acetylene mixed gases. SiO2 retained in the solid was determined by gravimetry (oxidation with HClO4). Determination of sulphur and total carbon was carried out using a LECO CS-244 analyzer (LECO, St. Joseph, MI, USA). Complementary chemical analyses were performed using X-Ray fluorescence (XRF) using a Brucker AXS S4 PIONEER spectrophotometer (Brucker, Billerica, MA, USA), following the Li-tetraborate fused bead technique. X-ray diffraction analysis was performance using a Brucker D8 diffractometer (Brucker, Billerica, MA, USA) and monochromatized CuKa radiation. Scanning electron microscopy analysis (Hitachi 5500 microscope; Hitachi, Tokyo, Japan) was realized using BFS powder dispersed in ethanol, sonicated and deposited on the carbon coated grid.

#### *2.2. Experimental Procedure*

Experimental tests were performed using a reactor (glass vessel) (Pyrex glass Vessel; Corning Inc., New York, NY, USA) with mechanic agitation. The BFS was mixed with a solution previously prepared, which could be a strong acid (0.1 M H2SO4 and 0.2 M HCl, pH = 0.9 and 1.1, respectively) or weak acid (0.2 M NH4Cl, pH = 5), prepared in deionized pure water (conductivity = 0.06 μS). The material and solution were poured in the reactor and were mixed using an agitation rate of 500 rpm. Solutions 0.1 M of ferric ion as ferric chloride (FeCl3 reagent grade, 97%, Sigma-Aldrich or ferric sulphate (Fe2(SO4)3, reagent grade 97%, Sigma-Aldrich), oxygen (industrial grade, 99.5%) and ozone (1% *v*/*v* O3 in O2, generated by a Pacific Ozone L22; Pacific Ozone Technology, Benicia, CA, USA) were used to have an oxidant condition in the aqueous media. Oxygen and ozone were injected to the solution through a bubbler positioned at the bottom of the reactor. The treatment duration was of 60 minutes, at room temperature (27 ◦C) and 80 ◦C.

Table 3 shows the experimental design used in this work. The variables studied were three different leaching media: (1) NH4Cl, (2) HCl and (3) H2SO4; three different oxidant agents: (1) Ferric ion (FeCl3 for NH4Cl and HCl, and Fe2(SO4)3 for H2SO4, respectively), (2) oxygen and (3) ozone; and two different temperature levels: (1) 27 ◦C and (2) 80 ◦C.

Samples of the solid material were withdrawn as well as samples of the aqueous solution. Chemical analyses were carried out in the collected samples to determine the final composition of solid residues and the barren solution.


**Table 3.** Experimental design implemented to study the different aqueous media, oxidant agents and temperature.

#### **3. Results**

#### *3.1. BFS Characterization*

Table 4 shows the chemical composition and the main mineral species contained in the BFS. This reconstruction was based in the chemical analysis and the X-ray diffraction analysis (Figure 1) and elemental analysis by Scanning Electron Microscopy/Energy Dispersive X-Ray Spectroscopy (SEM/EDS) (Hitachi 5500 microscope, Hitachi, Tokyo, Japan), Figure 2.

**Figure 1.** X-Ray diffraction pattern obtained for the BFS.

**Figure 2.** SEM/EDS micrograph of the BFS.



Figure 1 shows the XRD pattern of the initial sample, indicating that the residue is formed by multiple compounds. We want to show only those of Fe to determine its phases and their possible applications. As can be seen, iron metallic, wustite (FeO) and magnetite (Fe3O4) the main phases observed.

Figure 2 presents the SEM image showing the size and morphology of BFS particles. These microstructures exhibit spherical-shaped particles and some side-striped eroded particles with octahedral shape, which result due to intense erosion that occurs during transportation through the dust collection system. On the other hand, can be observed that SEM shows a particle size below one micron.

Element mapping by scanning electron microscopy shown in Figure 3 indicate that zinc and alkalis are integrated to iron compounds (metallic and iron oxides). It is clear that BFS can have a homogeneous chemical composition, which suggests they are formed by multi-metallic oxides complex where iron, silicon, zinc, alkalis and minor elements could have joined by sintering due to high temperatures and atmospheric conditions prevailing during the extraction process. It is worth noting that liberated franklinite or sodium and potassium oxides were not observed.

**Figure 3.** SEM element mapping of BFS.

### *3.2. Removal of Zinc and Alkali from Blast Furnace Dust by Di*ff*erent Experimental Conditions*

The following results show the removal of zinc and alkali obtained treating the steelmaking process BFS at different experimental conditions. Figure 4 shows the extraction of Zn and alkalis (as Na2O and K2O) for each of the tests carried out. The extraction was obtained from the balance of the chemical analysis of the aqueous solutions and solid residues sampling in the tests. The results showed in the figure indicate that there are great differences in the zinc and alkalis contents obtained with each experimental treatment.

Figure 4 also shows that the best results for zinc removal (>70%) were obtained when ozone and sulphuric acid were used. When other oxidants were used, it was not observed a significant decrease in the extraction of zinc. Exception case was obtained when NH4Cl was used, with a zinc extraction of 40%, due possibly to the dissolution of zinc oxides. It is interesting to note that the zinc removal was not observed when ferric ions were used. Possibly, the quantity of iron contained in the sample, mainly composed by hematite, under acid conditions have already conditions oxidant, but that are not enough for it to separate and to remove the zinc ferrites contained in the BFS sample. These results are in accordance with the conclusions made by other researchers [16,17], although the Zn/Fe ratio of the samples was different. It is important bear in mind that a smaller Zn/Fe ratio diminishes the effect of the ion ferric as an oxidant.

**Figure 4.** Zinc and alkalis (Na2O and K2O) removal from BFS at different experimental tests.

Figure 5 shows the remaining zinc and alkalis (Na2O + K2O) content after the BFS leaching tests. In this figure it can be appreciated that the removal of zinc varied greatly depending on the conditions of the aqueous media and the oxidant employed, as well as with the temperature. The results indicate that it was possible to obtain a zinc removal of 0.45% to less than 0.1%, when the sulfuric acid and the ozone were used. In these tests, a removal of alkalis of 75% and 60% for K2O and 30% and 20% for Na2O, respectively, were obtained.

**Figure 5.** Zn and alkalis content in the BFS at different experimental conditions.

It is interesting to note in Figure 5 that there were conditions allowing at the same time the removal of zinc and a significant decreasing of the alkali content. They were precisely those performed in an aqueous media that allowed dissolution of the sulphate-type compounds.

However, the results shown in Figure 6 indicate that there is a reduction in the iron containing in BFS, of 41.6 (original sample) to a minimum of 36%. However, this loss in iron is smaller than the one that would be presented if the oxidizing conditions were not present. These conditions allow the maintenance of the stability of the most acid soluble iron oxides species in the aqueous medium, also considering the low solubility of the metallic iron (the main element of the Fe compounds containing in the BFS in the low acid sulphuric concentration media). In any case, the recovery of iron of the solution is possibly increasing the pH of the aqueous solution, also under oxidant conditions, to obtain the precipitation of ferric oxide, previous separation of the zinc and alkalis dissolved in solution.

**Figure 6.** Zn and Fe content in blast furnace sludge at a different experimental test.
