*3.3. 3* × *3* × *2 Experimental Design and Analysis of Variance*

The effect of the individual experimental variables on the Zn, Na2O and K2O removal was examined using the statistical analysis of data. Analysis of variance (ANOVA) of the experimental tests data at different conditions was used to evaluate the effect of each individual variable.

Table 5 shows the main effects, as determined by ANOVA, using NCSS software (NCSS 2000 software, NCSS, LLC, Kaysville, UT, USA) [30]. The table indicates the values of degree freedom (DF), sum of squares (SS), media of squares (MS), Fisher ratio (F), probability level (Prob Level) and the probability that a false null hypothesis can be rejected (Power) with a 95% level of confidence (α = 0.05). According to F, Prob Level and Power values, ANOVA shows that under the studied conditions, leaching media and the type of oxidant were the most important factors during the extraction of zinc. Results also indicated that within the analyzed range, the three variables studied had the lower effect in Na2O and K2O removal.


**Table 5.** Analysis of variance (ANOVA) for Zn, Na2O and K2O removal from blast furnace sludge.

\* α = 0.05.

These results are graphically shown in Figure 7, which shows the box graph for the individual effects of the experimental parameters or variables. Each plot shows the box with the minor and major values, the average value and percentiles for each parameter. Furthermore, the comparison between different oxidants shows that for the case of zinc, the ozone could generate slightly more favorable conditions for increased extraction of zinc BFS. For alkalis, although the average removal of these compounds between the three oxidants was similar, under certain conditions the ferric ion allowed a better removal. These observations are discussed below.

**Figure 7.** Box graphs for the individual effects of leaching media and oxidant in Zn and alkalis removal.

Figure 8 shows the average response curves for the individual parameters to evaluate the effect of the ozone (considering that the ferric ion has the least effect in the Zn removal). The graphs were constructed with data from the tests that used the ozone and oxygen. The most significant comparative results were found for the oxidant (see Figure 8a), which Zn removal increased from 16% to 40% using oxygen and the ozone, respectively. However, alkalis removal was not affected by the change of oxidant. On the other hand, Fe loss increased from 4% to 9% when the ozone was employed. In the case of the type of aqueous solution, Figure 8b shows that the best result was obtained with H2SO4 where Zn removal increased from 20% and 18% to 80% using NH4Cl, HCl and H2SO4 respectively. Alkalis removal and Fe loss also increased when sulfuric acid was employed. In the case of temperature (see Figure 8c), Zn removal decreased by increasing temperature from 27 ◦C to 80 ◦C, while alkalis and Fe increased. Generally, in a leaching process, at higher temperature the velocity of dissolution is increased. However, an increment of temperature diminishes the absorption of ozone in the solution. In this case, it is possible that the effect of temperature in the reaction of dissolution were annulled, by which a change in the removal of zinc was observed.

According to results, the best conditions to remove up to 70% of Zn and alkalis are using a solution of 0.1 M H2SO4 and ozone (1% *v*/*v*) as the oxidant, at 25 ◦C. Comparing with conditions shown in Table 1, the sulphuric acid concentration used in the present work was very low. Besides the additional advantage of using ambient temperature, it also has low loss of iron during leaching. Obtaining a zinc extraction, like those obtained in the studies cited in the table.

**Figure 8.** Response curves of individual effects: (**a**) Oxidant, (**b**) type of aqueous solution and (**c**) temperature.

#### **4. Discussion**

The fundament of zinc and alkalis dissolution into strong or weak acid media using ozone or oxygen is that aqueous media react with the oxides of zinc and alkalis, incorporating them with the media as aqueous ions, just like it is shown by the average chemical analysis of the BFS with and without ozone treatment.

The presence of an oxidizing agent, like oxygen or ozone, allows the creation of severe oxidizing conditions in aqueous media, keeping stable and practically insoluble the ferric oxides. This oxidizing environment also creates the necessary conditions to transform oxides or ferrites of zinc and alkalis, as sodium and potassium oxides, to sulfates, chloride or chlorates, compounds with high solubility in an oxidizing aqueous media.

Eh–pH diagram [31] for a Zn–O–S system (see Figure 9a), shows that zinc, which is in the form of zinc ferrite in the BFS, reacts in acid aqueous medium (pH < 5) under oxidant conditions (Eh > 0) to produce soluble species. According to the results discussed above, the aqueous medium appropriate is the sulfuric acid and ozone as an oxidizing agent. The following reactions can be possible:

**Figure 9.** Eh–pH diagrams. (**a**) Zn–O–S system and (**b**) Fe–O–S system; under standard conditions at 25 ◦C [31].

In the absence of the ozone:

$$\text{ZnFe}\_2\text{O}\_4 + \text{H}\_2\text{SO}\_4 \rightarrow \text{Fe}\_2\text{O}\_3 + \text{ZnSO}\_4\\(\text{a}) + \text{H}\_2\text{O} \rightarrow \Delta\text{G}^\circ\_{20}\text{ }\_\text{C} = -107 \text{ kJ}. \tag{1}$$

In the presence of the ozone:

$$\text{ZnFe}\_2\text{O}\_4 + \text{H}\_2\text{SO}\_4 + 2\beta\text{O}\_3(\text{g}) \rightarrow \text{Fe}\_2\text{O}\_3 + \text{ZnSO}\_4(\text{a}) + \text{H}\_2\text{O} + \text{O}\_2(\text{g}) \rightarrow \Delta\text{G}^\circ\_{20}\text{ }^\circ\_{\text{C}} = -215 \text{ kJ}. \tag{2}$$

ΔG◦ <sup>20</sup> ◦<sup>C</sup> values were obtained from HSC Chemistry 6 [31].

According to the Eh–pH diagram for a Fe system (Figure 9b), under very acid and oxidant conditions, iron oxides could be dissolved. Then, low concentrations of acid (pH > 5) decrease iron leaching. In the case of iron metallic or iron species as magnetite or wustite, they could dissolve to a pH between 1 and 5 (weak acid conditions). However, an oxidant acid media (Eh potential > 0.5 for pH > 2) could inhibit iron dissolution.

The alkalis contained in BFS are found as ferrites or sodium (potassium)-iron jarosites. These compounds could react in different aqueous medium to obtain soluble species. When an acid medium and oxidant agent is used, the alkalis can be dissolved and separated from the BFS. The possible reactions involved are:

In the absence of the ozone:

$$\text{[K]}\text{Na}\_2\text{O}^\*\text{Fe}\_2\text{O}\_3 + \text{H}\_2\text{SO}\_4 \rightarrow \text{[K]}\text{Na}\_2\text{SO}\_4\\\text{(a)} + \text{Fe}\_2\text{O}\_3 + \text{H}\_2\text{O} \rightarrow \text{Al}\text{C}^\circ\_{20} \cdot \text{C} = -277 \text{ kJ}.\tag{3}$$

In the presence of the ozone:

$$\text{[K]}\text{Na}\_2\text{O}^\*\text{Fe}\_2\text{O}\_3 + \text{H}\_2\text{SO}\_4 + 2\text{\%O}\_3 \rightarrow \text{[K]}\text{Na}\_2\text{SO}\_4\\\text{(a)} + \text{Fe}\_2\text{O}\_3 + \text{H}\_2\text{O} + \text{O}\_2 \rightarrow \Delta\text{G}^\circ\_{20}\text{}^\circ\_{\text{C}} = -385\text{ kJ}. \quad \text{(4)}$$

According to the results, the removal of alkalis only was possible in the test where high oxidation conditions were employed, obtaining the best results when the ozone was employed as the oxidizing agent. The use of other oxidizing agents, as oxygen or ferric chloride, did not decrease the zinc and alkalis content in the blast furnace sludge.

The Zn/Fe and [K]Na/Fe ratios in the BFS allow zinc and alkalis to be as stable species (e.g., oxides/ferrites) under alkaline and acid non-oxidizing conditions, being necessary to increase the potential of oxidation of the aqueous medium to obtain the dissolution of these compounds. The use of the Fe3<sup>+</sup> ion could inhibit the Zn dissolution. This is possible if some Fe dissolution from zinc ferrite can be necessary to zinc leaching. Then, an initial Fe concentration in solution could decrease the ferrite dissolution due to equilibrium between the Fe concentration in products. In the case of oxygen or, better, the ozone, the oxidant condition for zinc ferrite and alkalis removal is possible.

### **5. Conclusions**

The high content of undesirable elements (alkalis and zinc) in iron blast furnace sludge (with respect to the 0.03% Zn and 0.12% alkalis required by iron and steelmaking industry) generated a loss of valuable units of iron, additionally to the potential environment hazards and the cost by handling and disposal/confinement. In accordance with the results presented, the BFS recycling by hydrometallurgical process is possible. The best conditions were obtained when the BFS was subjected to an acid leach process with sulfuric acid (0.1 M H2SO4), employing ozone (1% *v*/*v* O3) to obtain high oxidizing conditions. Only under these conditions, an important removal of zinc (up to 85%) and alkalis (up to 35% Na2O and 75% K2O) was possible to decrease the content of these compounds in the BFS. The fact that similar results were obtained under ambient temperature and at 80 ◦C indicates a compensation of two factors; the increment in the iron dissolution rate under high temperature and the diminishing of the oxidizing conditions due to the ozone decomposition. Although the zinc and alkalis concentration are still high according to the maximum values recommended, their lower content would allow it to increase the percentage of BFS added to the raw material of the blast furnace. On the other hand, the Fe loss (up to 13%) is still important, if blast furnace sludge volume is considered. Then, a Fe recovery treatment of the leaching solutions (e.g., precipitation) will be required.

**Author Contributions:** Data curation, F.R.C.P.; Formal analysis, F.R.C.P.; Funding acquisition, M.d.J.S.-A.; Investigation, M.d.J.S.-A., N.P.-R. and J.R.-C.; Methodology, F.R.C.P. and J.R.-C.; Project administration, M.d.J.S.-A.; Supervision, F.d.J.L.-S.; Validation, G.I.D.-P. and A.A.G.-I.; Visualization, F.d.J.L.-S.; Writing—original draft, F.R.C.P.; Writing—review & editing, G.I.D.-P. and A.A.G.-I.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors would like to thank to International Center for Nanotechnology and Advanced Materials–Kleberg Advanced Microscopy Center–at the University of Texas at San Antonio (ICNAM-UTSA) for their technical assistance with the SEM characterization.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
