**3. Experimental**

#### *3.1. Materials and Instrumentation*

EAF dust samples were kindly supplied by local smelters around Jakarta and Banten Province, Indonesia. A sieving test showed that more than 90% of dust passed through a 270 mesh (53 μm), and this size fraction was used in further leaching tests. All leaching experiment and characterization studies were performed in the Research Unit for Mineral Technology—Indonesian Institute of Sciences, Bandar Lampung, Indonesia, or otherwise stated.

XRF characterization (Panalytical X'Pert 3 Powder with Omnian Standard) was carried out to determine the major chemical components in EAF dust (Table 2), and shows the most dominant constituents to be zinc and aluminum. The mineralogical phases were determined using X-ray powder diffraction (Panalytical, Expert3 Powder). Quantitative determination of metal contents in EAF dust (Table 3) and pregnant leach solution after leaching for recovery and separation factor calculation were carried out using atomic absorption spectrophotometry (AAS, Shimadzu AA7000, Japan) and Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES, Analytik Jena, Plasma Quant 9000 Elite, Germany). SEM-EDS characterization to investigate the change of morphology and elemental distribution on the surface of EAF dust grain before and after leaching was performed using a SEM (Hitachi SUM 3500, Japan) at the Research Unit for Natural Product Technology—Indonesian Institute of Sciences, Yogyakarta, Indonesia. The advanced mineral identification and characterization system (AMICS) analysis to determine and map the mineral phases was carried out by Eagle Engineering, Butte, Montana, USA.

Chemicals such as sodium hydroxide, sulfuric acid, nitric acid, hydrochloric acid, and standard solutions were obtained from Merck, Darmstadt all in analytical grade, while monosodium glutamate was produced by PT Ajinomoto Indonesia (99% purity) and used as received. The solution pH was adjusted using dilute sulfuric acid or sodium hydroxide and monitored using a pH meter (Oakton 45, Vernon Hills, IL, USA), while deionized water (MilliQ) was used throughout the experiment.


**Table 3.** Chemical composition of important base metals in EAF dust (after aqua regia digestion and AAS/ICP-OES determination).


#### *3.2. Leaching Procedure*

Leaching studies were carried out using a batch method. In general, 1 g of EAF dust was mixed with 20 mL lixiviant in 250 mL sealed flask. The mixture was homogenized using an orbital shaker at 200 rpm. After the leaching was completed, the supernatant solution was separated using centrifugation and filtration (Whatman 42). The metal concentration was determined using AAS or ICP-OES. Metal recovery (*R*) was calculated using Equation (1). To evaluate the selectivity of the leaching scheme, a separation factor (SF) between Zn or Cu and other metals was evaluated using Equation (2). All leaching data were obtained in duplicates.

$$R = \frac{C\_E \times V}{C\_o \times m} \times 100\% \tag{1}$$

where:

*CE* Zn or Cu concentration in supernatant solution (mg/L) *Co* Zn or Cu content in EAF dust (mg/g) *m* mass of EAF dust used in leaching (g) *V* leaching agent volume (L)

$$SF = \frac{\text{C}\_{\text{Z}}}{\text{C}\_{\text{M}}} \tag{2}$$

where:

*CZ*, Zn or Cu concentration in pregnant leach solution (mg/L). *CM*, other metal concentration in pregnant leach solution (mg/L).

#### **4. Results and Discussion**

#### *4.1. Characterization*

EAF dust characterization using XRD showed the mineral phase composition was dominated by zinc oxide (ZnO). XRD of the dust residue obtained after leaching (leaching condition pH 9, MSG concentration 1 M, pulp density 50 g/L and 12-hour homogenization) revealed the phases were dominated by spinel (Mg-Al oxide), calcium aluminum oxide and zinc oxide phases (Figure 2). These elements such as Al, Mg, Ca and Fe were found to be relatively retained in the solid phase during the leaching process. The XRD characterization results are supported by AMICS analysis results (Supplementary file: Tables S1 and S2 and Figure S2), which revealed the major Zn phase to be ZnO and gahnite (ZnAl2O4) in original material.

**Figure 2.** Phases present in EAF dust (**a**) before and (**b**) after leaching.

The change in elemental distribution of EAF dust before and after leaching was also explored by SEM-EDS characterization (Figure 3). Surface morphology of EAF dust changed from porous and grainy to relatively smooth. Intensity decrease of Zn and Cu peaks on the EDS profile after leaching confirmed the efficacy of MSG as lixiviant in EAF dust leaching. Qualitative mapping on the surface reveals that elements such as Al, Fe and Ca were relatively retained and even enriched after leaching (Table 4), indicating the selective nature of the leaching using MSG.

**Figure 3.** Morphology (**a**) before and (**b**) after leaching, including EDS intensities of elements based on mapping on the respective field (**c**) before and (**d**) after leaching.


**Table 4.** Qualitative mapping of elements by EDS on grain surface before and after leaching.

#### *4.2. E*ff*ect of pH*

The effect of pH on the Zn and Cu leaching efficiency and selective recovery was studied between 6 and 11. The constant variables included MSG concentration 1 M, pulp density 50 g/L and an agitation time of 12 h at room temperature. pH ranges between 6–11 were chosen according to the modeling in Section 2 in order to suppress the solubility of other elements (Mg, Ca, Al and Fe). The leaching efficiency and leaching selectivity are shown in Figures 4 and 5, respectively.

Based on Figure 4, the optimum pH to recover Zn and Cu are 9. Low recovery at lower pH was due to the weaker or repulsive interaction between glutamate and Zn or Cu since the glutamate species was dominated as a protonated species, e.g., H3Glu<sup>+</sup>, HGlu−. Lower recovery at pH higher than 9 was probably due to the hydrolysis of Zn and Cu, which was lower compared to the modeling results using Visual Minteq (pH > 12 for Zn and pH > 13 for Cu). In the case of other elements, all but Al extraction decreased as pH became alkaline. Increasing recovery of Al in alkaline conditions (0.5% at pH 11) was due to the amphoteric characteristic of Al, which is soluble in excess of alkali as Al(OH)4 −.

Increasing Al recovery at higher pH and decreasing Zn or Cu contributed to the sharp decrease of the selectivity factor of Zn or Cu toward Al (Figure 5). The figure also shows the optimum pH to separate Zn and Cu are 9 and 10, respectively. The highest separation factor is obtained toward iron, which could reach an order of 10<sup>4</sup> (with Zn) and 400 (with Cu). Lower separation factors were obtained between Zn or Cu and Mg and Ca (reaching an order of 700).

**Figure 4.** The recovery of (**a**) Zn, Cu, Mg and Ca and (**b**) Fe and Al as function of pH.

**Figure 5.** Separation factor of (**a**) Zn: [Zn]/[Fe] (squares), [Zn]/[Al] (diamonds), [Zn]/[Mg] (circles), [Zn]/[Ca] (triangles) and (**b**) Cu: [Cu]/[Fe] (squares), [Cu]/[Al] (diamonds), [Cu]/[Mg] (circles), [Cu]/[Ca] (triangles) in pregnant leach solution as function of pH.

#### *4.3. E*ff*ect of MSG Concentration*

The effects of MSG concentration to metal recovery and Zn and Cu selectivity over other elements were studied between 0.1 M and 2 M, with constant variables pH 9, pulp density 50 g/L and 12 h leaching time at room temperature. The results in Figure 6 show that MSG concentration has positive effects on metal recovery. Based on modeling using Visual Minteq (Supplementary file: Figure S3), the minimum MSG concentration to completely solubilize Zn is 0.7 M, and at lower MSG concentration, Zn speciation is dominated by hydroxide precipitate. In the case of Cu, the model shows that MSG concentration has little effect on speciation, which contradicts the results in Figure 6. This is probably due to competition with Zn, which has a 40 times larger concentration. The leaching process at higher MSG concentration increased the recovery of other elements, e.g., Fe, due to increasing soluble species (FeGlu<sup>+</sup> complex) (Supplementary file: Figure S3). The results in Figure 6 show the optimum recovery of Zn and Cu are attained at 1.2 and 0.7 M, respectively, which at higher MSG concentration did not significantly increase the recovery of both metals. At higher MSG concentration, the recovery of other metals, especially Mg and Ca became substantial, decreasing the selectivity (Figure 7). The optimum MSG concentration based on data in Figure 6 is well supported by results in Figure 7.

**Figure 6.** The recovery of (**a**) Zn, Cu, Mg, Ca and Fe and (**b**) Al as function of MSG concentration.

**Figure 7.** Separation factor of (**a**) Zn: [Zn]/[Fe] (squares), [Zn]/[Al] (diamonds), [Zn]/[Mg] (circles), [Zn]/[Ca] (triangles) and (**b**) Cu: [Cu]/[Fe] (squares), [Cu]/[Al] (diamonds), [Cu]/[Mg] (circles), [Cu]/[Ca] (triangles) in pregnant leach solution as function of MSG concentration.

#### *4.4. E*ff*ect of Pulp Density*

To evaluate the effect of pulp density (ratio between lixiviant and EAF dust) on the recovery and selectivity, leaching was carried out at constant variables pH 9, MSG concentration 1 M, 12 h leach at room temperature, while pulp density was varied between 33 and 200 g/L. The results in Figure 8 show the recovery decreased as pulp density increased. Maximum pulp density to obtain optimum recovery of Zn and Cu were 50 and 100 g/L, respectively. In the case of Mg and Ca the optimum pulp density was 66.67 g/L, while in the case of Fe and Al, the recovery was very low (less than 1 percent).

The plot of separation factor value as a function of pulp density (Figure 9) reveals that pulp density has a negative effect on the separation factor. This was probably caused by the increase of solubility of elements especially Mg, Ca, Fe and Al at larger volumes in alkaline conditions. This also indicates the leaching of EAF dust using MSG as lixiviant in alkaline conditions should be performed using continuous methods, e.g., column leaching or heap leaching instead of a batch method, since in the continuous method the volume could be minimized to optimize the selectivity.

**Figure 8.** The recovery of (**a**) Zn, Cu, Mg and Ca and (**b**) Fe and Al as function of pulp density.

**Figure 9.** Separation factor of (**a**) Zn: [Zn]/[Fe] (squares), [Zn]/[Al] (diamonds), [Zn]/[Mg] (circles), [Zn]/[Ca] (triangles) and (**b**) Cu: [Cu]/[Fe] (squares), [Cu]/[Al] (diamonds), [Cu]/[Mg] (circles), [Cu]/[Ca] (triangles) in pregnant leach solution as function of pulp density.

#### *4.5. Kinetic Studies (E*ff*ect of Leaching Time and Temperature)*

Kinetic studies were carried out at a time range up to 12 h at three different temperatures (30, 55 and 80 ◦C), with constant variables pH 9, MSG concentration 1 M and pulp density 50 g/L. The results in Figure 10 show that maximum leaching efficiency was attained within 120 min for Zn and 4 h for Cu at 30 ◦C. At higher leaching temperature, recovery reached saturation in a shorter period. Figure 10b shows the recovery of Cu is delayed about 1 h. Based on Eh monitoring, in the early stage of leaching the aqueous phase was reductive, which inhibited the oxidation and complexation of Cu by lixiviant. As leaching progressed, the aqueous phase became oxidative, favoring the oxidation and complexation of Cu.

To describe the kinetic process of leaching, three kinetic models were tested to model experimental data: shrinking core model (chemical reaction control), shrinking particle model (film diffusion) and interface transfer and diffusion by Dickinson and Heal (1999) [36]. The fitting results (*R2*, coefficient of correlation) of each model are listed in Table 5.

**Figure 10.** The effect of leaching time and temperature on recovery of (**a**) Zn and (**b**) Cu. In b, the change of Eh in aqueous phase during leaching was obtained at 30 ◦C (single measurement).


**Table 5.** Correlation coefficient for experimental data fitting using three kinetic models (SCM, shrinking core model; SPM, shrinking particle model).

*k*, apparent rate constant (min<sup>−</sup>1); *t*, time (min); *R*, metal recovery.

Based on the coefficient of the correlation value in Table 5, the interface transfer and diffusion model is the best model to describe the experimental data. Although the shrinking core model and shrinking particle model are more popular models to explain kinetic data, both models are generally based on the assumption of constant lixiviant concentration during leaching [37]. In the leaching process, the concentration of glutamate would decrease due to complex formation. In this study, the model developed by Dickinson and Heal (1999) [36] was adequately applied to analyze the dissolution of Zn and Cu in EAF dust using glutamate.

The value of the apparent rate constant (*k*) obtained for each temperature using the interface transfer and diffusion model (Table 6) was used to determine the activation energy (*Ea*, kJ/mol) using the Arrhenius Equation (3), where *A*, *T*, and *R* are the frequency factor, temperature, and gas constant, respectively. Relatively low energy activation for both Zn and Cu indicates that the effect of temperature on the leaching process is minor.

$$k = Ae^{-\frac{E\_a}{RT}} \text{ or } ln = lnA - \frac{E\_a}{RT} \tag{3}$$

**Table 6.** Apparent rate constant (*k*) attained from fitting of experimental data using interface transfer and diffusion model, which was used to obtain the activation energy (*Ea*) using linear regression, including the coefficient of correlation (*R2*).


#### *4.6. Monosodium Glutamate Recovery*

Due to the significant amount of MSG used in leaching and to sustain the leaching scheme, the regeneration of glutamate is required. Glutamate could be recovered from the pregnant leach solution as precipitate by acidifying the solution to the pH between 2–4.5. In this pH range glutamate is precipitated as neutral species glutamic acid, H2Glu (Supplementary file: Figure S4). The addition of more acid (pH less than 1) caused the formation of soluble cationic species H3Glu<sup>+</sup>. The experiment performed to recover glutamate from the pregnant leach solution showed that the optimum pH was 3, which in this pH more than 90% of the glutamate was recovered as glutamic acid (Figure 11). The recovery efficiency of glutamate was determined gravimetrically.

**Figure 11.** Monosodium glutamate (MSG) recovery efficiency as function of pH.

#### **5. Conclusions**

Monosodium glutamate effectively and selectively recovered Zn and Cu from EAF dust, based on batch leaching studies according to leaching efficiency and separation factor values. The optimum conditions for the leaching scheme are pH 9, lixiviant concentration 1 M and pulp density 50 g/L. Studies on the effect of pulp density on the metal recovery and separation factor showed that pulp density correlated negatively to the metal recovery and positively to the separation factor of Zn and Cu to the other elements. This indicates the leaching is better performed using a continuous method, considering the selective recovery of Zn and Cu from other elements. Kinetic studies showed the leaching efficiency reached a saturation value in less than 2 and 4 h for Zn and Cu, respectively. The activation energy obtained from the experimental data modeling revealed the effect of temperature on the leaching process was minor. Further the use of MSG as lixiviant offers a sustainable leaching scheme since MSG is recoverable from the pregnant leach solution and reusable for the next leaching cycle.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2075-4701/10/5/644/s1, Figure S1: Species distribution of (a) Al, (b) Fe, (c) Mg and (d) Ca in glutamate-H2O system as function of pH. Glutamate concentration 1 M, Al2<sup>+</sup> 350 mM, Fe3<sup>+</sup> 10 mM, Mg2<sup>+</sup> 10 mM and Ca2<sup>+</sup> 20 mM, Table S1: AMICS Mineralogy for Zinc EAF Dust Sample, Figure S2: Identified AMICS Minerals for Zinc EAF Dust Sample, Table S2, AMICS Color Scheme, Figure S3: Species distribution of (a) Zn and (b) Fe in glutamate-H2O system as function of glutamate concentration at pH 9. Zn2<sup>+</sup> 300 mM and Fe3<sup>+</sup> 10 mM, Figure S4: Species distribution of glutamate in glutamate-H2O system as function of pH. Glutamate concentration 1 M.

**Author Contributions:** Conceptualization, E.P.; data curation, E.P.; formal analysis, E.P. and C.A.; funding acquisition, E.P. and C.A.; investigation, E.P. and C.A.; methodology, E.P.; project administration, F.R.M., E.P.; resources, E.P., C.A., F.N., M.A.M., A.S.H., F.R.M. and F.B.; validation, E.P.; visualization, M.A.M. and F.B.; writing–original draft, E.P.; writing–review & editing, E.P. and C.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded and supported by Indonesian Institute of Sciences FY 2019, Fulbright Program through Visiting Scholar Scheme 2019–2020, Indonesian Ministry of Research and Technology Insinas Program FY 2020, PRN (National Priority Research) of Indonesia FY 2020.

**Acknowledgments:** We would like to express our gratitude to the two reviewers for improving this manuscript.

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

#### **References**


© 2020 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/).
