Nonisothermal Investigation of Reaction Kinetics between Electric Arc Furnace Dust and Calcium Chloride under Carbon-Containing Conditions

Abstract

The volatilization of zinc in the electric arc furnace dust–CaCl₂ and ZnFe₂O₄–ZnO–CaCl₂–C reaction systems was investigated. Experiments were conducted under an N₂ atmosphere in an infrared lamp heating furnace, and the activation energy of the reaction rate of zinc volatilization was determined by the Ozawa method. The activation energy in the dust–CaCl₂ reaction system was 123 ± 27 kJ/mol; when the dust was decarburized, the activation energy was reduced to 84 ± 4 kJ/mol. Further, the simultaneous carbothermic reduction by the carbon contained in the dust and chloride volatilization of zinc improved the reaction rate. In the temperature range where carbothermic reduction and chlorination occur simultaneously, carbothermic reduction is favored. The reduction of metal oxides in the dust inhibits the chlorination and carbothermic reduction of zinc, reducing the reaction rate and activation energy.

1. Introduction

Zinc-coated steel sheets account for approximately 40% of domestic zinc demand, representing the highest share.¹⁾ Zinc-coated steel sheets are recovered as scrap steel and used as a raw material in electric furnace steelmaking. Zinc accumulates as ZnO and ZnFe₂O₄ in electric arc furnace (EAF) dust, a waste product of the steelmaking process; further, the concentration of zinc in EAF dust is approximately 30%, thus making electric arc furnaces a valuable resource for producing recycled zinc.

However, EAF dust must be properly treated to recover zinc. Most commonly, zinc is recovered using the Waelz method,²⁾ in which carbon is added to the dust and the zinc is recovered as zinc vapor by carbothermic reduction. However, in this process, the zinc, originally existing as an oxide, is reduced and then re-oxidized in the air, to be recovered as zinc oxide. Thus, this process is energetically inefficient;³⁾ further, the process requires an excess of carbon, which raises environmental concern.

Researchers have thus aimed to develop new methods of zinc recovery. In one such method, the zinc in the dust is chlorinated to produce zinc chloride and recovered by vaporization. The chlorinated zinc can then be converted to metallic zinc by electrolysis, and the residue is expected to be reused in steelmaking because of its high iron concentration. Although prior researchers have investigated this method using chlorine gas^4,5,6,7,8,9) and PVC,^10,11,12) this study aims to recover zinc from dust using calcium chloride, which is easier and safer to handle. Sato and Okumura¹³⁾ reported zinc oxide can be reduced by carbon in EAF dust and proposed a zinc recovery method combining carbothermic reduction by carbon and chloride volatilization by calcium chloride; however, no reaction rate analyses have been performed. This work therefore aims to determine the activation energy of zinc volatilization by carbothermic reduction and chloride volatilization by a nonisothermal method. The activation energy of the reaction rate constant cannot be measured by the isothermal method. The nonisothermal method is suitable when the reaction starts during the temperature rise and the reaction progresses considerably before reaching the target temperature.

2. Thermodynamic Investigation

To predict the reaction trends, the change in standard Gibbs free energy of the expected carbothermic reduction and chloride volatilization reactions within the expected temperature range was first calculated.¹⁴⁾

2.1. Carbothermic Reduction

The change in standard Gibbs free energy was calculated for the following reactions representing the carbothermic reduction of carbon in the EAF dust in the temperature range of 1073 K–1773 K.   

$$\mathrm{Z}\mathrm{n}\mathrm{O}(\mathrm{s})+\mathrm{C}(\mathrm{s})=\mathrm{Z}\mathrm{n}(\mathrm{g})+\mathrm{C}\mathrm{O}(\mathrm{g})$$(1)

  

$$\mathrm{Z}\mathrm{n}\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_4(\mathrm{s})+\mathrm{C}(\mathrm{s})=\mathrm{Z}\mathrm{n}(\mathrm{g})+\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_3(\mathrm{s})+\mathrm{C}\mathrm{O}(\mathrm{g})$$(2)

  

$$\frac{1}{2}\mathrm{\hspace{0.17em} }\mathrm{Z}\mathrm{n}\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_4(\mathrm{s})+\mathrm{C}(\mathrm{s})=\frac{1}{2}\mathrm{\hspace{0.17em} }\mathrm{Z}\mathrm{n}(\mathrm{g})+\mathrm{F}\mathrm{e}\mathrm{O}(\mathrm{s})+\mathrm{C}\mathrm{O}(\mathrm{g})$$(3)

  

$$\frac{3}{4}\mathrm{\hspace{0.17em} }\mathrm{Z}\mathrm{n}\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_4(\mathrm{s})+\mathrm{C}(\mathrm{s})=\frac{3}{4}\mathrm{\hspace{0.17em} }\mathrm{Z}\mathrm{n}(\mathrm{g})+\frac{1}{2}\mathrm{\hspace{0.17em} }\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4(\mathrm{s})+\mathrm{C}\mathrm{O}(\mathrm{g})$$(4)

The change in standard Gibbs free energy is negative for all reactions at approximately 1373 K, as shown in Fig. 1, suggesting that the reactions proceed spontaneously. As its change in standard Gibbs free energy is the most negative, the reduction of ZnO is expected to proceed preferentially at high temperatures.

Fig. 1.

Change in standard Gibbs free energy, 1073 K–1773 K, for the reaction of ZnO and ZnFe₂O₄ with C.

2.2. Volatilization of Chloride

Next, the change in standard Gibbs free energy was calculated for the following reactions representing the chloride volatilization in the EAF dust in the temperature range of 1073 K–1773 K.   

$$\frac{1}{3}\mathrm{\hspace{0.17em} }\mathrm{Z}\mathrm{n}\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_4(\mathrm{s})+\mathrm{C}\mathrm{a}\mathrm{C}{\mathrm{l}}_2(\mathrm{l})=\frac{2}{3}\mathrm{\hspace{0.17em} }\mathrm{F}\mathrm{e}\mathrm{C}{\mathrm{l}}_3(\mathrm{g})+\frac{1}{3}\mathrm{\hspace{0.17em} }\mathrm{Z}\mathrm{n}\mathrm{O}(\mathrm{s})+\mathrm{C}\mathrm{a}\mathrm{O}(\mathrm{s})$$(5)

  

$$\mathrm{Z}\mathrm{n}\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_4(\mathrm{s})+\mathrm{C}\mathrm{a}\mathrm{C}{\mathrm{l}}_2(\mathrm{l})=\mathrm{Z}\mathrm{n}\mathrm{C}{\mathrm{l}}_2(\mathrm{g})+\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_3(\mathrm{s})+\mathrm{C}\mathrm{a}\mathrm{O}(\mathrm{s})$$(6)

  

$$\mathrm{Z}\mathrm{n}\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_4(\mathrm{s})+\mathrm{C}\mathrm{a}\mathrm{C}{\mathrm{l}}_2(\mathrm{l})=\mathrm{Z}\mathrm{n}\mathrm{C}{\mathrm{l}}_2(\mathrm{g})+\mathrm{C}\mathrm{a}\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_4(\mathrm{s})$$(7)

  

$$\frac{1}{3}\mathrm{\hspace{0.17em} }\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_3(\mathrm{s})+\mathrm{C}\mathrm{a}\mathrm{C}{\mathrm{l}}_2(\mathrm{l})=\frac{2}{3}\mathrm{\hspace{0.17em} }\mathrm{F}\mathrm{e}\mathrm{C}{\mathrm{l}}_3(\mathrm{g})+\mathrm{C}\mathrm{a}\mathrm{O}(\mathrm{s})$$(8)

  

$$\mathrm{Z}\mathrm{n}\mathrm{O}(\mathrm{s})+\mathrm{C}\mathrm{a}\mathrm{C}{\mathrm{l}}_2(\mathrm{l})=\mathrm{Z}\mathrm{n}\mathrm{C}{\mathrm{l}}_2(\mathrm{g})+\mathrm{C}\mathrm{a}\mathrm{O}(\mathrm{s})$$(9)

The calculated change in standard Gibbs free energy of reactions (5)–(7) and reactions (8)–(9) are shown in Figs. 2 and 3,¹³⁾ respectively; the change in standard Gibbs free energy was positive for all reactions, indicating that they do not proceed spontaneously in the temperature range studied. However, as the actual experiment was conducted in an open system and the gas products were immediately removed, the decreasing partial pressure of the gaseous product allows the reaction to proceed; i.e., the reaction is likely to proceed if the partial pressure of the gas product is sufficiently low, despite the positive change in standard Gibbs free energy. The smaller change in standard Gibbs free energy of the chloride volatilization of zinc compared with the chloride volatilization of iron indicates that the former occurs preferentially, allowing the iron and zinc to be separated.

Fig. 2.

Change in standard Gibbs free energy, 1073 K–1773 K, for the reaction of ZnFe₂O₄ with CaCl₂.

Fig. 3.

Change in standard Gibbs free energy, 1073 K–1773 K, for the reaction of Fe₂O₃ and ZnO with CaCl₂.

3. Experiment

3.1. Materials

The composition of the actual EAF dust used is shown in Table 1. Zinc was present in the dust in the ratio of 69:31 mass% ZnFe₂O₄:ZnO.¹³⁾ In the simulated dust, the reagents used were ZnFe₂O₄ (purity 99.5%, average particle size 180 μm), ZnO (purity 99.9%, average particle size 1 μm), and graphite (average particle size 30 μm), assuming the carbon contained in the dust. Solid CaCl₂ (purity 95.0%) was used as a chlorinating agent for zinc.

Table 1. Chemical composition of electric arc furnace dust.

Element	C	O	Si	Cl	Ca	Mn	Fe	Zn	Pb	Others
mass%	2.6	29.9	1.3	4.4	2.0	1.6	22.5	30.5	1.6	3.60

3.2. Experimental Procedure

A schematic of the experimental apparatus used, an infrared lamp heating furnace, is shown in Fig. 4. The sample was placed in an alumina crucible (OD: 14 mm, ID: 10 mm, L: 10 mm) and heated in a quartz protective tube. The ambient gas in the furnace was then fully replaced using a flow rate of 2.3 × 10⁻⁶ m³/s of N₂ gas.

Fig. 4.

Experimental apparatus.

To investigate the effect of carbon contained in the dust on the volitization of zinc by calcium chloride, a non-constant temperature method, the Ozawa method,^15,16) was used to determine the activation energy of the volatilization of zinc. Experimental samples of EAF dust with eight Zn:CaCl₂:C ratios, detailed in Table 2, were heated in air at 600°C for 1 h to gasify and remove the carbon. In this table, the values of activation energy obtained in experiments are also shown at the same time.

Table 2. Material mixing ratio and activation energy of zinc volatilization.

Materials	Zn:CaCl2:C (mol)	Activation energy (kJ/mol)
dust–CaCl2	1:1:0.464	123 ± 27
decarburized dust–CaCl2	1:1:0	84 ± 4
ZnFe2O4–ZnO–CaCl2–C	1:1:0.464	217 ± 3
	1:1:0.232	171 ± 6
	1:1:0.116	148 ± 9
	1:1:0	115 ± 12
	1:0.5:0.464	198 ± 14
	1:0.25:0.464	242 ± 2

As other substances in the EAF dust may affect the volatilization of zinc, experiments were also conducted using simulated dust. In the simulated dust, ZnFe₂O₄ and ZnO were mixed in the same ratio found in the actual dust (69:31 mass% ZnFe₂O₄:ZnO). The ratio of graphite was determined based on the desired ratio of carbon to zinc in the dust.

Each sample was heated in the infrared lamp furnace at a heating rate, β, of 0.65, 0.50, 0.35, or 0.25 K/s until reaching the target temperature. After reaction, the samples were subjected to quantitative analysis by scanning electron microscopy with energy-dispersive X-ray analysis (SEM-EDX) to calculate the reaction rate of zinc (α_Zn) as   

$${\alpha }_{\text{Zn}}=\frac{W_{\text{bef}}-W_{\text{aft}}}{W_{\text{bef}}}$$(10)

where W_bef and W_aft represent the the mass of zinc before and after the reaction, respectively.

3.3. Nonisothermal Method (Ozawa Method)¹⁵⁾

According to the Ozawa method, the model equation for the reaction is expressed as   

$$\text{mA(s)}+\text{nB(l)}=\text{pC(s)}+\text{qD(g)}$$(11)

for which the reaction rate may be expressed as   

$$\frac{d\alpha }{dt}=k(T)A(\alpha )a_A^m{a}_B^n$$(12)

here, α is reaction rate (-); A(α), a function of the reaction rate, is the reaction interface area (in m²); a_A and a_B are the activities of A and B, respectively, (for pure A and B, a_A = a_B = 1); and k(T) is the rate constant (in s⁻¹·m⁻²), whose relationship with temperature can be given by the Arrhenius equation as   

$$k(T)=f \text{exp}\left(-\frac{E}{RT}\right)$$(13)

where f is the pre-exponential constant (in s^–1·m⁻²), R is the gas constant (in J/mol·K), and E is the activation energy (in J/mol).

For a linear heating rate,   

$$T=T_0+\beta t$$(14)

  

$$\frac{dT}{dt}=\beta$$(15)

where T₀ is the initial temperature (in K) and β is the linear heating rate (in K/s).

Combining Eqs. (12), (13), and (15), and taking a_A = a_B = 1, the following equation is obtained.   

$$\frac{1}{f\cdot A(\alpha )}\frac{d\alpha }{dT}=\frac{1}{\beta }exp\left(-\frac{E}{RT}\right)$$(16)

Here, f·A(α) should be constant for a given reaction rate. In Eq. (16), if dα/dT is constant, i.e., the reaction rate increases linearly, the Ozawa method can be used for analysis. Then, Eq. (16) can be transformed into following equation.   

$$\text{ln}\beta +\frac{E}{RT}\cong \text{const}$$(17)

Thus, the relationship between lnβ and 1/T is linear, and the slope of this line represents the activation energy.

4. Results and Discussion

4.1. Electric Furnace Dust

As shown in the resulting relationship between the temperature, heating rate, and zinc reaction rate in Fig. 5, the reaction rate increased with temperature, reaching a maximum of 0.86, but decreased with the increase in heating rate. The broken lines in the figure are drawn for the range where the slopes are almost equal. In the following similar graphs, the broken lines are drawn in the same way. As shown in Fig. 5, the reaction rate increased linearly in the range of reaction rate from 0.258 to 0.710. Thus, the relationship between lnβ and 1/T is linear, as shown in Fig. 6, the slope of this line represents the activation energy and was calculated as 123 ± 27 kJ/mol. Here the reaction rates with maximum and minimum activation energy are shown.

Fig. 5.

Effect of heating rate on zinc reaction rate from dust-CaCl₂ reaction system (Zn:CaCl₂:C = 1:1:0.464 (mol)).

Fig. 6.

Relation between heating rate and temperature from dust-CaCl₂ reaction system (Zn:CaCl₂:C = 1:1:0.464 (mol)).

The resulting experimental relationship between the temperature, heating rate, and zinc reaction rates using the decarburized dust are shown in Fig. 7, here, the slope is smaller than that shown in Fig. 5, and the final reaction rate was approximately 0.2 lower. Using the Ozawa method, the calculated activation energy in the range of reaction rate from 0.287 to 0.402 was 84 ± 4 kJ/mol, as shown in Fig. 8.

Fig. 7.

Effect of heating rate on zinc reaction rate from decarbonization dust-CaCl₂ reaction system (Zn:CaCl₂:C = 1:1:0 (mol)).

Fig. 8.

Relation between heating rate and temperature from decarbonization dust-CaCl₂ reaction system (Zn:CaCl₂:C = 1:1:0 (mol)).

The high reaction rate of zinc indicates the efficacy of the proposed zinc recovery method using calcium chloride. The reaction rate decreased after decarburization, suggesting that the carbon in the EAF dust accelerated the volatilization of zinc, i.e., the volatilization of chloride by calcium chloride and the reduction volatilization by carbon occur simultaneously. This may result in a reduced consumption of calcium chloride in practical applications, thereby reducing costs.

The activation energy increased in the presence of carbon, as the activation energy represents the temperature dependence, and this increase may have been caused by the increased reaction rate of zinc.

4.2. Simulated Dust

The resulting relationship between the temperature, heating rate, and reaction rate of zinc for samples with varying carbon content are shown in Figs. 9, 10, 11, 12. Overall, the reaction rate increased with time and with an increase in carbon concentration, reaching a maximum of 0.848. The calculated activation energy when the reaction rate was between 0.326 and 0.666, 0.279 and 0.575, 0.270 and 0.508, and 0.319 and 0.536 was 217 ± 3, 171 ± 6, 148 ± 9, and 115 ± 12 kJ/mol, respectively, as shown in Figs. 13, 14, 15, 16, respectively. Thus, the activation energy also decreased with a decreasing carbon content.

Fig. 9.

Effect of heating rate on zinc reaction rate from ZnFe₂O₄–ZnO–CaCl₂–C reaction system (Zn:CaCl₂:C = 1:1:0.464 (mol)).

Fig. 10.

Effect of heating rate on zinc reaction rate from ZnFe₂O₄–ZnO–CaCl₂–C reaction system (Zn:CaCl₂:C = 1:1:0.232 (mol)).

Fig. 11.

Effect of heating rate on zinc reaction rate from ZnFe₂O₄–ZnO–CaCl₂–C reaction system (Zn:CaCl₂:C = 1:1:0.116 (mol)).

Fig. 12.

Effect of heating rate on zinc reaction rate from ZnFe₂O₄–ZnO–CaCl₂ reaction system (Zn:CaCl₂:C = 1:1:0 (mol)).

Fig. 13.

Relation between heating rate and temperature from ZnFe₂O₄–ZnO–CaCl₂–C reaction system (Zn:CaCl₂:C = 1:1:0.464 (mol)).

Fig. 14.

Relation between heating rate and temperature from ZnFe₂O₄–ZnO–CaCl₂–C reaction system (Zn:CaCl₂:C = 1:1:0.232 (mol)).

Fig. 15.

Relation between heating rate and temperature from ZnFe₂O₄–ZnO–CaCl₂–C reaction system (Zn:CaCl₂:C = 1:1:0.116 (mol)).

Fig. 16.

Relation between heating rate and temperature from ZnFe₂O₄–ZnO–CaCl₂ reaction system (Zn:CaCl₂C = 1:1:0 (mol)).

Similarly, the resulting relationships between temperature, heating rate, and zinc reaction rate for samples with reduced calcium chloride content are shown in Figs. 17 and 18, and their corresponding relationships between lnβ and 1/T are shown in Figs. 19 and 20, respectively. On the other hand, when the amount of calcium chloride used was halved, there was no change in the reaction rate. Further reducing the amount of calcium chloride to a quarter of that used originally, the reaction rate decreased significantly. The corresponding activation energy when the reaction rate ranged from 0.315 to 0.639 and 0.160 to 0.265 was 198 ± 14 and 242 ± 2 kJ/mol, respectively.

Fig. 17.

Effect of heating rate on zinc reaction rate from ZnFe₂O₄–ZnO–CaCl₂–C reaction system (Zn:CaCl₂:C = 1:0.5:0.464 (mol)).

Fig. 18.

Effect of heating rate on zinc reaction rate from ZnFe₂O₄–ZnO–CaCl₂–C reaction system (Zn:CaCl₂:C = 1:0.25:0.464 (mol)).

Fig. 19.

Relation between heating rate and temperature from ZnFe₂O₄–ZnO–CaCl₂–C reaction system (Zn:CaCl₂:C = 1:0.5:0.464 (mol)).

Fig. 20.

Relation between heating rate and temperature from ZnFe₂O₄–ZnO–CaCl₂–C reaction system (Zn:CaCl₂:C = 1:0.25:0.464 (mol)).

Grillo et al.¹⁷⁾ determined the activation energy of chlorination of zinc in the ZnO–Fe₂O₃–CaCl₂ reaction system to be 107–172 kJ/mol, whereas Kim¹⁸⁾ reported that the activation energy required of the carbothermic reduction of zinc in the ZnO–C reaction system is 224 kJ/mol. Thus, the activation energy of the proposed ZnFe₂O₄–ZnO–CaCl₂–C reaction system approaches the activation energy of chlorination from carbothermic reduction as the amount of carbon is reduced. At Zn:CaCl₂:C = 1:1:0.464 (mol), the reduction of zinc oxide by carbon is more favorable and thus requires a higher activation energy. Further decreasing the carbon concentration may cause the reaction between calcium chloride and zinc oxide to become more favorable, reflected here as a lower activation energy of chlorination. No change in the activation energy was seen when the amount of calcium chloride was cut in half, due to the higher favorability of the reaction between carbon and zinc oxide, and the activation energy of carbothermic reduction was reflected. When the amount of calcium chloride was reduced further, the carbothermic reduction had a higher activation energy and the reaction rate decreased significantly. The reaction between carbon and zinc oxide is an inter-solid reaction, in which the reaction normally stops when the particles separate from each other after reaction. However, in this experiment, the carbon particles flowed through the liquid calcium chloride and came into contact with the reactants again, thereby advancing the reaction. As the amount of calcium chloride decreases, it becomes difficult for the carbon particles to flow and for the carbothermic reduction to occur. As a result, the reaction rate decreases significantly. Thus, using calcium chloride allows the carbon contained in the dust to be used efficiently and the zinc to be removed synergistically.

4.3. Comparison of Electric arc Furnace Dust and Simulated Dust

The calculated activation energies obtained for each of the material ratios are summarized in Table 2. A similar decrease in activation energy with the decrease in carbon content was observed for both the dust–CaCl₂ and ZnFe₂O₄–ZnO–CaCl₂–C reaction systems. However, under the same ratios of zinc, calcium chloride, and carbon, the activation energy of the dust–CaCl₂ reaction system was lower than that of the ZnFe₂O₄–ZnO–CaCl₂–C reaction system, possibly owing to the influence of other substances in the EAF dust. EAF dust comprises various metals, many of which are present as oxides; here, the carbon and calcium chloride may have reacted with other metal oxides, thereby inhibiting the volatilization reaction of zinc. It can be understood that the activation energy is reduced if the reaction of the metal oxides excluding zinc oxide occurs in advance and the reaction product acts as a catalyst.

5. Conclusions

By investigating the reaction rate of zinc by chlorination of EAF dust using calcium chloride, the following results were obtained.

(1) The activation energy of the volatilization of zinc during chlorination of EAF dust was 123 ± 27 kJ/mol. After decarburization, the activation energy was reduced to 84 ± 4 kJ/mol.

(2) The simultaneous carbothermic reduction by the carbon contained in the dust and chloride volatilization of zinc improved the reaction rate and suppressed the consumption of calcium chloride, which may lead to reduced costs.

(3) In the temperature range where carbothermic reduction and chlorination of zinc occur simultaneously, the carbothermic reduction of zinc was favored.

(4) Adding calcium chloride was demonstrated to efficiently utilize the carbon originally contained in the dust and synergistically remove zinc when compared with the conventional Waelz method, which requires the addition of excess carbon.

(5) The reduction of metal oxides other than zinc in the EAF dust likely inhibits the chlorination and carbothermic reduction of zinc, reducing the reaction rate and activation energy of the zinc volatilization.

References

• 1)  JOGMEC: Mineral Resources Information, Mineral resource material flow 2019, http://mric.jogmec.go.jp/wp-content/uploads/2020/05/material_flow2019.pdf, (accessed 2021-02-19) (in Japanese).
• 2)   H.  Sugimoto,  H.  Takizawa and  H.  Ueda: J. MMIJ, 123 (2007), 723 (in Japanese).
• 3)   T.  Yoshida: Shigen-to-Sozai, 113 (1997), 967 (in Japanese).
• 4)   F.  Santos,  E.  Brocchi,  V.  Araújo and  R.  Souza: Metall. Mater. Trans. B, 46 (2015), 1729.
• 5)   H.  Matsuura,  T.  Hamano and  F.  Tsukihashi: ISIJ Int., 46 (2006), 1113.
• 6)   H.  Matsuura,  T.  Hamano and  F.  Tsukihashi: Mater. Trans., 47 (2006), 2524.
• 7)   H.  Matsuura and  F.  Tsukihashi: Metall. Mater. Trans. B, 37 (2006), 413.
• 8)   C. A.  Pickles: J. Hazard. Mater., 166 (2009), 1030.
• 9)   G.  Sun,  H.  Matsuura and  F.  Tsukihashi: Mater. Trans., 58 (2017), 59.
• 10)   B.  Zhang,  X. Y.  Yan,  K.  Shibata,  T.  Uda,  M.  Tada and  M.  Hirasawa: Mater. Trans., JIM, 41 (2000), 1342.
• 11)   N.  Menad and  Bo.  Björkman: Resour. Conserv. Recycl., 24 (1998), 257.
• 12)   G.-S.  Lee and  Y. J.  Song: Miner. Eng., 20 (2007), 739.
• 13)   H.  Sato and  K.  Okumura: Tetsu-to-Hagané, 106 (2020), 858 (in Japanese).
• 14)   A.  Roine: Outokumpu HSC Chemistry for Windows, Ver. 5.1, Outokumpu Research Oy, Pori, Finland, (2002).
• 15)   T.  Ozawa: Bull. Chem. Soc. Jpn., 38 (1965), 1881.
• 16)   L.  Hong,  K.  Okumura and  M.  Sano: Metall. Mater. Trans. B, 30 (1999), 1003.
• 17)   F. F.  Grillo,  J. L.  Coleti,  D. C. R.  Espinosa,  J. R.  Oliveira and  J. A. S.  Tenório: Mater. Trans., 55 (2014), 351.
• 18)   B. S.  Kim,  J. M.  Yoo,  J. T.  Park and  J. C.  Lee: Mater. Trans., 47 (2006), 2421.