Reduction of Volatile Impurities in Zinc Chloride Melt with Metallic Iron and Its Effect on Vacuum Separation Behavior

Abstract

Metallic Zn production currently generates toxic gases and residues that cannot be recycled, and consumes a large amount of energy. To develop more efficient process using chlorination reaction of Zn resources, purification methods for crude ZnCl₂ melt have been developed with limited success. In this study, we have investigated a new type of pyrometallurgical purification that combines reduction reactions of FeCl₃ and CuCl₂ in ZnCl₂ melt using metallic Fe with vacuum distillation. Metallic Fe reduced FeCl₃ and CuCl₂ into FeCl₂ and CuCl/Cu, respectively, which show lower vapor pressures than ZnCl₂. Vacuum distillation of the crude ZnCl₂ melt after the Fe reduction successfully produced a high-purity ZnCl₂ deposit with less contamination by Fe and Cu. This study reveals that reduction treatment with metallic Fe converts FeCl₃ and CuCl₂ into less volatile substances and suppresses contamination by such substances into ZnCl₂ recovery.

1. Introduction

Producing metallic Zn from sulfide concentrates now requires environmentally friendly processes with low energy consumption, replacing the current process that generates toxic SO₂ gas and uses aqueous solution electrolysis with a large amount of electrical energy. Recycling Zn from electric arc furnace (EAF) dust also requires high separation efficiency between Zn and Fe unlike Waelz kiln, which generates CO₂ gas in carbothermic reduction but shows the incomplete recovery ratio for Zn. To solve these problems, chlorination processes for such Zn resources have been the focus of researches.^1–24) Chlorinating Zn concentrates produces ZnCl₂ and elemental S rather than SO₂ gas,^1–6) whereas chlorinating EAF dust completely separates Zn from Fe-based clinker.^7–24) Molten salt electrolysis of this ZnCl₂, which consumes lower energy than electrowinning from an aqueous solution, extracts liquid Zn to be easily transported without being stripped from the cathode.¹⁷⁾

Producing high-purity Zn requires purification of ZnCl₂ before the electrolysis, which is conducted through oxidation with air,^25,26) chlorination with Cl₂ gas,²⁷⁾ distillation through gas phase,^17,28,29) reduction with Zn metal,^17,30–33) and pre-electrolysis.³⁰⁾ The oxidation method is effective only for Fe chlorides in ZnCl₂ melt, and requires separation of formed Fe oxides and the melt, which is technically difficult especially at high temperature. The chlorination method converts FeCl₂ in ZnCl₂ melt into FeCl₃, which volatilizes from the melt, but a part of ZnCl₂ gas simultaneously volatilizes. The distillation method evaporates ZnCl₂ more preferentially than many chlorides and oxides; however, the separability is low under atmospheric pressure. The reduction method, what is called cementation in hydrometallurgy, removes nobler metallic ions than Zn as metallic forms, which also requires separation of the reduced elements and the molten salt. We have demonstrated that vacuum distillation separates ZnCl₂ from chlorides with lower vapor pressure like FeCl₂, PbCl₂, CuCl, CdCl₂, and MnCl₂, and produces a purified ZnCl₂ deposit.²⁹⁾ The concentrations of FeCl₂ in the purified ZnCl₂ deposits decreased to around 0.2 mol%, lower than the initial one of around 5 mol%. The vacuum distillation at a low temperature of 700 K for a short period of 30 min has produced a highly purified ZnCl₂ deposit with 5 mol ppm FeCl₂. Our study has also shown the good separability between ZnCl₂ and CuCl, which is unexpected because of their relatively close vapor pressures, suggesting that ZnCl₂ may suppress CuCl evaporation from the melt.

Transition metals typically have multiple ionic valences, such as Fe²⁺/Fe³⁺ and Cu⁺/Cu²⁺. Chlorinating Zn resources under high Cl₂ chemical potential generates chlorides composed of these high-valent ions like FeCl₃ and CuCl₂; however, few previous studies have directly included separation of ZnCl₂ and FeCl₃/CuCl₂. These chlorides show high vapor pressures as shown in Fig. 1, indicating that they would contaminate ZnCl₂ vapor during vacuum distillation.^34–36) FeCl₂ has an ionic structure with a high melting point of 950 K and a high boiling point of 1293 K, whereas FeCl₃ has a molecular structure with a low melting point of 577 K and a low boiling point of 604 K.^36,37) Vacuum distillation separating ZnCl₂, more volatile substances, and the others requires a multistep process using more than two distillation towers or an equipment for exhaust gas treatment, which inevitably increases the capital investment and operating cost. Nanjo et al. has proposed reducing FeCl₃ into FeCl₂ with metallic Fe to purify AlCl₃ obtained by chlorinating bauxites.³⁷⁾ Cooling the AlCl₃–FeCl₂ melt settles the formed FeCl₂ and the remaining Fe in the bottom of the melt, which produces a purified AlCl₃ melt. To separate a trichloride mixture of NdCl₃ and PrCl₃, Uda et al. has proposed reducing NdCl₃ into NdCl₂ with lower vapor pressure using metallic Nd, followed by vacuum distillation of the NdCl₂–PrCl₃ mixture.³⁸⁾

Fig. 1

Vapor pressure of several chlorides and metals.^34–36) p_i and p° represent vapor pressure of chemical species i in atm and a standard pressure of 1 atm, respectively.

To separate ZnCl₂ from FeCl₃ and CuCl₂ and recover high-purity ZnCl₂, this study focuses on reduction reactions of FeCl₃ and CuCl₂ in ZnCl₂ melt using metallic Fe, which is cheaper than any other metals, followed by vacuum distillation of the formed ZnCl₂–FeCl₂/Fe–CuCl/Cu mixture. First, we have thermodynamically analyzed these Fe reduction reactions in ZnCl₂ melt and investigated the reactions experimentally. Then, we have conducted vacuum distillation for such ZnCl₂ melt treated by the Fe reduction.

2. Thermodynamic Analysis

2.1 Fe reduction of FeCl₃ in ZnCl₂ melt

Metallic Fe reduces FeCl₃ into FeCl₂ following reaction (1), which is a reverse reaction of a disproportionation reaction.   

\begin{equation} \text{2 FeCl$_{3}$ (l)} + \text{Fe (s)} = \text{3 FeCl$_{2}$ (s)} \end{equation} (1)

  

\begin{equation*} \Delta_{\text{r1}}G^{\circ} = \text{$-248.5{}\,$kJ/mol at 600$\,$K}.^{34)} \end{equation*}

  

\begin{equation*} \Delta_{\text{r1}}G^{\circ} = \text{$-236.7{}\,$kJ/mol at 700$\,$K}.^{34)} \end{equation*}

Standard states of FeCl₃, Fe, and FeCl₂ are liquid FeCl₃, α-Fe, and solid FeCl₂, respectively. Reduction temperature should be as low as only slightly above a melting point of ZnCl₂ (591 K)^34,36) to avoid volatilization of FeCl₃ before the reduction. The standard Gibbs energy changes of reaction (1) at 600 to 700 K are largely negative, indicating the spontaneous progress of reaction (1). Notably, Fe content in chlorides increases by up to 1.5 times through reaction (1).

Figure 2 shows three-dimensional chemical potential diagrams of the Fe–Zn–Cl₂ system at (a) 600 K and (b) 700 K.^34,39) Dashed lines and dashed dotted lines correspond to an FeCl₂ activity of 0.01 and an FeCl₃ activity of 0.01, respectively, roughly assuming that they dissolve in ZnCl₂ melt. FeCl₃ is stable under high Cl₂ partial pressure and can coexist with ZnCl₂ at 600 K and 700 K. Metallic Fe is stable under low Cl₂ partial pressure. Through reaction (1), an FeCl₃/ZnCl₂ equilibrium state shifts to an Fe/FeCl₂/ZnCl₂ equilibrium state. The formed FeCl₂ dissolves into ZnCl₂ melt up to solubilities of 4 mol% at 600 K and 23 mol% at 700 K.⁴⁰⁾ This dissolution contributes to not only decreasing the FeCl₂ activity but also removing solid FeCl₂ from a surface of the metallic Fe, leading to the continuous progress of reaction (1).

Fig. 2

Three-dimensional chemical potential diagrams for the Fe–Zn–Cl₂ system at (a) 600 K and (b) 700 K.^34,39) Γ, Γ₁, δ, and ζ phases are Fe–Zn intermetallic compounds.

When excess Fe reacts with FeCl₃, eq. (2) determines the FeCl₃ activity under an Fe/FeCl₂ equilibrium.   

\begin{align} \log (a_{\text{FeCl${_{3}}$(l)}}) &= \frac{3}{2}\log (a_{\text{FeCl${_{2}}$(s)}}) - \frac{1}{2}\log (a_{\text{Fe(s)}}) \\ &\quad + \frac{\Delta_{\text{r1}}G^{\circ}}{2(\ln 10)RT} \end{align} (2)

Here, R is the gas constant. Pure Fe can exist and the activity is unity. Similarly, reactions (3) and (4) determine FeCl₃ partial pressure and Fe₂Cl₆ partial pressure under an Fe/FeCl₂ equilibrium, respectively.   

\begin{equation} \text{2 FeCl$_{3}$ (g)} + \text{Fe (s)} = \text{3 FeCl$_{2}$ (s)} \end{equation} (3)

  

\begin{equation*} \Delta_{\text{r3}}G^{\circ} = \text{$-308.3{}\,$kJ/mol at 600$\,$K}.^{34)} \end{equation*}

  

\begin{equation*} \Delta_{\text{r3}}G^{\circ} = \text{$-276.2{}\,$kJ/mol at 700$\,$K}.^{34)} \end{equation*}

  

\begin{equation} \text{Fe$_{2}$Cl$_{6}$ (g)} + \text{Fe (s)} = \text{3 FeCl$_{2}$ (s)} \end{equation} (4)

  

\begin{equation*} \Delta_{\text{r4}}G^{\circ} = \text{$-248.9{}\,$kJ/mol at 600$\,$K}.^{34)} \end{equation*}

  

\begin{equation*} \Delta_{\text{r4}}G^{\circ} = \text{$-230.6{}\,$kJ/mol at 700$\,$K}.^{34)} \end{equation*}

The FeCl₃ activity, the FeCl₃ partial pressure, and the Fe₂Cl₆ partial pressure decrease as the FeCl₂ activity decreases in ZnCl₂ melt as shown in Fig. 3.³⁴⁾ Vacuum distillation for such ZnCl₂ melt generates ZnCl₂ vapor with less contamination by Fe chlorides.

Fig. 3

Activity of FeCl₃ and dimensionless partial pressures of FeCl₃ and Fe₂Cl₆ equilibrated with α-Fe and FeCl₂ at (a) 600 K and (b) 700 K under 1 atm.³⁴⁾

2.2 Fe reduction of CuCl₂ in ZnCl₂ melt

Metallic Fe reduces CuCl₂ into Cu following reaction (5).   

\begin{equation} \text{CuCl$_{2}$ (s)} + \text{Fe (s)} = \text{Cu (s)} + \text{FeCl$_{2}$ (s)} \end{equation} (5)

  

\begin{equation*} \Delta_{\text{r5}}G^{\circ} = \text{$-145.8{}\,$kJ/mol at 600$\,$K}.^{34)} \end{equation*}

  

\begin{equation*} \Delta_{\text{r5}}G^{\circ} = \text{$-147.5{}\,$kJ/mol at 700$\,$K}.^{34)} \end{equation*}

Standard states of CuCl₂ and Cu are solid CuCl₂ and solid Cu, respectively. The standard Gibbs energy changes of reaction (5) at 600 to 700 K are largely negative, indicating the spontaneous progress of reaction (5). Cu content in ZnCl₂ melt decreases through the reduction.

Figure 4 shows three-dimensional chemical potential diagrams of the Cu–Fe–Cl₂ system at (a) 600 K and (b) 700 K.³⁴⁾ Dashed dotted lines correspond to a CuCl₂ activity of 0.01 or a CuCl activity of 0.01, roughly assuming that they dissolve in ZnCl₂ melt. Pure Cu can coexist with pure Fe at 600 K and 700 K.^41,42) CuCl₂ is stable under high Cl₂ partial pressure and can coexist with ZnCl₂, FeCl₂, and FeCl₃ at 600 K and 700 K. Metallic Cu is stable under low Cl₂ partial pressure. Through reaction (5) with excess Fe, a CuCl₂/FeCl_n/ZnCl₂ equilibrium state shifts to an Fe/Cu/FeCl₂/ZnCl₂ equilibrium state.

Fig. 4

Three-dimensional chemical potential diagrams for the Cu–Fe–Cl₂ system at (a) 600 K and (b) 700 K.³⁴⁾

A stable region of CuCl in Fig. 4 is large, implying that CuCl partly remains in ZnCl₂ melt even under Fe coexistence. Equation (6) determines the CuCl activity under an Fe/Cu/FeCl₂ equilibrium.   

\begin{equation} \text{2 CuCl (s)} + \text{Fe (s)} = \text{2 Cu (s)} + \text{FeCl$_{2}$ (s)} \end{equation} (6)

  

\begin{equation*} \Delta_{\text{r6}}G^{\circ} = \text{$-54.55{}\,$kJ/mol at 600$\,$K}.^{34)} \end{equation*}

  

\begin{equation*} \Delta_{\text{r6}}G^{\circ} = \text{$-51.95{}\,$kJ/mol at 700$\,$K}.^{34)} \end{equation*}

The formed CuCl dissolves into ZnCl₂ melt up to solubilities of 53 mol% at 600 K and 97 mol% at 700 K.^43,44)

Reactions (5) and (6) determine CuCl₂ activity and CuCl activity under an Fe/Cu/FeCl₂ equilibrium, respectively. The CuCl₂ activity and the CuCl activity decrease as the FeCl₂ activity decreases in ZnCl₂ melt as shown in Fig. 5.³⁴⁾ Fe reduction of CuCl₂ produces metallic Cu, whereas the partial reduction generates CuCl, which remains in ZnCl₂ melt. The vapor pressure of CuCl is lower than those of ZnCl₂ and CuCl₂. Vacuum distillation for such ZnCl₂ melt containing CuCl produces ZnCl₂ vapor with less contamination by Cu chlorides.

Fig. 5

Activities of CuCl₂ and solid CuCl equilibrated with α-Fe, solid Cu and FeCl₂ at 600 K and 700 K under 1 atm.³⁴⁾

3. Experimental

3.1 Fe reduction experiments

Figure 6 shows a schematic diagram of an experimental apparatus used for the Fe reduction of FeCl₃ and CuCl₂ in ZnCl₂ melt. Table 1 shows the initial sample information and the experimental conditions. Fe powder (99.9%, <45 µm, FUJIFILM Wako Pure Chemicals Corporation) or Fe wires (99.5%, Φ 1 mm, The Nilaco Corporation), FeCl₃ powder (99%, Kojundo Chemical Laboratory Co., Ltd.), and CuCl₂ (98%, Nacalai Tesque, Inc.) were placed in a PYREX test tube crucible (O.D. 15.8 mm × I.D. 14.2 mm × H. 100 mm) in a globe box filled with Ar gas. The Fe wire was ultrasonically cleaned in the order of pure water, ethanol, and acetone in advance. With its hygroscopicity, ZnCl₂ reagent usually contains water and oxides. The reagent ZnCl₂ powder (99.9%, FUJIFILM Wako Pure Chemicals Corporation) was firstly purified by vacuum distillation at about 763 K and then dried at 393 K under vacuum for more than 48 hours to remove these impurities from ZnCl₂, although some of the impurities apparently remained in the ZnCl₂. This purified ZnCl₂ was put on the sample in the crucible. This crucible and three PYREX collecting tubes (O.D. 16.0 mm × I.D. 13.6 mm × H. 50 mm) were placed in a closed-end PYREX supporting tube (O.D. 20.0 mm × I.D. 17.6 mm × H. 300 mm). These PYREX supporting tubes were placed in a closed-end quartz reaction tube (O.D. 50 mm × I.D. 46 mm × L. 500 mm), and silica-gel-dehydrated Ar gas (99.9999%) was introduced by 100 cm³/min. Temperature was raised to 600 K or 700 K using a mantle heater, and then kept at these temperatures. After a prescribed time, the reaction tube was removed from the heater and cooled in air.

Fig. 6

Schematic diagram of an experimental apparatus used for Fe reduction of FeCl₃ and CuCl₂ in ZnCl₂ melt.

Table 1 Information about initial crude ZnCl₂ samples containing FeCl₃ and CuCl₂ and experimental conditions for Fe reduction.

3.2 Vacuum distillation experiments

Figure 7 shows an experimental setup used for Fe reduction and vacuum distillation of ZnCl₂ melt containing FeCl₃ and CuCl₂. Table 2 shows the initial sample information and the experimental conditions. Height of a PYREX test tube crucible containing a sample was 75 mm. To efficiently recover deposits, three PYREX collecting tubes (O.D. 16.0 mm × H. 25 mm, I.D. (top) 13.6 mm, I.D. (bottom) 6–8 mm) were piled up just above the crucible as shown in Fig. 7. After Fe reduction at 600 K for 6 h, which was conducted by the same process mentioned in the previous section, the temperature was raised to 700 K, and then the inside of the reaction tube was evacuated with an oil rotary pump to decrease the internal pressure to 1–5 Pa. To examine an effect of Fe reduction on the behaviors of evaporation and deposition during vacuum distillation, samples without metallic Fe were prepared as references.

Fig. 7

(a) Schematic diagram of an experimental setup used for Fe reduction and vacuum distillation of ZnCl₂ melt containing FeCl₃ and CuCl₂, and (b) a top view, (c) a side view, and (d) a schematic illustration of a recovery tube for deposit.

Table 2 Information about initial crude ZnCl₂ samples containing FeCl₃ and CuCl₂ and experimental conditions for Fe reduction and vacuum distillation.

3.3 Analysis

A residue in the crucible was all decomposed in pure water ultrasonically to dissolve only chlorides in the residue. Suction filtration using a membrane filter with a pore size of 0.45 µm (0.025 µm for Sam. No. D-1-1) was conducted to separate an insoluble residue from the solution. One milliliter of nitric acid was added in the filtrate to prevent hydrolysis. A deposit on the recovery tube was all dissolved in a mixture of hydrochloric acid, nitric acid, and pure water. Concentrations of metal elements in the solution were determined by inductively coupled plasma optical emission spectrometry (ICP-OES) (PS7800, Hitachi High-Tech Corporation). These Fe concentrations were corrected because it was confirmed in the preliminary analysis that about 3% of the initial amount of metallic Fe powder passed through a membrane filter with a pore size of 0.45 µm.

After dried in a desiccator, the water-insoluble residue in the crucible was separated into magnetic and non-magnetic materials with a magnet. Crystalline phases of the powder materials were identified using an X-ray diffractometer (XRD) (MiniFlex600, Rigaku Corporation). The wire material in the crucible was cut using a tabletop cutter, and then embedded in phenolic resin (PolyFast, Struers). The cross section was polished with SiC grinding papers and then buffed with polycrystalline diamond sprays. The cross section of the wire sample was observed using a scanning electron microscope (SEM) (JSM-6010LA, JEOL). The elemental distribution was analyzed using an energy dispersive X-ray spectrometer (EDS) equipped to the SEM.

4. Results and Discussion

4.1 Fe reduction of FeCl₃ in ZnCl₂ melt

Table 3 shows the experimental results about residues in crucibles after reduction treatment using metallic Fe under various experimental conditions. Figure 8 shows Fe concentrations in the crude ZnCl₂ melt after the reduction treatment. Reducing FeCl₃ in ZnCl₂ melt through reaction (1) increases the Fe concentration from an initial one of 1.72 mass% to up to its 1.5 times of 2.58 mass%. The reduction treatments using metallic Fe powder mostly increased the Fe concentrations of the crude ZnCl₂ melt. When 3.0 times of the stoichiometric amount of Fe powder was added to the ZnCl₂–FeCl₃–CuCl₂ melt, the Fe concentration increased to 2.38 mass% after 6 hours of reaction at 600 K (Sam. No. D-5-2), indicating the progress of reaction (1). In the current experimental conditions, the progress of reaction (1) may result in the production of FeCl₂ exceeding its solubility in ZnCl₂ melt at 600 K (4 mol%),⁴⁰⁾ which might prevent Fe metal covered by the undissolved FeCl₂ from reacting with FeCl₃ that remains in the ZnCl₂ melt. In contrast, when Fe wires were added, the Fe concentrations in the crude ZnCl₂ melt decreased (Sam. No. D-1-2 and Sam. No. D-4-2), implying that not only reaction (1) was inhibited but also Fe chlorides were partly removed from the molten salt.

Table 3 Experimental results about crude ZnCl₂ melt samples as residues in crucibles after reduction treatment using metallic Fe.

Fig. 8

Fe concentration in crude ZnCl₂ melt after reduction treatment using metallic Fe. Labels show temperature (600 or 700 K), Fe form (P: powder or W: wire), ratio to stoichiometric amount (1.5 or 3.0), and CuCl₂ addition or not (Cu).

Volatilization or oxidation of unreacted FeCl₃ loses Fe chlorides from the molten salt. Some of the FeCl₃ evaporate from the surface of the ZnCl₂ melt without being reduced by Fe metal located at the bottom of the molten salt because of its extremely high vapor pressure at 600 or 700 K. Ar gas stream transfers the evaporated FeCl₃ to the low-temperature area above the crucible. Assuming that the weight differences of the initial samples and the residues were attributed only to the volatilization of FeCl₃ from the crucibles, about 16% of FeCl₃ in Sam. No. D-5-2 (T_R = 600 K) volatilized, in which Fe powder was used as the reductant. In contrast, about 30% of FeCl₃ in Sam. No. D-1-2 (T_R = 600 K) and about a half of FeCl₃ in Sam. No. D-4-2 (T_R = 700 K) volatilized, in which Fe wires were used as the reductants. Figure 9 shows XRD patterns of the water-insoluble residues after the reduction treatments using Fe wires, showing the formation of FeOCl at 600 K and Fe₂O₃ at 700 K. Comparison of Sam. No. D-1-2 with Sam. No. D-1-1, and Sam. No. D-4-2 with Sam. No. D-4-1 in Table 3 shows that the masses of residues after the reduction treatments using Fe wires were larger than those using Fe powder. A chemical potential diagram of the Fe–Cl₂–O₂ system depicted by Zhang et al. shows a stable region of FeOCl at low temperature such as 573 K.⁹⁾ The XRD measurement shows that FeOCl may form during reaction at 600 K or during cooling. As mentioned in Chapter 3, ZnCl₂ can contain water and O²⁻ ion, which might act as the oxidant based on our previous analysis.²⁹⁾ Figure 10 shows XRD patterns of the water-insoluble residues after the reduction treatments using Fe powder. The magnetic samples contain the original Fe powder, and then the XRD patterns show large peaks corresponding to α-Fe. The XRD patterns also show that FeOCl and Fe₂O₃ form at 600 K and 700 K, respectively.

Fig. 9

XRD patterns of water-insoluble non-magnetic residues in (a) Sam. No. D-1-2 and (b) Sam. No. D-4-2 after reduction treatments using Fe wires.

Fig. 10

XRD patterns of water-insoluble residues in (a), (b) Sam. No. D-4-1, (c) Sam. No. D-5-1, and (d) Sam. No. D-5-2 after reduction treatments using Fe powder.

Figure 11 shows SEM images and elemental mapping by EDS analysis of the Fe wire in Sam. No. D-1-2. Zn-enriched layers, which are blurred in the SEM images because of its charge-up, formed around the original Fe wires, and the Zn distribution partly overlapped with the O distribution, showing the formation of metallic Zn layers or ZnO layers. Reducing ZnCl₂ with metallic Fe forms metallic Zn through reactions (7) and (8).   

\begin{equation} \text{ZnCl$_{2}$ (l)} + \text{Fe (s)} = \text{Zn (s)} + \text{FeCl$_{2}$ (s)} \end{equation} (7)

  

\begin{equation*} \Delta_{\text{r7}}G^{\circ} = \text{58.2$\,$kJ/mol at 600$\,$K}.^{34)} \end{equation*}

  

\begin{equation} \text{ZnCl$_{2}$ (l)} + \text{Fe (s)} = \text{Zn (l)} + \text{FeCl$_{2}$ (s)} \end{equation} (8)

  

\begin{equation*} \Delta_{\text{r8}}G^{\circ} = \text{58.6$\,$kJ/mol at 700$\,$K}.^{34)} \end{equation*}

Oxidation of such Zn layers seems to form ZnO layers. Coating a Fe wire with Zn or ZnO reduces the reaction interface between metallic Fe and the molten salt. This coating also requires diffusion of Fe in the Zn layers or Fe³⁺ ions in the ZnO layers, which can decrease the overall reaction rate; thus, more FeCl₃ volatilize or are oxidized to FeOCl or Fe₂O₃ without being reduced.

Fig. 11

(a) SEM image and elemental mapping by EDS analysis of an Fe wire in Sam. No. D-1-2 after reduction: (b) Fe, (c) Zn, (d) O, and (e) Cl.

4.2 Fe reduction of CuCl₂ in ZnCl₂ melt

Figure 12 shows Cu concentrations in the crude ZnCl₂ melt after the reduction treatment under various experimental conditions. Reducing CuCl₂ in ZnCl₂ melt through reaction (5) decreases the Cu concentration from an initial one of 0.47 mass%. When 3.0 times of the stoichiometric amount of Fe powder was added to the ZnCl₂–FeCl₃–CuCl₂ melt, the Cu concentration decreased to 0.0785 mass% after 6 hours of reaction at 600 K (Sam. No. D-5-2), indicating the progress of reaction (5). Figure 10 also shows the formation of metallic Cu in the water-insoluble residues. When Fe wires were added, the Cu concentrations in the crude ZnCl₂ melt remained almost constant from the initial one (Sam. No. D-4-2), implying that reaction (5) was inhibited.

Fig. 12

Cu concentration in crude ZnCl₂ melt after reduction treatment using metallic Fe. Labels show temperature (600 or 700 K), Fe form (P: powder or W: wire), ratio to stoichiometric amount (1.5 or 3.0), and CuCl₂ addition or not (Cu).

4.3 Vacuum distillation after Fe reduction

Table 4 shows the experimental results about crude ZnCl₂ melt samples as residues in crucibles after reduction treatment using Fe powder and vacuum distillation. The masses of the molten salt remaining in the crucibles decreased significantly after the vacuum distillation because a large amount of ZnCl₂ volatilized from the crucibles. After 0.5 h of vacuum distillation at 700 K in Exp. No. D-7, the Fe concentration in the molten salt residue with Fe reduction treatment (Sam. No. D-7-1) was larger than that without Fe reduction treatment (Sam. No. D-7-2). The Cu concentration in the molten salt residue with Fe reduction treatment (Sam. D-7-1) was smaller than that without Fe reduction treatment (Sam. D-7-2). After 2 h of vacuum distillation at 700 K in Exp. No. D-8, the Fe concentration in the Fe-reduced molten salt residue (Sam. No. D-8-1) increased, whereas that in the residue without being reduced (Sam. No. D-8-2) decreased. The Cu concentrations in both residues increased through the longer vacuum distillation.

Table 4 Experimental results about crude ZnCl₂ melt samples as residues in crucibles after reduction treatment using Fe powder and vacuum distillation.

Figure 13 shows deposits on recovery tubes for Exp. No. D-8. The heights of the recovery tubes, H (cm), show vertical distances from the bottom of the crucibles. A clump of the slightly yellowish white chloride deposited on the recovery tube placed at heights of 11.5–13.5 cm from the bottom of the crucible for Sam. No. D-8-1. Table 5 shows the experimental results about the deposits on the recovery tubes after reduction treatment using Fe powder and vacuum distillation. Figure 14 shows the weight distribution of (a) Fe and (b) Cu in the deposits. After 0.5 h of vacuum distillation at 700 K in Exp. No. D-7, regardless of Fe addition, a part of FeCl₃ moved onto the recovery tubes located at the low-temperature region, resulting from volatilization of unreacted FeCl₃. The amount of FeCl₃ moving to the low-temperature region without Fe reduction was larger than that after Fe reduction. After 2 h of vacuum distillation at 700 K in Exp. No. D-8, the Fe-enriched deposits were obtained in the low-temperature region with or without Fe reduction, but their Fe amounts were slightly larger than the results of vacuum distillation for 0.5 h (Exp. No. D-7). These results suggest that the volatilization process of unreacted FeCl₃ and its migration process to the low-temperature region ended in the early stage of distillation. Purified ZnCl₂, which is defined as the heaviest deposit, contained 0.119 mass% Fe with Fe reduction (Sam. No. D-8-1) and 0.0335 mass% Fe without Fe reduction (Sam. No. D-8-2). The vacuum distillation without Fe reduction produced higher-purity ZnCl₂ than that with Fe reduction; however, the larger amount of FeCl₃ transported to the lower-temperature regions above where the heaviest deposit was obtained. This result demonstrates that reducing FeCl₃ in ZnCl₂ melt with metallic Fe can suppress FeCl₃ volatilization and the contamination of purified ZnCl₂ deposit during vacuum distillation. The purified ZnCl₂ deposit shows a lower Cu concentration of 0.00233 mass% with Fe reduction than 0.0122 mass% without Fe reduction, suggesting that the Fe reduction pretreatment is effective for separation of ZnCl₂ and CuCl₂ through vacuum distillation.

Fig. 13

Deposits on recovery tubes for Exp. No. D-8. The heights of the recovery tubes, H (cm), show vertical distances from the bottom of the crucibles.

Table 5 Experimental results about deposits on recovery tubes after reduction treatment using Fe powder and vacuum distillation.

Fig. 14

Weight distribution of (a) Fe and (b) Cu in deposits after reduction treatment using Fe powder and vacuum distillation. H (cm): Height from bottom of crucibles.

Figure 15 shows a SEM image and EDS analysis results of the residue for Sam. No. D-8-1. The Cu-enriched phases exist as needles composed of many polygonal Cu grains, which contain 3.0 mol% Fe at most. The Zn concentration was below detection limit (<0.01 mol%). Interestingly, this finding implies the possibility of metal synthesis using reduction treatment in molten salt.

Fig. 15

(a) SEM image and elemental mapping by EDS analysis of a water-insoluble magnetic residue in Sam. No. D-8-1 after vacuum distillation: (b) Fe, (c) Zn, (d) Cu, (e) O, and (f) Cl. (g) Enlarged SEM image and EDS point analysis of the frame in (a).

5. Conclusions

In this study, reduction reactions of ZnCl₂–FeCl₃–CuCl₂ (Fe: 1.72 mass%; Cu: 0.47 mass%) melt with metallic Fe were analyzed thermodynamically and investigated experimentally. Reducing FeCl₃ into FeCl₂ in the crude ZnCl₂ melt with metallic Fe increased the Fe concentration in the melt. Reducing CuCl₂ in the crude ZnCl₂ melt decreased the Cu concentration in the melt and formed metallic Cu. In vacuum distillation after the Fe reduction, a high-purity ZnCl₂ deposit with a low Fe concentration of 0.119 mass% and a very low Cu concentration of 0.00233 mass% was successfully recovered. Although vacuum distillation without Fe reduction also a produced purified ZnCl₂ deposit with a lower Fe concentration of 0.0335 mass%, the total amount of FeCl₃ transferred to low-temperature regions was larger. However, in these experiments, the reduction with metallic Fe proceeded incompletely and volatilization of unreacted FeCl₃ contaminated ZnCl₂ vapor during the vacuum distillation. Our future work to improve the reduction efficiency and to optimize the vacuum distillation process is necessary.

Acknowledgements

We gratefully acknowledge Professor Kazuo Terashima, Professor Kazuki Morita, Professor Toru H. Okabe, and Lecturer Kazuaki Kato at The University of Tokyo to their suggestive comments and fruitful discussion on our study. This research was supported by the 59th (2018) Toray Science and Technology Grant of Toray Science Foundation.

REFERENCES

• 1)  E.A. Ashcroft: British Patent 10 827, (1897).
• 2)   R.L.  Herrick: Mines Miner. 30 (1910) 370–374.
• 3)   S.A.  Ionides: Min. Sci. Press 112 (1916) 781–787.
• 4)   E.A.  Ashcroft: Trans. Inst. Min. Metall. 43 (1933–1934) 151–255.
• 5)   A.  Biswas and  M.H.  Khundkar: J. Indian Chem. Soc. Ind. News Ed. 14 (1951) 29–35.
• 6)  H.G. Lacell: Internal report, Imperial Smelting Corporation, Ltd., (1928).
• 7)   D.J.  Fray: Trans. Ins. Min. Metall. Sect. C 95 (1986) C55–C57.
• 8)   F.  Tailoka and  D.J.  Fray: Trans. Ins. Min. Metall. Sect. C 106 (1997) C115–C122.
• 9)   B.  Zhang,  X.  Yan,  K.  Shibata,  M.  Tada and  M.  Hirasawa: High Temp. Mat. PR-ISR 18 (1999) 197–212.
• 10)   B.  Zhang,  X.-Y.  Yan,  K.  Shibata,  T.  Uda,  M.  Tada and  M.  Hirasawa: Mater. Trans. 41 (2000) 1342–1350.
• 11)   H.  Matsuura,  T.  Hamano and  F.  Tsukihashi: ISIJ Int. 46 (2006) 1113–1119.
• 12)   H.  Matsuura,  T.  Hamano and  F.  Tsukihashi: Mater. Trans. 47 (2006) 2524–2532.
• 13)   G.-S.  Lee and  Y.J.  Song: Miner. Eng. 20 (2007) 739–746.
• 14)   C.A.  Pickles: J. Hazard. Mater. 166 (2009) 1030–1042.
• 15)   T.  Guo,  X.  Hu,  H.  Matsuura,  F.  Tsukihashi and  G.  Zhou: ISIJ Int. 50 (2010) 1084–1088.
• 16)   M.  Al-harahsheh,  S.  Kingman,  L.  Al-Makhadmah and  I.E.  Hamilton: J. Hazard. Mater. 274 (2014) 87–97.
• 17)  Kinotech Solar Energy Corporation: WO 2015/030235, 5 March 2015.
• 18)   F.  Santos,  E.  Brocchi,  V.  Araújo and  R.  Souza: Metall. Mater. Trans. B 46 (2015) 1729–1741.
• 19)  K. Yoshiyama and H. Matsuura: Proc. MMIJ Autumn Meet. 3 (2016) p. 2410 (in Japanese).
• 20)   G.  Sun,  H.  Matsuura and  F.  Tsukihashi: CAMP-ISIJ 29 (2016) 661.
• 21)   G.  Sun,  H.  Matsuura and  F.  Tsukihashi: Mater. Trans. 58 (2017) 59–64.
• 22)   G.  Sun,  H.  Matsuura and  F.  Tsukihashi: CAMP-ISIJ 30 (2017) 245.
• 23)   M.  Al-harahsheh,  J.  A.-Nu’aira,  A.  A.-Otoom,  I.  Al-hammouri,  H.  Al-jabali,  M.  Al-zoubi and  S.A.  Al’asal: J. Environ. Chem. Eng. 7 (2019) 102856.
• 24)   H.  Sato and  K.  Okunuma: Tetsu-to-Hagané 106 (2020) 858–866.
• 25)   J.  Iwanec and  B.J.  Welch: Aust. J. Chem. 22 (1969) 1783–1784.
• 26)   H.  Monk and  D.J.  Fray: Trans. Ins. Min. Metall. Sect. C 82 (1973) C161–C163.
• 27)  M.C. Campbell and G.M. Ritcey: Extraction Metallurgy, ’81, (IMM, London, 1981).
• 28)   C.  Wang,  X.  Hu,  H.  Matsuura and  F.  Tsukihashi: ISIJ Int. 47 (2007) 370–376.
• 29)   G.  Kamimura and  H.  Matsuura: Metall. Mater. Trans. B 51 (2020) 1395–1405.
• 30)   H.  Monk and  D.J.  Fray: Trans. Ins. Min. Metall. Sect. C 82 (1973) C240–C242.
• 31)   R.A.  Devilee,  A.V.  Sandwijk and  M.A.  Reuter: Metall. Mater. Trans. B 30 (1999) 607–611.
• 32)   G.  Kamimura and  H.  Matsuura: J. Sustain. Met. 5 (2019) 341–349.
• 33)   G.  Kamimura and  H.  Matsuura: Metall. Mater. Trans. B 52 (2021) 156–162.
• 34)   C.W.  Bale,  E.  Bélisle,  P.  Chartrand,  S.A.  Decterov,  G.  Eriksson,  A.E.  Gheribi,  K.  Hack,  I.H.  Jung,  Y.B.  Kang,  J.  Melançon,  A.D.  Pelton,  S.  Petersen,  C.  Robelin,  J.  Sangster and  M.-A.V.  Ende: Calphad 54 (2016) 35–53.
• 35)   C.  Robelin and  P.  Chartrand: J. Chem. Thermodyn. 43 (2011) 377–391.
• 36)  O. Knacke, O. Kubaschewski and K. Heseelmann: Thermochemical Properties of Inorganic Substance, 2nd ed., (Springer-Verlag, Berlin, 1991).
• 37)   M.  Nanjo,  T.  Kanai,  Y.  Itoh and  K.  Taniuchi: J. JILM 34 (1984) 464–471.
• 38)   T.  Uda,  K.T.  Jacob and  M.  Hirasawa: Science 289 (2000) 2326–2329.
• 39)   W.  Xiong,  Y.  Kong,  Y.  Du,  Z.-K.  Liu,  M.  Selleby and  W.-H.  Sun: Calphad 33 (2009) 433–440.
• 40)   A.  Ferrari and  C.  Colla: Atti Accad. Naz. Lincei, Cl. Sci. Fis., Mat. Nat.: Rend. 12 (1930) 672.
• 41)   S.  Cui and  I.-H.  Jung: Calphad 56 (2017) 241–259.
• 42)   K.  Shubhank and  Y.-B.  Kang: Calphad 45 (2014) 127–137.
• 43)   H.  Viitala,  P.  Taskinen and  D.  Lindberg: Calphad 67 (2019) 101667.
• 44)  E.M. Levin, C.R. Robbins and H.F. McMurdie: Phase Disgrams for Ceramists, vol. I, (American Ceramic Society, Ohio, 1964).