Indium recovery from acidic aqueous solutions by solvent extraction with D2EHPA: a statistical approach to the experimental design

Indium recovery from acidic aqueous solutions by solvent extraction with D2EHPA: a statistical approach to the experimental design

M.C.B.Fortes^I; A.H.Martins^II; J.S.Benedetto^I

^ICentro de Desenvolvimento da Tecnologia Nuclear (CDTN/CNEN), Rua Prof. Mario Werneck, S/N - Cx. P. 941, CEP 30.161-970, Fax: [phone], Cidade Universitária, Pampulha, Belo Horizonte - MG, Brazil. E-mail: [e-mail] ^IIUniversidade Federal de Minas Gerais (UFMG), Departamento de Engenharia Metalúrgica e de Materiais, Rua Espírito Santo, 35 sala 206, CEP 30.160-030, Fax: [phone], Centro, Belo Horizonte - MG, Brazil. E-mail: [e-mail]

ABSTRACT

This experimental work presents the optimization results of obtaining a high indium concentration solution and minimum iron poisoning by solvent extraction with D2EHPA solubilized in isoparaffin and exxsol. The variables studied in the extraction step were D2EHPA concentration, acidity of the aqueous phase and time of contact between phases. Different hydrochloric and sulfuric acid concentrations were studied for the stripping step. The optimum experimental conditions resulted in a solution with 99% indium extraction and less than 4% iron. The construction of a McCabe-Thiele diagram indicated two theoretical countercurrent stages for indium extraction and at least six stages for indium stripping. Finally, the influence of associated metals found in typical sulfate leach liquors from zinc plants was studied. Under the experimental conditions for maximum indium extraction, 96% indium extraction was obtained, iron extraction was about 4% and no Ga, Cu and Zn were co-extracted.

Keywords: Solvent extraction, indium, D2EHPA, statistical approach.

INTRODUCTION

Commercial grade indium is usually produced as a by-product of leach liquors from zinc plants. Therefore, it is necessary to separate it from the other elements that are usually at present high concentrations (Harrower, 1998). Much research has studied this separation and suggests the solvent extraction process as one of the main techniques used.

Tomii and Tsuchida (1981) showed an effective process for recovery of indium present as an impurity in aqueous leached solution by adjusting the pH of the aqueous solution. An organic solution of di(2-ethylhexyl) phosphoric and tributylphosphoric acids diluted in paraffin was used. Back-extraction of the indium ions was carried out using an aqueous sulfuric acid solution. Fossi and Sambarino (1983) reported a process using a diesterphosphoric acid and indium stripping with hydrochloric acid or halide brines.

Krajewski and Hanush (1987) described a process for solvent extraction of gallium, germanium or indium from acidic or alkaline hydrous solutions, in which a mixture of at least two hydroxyquinolines in various quantities was employed. Asai et al. (1995) discussed a new process tested at Nippon Mining and Metals Company which obtained satisfactory results.

This paper is concerned with the statistical optimization of experimental results, aiming at determining the best experimental conditions for the main variables in solvent extraction with di(2-ethylhexyl) phosphoric acid (D2EHPA) solubilized in isoparaffin and exxsol for maximum indium recovery and minimum iron poisoning. The main variables studied in the extraction step were extractant concentration in the organic phase, acidity of the aqueous phase and time of contact between organic and aqueous phases. Different hydrochloric and sulfuric acid concentrations in the aqueous phase were studied for the stripping step. The optimization procedure adopted was the ascending path statistical method, which used the numeric values determined with the Yates algorithm of the replicated full factorial design method (Duckworth, 1968).

Extraction and stripping isotherms for indium recovery were obtained after determination of the optimum experimental conditions. The usage of a McCabe-Thiele diagram (Ritcey and Ashbrook, 1984) permitted prediction of the theoretical number of countercurrent stages for indium extraction.

Finally, the influence of the main associated metals (iron, gallium, copper and zinc) on the percentage of indium recovery was also studied. This study was carried out under optimum experimental conditions for maximum indium extraction.

EXPERIMENTAL

The aqueous feed solutions were prepared by solubilizing indium oxide (99.9% In₂O₃/Aldrich Chemical Co., USA) and iron sulfate (21-23% Fe²⁺, Fe₂(SO₄).xH₂O/Hoechst, USA) in sulfuric acid solution (95-97% H₂SO₄ PA/Vertec, Brazil). The hydrogen ion concentration of the solution was adjusted by adding sulfuric acid according to the optimum conditions determined by the statistical approach. The iron and indium concentrations in the aqueous feed solution were 2.5 g/L and 0.5 g/L, respectively.

The influence of iron, gallium, copper and zinc ions in solution on indium recovery was studied. The aqueous solutions containing metal ions were prepared with metal concentrations based on the chemical composition of the liquor generated by the sulfuric acid leaching of a zinc ore concentrate after a concentration step (see Table 1). Analytical grade reagents and distilled water were used throughout the experiments.

The organic phases were prepared by diluting di(2-ethylhexyl) phosphoric acid (D2EHPA, Daihachi Chemical Industry Co. Ltd, Japan) in Isoparaffin 22/25 (Unipar Química, Brazil) or Exxsol D100 (Exxon Chemical, USA). Tests were carried out mixing equal volumes (20 mL) in 100 mL open vessels (Pyrex^®, Brazil) under mechanical stirring at 400 rpm (IKA Labortechnik RW 20n, Germany) at 25 ± 2^oC.

Initially, the full factorial design method with three variables at two experimental levels was used for replicated responses. The main variables studied were D2EHPA concentration in the organic phase, acidity of the aqueous phase and contact time between organic and aqueous phases. The upper and lower levels for the main variables are shown in Table 2. Then, the ascending path statistical method, which used the numeric values determined with the Yates algorithm of the replicated full factorial design method, was adopted the optimization procedure.

The validity of the statistical model was confirmed using the Fisher test (Duckworth, 1968) where the statistical confidence level was higher than 95% and the experimental error was lower than 5%.

A sample of the aqueous phase after each test was collected to determine the indium, gallium, copper and iron concentrations in solution by atomic absorption spectrophotometry (computerized model 932 AA by GBC Scientific Equipment, Australia). Titration with potassium dichromate was used to determine Fe²⁺ and Zn²⁺ concentrations (Ohlweiler, 1981). The concentration of Fe³⁺ was determined by finding the difference between overall iron and Fe²⁺ concentrations. The metal content in the organic phase was determined from the difference between the concentrations of aqueous feed and raffinate.

The extraction isotherms were plotted using the optimum experimental conditions determined by the statistical approach. Successive contacts between the phases were made in open vessels (Pyrex^®, Brazil) and under mechanical stirring at 400 rpm (IKA Labortechnik RW 20n, Germany) at room temperature (25 ± 2^oC) and a 1/1 organic-to-aqueous phase ratio. Figure 1 shows a schematic diagram of the experimental setup used to obtain the numeric values to plot the extraction isotherms for indium. Four contacts between the organic loading and the exhaust stages were made.

The number of theoretical countercurrent stages required for indium extractions was determined from the McCabe-Thiele diagram. After that, batch-stripping tests were carried out using sulfuric and hydrochloric acid solutions at different concentrations.

The McCabe-Thiele diagram for indium stripping from the loaded organic phase was plotted using the same procedure as that adopted for the extraction stage.

The numeric values shown in the extraction and stripping equilibrium curves were experimental results obtained at 25 ± 2^oC.

RESULTS AND DISCUSSION

The influence of some main variables on indium and iron extraction by solvent extraction with D2EHPA diluted in isoparaffin and exxsol was studied. Table 3 shows the experimental results obtained by the replicated full factorial design method for D2EHPA concentration in the organic phase, acidity of the aqueous phase and contact time between organic and aqueous phases at two experimental levels.

Calculations with the Yates algorithm determined that all parameters studied had significant statistical effects on the experimental responses, but under different statistical intensities. The acidity of the aqueous phase had the greatest influence on indium extraction.

The ascending path method used in this work allowed determination of the optimum experimental conditions for achieving maximum indium extraction using the acidity of the aqueous phase as a parameter of technological limitation. Under these conditions, the maximum indium present in the system was recovered independently of the iron concentration in the solution. Table 4 shows the optimum experimental conditions determined for maximum indium extraction using the ascending path statistical method.

According to the statistical method, the predicted response for maximum indium extraction was 99.4%. Two laboratory bench-scale tests using the optimum experimental conditions were performed to confirm the statistical prediction. Table 5 shows the results of the experimental tests using the optimum conditions for maximum indium extraction.

In spite of the extraction data obtained to determine the optimum experimental conditions using D2EHPA diluted in isoparaffin, it can be observed from the experimental results that the better extraction and indium-iron separation was achieved with D2EHPA diluted in exxsol.

During extraction using D2EHPA diluted in isoparaffin, it was observed that the colorless aqueous solution changed to a light-red-color solution. The color change might indicate the presence of Fe³⁺ ions in solution. Chemical analysis of a sample showed that about 50% of the Fe²⁺ ions in solution were oxidized. This oxidation permitted competition between Fe³⁺ and In³⁺ ions, resulting in the replacement of In³⁺ adsorbed onto the organic phase by Fe³⁺ ions and a decrease in indium extraction.

Studying the influence of the diluent on iron oxidation and the decrease in indium extraction, tests were carried out using Fe²⁺ ion aqueous solution (0.25 N H₂SO₄) in contact for six minutes with the organic phase of different D2EHPA concentrations in isoparaffin or exxsol (see Figure 2).

Figure 2

The experimental conditions for minimum iron extraction were also studied to obtain indium aqueous solutions almost without iron poisoning. According to calculations with the Yates algorithm, the time of contact between phases had the greatest influence on iron extraction. The same procedure to determine the optimum experimental conditions for achieving maximum indium extraction was followed, but now using time of contact between the phases as a parameter of technological limitation. Table 6 shows the optimum experimental conditions for minimum iron extraction determined by the ascending path statistical method.

Under these optimum experimental conditions iron extraction was nearly 0.3%. Table 7 shows the experimental results for minimum iron extraction obtained from laboratory bench-scale tests using the optimum statistical conditions.

The experimental results for iron extraction shown in Table 7 did not achieve the predicted response when extraction was conducted with D2EHPA diluted in isoparaffin. Iron oxidation was also observed during the tests. However, using D2EHPA diluted in exxsol the experimental results achieved the predicted response and no iron was detected in the organic phase after extraction. Although the statistical calculations for iron extraction could not predict the values for indium extraction, they presented satisfactory results in just one contact between the phases.

The McCabe-Thiele diagram was plotted to determine the number of theoretical countercurrent stages required for indium extraction. Figures 3 and 4 show the isotherms of indium extraction under the optimum experimental conditions for maximum indium extraction and minimum iron extraction. An iron extraction of 4.0% was observed for the isotherm plot using the conditions for maximum indium extraction. This value was constant during all experiments for the isotherm construction. Iron was not detected in the organic phase for the isotherm curve using the conditions for minimum iron extraction.

The McCabe-Thiele diagrams in Figures 3 and 4 indicate two theoretical countercurrent stages required for full indium recovery. However, at least three stages were required for indium extraction with minimum or no iron extraction.

Indium stripping results from loaded D2EHPA diluted in exxsol under both optimum experimental conditions, i.e., maximum indium extraction and minimum iron extraction in different hydrochloric and sulfuric acid concentrations, are shown in Table 8.

Using the experimental conditions for maximum indium extraction, the stripping solution using a high hydrochloric acid concentration was not selective, since 4.0% of the iron from the feed was transferred to the loading organic phase and all loaded iron was stripped. On the other hand, the stripping was selective but less indium was recovered using a low hydrochloric acid concentration.

The sulfuric acid stripping solutions were more selective for any acid concentration. Indium recovery was not significant to the process for sulfuric acid concentrations lower than 0.5 M. Meanwhile, the process was very effective for concentrations equal to or higher than 0.5 M. The highest indium recovery was achieved using 5.0 M sulfuric acid. However, highly concentrated acidic solutions are not economical, especially when the excess of acid in the indium purification stage should be neutralized.

In Table 8, it can be observed that indium recovery is improved with increases in acid concentration. With sulfuric acid stripping it also can be observed that even iron is not desirable in stripped solution; indium recovery is better when there is a minimum concentration of iron. This may indicate that iron participates in the indium stripping process.

Hydrochloric acid stripping was high and selective for indium using the experimental conditions for minimum iron extraction. On the other hand, indium recovery using sulfuric acid was not significant. It is important to point out that the feed solution medium was acidic sulfate solution and using it to replace it for acidic chloride solution would be expensive for industrial purposes. So, the experimental conditions for minimum iron extraction were rejected.

The experimental conditions for maximum indium extraction were chosen from the McCabe Thiele diagram and experimental stripping results, firstly, because it was necessary to minimize the number of theoretical countercurrent stages while using the conditions for minimum iron extraction for a complete indium recovery and, secondly, due to the low iron stripping associated with a high indium extraction. Finally, using these conditions the indium stripping was satisfactory for both acidic media. Nevertheless, the sulfuric acidic medium was chosen for indium stripping because the process is simpler than that for the chloride medium and is also more feasible from economic viewpoint.

The indium stripping isotherm was plotted using the experimental conditions for maximum indium extraction with 3.0 M H₂SO₄, as shown in Figure 5. The McCabe Thiele diagram in Figure 5 shows that at least six theoretical countercurrent stages for indium stripping were necessary.

The influence of the main associated metals on indium recovery was also studied. Gallium, copper, zinc and iron were the metals present in a sulfate leach liquor from a typical zinc hydrometallurgical plant.

Under the experimental conditions for maximum indium extraction, 96% indium extraction was obtained. The percentage of iron extracted was the same as that using the previous solution and no Ga, Cu or Zn were co-extracted. Thus, selectivity for indium was retained using the conditions for maximum indium extraction, even when a leach liquor similar to an industrial one was used.

The stripping test was carried out using 3.0 M H₂SO₄ and indium recovery reached 60% with no other metal stripping for only one contact stage.

ACKNOWLEDGMENTS

The authors are thankful for the Centro de Desenvolvimento da Tecnologia Nuclear-CDTN/CNEN, PRONEX/MCT and FAPEMIG-Brazil for the financial support to conduct this work.

Received: January 15, 2001

Accepted: August 30, 2002

• Asai, K., Okamoto, H., Hino, J. and Kawabata, T., Non-Ferrous Smelting Proc. Int. Symp., 131 (1995).
• Benedetto, J.S., Tambougi, E.B. and Mingote, R.M., ISEC'99, Proc. Int. Solvent Extr. Conference (1999).
• Duckworth, W.E., Statistical Techniques in Technological Research, Methuenle Co. Ltd., London (1968), 1^st ed.
• Fossi, P. and Sambarino, E., Process for the Recovery of Indium, U.S. Patent 4,372,922 (1983).
• Harrower, M., Metals and Minerals Annual Review, Mining Journal Ltd., 91 (1998).
• Krajewski, W. and Hanush, K., Process for Fluid-fluid Extraction of Gallium, Germanium or Indium from Liquid Solutions, U.S. Patent 4,666,686 (1987).
• Ohlweiler, O.A., Química Analítica Quantitativa. vol. 2. Livros Técnicos e Científicos, Rio de Janeiro (1981), 3^rd ed.
• Ritcey, G.M. and Ashbrook, A.W., Solvent Extraction: Principles and Applications to Process Metallurgy. Part I, Elsevier, New York (1984).
• Tomii, K. and Tsuchida, H., Solvent Extraction Recovery Process for Indium, U.S. Patent 4,292,284 (1981).

Publication Dates

• Publication in this collection
  25 June 2003
• Date of issue
  June 2003

History

• Received
  15 Jan 2001
• Accepted
  30 Aug 2002