Elsevier

Hydrometallurgy

Volume 56, Issue 3, July 2000, Pages 293-307
Hydrometallurgy

Recovery of sulfuric acid from copper tank house electrolyte bleeds

https://doi.org/10.1016/S0304-386X(00)00081-5Get rights and content

Abstract

Many hydrometallurgical processes produce large amounts of acid waste. For example, copper SX/EW process plants typically bleed a concentrated sulphuric acid stream (electrolyte) from their tank house in order to limit the build up of impurities in the electrowinning stage. A process based on solvent extraction, to selectively recover up to 90% of the sulphuric acid from electrolyte bleed streams, has been developed. This is recovered as a pure aqueous acid stream at a concentration of up to 130 g/L that can be recycled back into the tank house circuit thus reducing both neutralisation and acid make up costs. The acid concentration of the electrolyte is reduced from 180 g/L to as low as 18 g/L. The extraction system involves the use of a branched long chain aliphatic tertiary amine tris(2-ethylhexyl)amine (TEHA) as the extractant, Shellsol 2046 as the diluent and octanol as a modifier. The equilibrium data and simulation results were also compared with an alternative extractant, CYANEX 923.

Introduction

In the solvent extraction and electrowinning process for the production of copper, a 175–180 g/L sulphuric acid solution (electrolyte) is used to strip the copper from the loaded organic. It is then sent to the electrowinning cells where copper is electrolytically deposited onto stainless steel cathodes. The spent electrolyte from the electrowinning tank house is returned to the SX process and used to strip the organic again. Thus the electrolyte is continuously regenerated. During normal operation, impurities such as iron build up in the electrolyte, mainly from aqueous entrainment in the organic. Impurities such as iron affect the current efficiency of the electrowinning process as well as the purity of the deposited copper. In order to control the concentration of impurities in the electrolyte, a bleed stream is often required. This bleed has to be neutralised before it can be released to tails. An equivalent quantity of fresh acid is added to the circuit as make up to ensure that an acid concentration of 175–180 g/L is maintained. If the bleed electrolyte could be purified and reused in the electrolyte circuit, this would reduce the costs associated with neutralisation and make up acid.

The proposed recovery process is based on solvent extraction, which has the advantage of being a well established technology in the mining industry. The process also does not increase the environmental impact of the copper recovery plant as all additional chemicals can be completely recycled. A simplified diagram of the process is shown in Fig. 1. The process involves contacting the bleed stream with an organic solvent which will selectively extract the sulphuric acid, leaving any contaminating iron and other impurities in the electrolyte.

The reaction between the tertiary amine extractant, tris (2-ethylhexyl) amine (TEHA) and sulphuric acid can be described by two equations. The first, represented by Eq. (1) occurs at low acid concentrations (below 1 M) and involves the formation of the amine sulphate salt, combining two amines for every sulphate.H2SO4(aq)+2R3N(org)⇔(R3NH)2SO4(org)

The second equation, represented by Eq. (2) involves the formation of amine bisulphate. This occurs at higher concentrations of sulphuric acid, when the concentration of amine sulphate is above 0.02 MH2SO4(aq)+(R3NH)2SO4(org)⇔2(R3NH)HSO4(org)

The loaded organic solvent is then contacted with pure water at an elevated temperature to strip the solvent of acid and produce a pure sulphuric acid stream that can be recycled back into the copper process. The stripping reaction is simply the reverse of ,

The solvent is made up of two other constituents in addition to the TEHA, a diluent and a modifier. The diluent used in the experiments was Shellsol 2046, which is necessary to reduce the high viscosity of the loaded solvent. A modifier is required to avoid the formation of a third phase. The third phase is formed at a high solvent loading. Octanol was used as modifier in the initial investigation and it was found that the minimum concentration required to prevent the third phase formation was 20%. Later, tridecanol was used as it has a higher flash point and therefore more cost effective in any industrial application. Tridecanol and octanol demonstrated similar properties as modifiers.

The conditions, such as extraction and stripping temperature, modifier concentration and solvent to aqueous phase ratio, have been optimised for the conditions described in Fig. 1. These results show that 90% of the sulphuric acid can be recovered.

An alternative to basic extractants, such as amines, is neutral or solvating extractants such as liquid phosphine oxides. Two alternative phosphine oxides have been proposed for the extraction of sulphuric acid. The first being tri-n-octylphosphine oxide with carbon tetrachloride as the diluent that was tested for the extraction of mineral acids such as hydrochloric acid, nitric acid and sulphuric acid [1].

The second is CYANEX 923 which is a mixture of four trialkylphosphine oxides [2], [3]. Rickelton proposed CYANEX 923 as a possible extractant for the recovery of sulphuric acid. He found that CYANEX 923 displayed a good compromise between its ability to extract sulphuric acid and to be effectively stripped with water. Rickelton's experiments also observed that CYANEX 923 has a very high selectivity for acids in preference to both copper and nickel. Rickelton's experiments were performed without a diluent. This extractant has the potential to be used in the recovery of acid.

More recently, Alguacil and Lopez [3] looked at the effect of diluents such as decane and toluene on the equilibrium of the CYANEX 923 extraction system. They found that the diluent did not seem to influence the acid extraction although they observed the formation of a third phase with aliphatic diluents at CYANEX 923 concentrations of 10% to 20% and above an aqueous acid concentration of 3 M. It was also found that the extent of extraction decreased at higher temperatures for sulfuric acid.

The extraction of mineral acids by CYANEX 923 under the conditions used in this investigation can be represented by the general reaction described by Eq. (3).mHaq++Xaqm−+Lorg⇔HmXLorg

Section snippets

Experimental

The equilibrium experiments were performed with two types of technical grade extractants: TEHA (greater than 95% TEHA, Fluka) and CYANEX® 923 (93% trialkylphosphine oxide). For TEHA extraction system, solvents were prepared by making 1 M (43%) solutions of TEHA in varying amounts of octanol and Shellsol 2046 (Shell Aust). The concentrations of octanol varied from 14.25% to 57%. For the CYANEX 923 system, neither a diluent nor modifier was used. All experiments were done at an A/O ratio of 10:1

Equilibrium data

The process of acid recovery requires a compromise between extraction and stripping. Under certain operating conditions, the equilibrium isotherm is such that the loaded solvent cannot be stripped with water. For this reason, it was necessary to obtain equilibrium isotherms over a wide range of conditions. This would enable optimum operating conditions to be identified.

Modelling the system

The TEHA equilibrium data consisting of 127 points over a range of octanol concentration, temperature and aqueous acid concentration, was fitted to the following equation:y=A+B1+XCD

This form of equation was chosen because it has an asymptote at higher aqueous acid concentrations (i.e above 2 M). This meant that any extrapolation of the data would not result in significant error. It was found that parameters A and B did not significantly change with octanol concentration or temperature. C was a

Process optimisation

, , in addition to steady state material balances around equilibrium stages, were used to determine the optimal process conditions for the extraction of 180 g/L H2SO4, to produce as high as possible pure sulphuric acid using water as a stripping agent. The following six parameters were optimised to maximize both the amount of acid recovered and product acid concentration.

  • Octanol concentration

  • Temperature of extracting stages

  • Temperature of stripping stages

  • Extraction and strip O/A ratio

  • Solvent to

Comparison between TEHA and CYANEX 923 extraction system using industrial aqueous acid feed

Using this approach, it is also possible compare 1 M TEHA in 25% w/w isotridecanol to CYANEX 923 using industrial acid solutions. The extraction and stripping equilibrium isotherms for both solvents at 20°C were used. The optimum recovery and product acid concentration at an operating temperature of 20°C are shown in Table 3.

The results show both CYANEX 923 and TEHA could be used for an acid recovery process. Both extractants have the ability to produce a product acid concentration of well over

Conclusions

The following conclusions result from this work:

  • The modifier concentration has a significant effect on the equilibrium for TEHA extraction system.

  • The temperature can significantly affect the equilibrium and so alter the optimum recovery conditions for both TEHA and CYANEX 923. The recovery of the acid is significantly improved if stripping is performed at an elevated temperature.

  • Both TEHA and CYANEX 923 extraction systems could be used for an acid recovery process.

Acknowledgements

Support from the Australian Research Council and Zeneca Specialties (UK) for their funding of this project is gratefully acknowledged. The authors would also like to thank the Advanced Mineral Products Special Research Centre (AMPC) for their support and use of equipment and Gunpowder Copper Co for supply of industrial solutions, Shell Chemicals Australia for supply of the Shellsol 2046 and Cytec for the CYANEX 923.

References (6)

  • F.J Alguacil et al.

    Hydrometallurgy

    (1996)
  • D.M. Petkovic et al.

    Solvent Extraction and Ion Exchange

    (1992)
  • W.A. Rickelton

    ISEC'93

    (1993)
There are more references available in the full text version of this article.

Cited by (0)

View full text