Modeling for Copper Solvent Extraction Schemes Design

Joseph Kafumbila HydrometallurgicalConsultant

Modeling forCopper the solvent extraction schemes design

Copyright © 2015 Joseph Kafumbila [email protected]

www.linkedin.com All rights reserved

2

Abstract

The solvent extraction process consists of two circuits, extraction

and stripping circuits. These two circuits are coupled by a common organic solvent. The copper solvent extraction schemes range from single organic circulation loops to schemes employing organic and aqueous bypass or intermediate organic recycle. For the engineering calculation purpose, the computer programs are used to design of these schemes. This paper gives a new way to design the Copper solvent extraction schemes ranged from simply to the complex one. The design results for all schemes can be plotted in form of MacCabe Thiele graph.

3

1. Introduction

In 1968, the solvent extraction operation introduced into the Copper metallurgy aimed to eliminate the Copper cementation stage on the metallic Iron [1]. The decrease of Copper grade in mines, the improvement of the oxide and sulphide ores leachingtechniques, the improvement of the Copper solvent extraction and electrowinning technology and the improvement of the Copperextractantproperties make that the part of the Copperproduction passing by the solvent extraction does not stop to increase [1,2].

The design of the Copper solvent extraction operation is more complex because,in the majority of hydrometallurgy applications, the solvent extraction operation consists of two circuits coupled by the common solvent. In the extraction step; the metal is extracted from the aqueous solution by the organic solvent. In the stripping stage, the metal is extracted from the organic solvent by the aqueous solution. The simple method for the equilibrium point determination between both units is the graphic method called usually the MacCabe Thiele diagram [3]. This method consists of the extraction and stripping isotherms determination from laboratory. The extractantconcentration in the organic phase for the isotherms establishment is obtained fromthe extraction efficiency and the Copper transfer capacity in the organic phase. This method is more time consuming and expensive, for scanning multiple Coppersolvent extraction configurations. So modeling of the extraction and stripping equilibrium will make easier the compilation of solvent extraction simulation results.

For engineering calculation purpose, several computer programs of the Copper solvent extraction equilibriums modeling came out. The equilibrium correlations used in the computer program modeling are the isotherm extrapolation curves [3] or base on the Copper solvent extraction chemical reaction constant [4]. Some of the computer

4

programs can handle the Copper solvent extraction complex schemes using organic or aqueous bypass and organic recycle.

The purpose of this paperis to obtain the empirical equilibrium correlations of the solvent extraction operation using Lix 984N from thermodynamic property for extraction side and from stripping isotherm characteristic for stripping side. These equilibrium correlations can be used for Copper solvent extraction design. This method can easily handle the Copper solvent extraction complex schemes and can make a MacCabe Thiele graphic representation of the complex schemes.

2. Extraction modeling

The extraction stage depends on the mass transfer kinetics between both phases (organic and aqueous) and the transfer stops at in Copperthermodynamic equilibrium in both phases. The Lix 984N is the mixing of aldoxime and ketoxime extractants and the Copper solvent extraction by Lix 984N follows the chemical reaction (a) [4].

Cu��+ 2HR↔CuR� + 2H� (a)

where the ionic species are in the aqueous phase, HR is the acid form of the Lix 984N and CuR�is the Coppercomplex form in the organic phase.

The equilibrium condition of the chemical reaction (a) is given by the mathematical expression (1) [5].

��� + 2� - ��� - 2� = 0 (1)

whereμ� is the chemical or electrochemical potential of the species i.

5

The chemical potential of the species i is given by the mathematical expression (2).

μ� = RTln([γ�]x[I]) (2) Where γ� is the chemical activity coefficient of the species i and its value are between 0 and 1, [I] is the species i concentration in mole/l, R is the perfect gases constant and T is the temperature.

Substituting the expression (2) for each species in the expression (1), it appears the expression (3).

[����]�[�]�

[��]�[��]�=

[���]�[���]�

��������[��]� = K (3)

The K value from the molar concentrations of the species is

giving by the expression (4).

K = [����]�[�]

�

[��]�[��]� (4)

The Coppermolar concentrations at equilibrium in the organic and

aqueous phases are respectively obtained by the expressions (5), (6) and (7).

[CuR�] =[��]���

�

��.��(5)

[Cu]���� = [Cu]���

� + ([Cu]��� - [Cu]��

� )x ��

�� (6)

[Cu] = [��]��

�

��.�� (7)

6

where[Cu]��� is the Copperconcentration in the aqueous phase at

equilibrium (g/l), [Cu]��� is the initial concentration of the Copper in

the aqueous phase (g/l), [Cu]���� is the initial concentration of the

Copper in the organic phase (g/l), [Cu]���� is the Copperconcentration

in the organic phase at equilibrium, V� is the volume of the aqueous phase and V�is the volume of the organic phase.

The molar concentration of the free extractant in the organic phase will be calculated from the mathematical expression (8).

[HR] = �/�%��.�������

������� - 2 x

[��]����

��.�� (8)

where V/V % is the extractantvolume percentage in the organic phase, the value 0.91 is the extractantdensity and the value 270 is the extractant mass molar.

The sulphuric acid dissociation will follow the reactions (b) and (c).

H�SO� ↔ HSO�� + H� K1 = 10�(b)

HSO�

� ↔ SO��� + H�K2 = 1.25 10�� (c)

If C1 (mol/l) is sulphuric acid concentration in the aqueous phase

and C2 (mol/l) is the concentration of SO��� anions associated to the

salts in aqueous phase. The dissociation reaction (b) will be complete because K1 is big. The following mass balance comes from the dissociation of the reaction (b).

7

H�SO� ↔ HSO�� + H�

Initial state C1 0 0 Final state -C1 C1 C1 Mass balance 0 C1 C1 The second dissociation reaction (c) is not complete because K2 is

not big enough.The following mass balance comes from the second dissociation reaction where X is themolar quantity of HSO�

� goes in the dissociation.

HSO�

� ↔ SO��� + H�

Initial state C1 C2 C1 Final state -X XX Mass balance (C1-X) (C2+X)(C1+X) The chemical reaction constant of the second dissociation reaction

gives the expression (9).

K2 = (����)(����)

(����)(9)

From the expression (9), the quantity X becomes zero when le

concentration C2 is more than K2. Therefore, at equilibrium, hydrogen ion concentration in aqueous phase will be given by the expression (10) which take account only the dissociation reaction (b). The expression (11) gives the sulphuric acid concentration at equilibrium in the aqueous phase.

[H] = ����

�

�� (10)

Ac��

� = Ac��� + (Cu��

� - Cu��� ) x 1,54 (11)

8

whereAC��� is the sulphuric acid concentration in the aqueous phase at

equilibrium (g/l) and Ac��� is the initial concentration of the sulphuric

acid in the aqueous phases(g/l).

By substituting the expressions (5), (7), (8) and (10) in the expression (4), it appears the expression (12) which gives the K valueas function of the extractant volume percentage in the organic phase and the concentrations of the species (g/l) in the organic and aqueous phases.

K = �����

�

����� x

[����� ]�

[�.������/�%��.����������� ]�

(12)

The K value from the chemical activity coefficients of the species

is given by the mathematical expression (13).

K = [���]�[���]

�

��������[��]� (13)

The K valueis the multiplication of two ratios. The first ratio is

the ratio of chemical activity coefficient of species in the organic phase and the second ratio is the ratio of chemical activity coefficients of the species in the aqueous phase.

In the organic phase, it was shown that the ratio of the chemical activity coefficients of the species is a function of the Copperconcentration in the organic phase [5].

In the aqueous phase, when the species concentrations are lower than 1 mole/l, the values of the chemical activity coefficients approachthe value 1 [6]. In the extraction stage, the molar concentrations of Copper and acid are generally lower than 1 mole/l,

9

the value of the ratio of the chemical activity coefficients in the aqueous phase approach the value of 1. The consequence of these assumptions is that in the extraction stage, the K value is a function of the Copper concentration the in the organic phase.

The example chosen for reporting the K value obtained from the expression (12) versus the Copper concentration the in the organic phase is resumed in the BASF Redbook [7]. The initial aqueous phase contains 2.5 g/l of Copper at pH 1.8. The organic phase is constituted by 8.7 % of Lix 984N in Escaid 100. The initial organic phase contains 1.8 g/l of Copper.The Table 1 gives the Copper concentrations the in the organic and aqueous phases at equilibrium obtained in the laboratory.

Table 1: Extraction isotherm experimental data

Cu (aq) Cu (org)

g/l g/l 0.070 2.040 0.090 2.280 0.170 2.980 0.280 3.280 0.510 3.700 1.240 4.190 1.940 4.350

The Figure 1 gives the K value coming from expression (12) versus Copper concentration in the organic phase. The results show that the K value is the linear function of the Copper concentration in the organic phase.

10

Figure 1: K value versus Copper concentration in the organic phase

By using this linear thermodynamic property on the extraction isotherms results obtained in the laboratory with the extractant Lix 984N, the empirical modelhas been obtained for the extraction stage. This model is given by the expression (14).

K=[3.4657ln(v/v%)-13.524]x(�����

�

�/�%) + [-1.553ln(v/v%) + 6.4051] (14)

The comparison between the experimental data from theTable 1

and the predicted value coming from extraction model is showing in the Figure 2 where the predicted Copper concentration in the organic phase has been plotted versus the experimental value. The results show that the model fits well the experimental values.

R² = 0,983

0,00

0,40

0,80

1,20

1,60

2,00

2,00 2,50 3,00 3,50 4,00 4,50

K v

alu

e

Organic phase : Metal concentration (g/l)

11

Figure 2: predicted versus experimental values of Copper concentration in the organic phase

3. Stripping stage

The extraction model can’t apply to the stripping stage because

the assumption on the chemical activity coefficients values in the aqueous phase which approach the value 1 is not applicable to the stripping stage. The considerations thermodynamics in the stripping stage show that the ratio of the chemical activity coefficients in the

y = 0,995x - 0,028R² = 0,996

2,00

2,50

3,00

3,50

4,00

4,50

2,00 2,50 3,00 3,50 4,00 4,50

Pre

dic

ted

val

ue

in o

rgan

ic p

has

e : C

u (

g/l)

Experimental value in organic phase: Cu (g/l)

12

aqueous phase is a function of the ionic strength and the acid concentration the in the aqueous phase [5].

On the other hand, it has been observed that the

Copperconcentration in the organic phase is a linear function of the Copperand acid concentrationsratio in the aqueous phase at equilibrium in the range of 30 to 60 g/l of Copper and 140 to 200 g/l of acid in the aqueous phase.

The example chosen for reporting the Copper concentration in the organic phaseversus the Copperand the acid concentrations ratio in the aqueous phase is resumed in the BASF Redbook [7]. The spent electrolyte contains 30.7 g/l of Copperand 190 g/l of sulphuric acid. The organic phase is constituted by 8.7 % of Lix 984N in Escaid 100. The loaded organic phase contains 3.9 g/l of Copper.The Table2 gives the Copper concentrations the in the organic and aqueous phases at equilibrium obtained in the laboratory.

Figure 3 shows the Copper concentration graph in organic phase

versus the ratio Copper and acid concentration in the aqueous phase. The results show the Copper concentration in the organic phase is linear function of the ratio Copper and acid in the aqueous phase.

By using the relation on the stripping isotherms results obtained

in the laboratory with the extractant Lix 984N, the empirical stripping model has been obtained. The stripping model is given by the expression (15).

�����

�

�/�% = (

��.���

���� xV/V% + 0.6183)x(

�����

����� ) + (

�.��

����xV/V% +

�.��

����) (15)

13

Table2: Stripping isotherm experimental data

Cu (aq) Cu (org) g/l g/l

51.3 1.78 43.2 1.38 37.7 1.21 33.8 1.07 32.3 1.01 31.2 0.98

Figure 3: Copper concentration in organic phase versus the ratio Copper and acid concentrations in the aqueous phase.

R² = 0,996

0,8

1

1,2

1,4

1,6

1,8

2

0,15 0,2 0,25 0,3 0,35

Org

anic

ph

ase

: Cu

(g/

l)

Aqueous phase : Ratio Cu/acid

14 The comparison between the experimental values from Table 2

and the predicted values coming from stripping model is showing in the Figure 4 where the predicted Copper concentration in the organic phase has been plotted versus the experimental values. The results show that the model fits well the experimental values.

Figure 4: predicted versus experimental values of Copper concentration in the organic phase

4. Design of Copper solvent extraction schemes

The design exercise will go on the copper cobalt plant operating with the split circuit [8]. PLS1 is the overflow from the primary thickener and PLS2 is the overflow from the first CCD.

y = 0,986x + 0,004R² = 0,996

0,90

1,10

1,30

1,50

1,70

1,90

0,9 1,1 1,3 1,5 1,7 1,9

Pre

dic

ted

val

ue

in t

he

org

anic

ph

ase

: Cu

(g/

l)

Experimental value in organic phase : Cu (g/l)

15

4.1. Interlaced copper solvent extraction scheme

The solvent extraction scheme will be the interlaced configuration. The Figure 5 shows the Copper solvent extraction interlaced scheme. The Table 3 gives the design results.

Figure 5: Interlaced Copper solvent extraction scheme

E4 E3 E2 E1

S2 S1

PLS1 PLS2

Raff1 Raff2

Stripped

organic

Loaded

organic

Spent

electrolyte

Advance

electrolyte

16

Table 3: interlaced copper solvent extraction design results PLS 1 flow 654 m3/h

Cu 13.8 g/l

Acid 1.89 g/l

PLS 2 Flow 714 m3/h

Cu 4.50 g/l

Acid 5.00 g/l

Organic flow 1142 m3/h

v/v% 28 %

Spent electrolyte Cu 35 g/l

Acid 180 g/l

Advance electrolyte Cu 50 g/l

Mixing efficiency E1 93 %

Mixing efficiency E2 95 %

Mixing efficiency E3 97 %

Mixing efficiency E4 99 %

Mixing efficiency S1 95 %

Mixing efficiency S2 85 %

Loaded organic Cu 14.11 g/l

Stripped organic Cu 4.18 g/l

Raffinat 1 Cu 1.243 g/l

Raffinat 2 Cu 0.122 g/l

Extraction efficiency PLS1 90.99 %

Extraction efficiency PLS2 97.29 %

Stripping efficiency 70.35 %

Net transfer 0.354 g/l/v/v%

The Figure 6 and 7 show the extraction and stripping MacCabe

Thiele diagram respectively. On the extraction side, the MacCabe Thiele shows two equilibrium isotherms for PLS1 and PLS2.

17

Figure 6: Extraction MacCabe Thiele diagram for the interlaced Scheme

Figure 7: Stripping Mac Cabe Thiele diagram for interlaced scheme

3,50

6,50

9,50

12,50

15,50

0,00 3,00 6,00 9,00 12,00 15,00

Org

anic

ph

ase:

Cu

(g/

l)

aqueous phase: Cu (g/l)

0,00

3,00

6,00

9,00

12,00

15,00

30,00 35,00 40,00 45,00 50,00 55,00

Org

anic

ph

ase:

Cu

(g/

l)

Aqueous phase: Cu (g/l)

18

4.2.Optimum double serie parallel copper solvent extraction scheme

The solvent extraction scheme will be the the optimum double serie parallel configuration. The Figure 8 shows the Copper solvent extraction optimum double serie parallel scheme. The Table 4 gives the design results.

Figure 8: optimum double serie parallel Copper solvent extraction scheme

E4 E3 E2 E1

S2 S1

PLS2 PLS1 Raff1 Raff2

Stripped

organic

Loaded

organic

Spent

electrolyte

Advance

electrolyte

19

Table 4: Double serie parallel copper solvent extraction design results PLS 1 flow 654 m3/h

Cu 13.8 g/l

Acid 1.89 g/l

PLS 2 Flow 714 m3/h

Cu 4.50 g/l

Acid 5.00 g/l

Organic flow 1142 m3/h

v/v% 28 %

Spent electrolyte Cu 35 g/l

Acid 180 g/l

Advance electrolyte Cu 50 g/l

Mixing efficiency E1 93 %

Mixing efficiency E2 95 %

Mixing efficiency E3 97 %

Mixing efficiency E4 99 %

Mixing efficiency S1 95 %

Mixing efficiency S2 85 %

Loaded organic Cu 14.13 g/l

Stripped organic Cu 4.18 g/l

Raffinat 1 Cu 1.153 g/l

Raffinat 2 Cu 0.176 g/l

Extraction efficiency PLS1 91.648 %

Extraction efficiency PLS2 96.090 %

Stripping efficiency 70.384 %

Net transfer 0.355 g/l/v/v%

The Figure 9 and 10 show the extraction and stripping MacCabe

Thiele diagram respectively. On the extraction side, the MacCabe Thiele shows two equilibrium isotherms for PLS1 and PLS2.

20

Figure 9: Extraction MacCabe Thiele diagram for optimum double serie parallel Scheme

Figure 7: Stripping Mac Cabe Thiele diagram for optimum double serie parallel scheme

3,50

6,50

9,50

12,50

15,50

0,00 3,00 6,00 9,00 12,00 15,00

Org

anic

ph

ase:

Cu

(g/

l)

aqueous phase: Cu (g/l)

0,00

3,00

6,00

9,00

12,00

15,00

30,00 35,00 40,00 45,00 50,00 55,00

Org

anic

ph

ase:

Cu

(g/

l)

Aqueous phase: Cu (g/l)

21

5. Conclusion

The design of Copper solvent extraction schemes is done by using computer program. Some of the computer program can design all complex schemes of Copper solvent extraction.

From the thermodynamics properties forthe extraction stage and the isotherms characteristics for the stripping stage, it appears two empirical equilibrium correlations. These correlations can be used for Copper solvent extraction design schemes. The design results can be plotted in form of MacCabe Thiele for all Copper solvent extraction schemes.

6. Reference

1. Kordosky G.A., Copper recovery using leach/solvent

extraction/electrowinning technology: Forty years of innovation, 2.2 million tonnes of Copper annually, SAIMM, 2002.

2. Katarzyna Rotuska, Tomasz Chmielewski, Growing role of solvent extraction in Copper ores processing, Physicochemical Problems of mineral processing, 2008.

3. Alonso,A.I., Lassahn. A., Grulm, G., Optimal design of non-

dispersive solvent extraction processes. Comput. Chem. Eng, 2001.

4. Hossen Aminian, Modélisation et simulation des opérations

d’extraction par solvant et d’électrolyse du Cuivre, thèse, université de Laval, Canada, 1999.

22

5. Gerald L. Bauer and Thomas W. Chapman, Measurement and correlation of solvent extraction equilibria. The extraction of Copper by kelex 100, Metallurgical transaction B, 1976.

6. AMEL A., Etude thermodynamique de l’extraction des Métaux de transition par la Salicylidèneaniline, thèse, Université Mohamed Khider, 2013.

7. BASF, Redbook mining solutions, 2014.

8. Nisbell A., Baxter K., Marte K., Urbani M., Flowsheet

considerations for Copper – Cobalt projects, SAIMM conference, 2009.