Hydrometallurgy is a metal processing technology used to separate a valuable metal from an ore by preparing an aqueous solution of a salt of the valuable metal and then recovering the valuable metal from that solution (Britannica Editors, 2016). Hydrometallurgy consists of three main processes, namely leaching, separation and concentration and product refinement, also referred to as metal recovery (Britannica Editors, 2016). Leaching is the most important process in hydrometallurgy and involves the dissolution of the valuable metal in the mineral present in the ore deposit. The process of separation and concentration usually entails chemical separations (Kamberovic, et al, 2009). The aim of this process is to remove impurities, to increase the concentration of the valuable metal. The aqueous solution, also referred to as the pregnant leach solution (PLS), consists of dissolved metals (Kamberovic, et al, 2009). This solution is treated to achieve a separation of the dissolved metals. There are various techniques that can be utilised to separate the dissolved metals in the PLS. These techniques include precipitation, crystallization, ion exchange, adsorption and solvent extraction (Habashi, 1999). Product refinement is the final step of the hydrometallurgy process. In this step, the valuable metal is recovered from the solution, which was produced in the separation and concentration process.
Solvent extraction
Solvent extraction, also referred to as liquid-liquid extraction, involves the mixing of a liquid phase consisting of two or more components also referred to as a PLS, with a second liquid phase referred to as the solvent, allowing for a desired compound to be isolated (Seader et al., 2006). If an aqueous PLS is used a liquid organic phase containing an extractant will typically be used as the solvent (Seader et al., 2006). The second liquid phase should be immiscible or partially miscible with one or more components in the feed stream, and completely or partially miscible with the other components in the feed stream in order for separation to take place (Seader et al., 2006). This technique utilises the relative solubility of the different immiscible/partially miscible phases present in the mixture (Littlejohn, 2007). The desired compound in the aqueous phase will bond with the extractant and typically the solute will be transferred from the aqueous phase to the organic phase, as the solute-extractant bond is more soluble in the organic phase (Seader et al., 2006).
The choice of solvent is critical to effective separation using the solvent extraction system The solvent will have a significant effect on the pH and the temperature of the system, therefore also on the cost (Littlejohn, 2007). A suitable solvent should be cheap, easy to recover, stable, nontoxic and have a high affinity for the desired compound (Seader et al., 2006). It is preferable for the solvent to be immiscible with all the components in the PLS, except for the solute (Seader et al., 2006). It is also important that there be a significant difference in specific gravity of the PLS and the solvent, to ensure an effective gravity separation (Seader et al., 2006). If the two phases are immiscible, gravity settling can occur. By convention the organic phase is lighter than the aqueous phase, as a result of a lower density, therefore the organic phase will stay at the top and the aqueous phase will settle out to the bottom (Seader et al., 2006).
Different types of solvent extraction can be utilized including, solvating extraction, cationic exchange, anionic exchange and chelating extraction (Habashi, 1999). Cationic exchange is commonly used for the recovery of copper from ore deposits (Littlejohn, 2007).
Solvent extraction used to recover copper
Typically oximes are used as extractants for copper due to their coordination chemistry (Spence, n.d.). Oximes can be separated into either aldoximes or ketoximes (Littlejohn, 2007). Aldoximes have been proven to be significantly stronger than ketoximes and therefore also much more difficult to strip using spent electrolytes (Littlejohn, 2007). Today the most commonly used extractants for copper recovery are ester modified aldoximes (Littlejohn, 2007).
Oximes are used as extractants as as it forms a complex when reacting with copper. The complex as well as the copper ion are both large, resulting in more stability and therefore other ions are rejected (Schlesinger et al, 2011). This mechanism is referred to as chelating extraction (Schlesinger et al, 2011). Equation 1 below is a general reaction which describes the typical ion exchange that takes place between the copper cation in the pregnant leach solution and the extractant, with RH representing the extractant with proton H+ (Schlesinger et al, 2011). The dissolved copper cation in the PLS bonds with extractant in the organic phase to form a copper-extractant complex bond which will transfer the copper ion from the PLS to the organic phase. When the copper-extractant complex is formed, two protons are released, as can be observed in Equation 1.
Cu2+aq + 2 RHorg ↔ CuR2,org + 2 H+aq (1)
The extraction constant for Equation 1, a chelating reaction, can be defined using Equation 2.
K_ex= ([CuR_2 ]_org [H^+ ]_org^2)/([Cu^(2+) ]_aq [RH]_org^2 ) (2)
It can be observed that the capacity of the extractant used for solvent extraction is dependent on the pH of the solution as well as the stoichiometry of the specific extraction reaction. As the reaction progresses, protons may transfer from the organic phase to the aqueous phase, which will result in an increase in acidity of the aqueous solution. It can be observed from Equation 1 that this will result in an equilibrium shift to the left, which will decrease the extent of extraction. This pH dependence can be utilised to strip copper from the loaded organic phase and into a spent electrolyte returning from electrowinning (Schlesinger et al, 2011).
It is important to manage the acidity levels of the solvent extraction circuit, as the equilibrium of loading/stripping curves are very dependent on the pH level (Littlejohn, 2007). The acidity of the PLS must be regulated in such a way that the copper will transfer from the aqueous phase to the organic phase (Littlejohn, 2007). It can also be observed from Equation 1, that 2 moles of protons are formed for every mole of copper removed from the aqueous phase. If the copper concentration in the solvent extraction circuit is very high, the acidity will also be very high, which will result in an equilibrium shift towards stripping (Littlejohn, 2007).
Performance parameters
The performance parameters of the solvent extraction circuit will be discussed in terms of the distribution coefficient, percentage extraction of copper and the separation factor.
Distribution coefficient
Solutes distribute themselves between the two immiscible liquid phases according to their differences in solubility in those liquids. When equilibrium is reached, the concentration of a compound in two immiscible phases of a mixture, will stay constant for a specific temperature (Littlejohn, 2007). This ratio is referred to as the distribution coefficient (Littlejohn, 2007). The definition of the distribution coefficient can be observed in Equation 3.
D=(concentration of metal ions in organic phase)/(concentration of metal ions in aqeuous phase)=[M^(n+) ]_org/[M^(n+) ]_aq (3)
As Equation 3 represents the distribution of the solute in two different phases, the distribution coefficient is also a measure of the difference in solubility for the phases involved. It can be observed from Equation 3 above that in order to decrease the solvent to feed ratio, the distribution factor should ideally be greater than 1 (Seader et al., 2006).
The distribution coefficient is very dependent on the pH of the mixture to be separated. According to the Henderson-Hasselbalch Equation, the percentage ionization for copper increases with increasing pH (Seader et al., 2006). An increase in ionization will result in more of the solute being entrained in the organic phase, while a lower pH will result in more of the solute being transferred to the organic phase (Seader et al., 2006).
Percentage extraction
The percentage extraction refers to the amount of solute in the aqueous phase that has been transferred to the organic phase. Equation 4 represents the percentage extraction. As displayed in Equation 4, the amount of moles solute left in the aqueous phase after equilibrium has been reached, is subtracted from the original amount of moles solute in the aqueous phase, to obtain the amount of moles solute extracted from the aqueous phase. This amount is divided by the total amount of solute in the PLS, to obtain the percentage extraction.
% Extraction=([M^(n+) ]_(aq,feed) V_(aq,feed)-[M^(n+) ]_aq V_aq)/([M^(n+) ]_(aq,feed) V_(aq,feed) ) (4)
Separation factor
The separation factor is the ratio of the distribution coefficients, or the ratio of the activity coefficients of the different components in the different liquid phases (King, 1980). This is an indication of the tendency of one component in the mixture to be extracted more readily from one liquid phase to another, than some other component present in the mixture (King, 1980). The ratio is given by Equation 5.
SF(β)=D_1/D_2 (5)
Theoretical number of stages in a solvent extraction plant
The theoretical number of stages will be discussed in terms of the McCabe-Thiele method as well as the mixer stage efficiency.
McCabe-Thiele method
In order to determine the theoretical number of stages required for a specific degree of separation, the McCabe-Thiele method can be used in conjunction with the liquid-liquid equilibrium curve of the binary solution. The McCabe Thiele method requires a 45° degree line as well as an operating line, on which the step-off theoretical stages will be determined. The equation of the operating line can be obtained by doing a material balance over the system, as can be observed from Figure 1 below.
Figure 1: Schematic representation of a material balance over a solvent extraction circuit
Equation 6 represents the operating line of a solvent extraction system, after the material balance over the circuit has been rearranged.
[M^(n+) ]_(O,N+1)=A/O 〖([M^(n+) ]〗_(A,N)-[M^(n+) ]_(A,i))+[M^((n+) ) ]_(O,1) (7)
The use of this method has some limitations. In order to approximate the operating line as a linear function, the concentration of the solute must be very low. VA must be completely immiscible in VO and there must be negligible entrainment of VA to ensure that VA and VO stay constant.
To construct the McCabe-Thiele diagram, a good approximation for a starting point, is the incoming concentration of the pregnant leach solution, [Mn+]A,i. The number of stages can be determined by stepping off stages between the operating line and the equilibrium curve, until the concentration of the baron leach solution, [Mn+]A,N, is reached. The number of triangles drawn between these two concentrations, represent the number of theoretical stages. It is important to realise that the actual number of stages will be more than the theoretical number of stages, as equilibrium between the two phases in the solution, is unlikely to be reached completely.
Mixer efficiency
The theoretical number of stages required for a specific degree of separation in a solvent extraction circuit, can be determine using a McCabe-Thiele diagram, as mentioned in section 2.3.1. The McCabe-Thiele method determines the theoretical number of stages required, as it is only valid under the assumption that complete equilibrium is reached between the two immiscible liquid phases. As equilibrium is never completely reached, the mixer efficiency can be incorporated to account for this inaccuracy. The mixer efficiency refers to the actual amount of solute extracted over the theoretical amount of solute extracted if equilibrium was reached. The mixer efficiency is given by Equation 8.
ɳ=([M^(n+) ]_(aq,i) V_(aq,i)-[M^(n+) ]_(aq,1) V_(aq,1))/([M^(n+) ]_(aq,i) V_(aq,i)-[M^(n+) ]_(aq,eq) V_(aq,eq) ) (8)
Effect of varying extractant concentration
A large number of parameters are dependent on the extractant concentration. The extractant concentration is one of the primary factors determining the degree of extraction possible for a solvent extraction circuit. As the O/A ratio increases, the extractant concentration also increases, therefore theoretically the more extractant concentration, or the higher the O/A ratio, the better the extraction of the valuable mineral from the aqueous phase into the organic phase (Spence, n.d.). This is due to the PLS being contacted with more extractant (Spence, n.d.). The extractant concentration is inversely proportional to the volumetric flow rate and has a significant influence on the theoretical number of stages required for a certain degree of extraction. Theoretically, a higher extractant concentration will require fewer equilibrium stages.
It is important to note that the extractant concentration is not constant and varies with O/A ratio (Spence, n.d.). At low O/A ratios, the increase in copper extraction from the aqueous phase to the organic phase, due to an increase in extractant concentration is significantly more than at high O/A ratios (Spence, n.d.). At high O/A ratios, the amount of copper to be extracted is significantly less than the amount of extractant present in the solution. Therefore, the effect of an increase in extractant concentration will have a very small effect on the extraction of copper (Spence, n.d.).
Selectivity can be regarded as a measure of efficiency of the separation. The extractant concentration also has an influence on the selectivity of the solvent extraction system. The selectivity of the copper-iron system is defined in Equation 9.
S=[Cu^(2+) ]_(o,1)/[Fe^(2+) ]_(o,1) (9)
Method
The method is discussed in terms of the preparation of the synthetic leach solution and the extraction test procedure as well as the methodology used to determine certain parameters.
Experimental Procedure
A PLS comprising CuSO4.5H2O and Fe2(SO4)3·xH2O was prepared and added to two separate organic phases consisting of 10 vol% LIX 984N and 15 vol% LIX 984N respectively, diluted in kerosene. The mixtures were mixed until equilibrium was reached and separated in a separating funnel. As expected, the aqueous phase reported to the bottom. This procedure was repeated for different organic phase-aqueous phase ratios (O/A) in order to prepare samples, which was diluted with a factor of a 100, to determine the concentration of dissolved metal ions in both phases.
Performance Parameters
The performance parameters, i.e. distribution coefficient, separation factor and percentage extraction, were determined by use of the concentration data. The aqueous phase concentration of copper and iron was measured and then the organic phase concentration of copper and iron was determined with Equation 7. The distribution coefficient, percentage extraction and separation factor were determined by use of Equations 3, 4 and 5 respectively.
Number of stages in a solvent extraction plant
The theoretical number of stages required for a specific degree of extraction, was determined with the measured aqueous phase concentration of copper and iron as input data. The organic phase concentration of copper and iron had to be determined at different O/A ratios, to construct an equilibrium curve of the liquid-liquid system. It should be noted that the average values have been used for any repeat tests in the experiment. The liquid-liquid equilibrium curves of the 10% v/v and 15%v/v organic phases can be observed in Table 6, Appendix 3. The McCabe-Thiele method was used to step off stages, in order to determine the required number of stages. The operating line used for the McCabe-Thiele method, can be observed in Equation 7.
Effect of varying extractant concentration
The effect of varying concentration was determined by evaluating the theoretical number of stages, copper concentration in the last stage, volumetric flow rate as well as selectivity for the different extractant concentrations. The theoretical number of stages was determined for two different organic phases with different extractant concentrations to compare the effect of extractant concentration. The concentration of copper in the aqueous phase of the last stage was determined by use of the McCabe-Thiele method described in section 3.3. The concentration was obtained by reading the aqueous phase concentration of copper of the last stage from the McCabe-Thiele diagram. The effect on volumetric flow rate was evaluated by use of the McCabe-Thiele operating line, Equation 7. The selectivity of the system was determined for both the 10% v/v and 15% v/v organic phases, by dividing the organic phase copper at the end of the solvent extraction by the organic phase concentration of iron at the end. This was done by using Equation 9.
Results and discussion
The results obtained are discussed in this section. The results and discussion section comprise the performance parameters of the extraction circuit, the theoretical number of stages required for the circuit, the effect of varying extractant concentration as well as an error analysis.
Performance Parameters
The performance parameters are discussed in terms of the equilibrium curve of the solvent extraction circuit, the distribution coefficient, percentage extraction of copper as well as the separation factor.
Distribution coefficient
Figure 2 below exhibits the relationship between the distribution coefficients of copper and iron and O/A ratio.
It can be observed that the distribution coefficient increases with increasing O/A ratio and then decreases again after an O/A ratio of 5 has been reached. This would indicate that the optimum O/A ratio for copper extraction with a 10% v/v extractant is 5. The decrease in distribution coefficient after the optimum ratio of 5 has been reached is expected, as the total amount of Cu(ll) ions available for extraction from the aqueous phase to the organic phase, has already been extracted at an O/A ratio of 5. Therefore, if the O/A ratio keeps increasing, the amount of Cu(ll) ions in the organic phase will stay constant for O/A ratios greater than 5. Because the organic volume keeps increasing, this would result in the copper concentration in the organic phase decreasing again.
The trend of the distribution coefficient increasing with increasing O/A ratio (up until the optimum O/A ratio) is expected. At smaller volumes of organic phase, less of the dissolved copper ions will transfer to the organic phase, since only a limited number of copper-extractant complexes are able to form due to a limited amount of extractant, meaning only a limited amount of Cu(ll) ions can transfer to the organic phase and the rest will remain in the aqueous phase. This will result in lower distribution coefficients. As the organic volume of the extraction mixture increases, more Cu(ll) ions are able to transfer from the aqueous phase to the organic phase, since more copper-extractant complexes can form due to more extractant being availbale.
According to a paper by John R. Spence, the extraction will increase with increasing O/A ratio, meaning more Cu(ll) will be positioned in the organic phase than in the aqueous phase, i.e. an increase in distribution coefficient, as the O/A ratio increases (Spence, n.d.). The results observed in Figure 3 is therefore in accordance with other studies previously done.
It can be observed from Figure 2 that the distribution coefficient of the iron stays relatively constant. This is because iron is not as dependent on the O/A ratio as copper, therefore, will not be greatly affected by a change in O/A ratio. Iron’s dependence on the O/A was however not evaluated thoroughly in this experiment and should be verified by doing a separate experiment specifically aimed at evaluating the O/A dependence of iron. Although it can be observed from Figure 2 that the distribution coefficients of iron are significantly less than the distribution coefficients of the copper. This phenomena will be discussed in section 4.1.3 below.
Note that a theoretical PLS concentration of 22 mg/L was assumed for the iron, as an iron hydroxide precipitation that formed in the PLS resulted in infeasible data.
The raw data used for the determination of the distribution coefficients of copper and iron, at different O/A ratios, can be observed in Table 3, Appendix 1.
Figure 2: Distribution coefficients of copper and iron at different O/A ratios
Percentage extraction
Figure 3 below is a plot of the percentage extraction of both copper as well as iron, for different O/A ratios.
It can be observed from Figure 3 that the percentage extraction of Cu(ll) increases with increasing O/A ratio and then reaches an equilibrium after an O/A ratio of 5 has been reached. This again indicates that an O/A ratio of 5 is an optimum ratio for this copper extraction circuit. The percentage extraction increases between ratios of 0 and 5, because of increasing organic phase volume and therefore also increasing extractant volume. An increase in extractant would increase the extraction of Cu(ll) ions, as more copper-extractant complex bonds will form. Since the copper-extractant complex is more soluble in the organic phase, this will allow for more Cu(ll) ions to transfer to the organic phase as copper-extractant complexes (Spence, n.d.). At an O/A ratio of 5, 98.2% of the Cu(ll) ions have already been extracted, therefore the percentage extraction cannot increase any further, as 100% extraction is very unlikely to be achieved.
Similarly to the distribution coefficient, a paper by John R. Spence, also suggests that the percentage extraction will increase with increasing O/A ratio, therefore this finding is in accordance with other studies previously done (Spence, n.d.).
Similarly to the distribution coefficients of copper and iron, as observed in section 4.1.2, the percentage extraction of copper is significantly higher than the percentage extraction of the iron, which is the desired result. The extractant used for this copper extraction, LIX 984N, is an oxime. It can be observed from Equation 1, that the extractant gives off protons, in order for the Cu(ll) ions to form the copper-extractant bond. The extractant is more selective towards copper, than iron, as a result of the pseudo-macrocyclic structure of the copper-extractant complex (Wilson, 2013). As also mentioned in section 2.1.3, the copper-extractant complex is much more stable, than an iron-extractant complex. This is because the Cu(ll) ions fit the pseudo-macrocyclic structure of the oximes much better than iron does (Wilson, 2013). The selectivity of metals is also very pH dependent. At a pH of 1.65, the extractant is clearly more selective towards Cu(ll) ions. The extractant is strong enough to extract the copper from the aqueous phase, without extracting the iron as well. The percentage extraction of the iron therefore stays relatively constant at a low percentage.
It can be observed from Figure 3, that it is preferable to operate at O/A ratios higher than or equal to 5, as this would result in a very high copper extraction of 98.2%, and an almost insignificantly low iron extraction of a 5.3% average. It should be noted that the maximum percentage extraction can not exceed 100%, but to illustrate the uncertainty of the percentage extraction of copper, the percentage extraction in Figure 3 extends to 120%.
The raw data used to obtain the percentage extraction for each of the two metals present, can be observed in Table 3, Appendix 1.
Figure 3: Percentage extraction of copper and iron at different O/A ratios
Separation factor
Figure 4 is a plot of the separation factor of copper over iron, for different O/A ratios.
It can be observed that the separation factor increases with increasing O/A ratio. This is expected, as the distribution coefficient of the copper increases with increasing O/A ratio, while the distribution coefficient of the iron stays relatively constant with increasing O/A ratio. As copper ions fit the extractant complex better than Fe(ll) ions do, the increase in O/A ratio, therefore also the increase in extractant, will only have an effect on the copper extraction, as the increase in extractant will cause the Cu(ll) to bond to the extractant to form a copper-extractant-complex. The extraction of the Fe(ll) will stay constant.
A high separation factor at high O/A ratios indicates that copper is extracted more readily than iron, at these high O/A ratios. It is therefore preferred to operate at higher O/A ratios, as the desired product is copper and not iron. The increase in separation factor with increasing O/A ratio is also in accordance with the percentage extraction discussed in section 4.1.3. The percentage extraction of copper is the highest at O/A ratios higher than 5 and the percentage extraction of iron is the lowest at O/A ratios higher than five. This substantiates the conclusion that at higher O/A ratios, copper is extracted from the aqueous phase more readily than iron.