Single-Stage Extraction and Separation of Co2+ from Ni2+ Using Ionic Liquid of [C4H9NH3][Cyanex 272]

The purpose of this study was to optimize the extraction conditions for separating Co2+ from Ni2+ using N-butylamine phosphinate ionic liquid of [C4H9NH3][Cyanex 272]. A Box–Behnken design of response surface methodology was used to analyze the effects of the initial pH, extraction time, and extraction temperature on the separation factor of Co2+ from sulfuric acid solution containing Ni2+. The concentrations of Co2+ and Ni2+ in an aqueous solution were determined using inductively coupled plasma-optical emission spectrometry. The optimized extraction conditions were as follows: an initial pH of 3.7, an extraction time of 55.8 min, and an extraction temperature of 330.4 K. The separation factor of Co2+ from Ni2+ under optimized extraction conditions was 66.1, which was very close to the predicted value of 67.2, and the error was 1.7%. The equation for single-stage extraction with high reliability can be used for optimizing the multi-stage extraction process of Co2+ from Ni2+. The stoichiometry of chemical reaction for ion-exchange extraction was also investigated using the slope method.


Introduction
Cobalt is a strategic metal that has many industrial applications, e.g., rechargeable batteries, superalloys, cemented carbide, and colorants [1,2]. This metal can be extracted or recycled from natural and secondary resources, e.g., Ni-Co laterites [3], stainless steel scarps [4], and spent lithium-ion batteries [5]. Therefore, to obtain the cobalt products, a leaching process is needed [6]. The multicomponent mixtures of metal ions, e.g., Co 2+ , Ni 2+ , Mn 2+ , Al 3+ , Zn 2+ , and others, are coexisting in the leaching solutions [4]. Efficient separation of Co 2+ from leaching solution containing Ni 2+ is the key technology for the final preparation of high-purity cobalt or nickel products [7,8].
As a matter of fact, two metals, cobalt and nickel, are difficult to completely separate due to similar physicochemical properties [9,10]. Based on a novel scale micro-reactor, the efficient separation of Co 2+ from Ni 2+ in sulfuric acid solution was achieved using the organic extractant of Cyanex 272, and this separation method was a two-stage countercurrent extracted process [11]. Using the five-stage extracted process, the separation factor of Co 2+ from Ni 2+ (β Co/Ni ) could be reached 9.5 × 10 3 using the new organic extractant of di-decylphosphinic acid [12]. About >99% of Co 2+ and 11% of Ni 2+ were extracted using the extractant of primene JMT-Cyanex 272 with the process of four-stage counter extraction [13], and the efficient separation of the two metal ions was achieved based on this process. In addition, extractants of MEXTRAL 507P [14], INET-3 [14], and PC88A [15] were also used for the separation of Co 2+ from Ni 2+ . According to the above published articles, the efficient separation of Co 2+ from Ni 2+ in sulfate solution was generally realized using multi-stage extraction, and the experimental conditions of single-stage extraction were one of the essential factors for designing multi-stage extraction.
However, solvent extraction using organic extractants has several drawbacks, such as the irreversible loss of the organic phase [16] and easy volatilization of extractants [17]. For example, the oxidation reaction of primary amine N1923 may occur in the separation of V(V) from the leaching solution, and this process results in irreversible loss of organic extractant [18]. In addition, some organic extractants are easily volatile during extraction, and the organic vapors that fill the workshop may harm workers and cause environmental pollution problems [19]. Therefore, the demerits of solvent extraction using common organic extractants may lead to increased production costs and environmental problems [20,21]. How to reduce or avoid these problems in scientific research is one of the current research directions [22].
To our best knowledge, ionic liquids (ILs) have been used in solvent extraction for separating metal ions [4,23,24] due to their unique properties, i.e., low vapor pressure, high dissolving capability, chemical/thermal stability [25], and designable structure [26,27]. For example, the efficient and sustainable separation of Ho, Er, and Lu from Y was achieved using the functionalized ionic liquid of [N 1888 ][NDA] [28]; the new ionic liquid of [N 14111 ][DBP] exhibited high efficiency in uranium separation [29]; and the extraction efficiencies of Cd(II), Cu(II), and Zn(II) were about 85, 67, and 69% using the hydrophobic trioctylammonium-based ionic liquid of [HTOA][adipate] [30]. If more ionic liquids are used as extracted agents for the separation of Co 2+ from Ni 2+ , the environmentally friendly separation process may be further investigated based on the single-stage extraction [31], and different kinds of ionic liquids may be applied in the recovery of metal ions from secondary waste resources.
With the aim of reducing the irreversible loss of the organic phase and decreasing the environmental pollution in the extraction process, the task-specific ionic liquid (TSIL) of [C 4 H 9 NH 3 ][Cyanex 272] was synthesized using a one-step reaction. The FT-IR, 1 H NMR, and 13 C NMR of TSIL were made with the aim of analysis of purity and impurities. The optimization of extraction conditions for the separation of the two metal ions was determined by the Box-Behnken design (BBD) [32], which is a typical method of response surface methodology (RSM) [32]. Optimization of the extraction conditions for metal ions, e.g., Cu 2+ [33], Cr 3+ [34], and V 4+ [35], were successfully performed using the BBD. The BBD was used to optimize the initial pH value (pH ini ), extraction time (t e ), and extraction temperature (T e ) with the aim of achieving the maximum separation factor of Co 2+ from Ni 2+ (β Co/Ni ). In addition, the extraction mechanism was also investigated using the slope method according to the published papers.

Chemicals and Reagents
The bis (2,4,4- (Figure 1), respectively. All the chemicals and reagents were of reagent grade and used without further purification unless otherwise stated. The simulated sulfuric acid solution containing 14 g/L of Co 2+ and 14g/L of Ni 2+ [4,23] was prepared by dissolving the NiSO 4 ·6H 2 O and CoSO 4 ·7H 2 O in deionized water. With the aim of obtaining a purity solution, the simulated solutions should pass through a filtration membrane with an aperture of 0.45 microns. The pH ini of the simulated solutions was adjusted by adding H 2 SO 4 or NaOH solutions.

Analytical Methods
The spectra of FT-IR were obtained using a SEPCTRUM GX1 instrument (PerkinElmer, Waltham, MA, USA) at room temperature. The 1 H and 13 C NMR was obtained using an NMR spectrometer (Avance-III 400 MHz, Bruker, Zurich, Switzerland). A pH meter (Delta 320, Zurich, Switzerland) with an uncertainty of 0.01 was used to measure the pH ini of the solutions. The concentrations of the two metal ions were analyzed using inductively coupled plasma-optical emission spectrometry (ICP-OES, Optima 8000, PerkinElmer, Waltham, MA, USA). The simulated solutions containing Co 2+ and Ni 2+ were diluted by deionized water to the concentrations of 0.1-40 ppm for the next tests.

Single-Factor Experiment
The experiments were carried out by mixing an equal volume of the organic phase and simulated solution (each 15 mL) in a separating funnel using a water-bath thermostatic oscillator (SHA-C, Jie Rui Er, Changzhou, China). When the extraction was completed, the two immiscible phases were separated for analysis. The pH values in the sulfuric acid solution before and after extraction were determined using a pH meter. The concentrations of two metals before and after extraction were determined by ICP-OES. The effects of pH ini , t e , and T e were first studied using single-factor experiments as follows: one extraction condition was changed when the other conditions were kept constant in every experiment, and the influences of each condition were studied by analyzing the concentration of metal ions in the solutions. All experiments were carried out three times, and the average values were obtained for the diagraph analysis under normal pressure unless otherwise stated.

Optimizing the Extraction Conditions by BBD
Design-Expert Ver. 8.0.6 (Trial version, Stat-Ease, Minneapolis, MN, USA) was used to optimize the extraction conditions [36]. According to the single-factor experimental results, three independent factors, i.e., pH ini (A), t e (B), and T e (C), were found to be responsible for the β Co/Ni . With the aim of analyzing the A, B, and C, three-factor and three-level Box-Behnken experimental design was used to conduct response surface analysis [37], and the three factors were optimized at three levels (−1, 0, −1) (Table 1). Based on the experimental data, regression analysis was made using the following second-order polynomial equation: The significance of the equations was studied according to analysis of variance (ANOVA). The significance of each coefficient and the interaction between the independent variables were also investigated using the obtained p-value. In addition, real experiments were made to evaluate the accuracy of the optimized extraction conditions.
where V 1 (L) and V 2 (L) represent the volumes of the stock and raffinate solution, respectively. C 1(Co) (g/L), C 2(Co) (g/L), C 1(Ni) (g/L), and C 2(Ni) (g/L) represent the concentrations of Co 2+ and Ni 2+ in the stock and raffinate solution.

Single-Factor Experiment
The concentration of [C 4 H 9 NH 3 ][Cyanex 272] was set as 0.24 mol/L considering the concentration of Co 2+ in the solution is 0.24 mol/L [4]. With the volume ratio of organic phase to the aqueous phase (O/A) = 1:1, the effect of pH ini on β Co/Ni is shown in Figure 3a when t e = 60 min and T e = 303.15 K. The β Co/Ni first increases and then decreases with the growth of pH ini . According to Equation (4), the β Co/Ni is proportional to E Co and inversely proportional E Ni , and the final value of β Co/Ni is determined by the ratio of E Co and E Ni .
The influence of t e on β Co/Ni is shown in Figure 3b when T e = 303.15 K, pH ini = 3.5, and O/A = 1:1. The β Co/Ni first increases and then decreases with the increase in t e due to the different extraction kinetics of Co 2+ and Ni 2+ [38]. The effect of T e on β Co/Ni is shown in Figure 3c when t e = 45 min, pH ini = 3.5, and O/A = 1:1. The β Co/Ni increases with the growth of T e . The reason is that higher temperature reduces the viscosity and surface tension of the extraction system, which is beneficial to mass transfer [37]. In addition to this, the extraction reaction of Co 2+ using extractant of [C 4 H 9 NH 3 ][Cyanex 272] is an endothermic reaction, and the higher T e is conductive to the reaction [10]. However, considering the cost of the extraction process, the highest temperature in the next experimental design was set at 333.15 K.

Optimization of the Conditions for Extraction Process
Seventeen experiments were first designed using BBD and then carried out one by one, and the results of these experiments are listed in Table 2, which was used to obtain the predicting equation. At least one variable is different in each group of experiments, and three levels of pH ini are 1.50, 3.75, and 6.00, and three levels of t e are 15.00, 52.50, and 90.00 min, and three levels of T e are 298.15, 315.65, and 333.15 K, and the biggest value of β Co/Ni is 65.51 when pH ini = 3.75, t e = 52.50, and T e = 315.65. According to a regression analysis from the experimental data, the β Co/Ni could be explained as Equation (5)   The variance of the quadratic prediction Equation (5) is summarized and listed in Table 3. The determination coefficient (R 2 = 0.99) of Equation (5) is close to 1, which indicates that the effective correlation between predicated and measured values are obtained using this equation. The adjusted determination coefficients (adj R 2 = 0.98) are also close to 1, which demonstrates that the equation is significant for accurately predicting the β Co/Ni . In addition to this, the F-value is 159.51 and the p-value < 0.0001, which implies that the equation is significant too. The p-value of the lack of fit is 0.0548, which indicates that the lack of fit is insignificant compared to the pure error. The lower coefficient of variation (C.V.% = 3.25) indicated that small deviation and high reliability of the experimental value. The regression equation can be used for predicting the experiment results instead of the real extractive experiments, and it can also be used to predict and analyze the β Co/Ni . The values of A, B, C, A 2 , B 2 , and C 2 are significant (p < 0.05), and the other values are not significant (p > 0.05). The influence degree of various factors on the β Co/Ni is ranked in the following order: C > B > A. The coefficients of two factors, i.e., AB, AC, and BC, are 0.8668, 0.8028, and 0.5835, respectively. The effect degree of interaction based on two factors on the β Co/Ni is ranked as BC > AC > AB. The BC has the biggest effect on the β Co/Ni compared with the other two interaction factors of AC and AB, and the influent of BC on β Co/Ni is shown in Figure 4. As shown in Figure 4b, the red zone represents the higher β Co/Ni under the different T e and t e , and the efficient separation of the two metal ions will be operated according to the equation prediction.

Verification of Equation
Based on the quadratic equation, the optimum extraction conditions of Co 2+ from Ni 2+ were calculated as follows: a pH ini of 3.7, a t e of 55.8 min, a T e of 330.4 K, and the predicted β Co/Ni of 67.2. The verification experiments ( Figure 5) were tested with the optimum extraction conditions, and the actual β Co/Ni was 66.1. The experimental value was close to the predicted value, and the experimental error was small. The experimental results indicate that Equation (5) can better predict the β Co/Ni for single-stage extraction of Co 2+ from Ni 2+ . With the optimized single-stage extraction conditions, the separation factors of Co 2+ from Ni 2+ using Cyphos IL 104 [4], Cyanex 301 [39], Primene JMT-Cyanex 272 [13],

Extraction Mechanism of [C 4 H 9 NH 3 ][Cyanex 272]
The slope method can be used to determine the stoichiometry of chemical reactions for metal ion extraction [40,41]. The extraction equilibrium constant (K) and distribution ratio of the two phases (D) can be obtained using balanced chemical reactions. The [C 4 H 9 NH 3 ][Cyanex 272] is a protic ionic liquid [42], and the cationic group of C 4 H 9 NH 3 ] + is very soluble in the aqueous phase (Supplementary Materials Figure S5), but the solubility of [C 4 H 9 NH 3 ][Cyanex 272] in the aqueous phase is very poor (Supplementary Materials Figure S6, and the solubility is about 0.45 ± 0.02 g/100 g H 2 O at the temperature of 298.2 K and normal pressure). It is assumed that protons can be completely transferred during the extraction reaction. The extractive mechanism of Co 2+ in the sulfuric acid medium by [C 4 H 9 NH 3 ][Cyanex 272] is the ion-exchange reaction [43], and the chemical reaction can be expressed as Equation (6).
The extractive equilibrium constant of Co 2+ (K Co ) can be listed as Equation (7) according to Equation (6).
where the C 1 and C 2 (mol/L) represent the concentrations of Co  (8).
Equation (9) can be obtained by taking the D Co into the K Co .
Equation (10) can be obtained by taking the logarithm of K Co The value of m in Equation (10) can be determined by the slope method ( Figure 6). The m is 2.02. The value of m is approximately 2.0. Therefore, the equation of chemical reaction can be listed as Equation (11)

Stripping for TSIL-Based Extraction Phase and Its Recycling Use
After single-stage extraction, the TSIL-based extraction phase containing Co[Cyanex 272] 2 can be stripped with 0.5 mol/L H 2 SO 4 ( Figure 7). The stripping percentage of Co 2+ from TSIL-based extraction phase was about 99.2% with the optimized stripping conditions, i.e., stripping phase ratio of 1:1, stripping time of 35 min, and stripping temperature of 308 ± 1 K. The product of CoSO 4 ·7H 2 O was obtained from the stripping phase using evaporation crystallization [44]. The product of Ni(OH) 2 can be obtained from the raffinate of the aqueous phase containing Ni 2+ using the precipitation method [45]. The TSIL-based extraction phase can be recycled for extraction after activation. The specific steps for activation are as follows: (1) the TSIL-based organic phase from stripping is firstly washed using the deionized water (the volume ratio of deionized water to ILs-based organic phase was 1:4), and the aqueous phase containing metal ions, which remained in the organic phase, is completely washed by new deionized water; (2) the reusable TSIL-based organic phase is obtained by the activating reaction of the washed organic phase and commercial n-butylamine solution (the reaction was similar to the Section 2.3), and the recycled TSILbased organic phase (about 96.4% yield) can be reused for the single-extraction process.

Conclusions
In this study, the optimal extraction conditions of Co 2+ from Ni 2+ were determined using the RSM method to be: pH ini of 3.7, t e of 55.8 min, and a T e of 330.4 K. The real yield of β Co/Ni obtained by the validation test was 66.1, and the error was 1.7%, which was very close to the predicted value of 67.2. The result shows that the equation is efficient and feasible, with practical significance. The obtained Equation (5) of single-stage extraction is useful to design the multi-stage extraction for efficient separation of Co 2+ from Ni 2+ , and the multi-stage extraction can be used in industrial applications. The chemical reactive equation was determined using the slope method. Additionally, the recycled TSIL-based organic phase with 96.4% yield can be reused for the single-extraction process.