稀有金属(英文版) 2015,34(11),818-828

收稿日期：10 October 2014

基金：financially supported by the Presidential Scholars and Technologists Fund of Iran;

Solvent extraction of rubidium from gold waste using conventional SX and new CFE methods

Mohammad Reza Tavakoli Mohammadi Seyed Mohammad Javad Koleini Sepideh Javanshir Hossein Abolghasemi Mahmoud Abdollahy

Mineral Processing Department, Tarbiat Modares University

Mining Engineering Department, University of Birjand

Oil and Gas Center of Excellence, School of Chemical Engineering, College of Engineering, University of Tehran

Abstract：

Solvent extraction(SX) of rubidium(Rb) from leaching filtrate of gold waste(GW) in Mouteh Processing Plant by 18-crown-6(18C6) was studied. High potential of new column flotoextraction(CFE) method in extraction of Rb from dilute solutions was also demonstrated. First, the factors affecting SX of Rb from a synthetic rubidium sulfate solution(containing 100 mg·L^-1Rb) were identified.0.05 mol·L^-118C6 in kerosene, 0.02 mol·L^-1 picric acid in aqueous phase, aqueous to organic(A/O) ratio of 1, p H7 and mixing time of 15 min are the optimum values for affective factors. The three-stage precipitation with sodium carbonate, p H adjustment with sulfuric acid and two-stage evaporation and crystallization were conducted to purify the leaching from impurities such as Fe, Al, Mn, Ca and especially competing cations of K and Na. Almost complete extraction of Rb and K from final filtrate containing 0.08 mol·L^-1picric acid was performed using 0.2 mol·L^-118C6 in kerosene. The Rb and K strippings of 99.12 % and 9.93 %, respectively, are obtained with 2 mol·L^-1nitric acid with A/O ratio of 2. Finally, the performance comparison of the CFE and SX methods in similar conditionsshows increased extraction of K and Rb by 16 % and 5 %,respectively, using CFE method.

Keyword：

Rubidium; Gold waste; Solvent extraction; Column flotoextraction; 18-crown-6;

Author： Mohammad Reza Tavakoli Mohammadi e-mail: mr.tavakolimohammadi@modares.ac.ir;

Received： 10 October 2014

1 Introduction

The silvery-white metallic, very soft and ductile element of rubidium (Rb) is the second most electropositive of the non-radioactive alkali metals and the fourth light metallic element. Low melting point (312.46 K) and suitable mal- leability properties make Rb a photoemission and strongly chemically active element. These characteristics lead to different chemical and electronic applications of Rb [1–3].

Seawater contains 125.0 lg L^-1Rb and 0.3 lg L^-1Cs (on average) [4]. Some potassium minerals and chlorides also contain this element in commercially significant quantities [5]. In addition, during magma crystallization and formation of pegmatite ore, Rb is concentrated along with its heavier analog Cs in the liquid phase and crystal- lizes at last. Therefore, these ores contain acceptable amounts of Cs in the form of pollucite [Cs₂O Al₂O₃4Si O₂] and Li in the form of lepidolite [KRb Li(OH,F)Al₂Si₃O₁₀], and Rb is produced as a by-product from their processing [2]. In addition to these two minerals, Rb is found in leucite [K(Al Si₂O₆)₂], carnallite [KMg Cl₃6(H₂O)₃] and zin- nwaldite [K(Li,Al,Fe)₃(Al,Si)₄O₁₀F₂)₄] in oxide form up to 1 % [6]. Therefore, this element is not found in any mineral as main constituent and is mainly produced as a by-product from Li mineral processing of lepidolite and zinnwaldite [7, 8].

The conventional hydrometallurgical methods for the processing of Li minerals include grinding, concentration,chemical roasting and leaching, during which the valuable minerals of Li and probably Rb and Cs in the ore are entered into leach liquor (filtrate) [9–14]. Numerous studies have not been reported to recover Rb from this filtrate.

Shan et al. [3] proposed solvent extraction (SX) method by 4-tert-butyl-2-(a-methylbenzyl) phenol (t-BAMBP) extractant for the Rb extraction from filtrate similar to Chen et al. [15] and Yan et al. [16]. In their study, the Rb extraction rate of 93 % was achieved with extractant con- centration of 1 mol L^-1, p H 14 and aqueous/organic (A/O) phase ratio of 1:1.

Arnold et al. [17] used 1 mol L^-1solution of 4-sec- butyl-2(a-methylbenzyl) phenol in kerosene for the Cs extraction from filtrate. Scrubbing of the extract with dilute acid and stripping with carbon dioxide plus water gave rise to cesium carbonate solution with less than 0.01 % com- bined alkali metal impurities. They suggested this method for the extraction and purification of Rb.

The ability of various extractants such as sodium te- traphenylborate (TBP), sodium perchlorate and perchloric acid into nitromethane, nitroethane, nitrobenzene, methyl isobutyl ketone and tributyl phosphate was evaluated for extraction of Na and small amounts of K, Rb and Cs by Sekin and Dyrssen [18]. In this study, the highest distri- bution ratios for TBP and perchlorate were achieved using nitrobenzene and nitromethane, respectively. Slater [19] studied the extraction of K, Rb, Sr and Ce polyiodides from different aqueous solutions into nitrobenzene. The extrac- tion coefficients for Na, K, Rb, Cs, Sr and Ce were 0.645, 7.600, 170.000, 659.000, 0.422 and 0.355, respectively, which indicated a high extraction rate of Rb and Cs by nitrobenzene. Sodium triphenylcyanoboron was investi- gated for extraction of Rb and Cs by Lee [20]. These re- searchers showed that this compound was stable for extraction of alkali metals inverse of their tendency to decompose in acidic solutions.

Discovery of crown ethers facilitated the extraction of alkali metals by the so-called guest–host extraction mechanism. These compounds are large ring polyether, and the desired cation is inserted as a guest within their ring. The formation of complexes with cations is caused by electrostatic ion–dipole interaction between the cation and the negatively charged oxygen donor atom in the crown ether. Then, cation-crown ether complex with an anion (which is electrically neutral) enters into the organic phase [21–25]. The stability constants of some of cation-crown ether complexes and schematic of Rb-18-crown-6 (18C6) complex are shown in Table 1 [26, 27] and Fig. 1, respectively.

Some researchers [28–32] studied the extraction of al- kali metal ions by variety derivatives of these compounds; however, great dispersion and small volume of studies re- ported show the need to develop researches in this field. The extraction of Rb from gold waste (GW) of Mouteh Processing Plant in Iran was conducted using the acid washing-sulfation roasting-water leaching process and 97.14 % of Rb transferred into filtrate in optimum condi- tions [33]. The aim of this study was SX of Rb from filtrate using crown ether extractant as the most effective extrac- tion agent in this field. For this purpose, after assessing the effective factors in the SX of Rb from a synthetic aqueous solution, the extraction of this valuable metal from filtrate was studied. Afterward, to demonstrate the high efficiency of the new column flotoextraction (CFE) method in ex- traction of valuable ions from dilute solutions, the extrac- tion rates of Rb in SX and CFE methods were compared.

2 Experimental

2.1 Materials

Synthetic solution (containing 100 mg L^-1Rb) was pre- pared by dissolving Rb₂SO₄(MERCK) in distilled water. 0.005–0.025 mol L^-1picric acid (ROYALEX) was added into the synthetic solution to produce bulky anions. The organicphasewas0.0005–0.1000 mol L^-118C6 (MERCK) extractant with chemical formula of C₁₂H₂₄O₆into kerosene, benzene or toluene as diluent. Silicone oil was procured from Shin-Etsu. Other chemicals such as sodium carbonate, sulfuric acid and nitric acid were ob- tained from Merck Company.

Table 1 Stability constants of cation-crown ether complexes  下载原图

Table 1 Stability constants of cation-crown ether complexes

Fig. 1 Rb-18C6 complex (guest–host mechanism)

The ion-pair extraction equilibrium of alkali cation (M?)-crown ether complex, ML?, with picrate ion (Pic^-) is as follows:

where O subscript indicates the organic phase. ML? Pic_O^-denotes the combination of the extracted ion-pair into the organic phase [22, 23].

2.2 Methods

2.2.1 Purification of filtrate and concentration methods

Acid washing of the GW was performed using 5 mol L^-1HNO₃at 85 °C for 5 h. Roasting of the residue at 910 °C for 30 min with sodium sulfate and calcium chloride wascarried out. Water leaching with solid/liquid ratio of 1.69 at 58.51 °C for 31.63 min resulted in 97.14 % Rb extraction into the filtrate. The analysis results of Rb and other ele- ments into the filtrate are shown in Table 2.

Table 2 Composition of filtrate (mg L^-1)  下载原图

Table 2 Composition of filtrate (mg L^-1)

The concentration of cations competing with Rb in Table 2 and stability constants of cation-crown ether complexes in Table 1 show that prior to the SX of Rb, Na and Ca concentrations should be significantly reduced. To do this, sodium carbonate equivalent of 1/2 stoichiometric Ca ions was gradually added to filtrate at room tem- perature to remove trace impurities such as Fe, Al, Mn and half of Ca. After filtration and partial washing of the residue, the precipitation process of filtrate was repeated twice to decrease the volume of filtrate to half by evaporation in each stage. After these three stages, a fil- trate with p H 11.93 and the composition mentioned in Table 3 were obtained.

Afterward, a fraction of carbonate was precipitated with the addition of 1 mol L^-1sulfuric acid and adjustment of p H equal to 7–8. The composition of filtrate after this stage is presented in Table 4. Evaporation and rapid cooling of filtrate up to 0–5 °C provided the conditions for crystal- lization and removal of a fraction of sodium compounds in solution. The final filtrate was obtained by repeating this stage, and its composition is shown in Table 5. It is notable that the significant concentration of Sr for extraction is in this table.

Table 3 Composition of filtrate after three precipitation stages (mg L^-1)  下载原图

Table 3 Composition of filtrate after three precipitation stages (mg L^-1)

Table 4 Composition of filtrate after p H adjustment (mg L^-1)  下载原图

Table 4 Composition of filtrate after p H adjustment (mg L^-1)

Table 5 Composition of filtrate after two evaporation and crystal- lization stages (mg L^-1)  下载原图

Table 5 Composition of filtrate after two evaporation and crystal- lization stages (mg L^-1)

2.2.2 SX and strip methods

According to the aqueous to organic (A/O) ratios desired, certain volumes of the aqueous (synthetic solution or final filtrate) and organic phases were mixed using magnetic agitator (500 r min^-1) for 15 min. The content of the agitator was transferred into a separatory funnel for separating. Complete separation occurred at 20 min. The equilibrium p H was measured. Concentration of Rb in the aqueous phase was determined by inductively coupled plasma (ICP) analysis, and consequently, organic phase concentration was obtained by mass balance. Loaded or- ganic phases with various concentrations of nitric acid solution were mixed for 10 min. The analysis procedure was the same as previous.

2.2.3 CFE method

In recent years, dissolved nitrogen predispersed solvent extraction (DNPDSE) method was invented for improving the conventional SX equipment performance, especially for dilute solutions. Mixing operation is based on bubble dis- persion of the organic phase, instead of its drop dispersion into the aqueous phase. This substitution is intended to increase the two phases’ contact area (to improve the re- covery of metal ions from the dilute solutions) and also to enhance the buoyancy force of organic phase (to improve separation of two phases). Organic phase bubbles in size range of \100 lm were produced during a long multi-step preparation process and a system under pressure similar to dissolved air flotation (DAF) method [34, 35]. CFE is the modified method of DNPDSE with the aim of improving the system performance, better separation of phases and facilitation of extraction process and boosting the safety level through changes in bubble production system. In this method, the main drawbacks of the DNPDSE method were resolved by pump and spargers similar to the downcomer of Jameson cell in contactor, rather than the system underpressure in DNPDSE contactor which produces organic phase bubbles in size range of 100–600 lm. The CFE contactor and its schematic diagram are shown in Figs. 2 and 3.

Fig. 2 CFE contactor used in experimental work

Fig. 3 Schematic diagram of CFE contactor

In this method, a certain volume of purified filtrate was transferred to CFE cell. Bubble injection of the organic phase was performed in eight steps through mixing tube with a flow rate of 1.5 L min^-1. At each step, after the rise of organic phase bubbles through the filtrate and simultaneous extraction and separation op- erations, sampling from the filtrate was performed in upper part of the cell. Extraction of Rb was calculated at each step and compared with SX method with the same A/O. In this study, downcomer nozzle was 2 mm in diameter.

3 Results and discussion

3.1 Solvent extraction of Rb from synthetic aqueous solution

3.1.1 Effects of extractant concentration, anion concentration and diluents type

In general, in the SX processes, extraction capacity of ex- tractants is improved by the appropriate diluents, and in- creased concentration of extractants could be beneficial to increase the extraction of the desired element. In the SX of Rb, the presence of bulky anions is necessary to facilitate the dissolution of cations in the organic phase, because large anions with low free hydration energy and charge density are extracted more easily with organic phase [21–25]. The effect of these three effective factors was first evaluated.

Figure 4 shows the effect of extractant concentration and type of diluents on the extraction of Rb and equilib- rium p H in the absence of picrate anion (Pic^-). According to Fig. 4a, Rb extraction rates are low, and the best result is obtained by 0.005 mol L^-118C6 in benzene. The inter- esting note is the decrease in Rb extraction with the in- crease in extractant concentration.

By comparing Fig. 4a and b, it can be concluded that in the absence of Pic^-, transfer of Rb-18C6 complexes to the organic phase is highly dependent upon equilibrium p H, which is affected by 18C6 concentration. Increased con- centration of 18C6 by decreasing equilibrium p H of the aqueous phase decreases the transfer of Rb-18C6 com- plexes to the organic phase. Therefore, p H has an important role in the extraction of Rb.

Figures 5, 6 and 7 show the effect of 18C6 and Pic^-concentrations on Rb extraction and equilibrium p H of theaqueous phase in presence various diluents. Comparison of the results indicates that the increase in Pic^-and 18C6 concentrations causes the improvement of Rb–18C6 com- plexes transfer into organic phase and the negligible in- crease in equilibrium p H.

Fig. 4 Effect of 18C6 concentration in various diluents on a Rb extraction and b equilibrium p H

Fig. 5 Effect of picric acid concentration in aqueous phase and 18C6 concentration in toluene on a Rb extraction and b equilibrium p H

Fig. 6 Effect of picric acid concentration in aqueous phase and 18C6 concentration in kerosene on a Rb extraction and b equilibrium p H

Fig. 7 Effect of picric acid concentration in aqueous phase and 18C6 concentration in benzene on a Rb extraction and b equilibrium p H

In Fig. 8, the Rb extraction results are compared be- tween 0.05 mol L^-118C6 and different concentrations of Pic^-, and the highest extraction capacity of Rb is obtained using kerosene as diluents.

The effect of increased concentrations of 18C6 and Pic^-on Rb extraction and equilibrium p H of the aqueous phase is shown in Fig. 9. According to Fig. 9, the increase in extraction is not significant. Picric acid concentration up to 0.025 mol L^-1decreases the extraction of Rb, indicating the impact of the aqueous phase p H on Rb extraction. Thus, 0.05 mol L^-118C6 and 0.02 mol L^-1picric acids were chosen as optimum values, and 87.10 % Rb extraction is obtained under this condition.

3.1.2 Effect of mixing time

To select the suitable time for mixing, extraction ex- periments with 0.05 mol L^-118C6 in kerosene and 0.02 mol L^-1picric acid in aqueous phase were repeated at different time. The results in Fig. 10 confirm the suitable mixing time to be 15 min.

3.1.3 Effect of A/O ratios

According to the results in Fig. 11, the increase in ex- traction resulting from the decrease in A/O ratio is notsignificant. Therefore, A/O ratio = 1 is preferred for the following experiments.

Fig. 8 Effect of diluent type on Rb extraction in 0.05 mol L-118C6 and various concentrations of picric acids

Fig. 9 Effect of 18C6 and picric acid concentrations on Rb extraction

3.1.4 Effect of p H solution

It was previously identified that p H of the aqueous solution was effective in the extraction process, as seen in Sect. 3.1.1. For this purpose, the effect of p H solution on Rb extraction was investigated by adding certain amount of sulfuric acid (Fig. 12). According to Fig. 12, the increase in p H up to *7 causes the increase in Rb extraction to *91 %.

Fig. 10 Effect of mixing time on Rb extraction

Fig. 11 Effect of A/O ratio on Rb extraction

Fig. 12 Effect of solution p H on Rb extraction (0.05 mol L-118C6 in kerosene,0.02 mol L-1picric acid)

Table 6 Filtrate chemical composition for SX and CFE experiments (mg L^-1)  下载原图

Table 6 Filtrate chemical composition for SX and CFE experiments (mg L^-1)

3.2 Solvent extraction of Rb from filtrate

Clearly, the efficiency of SX process is higher for dense solutions [36]. However, in this study, owing to the high cost of extractant and difficulty in preparation of the re- quired filtrate, the solution was prepared by tenfold dilution of the final filtrate. Chemical composition of leach liquor is given in Table 6.

In previous parts, 0.05 mol L^-118C6 and 0.02 mol L^-1picric acids were selected as optimum concentrations for 100 mg L^-1Rb in the synthetic solution. Taking into ac- count the concentrations of Rb, K, Na and Ca in the filtrate, several experiments were conducted as presented in Table 7. According to the results, with the addition of 0.08 mol L^-1picric acid to the filtrate, almost complete extraction of Rb and K is obtained by 0.2 mol L^-118C6 in kerosene.

3.3 Rb stripping

According to the research by Mohite and Khopkar [29], nitric acid was used for Rb stripping from organic phase. Of course, the concentrations of nitric acid used for separating Rb and K are low because of higher complex formation tendency of 18C6 with K than Rb, according to Table 1.

Table 8 shows experimental conditions and results ob- tained for Rb and K strippings in various concentrations ofnitric acid. According to Table 8, the satisfactory separa- tion of Rb and K is obtained by 2 mol L^-1nitric acid in A/O ratio = 2.

Table 8 Conditions and stripping results of Rb and K  下载原图

Table 8 Conditions and stripping results of Rb and K

3.4 Process flow sheet

Figure 13 shows the flow sheet suggested for extraction of valuable elements from GW.

3.5 Rb extraction using CFE method

3.5.1 Effect of silicone oil on Rb extraction

Silicone oil was used to improve the foaming property of the organic phase to produce bubbles. Since surfactant may degrade the extraction properties of the solvent [37], the effect of silicone oil on Rb extraction was assessed. Ex- perimental conditions and its results are given in Table 9. It is clear that silicone oil does not affect the extraction performance.

3.5.2 Comparison of extraction performance of SX and CFE methods

The extraction performance of the two methods is compared in Fig. 14. Clearly, the results show increased extraction rates of Rb and K in CFE method. However, as expected, increased extraction of K is more than that of Rb due to higher complex formation tendency of 18C6 with K. For example, with A/O ratio of 9, extractions of K by SX andCFE methods are 31.11 % and 38.81 %, and extractions of Rb by SX and CFE methods are 11.38 % and 13.22 %, respectively. Obviously, the higher extraction rates would be achievable by performing CFE experiments with lower A/O ratio due to greater impact of increased contact surface with volume of the organic phase increasing.

Table 7 Extraction and concentration of cations in filtrate (A/O ratio = 1, p H = 7, diluent: kerosene)  下载原图

Table 7 Extraction and concentration of cations in filtrate (A/O ratio = 1, p H = 7, diluent: kerosene)

Fig. 13 Proposed flow sheet of Rb extraction from GW

Another point is that the extraction time in CFE method (*1 min) is much less than that in SX method (*15 min). Regarding the extraction time of 1 min, extraction rates of K and Rb with A/O ratio of 9 using SX method are ob- tained as 23.33 % and 7.97 %, respectively, which repre- sents increased extraction rate of about 16 % for K and 5 % for Rb using CFE method. Increased extraction rate and reduced operation time by CFE contactor indicate thehigh potential of this new method in improving the SX equipment performance, especially for dilute solutions.

Table 9 Experimental conditions and results of silicone oil effect on Rb extraction (p H 7, 15 min,500 r min-1)  下载原图

Table 9 Experimental conditions and results of silicone oil effect on Rb extraction (p H 7, 15 min,500 r min-1)

Fig. 14 Performance comparison of SX and CFE methods in various A/O ratios

4 Conclusion

Development of commercial applications of the rare alkali metal of Rb in recent years due to the discovery of its specific physical and chemical properties has encouraged researchers for extraction of Rb from its sources. Owing to limited resources of minerals containing Rb in nature, Rb extraction from secondary sources was evaluated in this study. GW of Mouteh Processing Plant, which contains valuable elements such as Rb, Ti, Ce, Nd, and La, is transferred to the tailings dam. In the previous study, through optimization of acid washing-sulfation roasting- water leaching process of GW, 97.14 % extraction rate of Rb was achieved. The aim of this study was SX of Rb from the resulting filtrate using 18C6 extractant. The factors af- fecting extraction of Rb were studied from synthetic sulfate solution to ignore the effect of interfering elements. Ac- cording to the results, the concentration of picric acid plays a major role in the transfer of Rb-18C6 complexes to the or- ganic phase. Adding 0.02 mol L^-1picric acid to the aqueousphase containing 100 mg L^-1Rb follows *90 % extraction of Rb into organic phase containing 0.05 mol L^-118C6 in kerosene. Regarding the impurity concentration in leaching filtrate such as Fe, Al, Mn and Ca, especially competing cations of K and Na, concentration and purification pro- cesses are necessary before the SX process. The processes used include three-stage precipitation by sodium carbonate, p H adjustment with sulfuric acid and two-stage evaporation and crystallization. During these steps, removal of Fe, Al, Mn and Mg and a significant reduction in Na and Ca con- centration occur, and consequently, the concentration of Rb in the final filtrate increases from 51.34 to 1985.76 mg L^-1. However, increased concentration of potassium (from 35.62 to 1113.67 mg L^-1) causes the extra consumption of the extraction agent due to the higher complex formation ten- dency of K to 18C6 than that of Rb. Almost complete ex- traction of Rb from a filtrate containing 198.58 mg L^-1Rb, addition of 0.2 mol L^-118C6 in kerosene and 0.08 mol L^-1picric acid in filtrate, is necessary, which is four times more than used agent concentration in synthetic solution. Fortu- nately, less tendency of Rb relative to K for complex for- mation with 18C6 causes good separation of the two cations in the stripping step. Using 2 mol L^-1dilute nitric acid with A/O ratio of 2 results in 99.12 % and 9.93 % stripping of Rb and K, respectively. Notable achievement is the sig- nificant increase in Sr concentration in the final filtrate from 23.89 to 998.22 mg L^-1. The extraction of Sr is currently under study. Comparison of the extraction rates of Rb from filtrate using SX and CFE methods indicates better perfor- mance of this new method for extracting valuable ions from dilute solutions. At the same A/O ratio, the extraction rates of K and Rb increase by 16 % and 5 %, respectively, compared with conventional method.