18-Crown-6 and Dibenzo-18-crown-6 Assisted Extraction of Cesium from Water into Room Temperature Ionic Liquids and Its Correlation with Stability Constants for Cesium Complexes

The pH-profiles of the extraction of Cs+ into four conventional (1-butyl-3- methylimidazolium hexafluorophosphate and bis[trifluoromethyl)sulphonyl]imides of 1- butyl-3-methylimidazolium, 1-hexyl-3-methylimidazolium, and 1-(2-ethylhexyl)-3-methylimidazolium) and two novel (trioctylmethylammonium salicylate and tetrahexyl-ammonium dihexylsulfosuccinate) room temperature ionic liquids have been determined both in the absence and in the presence of crown ether (18-crown-6 or dibenzo-18-crown-6). The pH-profiles of distribution ratio of crown ethers have been established in the same conditions. The relationship of cesium extraction efficiency both with the stability of its complexes with crown ethers and crown ethers’ distribution ratio has been clarified.


Introduction
Hydrophobic crown ethers are applied as extractants in the separation and recovery of 137 Cs and 90 Sr radioactive isotopes from high-level nuclear waste solutions, containing high concentrations of acids and salts, into molecular solvents [1][2][3][4][5]. Room-temperature ionic liquids (RTILs) are attracting increasing attention in solvent extraction processes due to important advantages over conventional organic diluents such as negligible vapor pressure, low flammability, moisture stability, unusual extraction properties and possibility to eliminate aqueous phase acidification [6][7][8][9][10][11][12][13]. It was demonstrated that extraction efficiency of RTIL can be modulated by using complexing agents, for example, crown ethers. Dai et al. [8] discovered that highly efficient extraction of strontium ions can be achieved when dicyclohexane-18-crown-6 is combined with RTILs. Rogers et al. [9] and Bartsch et al. [12] reported the extraction of various alkali metal ions with crown ethers in RTILs. Visser and Rogers demonstrated that octyl(phenyl)-N,N-diisobutylcarbamoylmethylphosphine oxide dissolved in RTILs enhances the extractability of lanthanides and actinides in comparison to conventional organic solvents [13]. The extraction of silver ions was found to be greatly enhanced by a combined application of RTIL and calix [4]arene compared to that of chloroform [14]. In addition, the task-specific RTILs with complexing functionality built in the RTIL cation have been reported [15,16]. Recently the efficiency of chelate extraction of 3d-cations with 8-sulfonamidoquinoline [17], Pu(IV) with carbamoylmethylphosphine oxide [18] and uranyl ion with tributylphosphate (TBP) [19] from aqueous phase into RTILs was reported. The higher selectivity of dibenzo-18-crown-6 to K + over Na + in N-octadecylisoquinolinium tetrakis [3,5-bis(trifluoromethyl)phenyl]borate compared with that in molecular solvents suggests that RTIL provides a unique solvation environment for the complexation of crown ethers with the ions [20].
Besides the issues of cation, ligand and complex solubility in water and in RTIL, the relative stabilities of complex formation in both phases are of significant importance for extraction selectivity. In our recent communication [21] we have reported on the thermodynamics of complex formation of cesium ions with 18-crown-6 (18C6, L) in six hydrophobic RTILs (see Scheme  The present work aims to study the extraction capability of cesium by the above mentioned RTILs in the presence of 18C6 and dibenzo-18-crown-6 (DB18C6, L) and to clarify its relationship with the stability of complexes formed in aqueous and organic phases. Another objective is to study the impact of crown ether's nature on the extraction efficiency. For this reason the two crown ethers with similar cavity, but different hydrophobicity (18C6 and DB18C6) have been chosen. Scheme 1. Structures of the RTILs studied.

Results and Discussion
The results of the extraction experiments are presented in Tables 1 to 3  [DHSS] is the first example of a negative influence of crown ether on a cation extraction into organic phase. Probably, the main reason is that crown-ether competes with a RTIL's anion for cesium and, being better soluble in water, provides a partial transfer of the metal ion into an aqueous phase. This suggestion is indirectly supported by the low value of [Cs(18C6)] + stability constant in [THA][DHSS] (logK 1 ), Table 4, which appears to be even lower than in water.
For DB18C6 an enhancement of cesium extraction is much more pronounced then for 18C6, Table 3. Notably, increasing HNO 3 concentration in an aqueous phase decreases cesium's recovery for all RTILs, but at very low pH both logD Cs 18C6 and logD Cs DB18C6 start to increase. Thus, there appears a minimum at the curves logD Cs = f(pH), in between pH 0 and pH 7 ( Figure 1). Worthy of mention is the fact that the extraction efficiency of solutions of CE in conventional solvents is often unsatisfactory even in strongly acidic media. According to Yakshin et al. [22], in the case of Cs + (1 × 10 -3 M) extraction from 2 M HNO 3 aqueous solution into 0.1 M 18C6 in CHCl 3 the value of D Cs is extremely low (3 × 10 -3 ; logD Cs = -2.52). This value is lower than even the minimal values of D Cs obtained in the case of our RTILs-based systems. As a rule, the extraction efficiency of the systems based on conventional solvents may be raised to the level of CE/RTIL-systems only by addition of a bulk hydrophobic counter-ion (such as picrate) to the aqueous phase. So, the logD Cs  Table 4), while the hazardous properties of the former are incomparably higher.
The crown ether influence on cesium extraction by RTILs is evidently explained by a formation of rather stable complexes. Previously, we found [21] [21]. In an aqueous phase only [Cs(18C6)] + species are registered [24].  (1) can be neglected. The equation (1) is therefore transformed into (2): or: Thus, the plot of (logD Cs L -logD Cs -logD L ) versus logK 1 RTIL was expected to be linear with a slope of 1. In our case the initially stated requirement logK 1 RTIL ≥ 2; logK 1 w ≥ 2 does not take place for some RTILs and for water. Indeed, logK 1 w = 0.98 [24] (1), (2), (4) and (5) become more complicated due to formation of CsL 2 complexes and the lack of such complexes in water. However, all six RTILs fit equation (5) with a slope 0.89 and a correlation coefficient R 2 = 0.98, Figure 3. This can be explained by a domination of logK 1 RTIL values over logK 2 RTIL ones, Table 4.
Another reason of low contribution of K 2 RTIL to the extraction is in the absence of enough sufficient excess of crown re. cesium concentration. Anyway, the linear relationship of (logD Cs 18C6 -logD Cs -logD 18C6 Cs ) and logK 1 RTIL indicates clearly an importance of complex stability contribution for the extraction of cesium into hydrophobic RTILs. Besides logK 1 RTIL , also a contribution of logD L should be examined. Thus, along with cesium extraction we have studied the water/RTIL distribution of crown ethers, Tables 2 and 3. 18C6 demonstrates similar affinity towards as the RTILs, so the conventional molecular solvents: 1,2-dichloroethane, chloroform and dichloromethane [5]. The detaining of DB18C6 in extracts is sufficiently higher for all RTILs then that of 18C6, Tables 2 and 3. At the same time logD DB18C6 Cs values are one to two log unites lower then those found for benzene, 1,2-dichloroethane, chloroform and nitrobenzene [5].   It should be noted that the variation of logD Cs 18C6 and logD Cs DB18C6 with pH shows the same trend as the crown ether distribution between aqueous phase and RTIL, logD 18C6 Cs and logD DB18C6 Cs , Figure 1.

(b)
Analysis of the data presented in Table 4  A comparison of 18C6 and DB18C6 indicates a much higher extraction efficiency of the latter. The stability constants of cesium complexes with DB18C6 in the studied RTILs have not been measured yet, but for other than RTIL solvents they reveal 0.2 to 0.5 log units lower values than those for 18C6 [25]. However, bearing in mind that this is also valid for an aqueous solution, the difference in (logK 1 RTIL -logK 1 w ) in Equation (5) becomes negligible. Moreover, this unfavorable for extraction loss is excessively compensated by the fact that the values of logD DB18C6 Cs are 2 to 3 log units higher than logD 18C6 Cs values.

Synthesis and characterization of the RTILs
Trioctylmethylammonium salicylate. Aliquat ® 336 (~0.2 mol) was mixed with 30% excess of sodium salicylate in chloroform (200 mL). The mixture was shaken for 4 h and then rinsed 20 times with a large amount of distilled water, then the solvent was evaporated and the liquid residue was heated up to 100 °C under reduced pressure for 5 h. After cooling to room temperature a white solid matter was obtained with a density 0.943 g·cm −3 ; T m = (32. 8  Tetrahexylammonium dihexylsulfosuccinate was synthesized according to [26] as a transparent viscous liquid (yield: 85%), analyzed by NMR, and then used without further purification. 1  1-Butyl-3-methylimidazolium hexafluorophosphate. A 1-L, one-necked, round-bottomed flask was equipped with a magnetic stirrer and charged with 1-butyl-3-methylimidazolium chloride (367.2 g, 2.10 mol, 1 equiv.) and potassium hexafluorophosphate (387.3 g, 2.10 mol, 1 equiv.) in distilled water (700 mL). The reaction mixture was stirred at room temperature for 2 h, affording a two-phase system. The organic phase was separated and washed with water (5 × 40 mL) until the aqueous fraction observed to be free of chloride (AgNO 3 ). Then dichloromethane (400 mL) was added. The dichloromethane solution was mixed with activated charcoal, stirred for 2 h, filtered and dried over anhydrous magnesium sulfate. After 1 h, the suspension was filtered and the volatile material was removed by rotary evaporation. The resulting colourless or light-yellow viscous liquid was dried under reduced pressure (0.5 mm Hg) at 70 °C for 12 h. Yield 462 g (77.3%). 1

1-Butyl-3-methylimidazolium bis[trifluoromethyl)sulphonyl]imide.
A 500-mL, one-necked, roundbottomed flask was equipped with a magnetic stirrer and charged with lithium bis(trifluoromethylsulfonyl)imide (143.693 g, 0.5 mol, 1 equiv.) and 1-butyl-3-methylimidazolium chloride (87.25 g, 0.5 mol, 1 equiv.) dissolved in distilled water (80 mL). The reaction mixture was stirred at room temperature for 2 h affording a two-phase system. Then dichloromethane (200 mL) was added. The lower organic layer was separated and washed with distilled water (5 × 20 mL) until the aqueous fraction observed to be free of chloride (AgNO 3 ). The dichloromethane solution was mixed with activated charcoal, stirred for 2 h, filtered and dried over anhydrous magnesium sulfate. After 1 h, the suspension was filtered and the volatile material was removed by rotary evaporation. The resulting colourless viscous liquid was dried under reduced pressure (0.5 mm Hg) at 70 °C for 12 h. Yield 162.6 g (77.5%). 1

Extraction
For a sample preparation the weighted amount of solid CsNO 3 was dissolved in a distilled water to obtain the initial solution of salt (1.5·10 -2 mol·dm -3 ), which was further used for the preparation of the diluted samples. For adjusting pH values of aqueous phase, 0.01-3 M HNO 3 (analytical reagent grade) was used.
Cesium extraction process in the absence of crown ethers was performed by adding of watersaturated RTIL samples ). The biphasic systems were shaken for 2 h (adequate time for equilibrium state establishment for all RTILs studied at room temperature, 22 °C). After centrifugation for 15 min aqueous and organic phases were separated. An aqueous phase was analyzed for equilibrium pH (pH-meter pH-410; combined glass microelectrode ESLK-13.7, Aquilon, Russia) and cesium content (flame photometry, FPA-2; Zagorsk Optical and Mechanical Plant, Russia). The recovery (R(Cs), %) and the distribution constant (D(Cs)) of cesium were found according to eqns. (6) and (7): where C w 0 (Cs) and C w (Cs) are the initial and equilibrium concentrations of Cs + in an aqueous phase, respectively, V w и V o correspond to the total volumes of aqueous and organic phase, respectively (mL  [DHSS] was lower (C(DB18С6) 0 = 5 × 10 -2 mol·dm -3 ). This concentration was obtained after heating of these RTILs with the amounts of DB18C6 at 80-90 °C. Thus, these solutions are likely to be oversaturated, although no precipitation after cooling to room temperature was observed.
To ). Then the resulting solutions were analyzed spectrophotometrically (U-2900 UV-Vis Recording Spectrophotometer, Hitachi, Japan; 1 cm quartz cells) for DB18C6 content using crown ether absorption band λ max = 274 nm (ε max = 5,000 ÷ 5,500) and the reference solutions of the corresponding RTILs in ethanol with the same dilution as for DB18C6/RTILs. Detaining (R(CE), %) for DB18C6 has been calculated using equation (8) where C o 0 (CE) and C o (CE) denote initial and equilibrium concentrations of crown ether in organic phase (extract), respectively; while distribution constant (D(CE)) -using equation (7). Distribution of 18C6 was controlled by the method described elsewhere [27]. Bromothymol Blue (H-BTB) and KNO 3 were dissolved in an aliquot taken from an aqueous phase (to adjust 1.2 × 10 -2 mol·dm -3 concentration of both of them). Then the extraction of the complex [K(18С6) + ·BTB -] into CHCl 3 was performed with 2 h contact time and phases volume ratio 1:1; pH of an aqueous phase was kept constant (pH = 6.9) by phosphate buffer for all solutions other then neutral. Then the system was centrifuged for 15 min, an organic phase was separated and analyzed spectrophotometrically (λ max = 400 ÷ 450 nm, ε max = 1,500 ÷ 2,000; the chloroform extract obtained in the blank experiment was used as the reference solution). Detaining (R(CE), %) and distribution ratio (D(CE)) were calculated by the eqns., completely analogous to (1) and (2), taking into consideration that C o (CE) = C o 0 (CE) -10·C w (CE), where C w (CE) denotes an equilibrium concentration of crown ether in aqueous phase.

Conclusions
The data presented reveals that several ionic liquids are suitable for extraction of cesium in the presence of crown ethers. The liquids which show a higher retention of a crown ether and a higher stability of the complexes demonstrate better capabilities to extract Cs + . The former factor dominates evidently over the latter one. Among the crowns the ligand of a choice is the one that demonstrates higher hydrophobicity, and, therefore, a higher detaining in extract, e.g., DB18C6.