Zwitterions as novel phase forming components of aqueous biphasic systems

Synthesis of zwitterions

The selected zwitterions have been synthesized from various bases including triethylamine, N-Methyl imidazole, N-Vinyl imidazole, pyridine, N-Methyl pyrrolidine, N-Ethyl piperdine and 1,4-butane sultone by following earlier reported synthesis method [17], [37]. Typical synthesis method for the triethylamine based ZI is explained here. Triethylamine along with acetonitrile as solvent was taken in a round bottom flask fitted with reflux condenser and nitrogen gas supply to maintain inert conditions. The equimolar quantity of 1,4-butane sultone was added to the round bottom flask and the reaction was carried out at 80°C under constant stirring conditions for 12 h. After 12 h, the acetonitrile was removed under reduced pressure. The resultant zwitterion product was washed with diethyl ether and toluene to remove starting materials present if any. Then the product was dried under vacuum for 12 h to obtain the white colored solid compound and was characterized by ¹H and ¹³C NMR spectral techniques. Bruker Avance Spectrometer (400 MHz) was employed for NMR spectral analysis where D₂O and TMS were used as solvent and internal standard, respectively. The general synthesis scheme followed for ZI synthesis and the structures of resultant ZIs are given in Fig. 1. The water content of synthesized ZIs has been determined through Karl Fischer titrator (Metrohm 970 KF Titrino plus) and were found to be less than 2000 ppm. The list of zwitterions synthesized along with their abbreviations, molecular weights and purities are presented in Table 1.

Fig. 1:

(a) Synthesis scheme for zwitterions and (b) structures of synthesized ZIs.

Spectral characterization

Thermogravimetric analysis

The thermal stability of synthesized ZIs has been determined through thermogravimetric analysis. TGA instrument (TA instruments Hi-Res. TGA Q500) was used for the analysis. It measures the difference in sample weight taken on aluminum pan maintained under inert nitrogen atmosphere with the weighing precision of ±0.01%. The heating is performed from room temperature to 500°C at the heating rate of 10°C min⁻¹. The heating rate is controlled using thermocouples kept adjacent to the sample. Prior to the measurements, instrument was calibrated with Nickel metal.

Cloud point titration

Cloud point titration method was used for experimental determination of binodal curves and phase diagrams for polymer based [38] as well as IL based ABS [39]. All binodal curves were determined at atmospheric pressure and fixed temperature of 298.15 K that was maintained using temperature controller bath (F25-ME, Julabo, Germany) with an uncertainty of ±0.1 K. For titration experiments, aqueous solutions of ZIs (50–60 wt%) and aqueous solutions of potassium salts (40 wt%) were prepared on gravimetric basis. Analytical balance (Sartorius CPA225D) with the precision of ±0.01 mg was used for measuring weights during the course of experiments. The process of titration involved the drop wise addition of aqueous salt solution to the aqueous ZI solution maintained at constant temperature and constant stirring, till the appearance of turbidity. The turbidity indicates the formation of biphasic system and the amount of salt solution that was added to obtain the turbidity was recorded. In next step, ultrapure Millipore quality water was added drop-wise till the solution turned clear suggesting the formation of monophasic region. The weight of water added was also recorded. The steps were repeated sufficient number of times to construct the complete binodal curve that extends over range of concentrations. Based on the recorded weights, exact compositions of all three components i.e. ZI, salt and water, involved in phase formation were calculated. All experiments were performed at least for three times in order to obtain reliable and accurate results.

The experimental data derived from cloud point titration was fitted to Merchuks’ equation [5] through least squares regression. Merchuks’ equation that was initially proposed for polymer based ABS, later successfully applied for IL based ABS, was used in present work also. The Merchuks’ equation is given as below:

X and Y in the equation represent concentration in weight fractions for salt and ZI, respectively and constants A, B and C are regression coefficients.

Tie-line determination

The selected ternary systems containing ZI, salt and water were observed to form biphasic systems at specific concentrations resulting in visibly separated top and bottom phases. Tie line data helps to know the concentration of each individual component in top and bottom phases separately. In order to determine tie line data, ternary mixtures with composition in biphasic region is selected and prepared in weight basis. These mixtures were stirred vigorously for half an hour and then allowed to equilibrate for 12 h at 298.15 K. Top and bottom phases’ separation could be clearly observed after sufficient time of equilibration, which were then separated and weighed accurately. By using the measured weights and other known parameters including fitting parameters of Merchuks’ equation, the tie line compositions were determined using Lever arm rule [5]. The system of following four equations (eqs. 2–5) was employed to get the tie line compositions.

In present system, X and Y represent the concentration of salt and ZI phases, respectively while subscripts T, B and M denote top phase, bottom phase and mixture, respectively. The α parameter is the ratio between weight of top phase to weight of total mixture. On obtaining compositions of salt and ZI in both phases, tie line length (TLL) can also be determined through the eq. 6.

TLL is numerical quantification of the difference in concentration of ZI and salt in two co existing phases.

Partitioning of caffeine

The preferential partitioning of caffeine to the ZI rich phase has been explored through extraction experiments executed at 298.15 K and atmospheric pressure. Caffeine aqueous solution of 5×10⁻³ m was used for all partitioning experiments. Specific compositions of ternary systems belonging to the biphasic region are selected and solutions were prepared on weight basis. Prepared solutions were stirred vigorously and left for 12 h at 298.15 K to attain equilibrium. The temperature controller bath (F25-ME, Julabo, Germany) was used for maintaining the constant temperature. UV-Visible spectroscopy was employed to quantify the amount of caffeine present in each phase separated after equilibrium. Samples from each phase were taken and diluted sufficiently and their absorbance was measured using UV-Vis spectrophotometer (Shimadzu UV-1300) at 274 nm. Concentration of caffeine was determined using the previously established calibration plots. Three sets of individual samples were prepared to obtain reliable and accurate data. To account for the interventions caused by ZI or salt, blank samples of same ternary mixture compositions were prepared using water instead of aqueous caffeine solutions. For calculation of partition coefficients for caffeine, the following equation was used

here the concentration of caffeine is given by [Caf]. Further, the extraction efficiencies in percentage can also be calculated using the given equation

the EE_caf is percentage ratio of quantity of caffeine in ZI rich phase, WCafZI to the quantity present in total mixture, WCafZI+WCafSalt, where WCafSalt is the quantity of caffeine present in salt rich phase.

TGA analysis of synthesized zwitterions

Thermogravimetric analysis evaluates the thermal stability of ZIs through their decomposition temperatures. The experimentally obtained onset decomposition temperatures for all ZIs are given in Table 1 and the values indicated that all studied ZIs are stable up to 250°C. The graphical representation of the thermogravimetric analysis is given in Fig. 2. The overall thermal stability order for ZIs is as follows: [ViIm(C₄SO₃)]>[MeIm (C₄SO₃)]>[MePyr(C₄SO₃)]>[N₂₂₂(C₄SO₃)]>[EtPip(C₄SO₃)]>[Py(C₄SO₃)]. The order observed can be attributed to the structure and aliphatic moieties present in ZIs. From the order, ZIs with imidazole group were observed to show highest stability in comparison to those containing other groups like pyrrolidine, pyridine, piperidine etc. Similar observation was found in case of ILs with imidazolium cations. In addition, thermal stability of ZIs was found to be higher than structurally similar acidic ILs containing HSO₄⁻ and CF₃SO₃⁻ anions [40].

Fig. 2:

Thermogravimetric analysis of synthesized zwitterions where (a) [Py(C₄SO₃)], (b) [EtPip(C₄SO₃)], (c) [N₂₂₂(C₄SO₃)], (d) [MePyr(C₄SO₃)], (e) [MeIm(C₄SO₃)] and (f) [ViIm(C₄SO₃)].

Phase behavior of zwitterions

The ability of each ZI to form biphasic systems when mixed with potassium based inorganic salt solutions was evaluated. It was observed that all synthesized ZIs were able to form ABS in combination with selected three potassium based inorganic salts. The evaluated phase behavior of all synthesized ZIs in combination with K₃PO₄, K₂HPO₄ and K₂CO₃ are presented in Fig. 3a–c respectively. Binodal curves are graphical representation of the concentrations of ZI and salts at which the ABS are formed and they are presented in molality units for better analysis of results and to avoid the interference of difference in molecular weights of ZIs. In phase diagrams, the position of binodal curve gives idea about the hydrophobic/hydrophilic nature of the system. The region in phase diagram that lies below the binodal curve is monophasic in nature while the concentration region above binodal curve is biphasic in nature. The nearer the binodal curve is to the origin, the higher is the ability to form biphasic systems suggesting that particular system is relatively hydrophobic.

Fig. 3:

Phase diagrams for systems composed of ZIs, water and (a) K₃PO₄ (b) K₂HPO₄ (c) K₂CO₃ at 298.15 K and atmospheric pressure.

From the Fig. 3a, it can be observed that highest phase formation ability among ZIs was shown by [EtPip(C₄SO₃)] followed by triethylamine based ZI i.e. [N₂₂₂(C₄SO₃)]. Figure 3a–c assist in evaluating the influence of ZI structure on their phase formation ability. From Fig. 3a, the overall order for studied ZIs can be inferred as [EtPip(C₄SO₃)]>[N₂₂₂(C₄SO₃)]>[MePyr(C₄SO₃)]≈[ViIm(C₄SO₃)]>[MeIm(C₄SO₃)]≥[Py(C₄SO₃)].

As per the observed order, the ZI with highest phase formation ability is [EtPip(C₄SO₃)] and it is the most hydrophobic among studied ZIs. The observed order reflects the extent of interactions between ZIs and the solvent i.e. water or the ability of these ZIs to be solvated by water. The solubility of ZIs in water depends on various interactions including hydrogen bonding, steric effect, entropic contributions etc. ZIs containing larger nitrogen containing groups like piperidinium were able to cause formation more easily in comparison to the smaller ring containing ZIs like imidazolium and pyrrolidinium groups. However, the discrepancy observed in case of pyridine containing ZI showing the least ability to induce phases among all studied ZIs could be attributed to the steric effects caused by the alkyl substituents present in all other ZIs. [Py(C₄SO₃)] could interact with water through the aromatic hydrogen bonding leading to its higher solubility in water and consequential difficulty in phase formation. Further aromatic moieties are known to present higher solvation capabilities for water in comparison to aliphatic ones thus making aromatic group containing ZIs hydrophilic. These hydrophilic ZIs require higher quantities of water to show phase separation. Similar observations were reported in case of ABS formed by ILs belonging to various cation classes in combination with potassium salts with primary focus on evaluating ability of IL cations towards phase formation [41].

The phase diagrams for all ZIs in presence of other two potassium salts i.e. K₂HPO₄ and K₂CO₃ were determined and graphically presented in Fig. 3b and c, respectively. The phase formation trends of ZIs in presence of K₂HPO₄ and K₂CO₃ are almost similar to that observed in case of K₃PO₄. This suggests that irrespective of identity of salt, the phase formation ability of studied ZIs remain almost same.

In addition, binodal curves obtained in present work were also compared with reported ammonium based ZIs containing propane sulfonate and butane sulfonate groups [26]. K₃PO₄ was the common salt for all ZIs and the comparison is graphically presented in Fig. 4. The dependence of phase formation ability on the alkyl chain length and consequent hydrophobicity was further confirmed form this comparison. The ZI constituting tripentyl groups i.e. [N₅₅₅(C₃SO₃)] showed highest phase formation ability among all ZIs, due to longer alkyl chain length and hydrophobic nature. The position of binodal curves of ammonium ZIs with intermediate chain length were comparable to those obtained for ZIs in present study. Again, the ZI of shortest alkyl chain i.e. [N₁₁₁(C₃SO₃)] was most hydrophilic among all ZIs including earlier reported as well as ZIs of current study.

Fig. 4:

Phase diagrams for synthesized ZIs in comparison with reported ammonium based ZIs.

Further, the effect of potassium salts on the phase behavior has also been systematically studied and phase diagrams of all six ZIs in combination with K₃PO₄, K₂HPO₄ and K₂CO₃ are given in Fig. 5. The observed trends with respect to salts could be mainly ascribed to salt anions since the potassium cation remains same in all cases. Thus the experimentally determined trends for phase formation by salts can be given as PO₄³⁻>HPO₄²⁻>CO₃²⁻. The trend observed was same for all ZIs. The observed salting-out strength of salts can be explained based on Hofmeister series [42]. During the ABS formation, there exists mutual competition between ions of salts and ZIs for the amount of water available in order to get hydrated. As per the Hofmeister series, the ions of higher valence gets hydrated to larger extent thus reducing the amount of water available for the ZIs to get hydrated. Potassium salts in present case acts as salting-out agents and salt with higher valency showed higher salting-out as well phase separation ability. Further, salting-out ability of ions could also be explained based on their Gibbs free energy of hydration, ΔG_hyd values for the anion studied: PO₄³⁻=−2835 kJ·mol⁻¹, HPO₄²⁻=−1125 kJ·mol⁻¹, CO₃²⁻=−1300 kJ·mol⁻¹ [43]. The larger negative ΔG_hyd values suggest the more structuring of water molecules surrounding them and consequent kosmotropic nature of salts. Though this analysis of observed salt behavior was proposed for ABS composed of ILs and salts [5], it was found to be equally applicable for ZI based ABS also [26].

Fig. 5:

Phase diagrams for systems composed of ZIs, water and (a) [N₂₂₂(C₄SO₃)] (b) [MeIm(C₄SO₃)] (c) [ViIm(C₄SO₃)] (d) [Py(C₄SO₃)] (e) [MePyr(C₄SO₃)] (f) [EtPip(C₄SO₃)] at 298.15 K and atmospheric pressure.

On formation of two phases in case of all ZI based ABS, the upper phase was rich in ZI and the bottom phase was salt rich and this behavior matched with most cases of IL based ABS where top phase was IL rich and bottom salt rich [5].

The binodal curve data obtained through experiment is fitted to the Merchuks’ equation as given in eq. 1 and obtained fitting parameters A, B and C for all combinations of ZIs and salts are given in Table 2 along with corresponding standard deviations and correlation coefficients. The correlation used helps in predicting the binodal data in the regions where experimental data is unavailable and appreciable correlation coefficients have been obtained in present study suggests the applicability of Merchuks’ equation for ZI based ABS too. Additionally, tie line data has been determined using eqs. 2–5. Experimental details of mixture compositions selected for the determination of tie lines and calculated concentrations of ZI and salt in each phase along with resultant tie line lengths have been presented in Table 3. A representative tie line data of system containing [Py(C₄SO₃)] and K₃PO₄ is given in Fig. 6. Tie lines for two different mixture compositions were almost parallel to each other as can be seen from Fig. 6. For representative purpose, tie line data determined for other two salts with [Py(C₄SO₃)] are also given in Table 3. Tie line length is a numerical expression for the difference in concentration of ZIs and salts in top and bottom phases. The higher TLL values as seen in case of [Py(C₄SO₃)] and K₃PO₄ indicates the higher concentration of ZI and salt in top and bottom phases, respectively.

Fig. 6:

Phase diagram representing tie-line compositions for system containing [Py(C₄SO₃)], K₃PO₄ and water at 298.15 K and atmospheric pressure.

Effect of temperature on phase behavior of zwitterions

In order to analyze the effect of temperature on the phase behavior of ZIs, phase diagrams have been determined at three different temperatures i.e. 298.15, 308.15, 318.15 K and common salt K₃PO₄. The obtained representative binodal curves for [Py(C₄SO₃)], [MePyr(C₄SO₃)] and [EtPip(C₄SO₃)] are given in Fig. 7a–c. The position of binodal curve suggested that with increase in temperature, the biphasic region narrowed down and monophasic region increased. This indicates that phase formation is more favorable at lower temperatures and thus exhibits UCST type phase behavior.

Fig. 7:

Phase diagrams for systems containing K₃PO₄, water and (a) [Py(C₄SO₃)] (b) [MePyr(C₄SO₃)] and (c) [EtPip(C₄SO₃)] at temperatures of 298.15, 308.15 and 318.15 K and atmospheric pressure.

Recent reports on ABS formed by ammonium based ZIs with potassium salts and PEG have explored their temperature reversible behavior. Hydrophilic ZIs based on trimethylamine, [N₁₁₁(C₃SO₃)] were reported to show UCST type phase behavior while hydrophobic ZIs like tripentylamine based ZIs, [N₅₅₅(C₃SO₃)] showed LCST type behavior in presence of salts [26]. On the other hand, a reversal in trends was observed while they were studied in combination with polymers like PEG [44]. The transition from LCST to UCST or vice versa depends on the structure of ZI and the other phase forming component of the system.

In present case, the study on thermal effects on phase behavior suggested that all ZIs exhibited UCST behavior similar to the case of polymer-polymer ABS [45] and the phase formation was favorable at lower temperatures. Nevertheless, the extent of separation between constructed binodal curves for each system gives some idea about the hydrophilic/hydrophobic character of the ZI. In comparison to reported ammonium based ZIs, the ZIs of present work were hydrophilic enough to show same kind of thermal dependence i.e. UCST type behavior. For most hydrophobic ZI among all studied ZIs i.e. for [Py(C₄SO₃)], the appreciable separation between binodal curves at different temperatures was observed. Whereas the distance between binodal curves significantly decreased for ZI of intermediate hydrophilicity i.e. [MePyr(C₄SO₃)], while binodal curves started overlapping in case of [Etpip(C₄SO₃)]. The overlapping observed indicates that there is no significant dependence of phase behavior on temperature.

Partitioning of caffeine

The synthesized ZIs have been studied for their efficiency with respect to partitioning of an alkaloid i.e. caffeine. For partition studies we have considered the combination of synthesized ZIs with salt of highest salting-out ability i.e. K₃PO_4. The ternary mixtures of 20 wt% of K₃PO₄, 24 wt% of ZI and 56 wt% of caffeine aqueous solution were prepared for the partitioning study. The weight percentages of each component were selected based on the phase diagrams. The partition coefficients obtained for each system and the extraction efficiencies are graphically presented in Fig. 8a and b.

Fig. 8:

Bar diagram representing (a) partition coefficients, K_caf and (b) extraction efficiencies in ZI based ABS at 298.15 K and atmospheric pressure.

In the present study, caffeine was observed to preferably partition to the ZI rich top phase in all cases. However, the partition coefficient values obtained were different depending on the structure of ZIs studied. The extraction of caffeine into the ZI rich phase seems to be dependent on various factors like π-π interactions between the aromatic groups of caffeine and that of ZIs, dispersive interactions among alkyl groups of caffeine and those present in ZIs, hydrogen bonding etc. [46]. The partition coefficient values obtained in case of pyridine and imidazole based ZIs suggested the favorable π-π interactions between the aromatic ring of pyridine and imidazole with the purine ring present in caffeine. In case of aliphatic cyclic groups like pyrrolidine and piperidine containing ZIs, dispersive interactions seem to be the dominating factor. The lowest partitioning among studied ZIs was observed for the triethylamine based ZI that could be attributed to the absence of favorable interactions with caffeine like π-π interactions.

The partition coefficient values of present study were higher in comparison to those obtained for ABS formed by phosphonium based ILs with inorganic salts [47] and imidazolium based ILs with amino acids [33]. However, these were comparable to values observed for ABS formed by ILs belonging to imidazolium cation various anions in combination with K₃PO₄ suggesting that ABS based on ZIs of present study could be potential extraction systems. The extractive potential could be manipulated based on the structures and type of phase forming components and thus can find applications in separation and extraction processes as replacements for traditional systems.