Synthesis of Purine-Based Ionic Liquids and Their Applications

Bio-based ionic liquids (ILs) are being increasingly sought after, as they are more sustainable and eco-friendly. Purines are the most widely distributed, naturally occurring N-heterocycles, but their low water-solubility limits their application. In this work, four purines (theobromine, theophylline, xanthine, and uric acid) were combined with the cation tetrabutylammonium to synthesize bio-based ILs. The physico–chemical properties of the purine-based ILs were characterized, including their melting and decomposition temperatures and water-solubility. The ecotoxicity against the microalgae Raphidocelis subcapitata was also determined. The ILs show good thermal stability (>457 K) and an aqueous solubility enhancement ranging from 53- to 870-fold, in comparison to their respective purine percursors, unlocking new prospects for their application where aqueous solutions are demanded. The ecotoxicity of these ILs seems to be dominated by the cation, and it is similar to chloride-based IL, emphasizing that the use of natural anions does not necessarily translate to more benign ILs. The application of the novel ILs in the formation of aqueous biphasic systems (ABS), and as solubility enhancers, was also evaluated. The ILs were able to form ABS with sodium sulfate and tripotassium citrate salts. The development of thermoresponsive ABS, using sodium sulfate as a salting-out agent, was accomplished, with the ILs having different thermosensitivities. In addition, the purine-based ILs acted as solubility enhancers of ferulic acid in aqueous solution.


Introduction
Ionic liquids (ILs) are alternative solvents, composed of a large organic cation (e.g., tetraalkylammonium, tetraalkylphosphonium, imidazolium, cholinium, pyridinium) and an organic or inorganic anion [1]. Cation-anion asymmetry and charge dispersion, in at least one of the ions, reduces the electrostatic interactions and crystallization ability, creating salts with low melting points. In fact, the first definition of ILs was proposed by Walden et al. [2] as salts with melting temperatures below an arbitrary threshold of 373 K, in order to differentiate these from typical salts. More recently, the definition of ILs was refined by Mariani et al. [3], who proposed that ILs are water-free organic salts with a melting temperature lower than decomposition temperature.

Synthesis and Characterization of Purine-Based ILs
Four purines, namely theobromine, theophylline, xanthine, and uric acid, were converted into salts through a well-established neutralization protocol [35]. [N4444]OH (1 equivalent, 40 wt% in aqueous solution) was slowly added to an aqueous solution of the purine, with a molar excess of 1.1 equivalents. The reaction mixture was stirred for 2 h at room temperature. As a result, the respective salt was obtained, and water was formed as a byproduct. The water was removed under reduced pressure. The resultant product was dissolved in acetonitrile and filtered to remove the excess purine. Lastly, acetonitrile was removed under reduced pressure, and the obtained compound was dried under high vacuum for at least 48 h. All the salts were obtained as white solids, except for [N4444] [Xan], which was obtained as a pale beige solid. The same protocol was used, in the attempt to produce fully bio-based ILs by using cholinium cation, instead of tetrabutylammonium and theophylline as the anion; however, this was not successful. The structure of the synthesized salts was confirmed by 1 H and 13 C nuclear magnetic resonance (NMR) spectroscopy, as reported in the Supplementary Materials. The NMR spectra were recorded with 300. 13 MHz ( 1 H) and 75. 47 MHz ( 13 C) on a Bruker Avance III NMR spectrometer (Spectrum S1 to S8). Tetramethylsilane was added as an internal reference, and D2O was used as a solvent.   [Ur]. Tetrabutylammonium hydroxide ([N4444]OH, in aqueous solution at 40 wt%), tetrabutylammonium chloride (97 wt% pure), theobromine (99 wt% pure), and xanthine (99 wt% pure) were acquired from Sigma Aldrich (Saint Louis, MO, USA) Theophylline (99 wt% pure) and tripotassium citrate monohydrate (K3C6H5O7, 99 wt% pure) were acquired from Acros Organic (Waltham, MA, USA); uric acid (99 wt% pure) from Alfa Aesar (Haverhill, MA, USA); and sodium sulfate anhydrous (Na2SO4, 99 wt% pure) from José Manuel Gomes dos Santos (Odivelas, Portugal). The 4-hydroxy-3methoxycinnamic acid (hereinafter referred to as ferulic acid, 99 wt% pure) was purchased from TCI. The used water was double distilled, passed through a reverse osmosis system and further treated with a Milli-Q plus 185 water purification apparatus.

Synthesis and Characterization of Purine-Based ILs
Four purines, namely theobromine, theophylline, xanthine, and uric acid, were converted into salts through a well-established neutralization protocol [35]. [N4444]OH (1 equivalent, 40 wt% in aqueous solution) was slowly added to an aqueous solution of the purine, with a molar excess of 1.1 equivalents. The reaction mixture was stirred for 2 h a room temperature. As a result, the respective salt was obtained, and water was formed as a byproduct. The water was removed under reduced pressure. The resultant product was dissolved in acetonitrile and filtered to remove the excess purine. Lastly, acetonitrile was removed under reduced pressure, and the obtained compound was dried under high vacuum for at least 48 h. All the salts were obtained as white solids, except for [N4444] [Xan] which was obtained as a pale beige solid. The same protocol was used, in the attempt to produce fully bio-based ILs by using cholinium cation, instead of tetrabutylammonium and theophylline as the anion; however, this was not successful. The structure of the synthesized salts was confirmed by 1 H and 13 C nuclear magnetic resonance (NMR) spectroscopy, as reported in the Supplementary Materials. The NMR spectra were recorded with 300.13 MHz ( 1 H) and 75.47 MHz ( 13 C) on a Bruker Avance III NMR spectrometer (Spectrum S1 to S8). Tetramethylsilane was added as an internal reference and D2O was used as a solvent.   , tetrabutylammonium chloride (97 wt% pure), theobromine (99 wt% pure), and xanthine (99 wt% pure) were acquired from Sigma Aldrich (Saint Louis, MO, USA) Theophylline (99 wt% pure) and tripotassium citrate monohydrate (K3C6H5O7, 99 wt% pure) were acquired from Acros Organic (Waltham, MA, USA); uric acid (99 wt% pure from Alfa Aesar (Haverhill, MA, USA); and sodium sulfate anhydrous (Na2SO4, 99 wt% pure) from José Manuel Gomes dos Santos (Odivelas, Portugal). The 4-hydroxy-3 methoxycinnamic acid (hereinafter referred to as ferulic acid, 99 wt% pure) was purchased from TCI. The used water was double distilled, passed through a reverse osmosis system and further treated with a Milli-Q plus 185 water purification apparatus.

Synthesis and Characterization of Purine-Based ILs
Four purines, namely theobromine, theophylline, xanthine, and uric acid, were converted into salts through a well-established neutralization protocol [35]. [N4444]OH (1 equivalent, 40 wt% in aqueous solution) was slowly added to an aqueous solution of the purine, with a molar excess of 1.1 equivalents. The reaction mixture was stirred for 2 h a room temperature. As a result, the respective salt was obtained, and water was formed as a byproduct. The water was removed under reduced pressure. The resultant product was dissolved in acetonitrile and filtered to remove the excess purine. Lastly, acetonitrile was removed under reduced pressure, and the obtained compound was dried under high vacuum for at least 48 h. All the salts were obtained as white solids, except for [N4444] [Xan] which was obtained as a pale beige solid. The same protocol was used, in the attempt to produce fully bio-based ILs by using cholinium cation, instead of tetrabutylammonium and theophylline as the anion; however, this was not successful. The structure of the synthesized salts was confirmed by 1 H and 13 C nuclear magnetic resonance (NMR) spectroscopy, as reported in the Supplementary Materials. The NMR spectra were recorded with 300.13 MHz ( 1 H) and 75.47 MHz ( 13 C) on a Bruker Avance III NMR spectrometer (Spectrum S1 to S8). Tetramethylsilane was added as an internal reference and D2O was used as a solvent.  , tetrabutylammonium chloride (97 wt% pure), theobromine (99 wt% pure), an xanthine (99 wt% pure) were acquired from Sigma Aldrich (Saint Louis, MO, USA Theophylline (99 wt% pure) and tripotassium citrate monohydrate (K3C6H5O7, 99 wt pure) were acquired from Acros Organic (Waltham, MA, USA); uric acid (99 wt% pur from Alfa Aesar (Haverhill, MA, USA); and sodium sulfate anhydrous (Na2SO4, 99 wt pure) from José Manuel Gomes dos Santos (Odivelas, Portugal). The 4-hydroxymethoxycinnamic acid (hereinafter referred to as ferulic acid, 99 wt% pure) was purchase from TCI. The used water was double distilled, passed through a reverse osmosis system and further treated with a Milli-Q plus 185 water purification apparatus.

Synthesis and Characterization of Purine-Based ILs
Four purines, namely theobromine, theophylline, xanthine, and uric acid, we converted into salts through a well-established neutralization protocol [35].
[N4444]OH equivalent, 40 wt% in aqueous solution) was slowly added to an aqueous solution of th purine, with a molar excess of 1.1 equivalents. The reaction mixture was stirred for 2 h room temperature. As a result, the respective salt was obtained, and water was formed a byproduct. The water was removed under reduced pressure. The resultant product w dissolved in acetonitrile and filtered to remove the excess purine. Lastly, acetonitrile w removed under reduced pressure, and the obtained compound was dried under hig vacuum for at least 48 h. All the salts were obtained as white solids, except for [N4444][Xan which was obtained as a pale beige solid. The same protocol was used, in the attempt produce fully bio-based ILs by using cholinium cation, instead of tetrabutylammoniu and theophylline as the anion; however, this was not successful. The structure of th synthesized salts was confirmed by 1 H and 13 C nuclear magnetic resonance (NMR spectroscopy, as reported in the Supplementary Materials. The NMR spectra we recorded with 300.13 MHz ( 1 H) and 75.47 MHz ( 13 C) on a Bruker Avance III NM spectrometer (Spectrum S1 to S8). Tetramethylsilane was added as an internal referenc and D2O was used as a solvent.  ]OH, in aqueous solution at 40 wt%), tetrabutylammonium chloride (97 wt% pure), theobromine (99 wt% pure), and xanthine (99 wt% pure) were acquired from Sigma Aldrich (Saint Louis, MO, USA). Theophylline (99 wt% pure) and tripotassium citrate monohydrate (K 3 C 6 H 5 O 7 , 99 wt% pure) were acquired from Acros Organic (Waltham, MA, USA); uric acid (99 wt% pure) from Alfa Aesar (Haverhill, MA, USA); and sodium sulfate anhydrous (Na 2 SO 4 , 99 wt% pure) from José Manuel Gomes dos Santos (Odivelas, Portugal). The 4-hydroxy-3-methoxycinnamic acid (hereinafter referred to as ferulic acid, 99 wt% pure) was purchased from TCI. The used water was double distilled, passed through a reverse osmosis system, and further treated with a Milli-Q plus 185 water purification apparatus.

Synthesis and Characterization of Purine-Based ILs
Four purines, namely theobromine, theophylline, xanthine, and uric acid, were converted into salts through a well-established neutralization protocol [35]. [N 4444 ]OH (1 equivalent, 40 wt% in aqueous solution) was slowly added to an aqueous solution of the purine, with a molar excess of 1.1 equivalents. The reaction mixture was stirred for 2 h at room temperature. As a result, the respective salt was obtained, and water was formed as a byproduct. The water was removed under reduced pressure. The resultant product was dissolved in acetonitrile and filtered to remove the excess purine. Lastly, acetonitrile was removed under reduced pressure, and the obtained compound was dried under high vacuum for at least 48 h. All the salts were obtained as white solids, except for [N 4444 ][Xan], which was obtained as a pale beige solid. The same protocol was used, in the attempt to produce fully bio-based ILs by using cholinium cation, instead of tetrabutylammonium and theophylline as the anion; however, this was not successful. The structure of the synthesized salts was confirmed by 1 H and 13 C nuclear magnetic resonance (NMR) spectroscopy, as reported in the Supplementary Materials. The NMR spectra were recorded with 300.13 MHz ( 1 H) and 75.47 MHz ( 13 C) on a Bruker Avance III NMR spectrometer (Spectrum S1 to S8). Tetramethylsilane was added as an internal reference, and D 2 O was used as a solvent.

Melting and Degradation Temperatures
The melting temperatures (T m ) were measured in a differential scanning calorimetry (DSC), using a Hitachi DSC7000X equipment at atmospheric pressure, with hermetically sealed aluminum crucibles and a constant flow of nitrogen. The equipment was previously calibrated using compounds with mass fractions purities higher than 99%. Each sample underwent three cycles of cooling and heating, at 5 K·min −1 and 2 K·min −1 , respectively. The melting point of each compound was taken as the peak temperature. The determined melting points are estimated to have a 1 K uncertainty. When necessary, melting points were also measured in a Bücchi device, model M-565, with a temperature resolution of 0.1 K. The degradation temperature (T d ) of the purine-based salts was measured using a Setsys Evolution 1750 (SETARAM) instrument. Each sample was heated in an aluminia pan, under a nitrogen atmosphere, with a heating rate of 2 K·min −1 . The degradation temperature-associated uncertainty is estimated to be 0.01 K. Degradation temperature was taken as the peak temperature of the derivative.

Solubility Assays
IL solubility studies were performed using the isothermal shake-flask method, as reported in the literature [36] were added in excess of 0.5 mL of water, forming a saturated aqueous solution. The samples were stirred at 1050 rpm at (298 ± 1) K for 72 h using the Eppendorf Thermomixer comfort equipment. After that time, equilibrium was reached, the stirring was turned off, and samples were left undisturbed at 298 K, in order to enable the separation of the excess undissolved solute and the liquid phase. If needed, samples were centrifuged for 2 min at 12,000 rpm. Samples of the liquid phase were collected and diluted in distilled water, followed by the preparation of a calibration curve for each compound of interest. The amount of compound in the liquid phase was quantified using a UV-Vis Synergy HT microplate reader from BioTek (Winooski, VT, USA  [N 4444 ]Cl, with the following concentrations: 0.05, 0.1, 0.25, 0.5, 1.0, and 1.5 mol L −1 . The samples were stirred at 1150 rpm at (303 ± 1) K for 72 h. After that time, equilibrium was reached, the stirring was turned off, and samples were left undisturbed at 303 K. If needed, samples were centrifuged for 2 min at 12,000 rpm. Samples were quantified, as previously described, at 316 nm (see Figure S1). All assays were performed in triplicates. To verify the integrity of the IL, the liquid phase was analyzed by 1 H and 13 C NMR.
The effect of pH on the water-solubility of ferulic acid was conducted by adjusting the initial pH of the aqueous solutions with NaOH, followed by the addition of ferulic acid until saturation was completed. The samples were stirred at 1150 rpm at (303 ± 1) K for 72 h, and pH was remeasured. When necessary, NaOH was re-added to the solutions, aiming to achieve a wide range of pH values. The assays were conducted in duplicates.

Microalgae Ecotoxicity Assays
The microalgae ecotoxicity assays were conducted, as previously described in detail [37], using the microalgae Raphidocelis subcapitata, i.e., following the guidelines of OECD [38], and adapted to the use of 24-well microplates [39]. Briefly, the initial test cell density of 10 4 cells·mL −1 was obtained by microscopic cell counting (Neubauer hemocytometer) of a microalgae inoculum culture, grown in MBL medium [40], under controlled incubation conditions (296 ± 1) K and permanent illumination of 7000 lux. The microalgae were exposed to a range of concentrations of each salt, and all the assays, including the MBL blank control, were conducted in triplicate. The microplates were incubated for Molecules 2021, 26, 6958 5 of 18 96 h, under controlled incubation conditions, as used for the inoculum. After that time, the microalgae yield in each treatment was calculated as the difference between the cell densities at the end and beginning of the test. EC 50 , and the corresponding 95% confidence intervals, were calculated through a non-linear regression, using the least squares method to fit the data to the logistic equation.

Aqueous Biphasic Systems
The ABS ternary phase diagrams, composed of [N 4444 ] [Theob] or [N 4444 ][Theop], Na 2 SO 4 or K 3 C 6 H 5 O 7 , and water were determined through the cloud point titration method [41,42] at atmospheric pressure, continuous stirring, and in a temperature-controlled cell at (298, 323 or 353) K ± 1 K for the Na 2 SO 4 salt and at (298 ± 1) K for the K 3 C 6 H 5 O 7 salt. [N 4444 ]Cl ternary phase diagrams, using Na 2 SO 4 as a salting-out agent, were assessed under the same conditions described above, as a means of comparison. Each IL solution was prepared at 60 wt%, and the salting-out agents Na 2 SO 4 and K 3 C 6 H 5 O 7 solutions were prepared at 17 and 40 wt%, respectively. Briefly, the salt aqueous solution was added dropwise to a known amount of IL aqueous solution, until the mixture became turbid. Then, water was added to the cloudy mixture until it became clear, fitting the monophasic region. The inverse methodology was also used, when necessary, by adding an aqueous solution of the IL dropwise to a known aqueous solution of salt, to further complete the profile of the binodal curve. The ternary systems composition was determined by weight quantification (±10 −4 g) of all the added components. Tie-lines (TL) and tie-line lengths (TLL) were determined by applying the gravimetric method (±10 −4 g), proposed by Merchuk et al. [43]. Biphasic mixtures, composed of IL, salt, and water, were prepared and vigorously stirred. After being left to equilibrate at (298, 323 or 353) K ± 1 K overnight, the two phases were carefully separated and weighed, allowing for the determination of the mentioned parameters. The pH of each phase was also determined. The fitting of the experimental binodal curves and calculation of the TLLs were accomplished according to the literature [44]. Studies on the ferulic acid partition were conducted using biphasic mixtures, and the experimental details can be found in the Supplementary Materials.

COSMO-RS
The purine-based IL σ-profile and hydrogen-bonding interaction energies (E HB ) were estimated using COSMO-RS (conductor-like screening model for real solvents) thermodynamic model. Details about the employed methodology can be found elsewhere [45,46]. The quantum chemical COSMO calculations were performed with the TURBOMOLE 6.1 program package on the density functional theory (DFT) level, applying the BP functional B88-P86 with a triple-2 valence polarized basis set (TZVP) and the resolution of identity standard (RI) approximation [47]. The COSMOthermX program, using the parameter file BP_TZVP_C20_0111 (COSMOlogic GmbH & Co KG, Leverkusen, Germany), was used in all calculations [48].

Results and Discussion
Four novel purine-based salts, including three ILs, composed of tetrabutylammonium cation and the anions theobromine, theophylline, xanthine, and uric acid, were successfully synthesized in this work (Table 1). Aiming to develop fully bio-based compounds, the preparation of purine-based ILs with cholinium as a cation and theophylline as an anion was also tested, following the same synthesis protocol. It was not possible to obtain the aimed IL structure, but no protocol adaptations were carried out. Thus, with proper optimization, it could be possible to synthesize purine-based ILs with cholinium as a cation.

Synthesis and Characterization of Purine-Based ILs
The purine-based salts were obtained as solids at yields above 94% (see Table 1 [49]. The high melting temperature of the purines is likely due to the strong intramolecular interactions between the N-H groups [50]. The introduction of the bulky [N 4444 ] + cation disrupts this efficient packing, lowering the lattice energy and melting point of the salts, in comparison to the purines. Nevertheless, the melting point of the purine-based ILs is still higher than the melting point of other conventional ILs, such as [N 4444 ]Cl (344 K) [51]. As for the decomposition temperature of the synthesized salts, this property increased in the following order: . The degradation order of these ILs matches the order of the catabolic pathway of caffeine: caffeine → theobromine/theophylline → xanthine → uric acid [52]. This further supports the theory that N-demethylation is connected to the decomposition temperature values of the purine-based salts. This is not valid for uric acid, since the conversion of xanthine to uric acid occurs via oxidation. Still, [N 4444 ][Ur] has the highest T d value, while simultaneously being the last metabolite of the excerpt of this catabolic pathway. The aqueous solubility enhancement (S/S 0 ), where S is the molar water-solubility of the purine-based salt and S 0 is the molar water-solubility of its free corresponding purine, is depicted in Figure 1. Detailed data, including the pH of the saturated IL solutions, can be consulted in the Supplementary Materials, Table S1.
All the salts presented significant water-solubility improvements, when compared to the solubility of the purines, with the aqueous solubility enhancement ranging from 53-to 870-folds. The S/S 0 of the purine-based salts increases as follows: [ are isomers, their solubility is significantly different. The same occurs in their free purines, with theobromine and theophylline having a water-solubility of 1.83 × 10 −3 mol·L −1 and 4.09 × 10 −2 mol·L −1 , respectively [54]. Theobromine has relatively strong intermolecular hydrogen bonding between its carbonyl groups and strong π-π stacking interactions [18,55,56]. In theophylline, the hydrogen between the two carbonyl groups is replaced by a methyl group, causing the previously mentioned interactions to be weaker. Ultimately, this leads to theophylline being more soluble than theobromine [18,55,56] Table S1). In the ILs case, the anionic form of theobromine may be responsible for reducing the occurrence of π-π stacking, contributing to the higher solubility of the IL, when compared to its respective free purine. When comparing the structures of the four purine-based salt (  [57]. However, this was not observed in the present work. The overall poor solubility of purines is connected to their inter-base hydrogen bonding [56]. The insertion of the methyl groups replaces protons that would otherwise be available to form inter-base hydrogen bonds [56]. Consequently, the insertion of the two methyl groups increases the solubility of the methylated purines, in comparison to the non-methylated ones. This is in agreement with our results, with [ Table S2). Sodium theophyllinate has better watersolubility than theophylline, but its solubility is not improved in aqueous solutions of [N 4444 ]Cl. This shows that the solubility improvements seen here are due to the formation of the salt, rather than the hydrotropic effect of the cation.   Table S3. hydrotropic effect of the cation.
The ecotoxicity of the synthesized purine-based salts and [N4444]Cl was determined using the microalgae Raphidocelis subcapitata. The obtained EC 50 values (μmol•L −1 ), after 96 h of incubation, and their correlation with the octanol-water partition coefficient of the salt anions is shown in Figure 2. The EC50 values in mg•L −1 are available for consultation in Supplementary Materials, Table S3. . According to Egorova et al. [61], the ecotoxicity of ILs is mostly influenced by the alkyl chain length and side-chain functionalization of the cation, both the cation and anion nature, and the mutual influence of the anion-cation combination. In our work, the anion has little impact on the ecotoxicity of the ILs, suggesting that the ecotoxicity is dominated by the cation, which is in close agreement with other reports [62,63]. The tetrabutylammonium cation has four butyl side-chains, linked to a central heteroatom, causing it to have a higher hydrophobic profile. As a consequence, this cation is more prone to interact with the hydrophobic domains present in the microalgae's cell wall and potentially cause it to disrupt [63]. Although the anion has a small effect on the salts ecotoxicity, it is not insignificant. It was found that purine-based salts ecotoxicity is correlated with the logarithm of their anion octanol-water partition coefficient, log (Kow) ( Figure 2B). The obtained results suggest that salts, composed of anions with lower log (Kow) values, and, thus, a higher affinity for water, have lower ecotoxicity values, which is in agreement with the literature [64]. However, [N4444]Cl ecotoxicity is out of this trend.   [61], the ecotoxicity of ILs is mostly influenced by the alkyl chain length and side-chain functionalization of the cation, both the cation and anion nature, and the mutual influence of the anion-cation combination. In our work, the anion has little impact on the ecotoxicity of the ILs, suggesting that the ecotoxicity is dominated by the cation, which is in close agreement with other reports [62,63]. The tetrabutylammonium cation has four butyl side-chains, linked to a central heteroatom, causing it to have a higher hydrophobic profile. As a consequence, this cation is more prone to interact with the hydrophobic domains present in the microalgae's cell wall and potentially cause it to disrupt [63]. Although the anion has a small effect on the salts ecotoxicity, it is not insignificant. It was found that purine-based salts ecotoxicity is correlated with the logarithm of their anion octanol-water partition coefficient, log (K ow ) ( Figure 2B). The obtained results suggest that salts, composed of anions with lower log (K ow ) values, and, thus, a higher affinity for water, have lower ecotoxicity values, which is in agreement with the literature [64]. However, [N 4444 ]Cl ecotoxicity is out of this trend.
The incorporation of naturally occurring purines as anions was not sufficient to balance the ILs and salt ecotoxicity. As also shown by other reports, employing biocompatible feedstock for the design of ILs does not always translate into eco-friendly ILs [65,66]. There is no simple answer for the synthesis of biocompatible ILs, which reinforces the need to do (eco)toxicity screening evaluations and consequent structure optimizations, aiming to achieve more benign ILs.

Aqueous Biphasic Systems Formation
To unlock new applications for the water-soluble purine-based salts synthesized in this work, their ability to form ABS was evaluated. To evaluate the potential ability to form thermoresponsive ABS, three different temperatures were studied in the ILs-Na 2 SO 4 systems: 298, 323, and 353 (±1) K. The determined binodal curves, using Na 2 SO 4 as salting-out agents, are shown in Figure 3 and Supplementary Materials, Figure S2. The comparison of binodal curves using different salts (Na 2 SO 4 and K 3 C 6 H 5 O 7 ) at 298 K is depicted in Figure 4. The experimental weight fraction data, Merchuk equation parameters, and TL data are reported in the Supplementary Materials, Tables S4-S11. The biphasic region is located above the binodal curve.
To unlock new applications for the water-soluble purine-based salts synthesized in this work, their ability to form ABS was evaluated. Cl ILs with Na2SO4 and K3C6H5O7 as salting-out agents were determined. To evaluate the potential ability to form thermoresponsive ABS, three different temperatures were studied in the ILs-Na2SO4 systems: 298, 323, and 353 (± 1) K. The determined binodal curves, using Na2SO4 as salting-out agents, are shown in Figure 3 and Supplementary Materials, Figure S2. The comparison of binodal curves using different salts (Na2SO4 and K3C6H5O7) at 298 K is depicted in Figure 4. The experimental weight fraction data, Merchuk equation parameters, and TL data are reported in the Supplementary Materials, Tables S4-S11. The biphasic region is located above the binodal curve.       [67]. This correlates with both the hydrogen-bond basicity of the IL and the molar entropy of the salt. High charge density salts are more prone to establish water hydration complexes, inducing the formation of the ABS, by salting-out the IL from the aqueous media. ILs with higher affinity for water and, therefore, higher hydrogen-bond basicity, difficult their salting-out and, consequently, ABS formation. To better understand the obtained experimental results, the σ-profiles and hydrogen-bonding interaction energies (E HB ) of each IL were determined using COSMO-RS. The σ-profiles and surfaces of the IL anions are presented in Figure 5. As expected, the σ-profiles of the theophylline and theobromine anions are similar, presenting the main peaks at the non-polar and H-bond acceptor regions, while chloride anion has a single peak at the H-bond acceptor region. In what concerns the hydrogen-bonding interaction energies of the studied ILs, the E HB increased as follows:  Experimental results, presented in Figure 4, show that the salt influence on the formation of the ABS depends on the IL. In systems using [N4444][Theop], the binodal curves are similar, with K3C6H5O7 being a slightly better ABS inducer than Na2SO4. The same does not apply for [N4444][Theob] binodal curves, with Na2SO4 being a better ABS promoter than K3C6H5O7. This change in tendency was also observed with other ILs [67]. Take the formation of ABS with [N4444]Cl/[N4444]Br and Na2SO4/K3C6H5O7 as an example, for which the K3C6H5O7 salt ability to form ABS with [N4444]Br is slightly better than the one reported for Na2SO4 [68]. However, the anion change from bromide to chloride leads to the modification of this tendency, with K3C6H5O7 being a stronger salting-out agent than Na2SO4 [67,68]. Overall, the salt seems to have a small impact on the binodal curves, under the tested conditions.
As  Experimental results, presented in Figure 4, show that the salt influence on the formation of the ABS depends on the IL. In systems using [N 4444 ][Theop], the binodal curves are similar, with K 3 C 6 H 5 O 7 being a slightly better ABS inducer than Na 2 SO 4 . The same does not apply for [N 4444 ][Theob] binodal curves, with Na 2 SO 4 being a better ABS promoter than K 3 C 6 H 5 O 7 . This change in tendency was also observed with other ILs [67]. Take the formation of ABS with [N 4444 ]Cl/[N 4444 ]Br and Na 2 SO 4 /K 3 C 6 H 5 O 7 as an example, for which the K 3 C 6 H 5 O 7 salt ability to form ABS with [N 4444 ]Br is slightly better than the one reported for Na 2 SO 4 [68]. However, the anion change from bromide to chloride leads to the modification of this tendency, with K 3 C 6 H 5 O 7 being a stronger salting-out agent than Na 2 SO 4 [67,68]. Overall, the salt seems to have a small impact on the binodal curves, under the tested conditions.
As ]Cl has a smaller response to the tested temperature stimulus. Altogether, the purine-based ILs seem to have better thermoresponsive behavior than [N 4444 ]Cl. All the systems present a lower critical solution temperature (LCST) phase transition type, meaning that the formation of the ABS is favored by the increase of temperature. In addition, the thermoresponsive behavior of the ABS was found to be reversible. If a monophasic ternary mixture at 298 K is heated up to 323 or 353 K, a biphasic ternary mixture is formed. However, when that mixture is cooled down to 298 K, the system turns monophasic again. The thermoresponsive character of these systems unlocks more potential applications of the purine-based ILs, enabling not only the extraction and separation of molecules but also their recovery and reuse [69]. The [N 4444 ][Theob]-Na 2 SO 4 -water system is particularly promising, since smaller temperature increases (from 298 to 323 K) lead to larger differences in the binodal curves. Although the thermoresponsive behavior of these ILs is less pronounced than in other reported ILs [70], the LCST behavior is usually found in IL-polymer-water systems and not in IL-salt-water systems [69,70]. Herein, the ability to form thermoresponsive systems using Na 2 SO 4 as a salting-out agent may be related to the self-aggregation of the tetrabutylammonium cation [71]. Tie-lines (TL) were determined, as detailed in the Supplementary Materials (in Tables S10 and S11).
To further reinforce the relevance of purine-based ILs application in the formation of ABS, a brief study on the ability to partition a biomolecule (ferulic acid), using mixture points from the determined systems, was conducted at 298 K (data detailed in the Supplementary Materials, Figure S3). The purine-based ILs and [N 4444 ]Cl had a similar extraction performance, with 96 to 100% of ferulic acid being extracted to the top phase, regardless of the used salting-out agent (K 3 C 5 H 6 O 7 or Na 2 SO 4 ).

Solubility Enhancement in Aqueous Solutions
To study the solubility enhancement capacity of the purine-based ILs, aqueous solutions of [ ]Cl solubility enhancement capacity was also studied, in order to compare its efficiency with the purine-based ILs. The molar solubility and the molar aqueous solubility enhancement (S/S 0 ) of ferulic acid at different concentrations of IL are depicted in Figure 6, and detailed data can be consulted in the Supplementary Materials (Table S12). The S 0 value (0.0041 ± 0.0002 mol·L −1 ) was obtained by saturating ferulic acid in pure water, which is in agreement with the reported ferulic acid water-solubility of 0.00470 ± 0.00001 mol·L −1 [72]. The IL integrity was confirmed by 1 H and 13  . At lower IL concentrations (the diluted region), it is possible to calculate the Setschenow constant, which correlates the ratio between the solubility enhancement and solubility enhancer concentration in the diluted region. The solubility data obtained at the diluted region was used to calculate the Setschenow constant, according to the following equation [73]: where S is the molar solubility of ferulic acid, K HYD is the empirical Setschenow constant, C IL is the molar concentration of the IL, and S 0 is the water-solubility of ferulic acid. respectively (see Figure S4 and Table S13). Regarding the S/S 0 of ferulic acid in IL solutions, all the ILs seem to enhance the solubility of ferulic acid (up to 146-folds). Yet, the pH-dependency of the solubility of ferulic acid was not taken into consideration in the previously presented data. To understand how pH influences the obtained solubility results, the water-solubility of ferulic acid at different pH values and pH of the saturated samples were studied. The pH of the samples was greatly dependent on the IL nature and its concentration (see Figure 7A and Table S14 in the Supplementary Materials). Since the solubility of ferulic acid depends on pH, different S 0 values were used to determine the S/S 0 , according to the pH value of the samples ( Figure 7B and  (Table S value (0.0041 ± 0.0002 mol•L −1 ) was obtained by saturating ferulic acid in p which is in agreement with the reported ferulic acid water-solubility of 0.0047 mol•L −1 [72]. The IL integrity was confirmed by 1 H and 13 C NMR. No data is pr [N4444][Theob] at 1.0 and 1.5 mol•L −1 since ionic exchange was verified at these trations.     Table S14). The addition of ferulic acid to the IL solutions lowered their pH, with this change being more evident in solutions with lower IL concentrations. Controlling the pH of the samples would require the addition of buffers, which would interfere with the solubility enhancement assays. For this reason, no attempts were made to adjust the pH values of the samples, and, for each sample, the initial and final pH values were registered. In the [N 4444 ]Cl solutions, the pH of the samples was similar to the different IL concentrations, ranging from 2.10 to 3.29. In this case, ferulic acid was in its neutral form and [N 4444 ]Cl improved its solubility (up to 117-folds) through hydrotropy in a pH-independent way ( Figure 7A) , the pH of the samples increases as the IL concentration increases (5.36 to 9.62 and 5.30 to 9.95, respectively), leading to the deprotonation of ferulic acid. When ferulic acid is deprotonated, its solubility is mostly ruled by pH, rather than hydrotropy. Due to the impact of pH in the solubility of ferulic acid, the S/S 0 was recalculated using the ferulic acid water-solubility at the most suitable pH, i.e., using a S 0 value determined to be at a similar pH value to that of the sample. To do so, the water-solubility (S 0 ) of ferulic acid was determined at different pH values (consult Table S15 in the Supplementary Materials). Since the addition of ferulic acid significantly lowers the pH of the aqueous solutions, it was necessary to add NaOH to the aqueous solutions, in order to measure its water-solubility at high pH values. The S/S 0 recalculation ( Figure 7C) showed that the solubility enhancement of ferulic acid, by [N 4444 ][Theob], is mainly ruled by pH, being its solubility improvement lower than the one provided by pH alone (S/S 0 < 1). Regarding [N 4444 ][Theop], at IL concentrations ranging from 0.5 to 1.5 mol·L −1 , the ferulic acid solubility enhancement is higher than the one caused by pH (S/S 0 > 1). This means that [N 4444 ][Theop] has a hydrotropic effect besides the pH-dependent solubility enhancement. The ferulic acid solubility seems to achieve its maximum peak at IL concentrations of 1.0 mol·L −1 .
Charged molecules are more soluble than neutral ones. Thus, ferulic acid speciation changes from 0 to −1 or −2 global net charge, results in higher S 0 values and a lower hydrotropic effect, causing pH to have a great influence on the solubility of charged molecules. However, this does not justify why the hydrotropic effect of the tetrabutylammonium cation significantly dropped in the samples containing higher pH values. The hydrotropic behavior of the tetrabutylammonium cation has been reported towards other molecules [4,32], suggesting that other effects may be influencing the solubility of ferulic acid at higher pH values. Briefly, despite pH having a great impact on the solubility of ferulic acid, [N 4444 ][Theop] was still able to demonstrate a hydrotropic effect at IL concentrations ranging from 0.5 to 1.5 mol·L −1 , increasing the solubility of ferulic acid up to two-folds. As a side note, it should be highlighted that the purine-based ILs enable an effortless solubilization of ferulic acid at basic pH values, which could be advantageous for potential applications requiring ferulic acid to be at alkaline pH values.

Conclusions
Three purine-based ILs and one salt having tetrabutylammonium as a cation, were synthesized and characterized, regarding their melting and degradation temperatures, water-solubility, and ecotoxicity against the green microalgal Raphidocelis subcapitata. The synthesized salts displayed good thermal stability, decomposing from 457 to 505 K. Ndemethylation seems to play a role in the thermal decomposition of the purine-based ILs. In comparison to their purine precursors, the purine-based ILs afforded much higher aqueous solubilities, ranging from 53- Figure S1, ferulic acid absorbance spectrum in an aqueous solution of [N 4444 ][Theob] at 0.5 mol·L −1 ; Table S1, molar water-solubility (S) of the purine-based ILs, pH of the saturated IL solution at (298 ± 1) K, and molar water-solubility of their respective purines (S 0 ); Table S2, Table S4, experimental weight fraction of the ternary systems composed of IL, water, and Na 2 SO 4 at (298 ± 1) K and atmospheric pressure (0.1 MPa); Table S5, experimental weight fraction of the ternary systems, composed of IL, water, and Na 2 SO 4 at (323 ± 1) K, as well as atmospheric pressure (0.1 MPa); Table S6, experimental weight fraction of the ternary systems composed of IL, water, and Na 2 SO 4 at (353 ± 1) K; Table S7, experimental weight fraction of the ternary systems, composed of IL, water, and K 3 C 6 H 5 O 7 at (298 ± 1) K; Table S8, Merchuk equation parameters (A, B, and C) and the respective standard deviations (σ) for the ternary systems composed of IL, water and Na 2 SO 4 at 298 K, 323 K and 353 K (±1 K); Table S9. Merchuk equation parameters (A, B and C) and the respective standard deviations (σ) for the ternary systems composed of IL, water, and K 3 C 5 H 6 O 7 at (298 ± 1) K; Table S10, experimental TLs and TLLs of the ternary systems composed of ILs + Na 2 SO 4 at 298 K or 323 K (±1 K) and atmospheric pressure (0.1 MPa). No TLs were determined at (353 ± 1) K; Table S11, experimental TLs and TLLs of the ternary systems composed of ILs + K 3 C 5 H 6 O 7 at (298 ± 1) K and atmospheric pressure (0.1 MPa); Figure S3, extraction efficiency (EE%) of each ternary mixture at (289 ± 1) K and atmospheric pressure (0.1 MPa). Blue bars represent data from systems containing the salting-out agent K 3 C 5 H 6 O 7 , and orange bars represent ternary mixtures with Na 2 SO 4 as the salting-out agent; Table S12, molar solubility (S) of ferulic acid in IL solutions with different concentrations; Figure S4 Table S13, Setchenow constant for each IL, given by the slope of the fitted regression to the diluted region solubility data and the respective coefficient of determination (R 2 ); Table S14, measured pH of the IL solutions, before the addition of ferulic acid and after obtaining a saturated solution; Table S15, ferulic acid water-solubility at different pH values.