Supramolecular solvents based on quaternary ammonium salts and perfluorinated compounds for dispersive liquid-liquid microextraction of phthalates

INTRODUCTION

In the  past two decades, there has been an increasing and fully justified emphasis on the miniaturisation of traditional sample preparation techniques [1]. The main advantages of the miniaturised extraction methods are the high sample throughput, the ease of operation, low costs, small sample amounts required and an extremely low solvent consumption.

Dispersive liquid-liquid microextraction (DLLME), introduced by Rezaee et al. in 2006 [2], is, perhaps, the  most popular microextraction technique used in the  last decade. In the  typical DLLME procedure, a hydrophobic and highdensity extraction solvent is quickly dispersed into a conical tube containing the aqueous sample by means of the so-called dispersant solvent, which is miscible in both the  extraction solvent and sample. The microemulsion formed by hundreds of microdroplets allows a large surface area to exist between the  different phases, favouring the instant transfer of the analytes from the donor phase to the extraction solvent. After extraction, phase separation is achieved by centrifugation, where the  settled phase is easily collected with a microsyringe for later measurement in the analytical instrument. However, one of the  weakest aspects of DLLME techniques is the limited number of the effective extractants [3, 4]. Heavier than water hydrophobic solvents are preferred because such solvents sediment at the bottom of the centrifuge tube and can be simply collected by using a microsyringe.

For that reason, during the  last years, several types of alternative solvents such as ionic liquids [5], deep eutectic solvents [6] and supramolecular solvents [7] were proposed for DLLME. Because most of these solvents consist of two components, some of their properties (e.g. viscosity, density and hydrophobicity) can be tuned by combining properly the different constituents.

Supramolecular (SUPRA) solvents are defined as water-immiscible nano/microstructured liquids formed from a sequential self-assembly of different amphiphilic molecules, such as surfactants and long-chain carboxylic acids  [8]. The  production of SUPRA solvents occurs through two steps including (i) aggregation of an amphiphile, providing supramolecular assemblies, micelles, or vesicles in a homogeneous solution and (ii) coacervation, producing water-immiscible liquids (SUPRA phase) that separate from the bulk solution. The  coacervation step requires the  action of external factors such as electrolyte, pH, temperature and organic solvent. In SUPRA selfassembly, various non-covalent interactions (hydrophobic, electrostatic, hydrogen bond, etc.) cooperatively control the  assembly process and the  final structure of the  solvent. These SUPRA solvents possess exceptional extraction capabilities due to their great surface and their different polarity environments  [7]. Moreover, they can be synthesised in situ during the  analysis, making them really attractive for microextraction approaches.

The first time that SUPRA solvents were employed for DLLME was in 2009. In this application, Ballesteros-Gómez et al. [9] used a SUPRA phase formed by decanoic acid and tetrahydrofuran (THF) as coacervation agent for the determination of benzo[a]pyrene, bisphenol A, and ochratoxin from beverages. Since then, a  lot of works have been published on the use of SUPRA solvents in DLLME  [10]. However, most of the works published employed similar approaches, using alkanoic acids (i.e. decanoic acid and octanoic acid) as amphiphilic substances and THF as coacervation agent.

The fluoroalcohol-induced coacervation of amphipiles in aqueous media was originally discovered and characterised by Khaledi and coworkers [11–13]. In recent years, several types of hexafluoroisopropanol-induced SUPRA solvents have been proposed and applied for conventional liquid-liquid extraction [14–16] and DLLME [17] of organic pollutants from aqueous samples. These studies demonstrate that due to their strong hydrogen bond donor ability and high density perfluorinated compounds should be very promising alternative coacervation-inducing agents.

In the present work, eight hydrophobic SUPRA solvents composed of tetrabutylammonium chloride and tetradecyltrimethylammonium bromide as amphiphiles and four perfluorinated compounds (hexafluoroisopropanol, trifluoroacetic acid, pentafluoropropionic acid and heptafluorobutyric acid) as coacervation-inducing agents were investigated and applied for DLLME of phthalates from wastewaters.

EXPERIMENTAL

Reagents and solutions

Tetrabutylammonium chloride (TBAC, purity ≥99%), tetradecyltrimethylammonium bromide (TDTMAB, purity ≥98%), hexafluoroisopropanol (HFIP, purity ≥99%), trifluoroacetic acid (TFA, for HPLC, purity ≥99%), pentafluoropropionic acid (PFPA, purity ≥97%), heptafluorobutyric acid (HFBA, purity ≥98%), naphthalene (NAPH, purity ≥99%), acetonitrile (for HPLC, gradient grade, purity ≥99.9%) and methanol (ACN, for HPLC, gradient grade, purity ≥99.9%) were purchased from Sigma-Aldrich (St. Lous, MO, USA). Dimethyl-(DMP), diethyl-(DEP), dibutyl-(DBP) and bis(2-ethylhexyl)-(DEHP) phthalates were used as model analytes. All of them were obtained from Sigma-Aldrich, and their purity was higher than 99%.

Stock solutions (100  mg/L) of all the  phthalates and the internal standard (naphthalene) were prepared in methanol. Stock solutions of TBAC and TDTMAB (200  mmol/L) were prepared in water. All solutions were stored in a refrigerator at 4°C.

Instrumentation

Chromatographic separations were performed on an Agilent 1290 Infinity II LC system (Agilent, Waldbronn, Germany) equipped with a  ternary pump, thermostatted column compartment, photodiode array detector and an autosampler. InfinityLab Poroshell 120 EC-C18 (3.0 × 150 mm, 2.7 μm, Agilent) column, maintained at 25°C, was used in the  experiments. Separations were performed with ACN/water mobile phases at a flow rate of 0.5  mL/min under linear gradient elution conditions. The  composition of the  mobile phase was changed from 70% ACN to 100% ACN in 5  min and held constant for further 7  min. 5 min post-run time was set to fully reequilibrate the systems. The injection volume was 5 μL and the detection wavelength was set at 220 nm (for naphthalene) and 230  nm (for phthalates). Data collection and management was performed by Agilent OpenLAB CDS software.

Procedures

The overall SUPRA phase formation procedure is illustrated in Fig.  1. Specifically, the  appropriate volume (5.0–10.0 mL) of an aqueous sample was placed into a glass centrifuge tube, and the specified amounts of NaCl (only for TDTMAB/HFIP solvent), quaternary ammonium salt solution and perfluorinated compound were sequentially added. The mixture was shaken manually for 1–30 s, resulting in the  formation of emulsion. The  phases were separated by centrifugation at 3000 rpm for 5 min and the volume of SUPRA phase was measured using a microsyringe. For the extraction efficiency measurements, the  upper aqueous phase was analysed using the HPLC technique.

For the  final DLLME of phthalates, 7.5  mL of the  wastewater sample was placed into a  12  mL glass centrifuge tube and spiked with naphthalene at 20 μg/L. Then 0.75 mL of 5.0 mol/L NaCl, 0.20 mL of 0.20 mol/L TDTMAB and 0.20 mL of HFIP were sequentially added and the mixture was shaken manually for 5 s. The phases were separated by centrifugation at 3000 rpm for 5 min, the upper aqueous phase was removed using a  syringe, the bottom SUPRA phase was dissolved in 50 μL acetonitrile and analysed using the  HPLC technique.

All measurements were performed in triplicate and mean values are reported. Before extraction, the wastewater samples were filtered through a 0.7 μm glass fiber filter.

RESULTS AND DISCUSSION

In situ formation of SUPRA phases

As already mentioned in the introduction, two quaternary ammonium salts (TBAC and TDTMAB) and four perfluorinated compounds (HFIP, TFA, PFPA and HFBA) were selected for preliminary experiments. Eight combinations of quaternary ammonium salt/perfluorinated compound were composed and examined as potential SUPRA solvents. In initial tests, different amounts of perfluorinated compound were added into a 5 mL aqueous solution containing quaternary ammonium salt, then the mixture was shaken for 30 s and centrifuged. It was found that the  mixtures of TDTMAB and HFIP did not form SUPRA phases at any concentrations in pure aqueous solutions. Further experiments showed that the TDTMAB/HFIP based SUPRA phase forms only in high ionic strength (≥100  mmol/L NaCl) solutions. Thus, all further experiments with this composition were performed using aqueous solutions containing 0.5  mol/L NaCl. The  other seven compositions formed colourless liquid SUPRA phases without the addition of NaCl. Due to the high density of perfluorinated compounds, all the in situ formed SUPRA phases were heavier than water.

In the  DLLME techniques, target analytes are usually extracted from 5–10  mL of water samples using small volumes (50–100 µL) of extraction solvents. In the  present work, the  extraction solvent was formed in situ. Therefore it is important to know what concentrations of SUPRA components generate required solvent volumes. For this reason, the effects of quaternary ammonium salt and perfluorinated compound concentrations on the  volume of the formed SUPRA phase were investigated. In this experiment, a small amount of alizarin dye (~0.02  mmol/L) was added to the  initial aqueous solution. Relatively hydrophobic alizarin was concentrated in the  SUPRA phase and coloured it in an intense red colour. This improved the  visibility of SUPRA solvents and facilitated their collection after phase separation. Figure 2 shows the effects of TBAC and TDTMAB concentration on the volume of the formed SUPRA phase. As expected, for all systems, the SUPRA phase volume increases gradually as the  concentration of quternary ammonium salt increases from 5 to 100 mmol/L. A similar behaviour has also been reported by Khaledi et al. [12, 13] in HFIP induced complex SUPRA phases of cationic-anionic surfactant  [13] and polyelectrolyte-surfactant mixtures [12]. They concluded that the increase in volume with the surfactant concen tration was indicative of a  growth of the  SUPRA phase that is in a bilayered bicontinuous structure rather than in discrete structures such as micelles. Increasing the  quaternary ammonium salt concentration would simply make the structure larger, as opposed to the more densely packed one. In both TBAC and TDTMAB based systems, the HFIP induced much larger SUPRA phase volumes than the perfluorinated acids at comparable concentrations. This may be attributed to the different charge state of perfluorinated compounds: alcohol has a net charge of zero, whereas strong acids are present in an anionic form. Although the structures of these phases are not known at this point, one can assume that the  oppositely charged quaternary ammonium cation-perfluorinated acid anion would form tighter assemblies resulting in lower volumes. It is worth noting that for the  TBAC salt (Fig.  2a) perfluorinated acids with a longer fluoroalkyl chain induce larger SUPRA volumes, whereas for the TDTMAB (Fig. 2b) this phenomenon has not been observed. However, the exact reason for this behaviour is not clear.

Figure  3a depicts the  behaviour when a  fixed concentration of TBAC salt is combined with increased amounts of appropriate perfruorinated compound. The  volume of the  TBAC/HFIP phase increased gradually with increasing the HFIP amount in a  range of 1–5  mmol. In contrast, the  SUPRA phase volume remains almost constant for TBAC/ HFBA, TBAC/PFPA and TBAC/TFA systems within a  range of 2–5, 3–5 and 4–5  mmol of appropriate perfluorinated acid, respectively. At lower amounts of perfluorinated acid, all three mixtures form white precipitates. The initial amount of acid that is needed to form the SUPRA phase decreases in the following order: TFA > PFPA > HFBA. This indicates that the SUPRA phase formation ability of perfluorinated acids with a longer fluoroalkyl chain is more efficient.

Figure  3 b illustrates the  behaviour when TDTMAB is combined with different amounts of perfluorinated compounds. As the amount of appropriate perfluorinated compound increases from 1 to 5 mmol, the volume of TDTMAB/HFIP and TDTMAB/PFPA phases first decreases and then gradually increases (for TDTMAB/HFIP) or remains almost unchanged (for TDTMAB/PFPA). The volume of TDTMAB/HFBA phase formed was nearly unaffected by the varied amount of HFBA. Finally, as in the  case of TBAC/TFA system, for the TDTMAB solution, the minimum amount of TFA required to form the SUPRA phase is 4 mmol.

The obtained results show that there are some differences among the two quaternary ammonium salts. One possible reason for such differences could be attributed to different initial states of these salts in an aqueus solution. In the concentration range used, TDTMAB forms micelles (critical micelle concentration 3.6–3.7 mmol/L [18]), whereas TBAC exists in the  non-micellar state. Additional information about the exact structures of these phases is needed to achieve a  better understanding of the  involved mechanisms. However, a  detailed investigation of these effects is beyond the scope of this study.

Extraction properties

The suitability of the  SUPRA solvents for liquidphase microextraction was tested for the  extraction of four phthalates from the  model aqueous solutions. The comparison of the extraction properties of different solvents should be performed under identical conditions, i.e. using equal or at least similar extraction solvent volumes. For each solvent we have selected such amounts of both SUPRA compounds that gave the final solvent volume of 150 ± 10 µL. The extraction efficiency (EE) of each analyte was calculated according to the following equation

where c_i is the  initial concentration of the  analyte in the aqueous phase before the extraction and c_f is the analyte concentration in the aqueous phase after the extraction. The final concentrations of the analytes in the  aqueous phase were measured using the  HPLC technique (for HPLC conditions see the Experimental section).

The obtained results are presented in Table 1. It was not possible to evaluate higher than 98% EE data for phthalates because their concentrations in the aqueous phase after the extraction were below the limit of quantification. In general, under the  employed extraction conditions all developed SUPRA solvents demonstrated a good or even excellent overall extraction efficiency. As expected, for all phthalates their EE values showed a good correlation with their log P values, listed in Table 1, indicating that hydrophobicity plays a crucial role in the extraction of these compounds from aqueous solutions. Both TFA-based solvents were found to be least effective extractants: the extraction of the most hydrophilic DMP reached efficiencies of 58.9 and 70.2% for TDTMAB/TFA and TBAC/TFA, respectively. Finally, the TDTMAB/ HFIP solvent demonstrated the  best extractability: using this solvent all phthalates were extracted with higher than 94% extraction efficiencies.

It should be taken into account that by the extraction with perfluorinated acid-based solvents the pH of aqueous phase considerably decreased. Although the pH decrease had no influence on the extraction efficiency of the compounds studied, this may have a  significant effect on the  extractability of weakly acidic and basics compounds. Thus, these solvents should be useful for the extractions in which acidification of the aqueous phase or the extractant is required.

Compatibility with RP-HPLC

HPLC in the  reversed-phase mode (RP-HPLC) is the most common analytical technique used in combination with DLLME sample pretreatment procedures [3–7]. It is well established that in RP-HPLC the  sample solvent with a  higher elution strength (i.e. higher hydrophobicity) than the mobile phase leads to peak broadening, extreme distortions or even fragmentation of peaks [19, 20]. In addition, the difference in viscosity between the sample solvent and mobile phase also may affect the separation performance. When the  sample viscosity greatly exceeded that of the mobile phase, flow instabilities could occur which lead to serious deformations and splitting of sample component peaks [21]. Another problem that may occur by the RP-HPLC analysis is the interferences from the solvent itself, its main compounds or impurities which are detected under conditions used.

In liquid-phase microextractions employing conventional hydrophobic extractants, this problem is usually avoided by simple evaporation/reconstitution steps prior to RP-HPLC analysis. In contrast to the common organic solvents, SUPRA solvents are nonvolatile and hence they cannot be removed by evaporation. Thus, before the application of the solvents for the  microextraction of phthalates with further RP-HPLC analysis it is worth to check their compatibility with RP-HPLC.

Figure 4 compares the  chromatograms of four phthalates dissolved in pure ACN and in four TDTMAB-based SUPRA solvents diluted (1:1, v/v) with ACN. As expected, the  sample dissolved in ACN shows four narrow and symmetric peaks. No significant broadening and distortion of the peak profiles was observed for the phthalates dissolved in ACN/SUPRA mixtures indicating that the  hydrophobicity of the  SUPRA solvents has no remarkable effect on the  peak shapes. Most likely, the dilution of the SUPRA solvent with ACN and/ or its interaction with the stationary phase during separation breaks the bond network of the solvent and, consequently, reduces the  hydrophobicity of the diluent.

However, two unexpected peaks appeared in the chromatograms of phthalates dissolved in perfluorinated acid containing SUPRA solvents. One of them corresponds to an unretained compound and does not interfere in the  detection of phthalates. The other broad peak, however, overlaps with the DEP peak or with both DMP and DEP peaks. Moreover, its intensity and retention time depend on the nature of perfluorinated acid and increase in the following order: TFA < PFPA < HFBA. The obtained results indicate that the extra peaks may be attributed to the perfluorinated acid itself, its impurity and/or the formation of a certain compound. This was confirmed by injecting samples of individual perfluorinated acids and their SUPRA solvents (data not shown). In the  chromatograms of perfluorinated acid solutions only one peak appeared at about 1.3  min indicating that free acids were unretained under the given conditions. In contrast, by the analysis of SUPRA solvent solutions (without phthalates) two peaks were detected and their chromatographic profiles were quite similar to the extra peak profiles obtained in the corresponding chromatograms of phthalates which are shown in Fig. 4. These results suggest that most likely in the acetonitrile-rich SUPRA solution perfluorinated acid exists in two forms: unretained hydrophilic free acid and retained less hydrophilic neutral ion pair with an oppositely charged quaternary ammonium cation.

It should be noted that by using UV detection at ≥250  nm, these interferences are completely avoided but with a significant loss of the detection sensitivity of phthalates. Similar chromatographic profiles were also observed for the phthalates dissolved in the  four TBAC-based SUPRA solvents (data not shown).

As a conclusion, it seems that at least for some phthalates the  perfluorinated acid-based SUPRA solvents are not suitable as extractants and sample diluents for the  RP-HPLC analysis. On the  other hand, because these solvents exhibited good extraction properties, they may be very promising alternative extractants for other analyte types or in combination with other analytical techniques.

DLLME of phthalates

Based on the  above results, the  TDTMAB/HFIP solvent was selected for the DLLME of phthalates from the  real water samples combining DLLME with RP-HPLC analysis. Naphthalene (NAPH) was used as an internal standard. In order to obtain high enrichment factors, the volume ratio of extraction solvent to sample solution should be minimised. For this reason, the effect of TDTMAB/ HFIP solvent volume on the extraction efficiency of phthalates and naphthalene was investigated and the obtained results are presented in Table 2. To obtain the  required solvent volumes, different amounts of TDTMAB and HFIP were added to a 7.5 mL of aqueous sample spiked at 1.0 mg/L of each phthalate and the  internal standard. As can be seen, by decreasing the SUPRA solvent volume up to 50 μL higher than 95% extraction efficiencies were obtained for all five compounds. Based on these results, 50 μL of TDTMAB/HFIP volume was selected for subsequent experiments.

Next, the effect of extraction time on the extraction efficiency was studied over a range of 1–30 s. No significant difference in the extraction efficiency of all compounds was observed in the time range studied indicating that the  SUPRA solvent-based DLLME is time-independent. The final extraction conditions were as follows: the aqueous sample volume 7.5 mL, SUPRA extractant volume 50 μL and extraction time 5 s.

The analytical performance of the method was investigated under the  optimised extraction conditions. The  results are summarised in Table  3. The  calibration curves were linear for a  concentration level between 1.0–2.0 and 50.0  μg/L with the  correlation coefficients (R²) ranging from 0.9965 to 0.9982. The  limits of detection (LODs) and the limits of quantification (LOQs), calculated based on signal to noise ratios of 3 and 10, respectively, were in the ranges 0.3–0.6 and 1.0–2.0 μg/L. The  enrichment factors for all studied phthalates were around 75.

Finally, the developed method was applied for the determination of phthalates in two influent and one effluent (i.e. after treatment) water samples collected from the  main wastewater treatment plant of Vilnius. The  influent samples were collected at different days. Only DBP and DEHP in the one influent water sample were detected. Unfortunately, the determination of DMP and DEP was impossible due to the interference from coeluting matrix compounds. Recovery studies were performed by spiking samples with the  known concentrations of DBP and DEHP. The results are summarised in Table 4. The average recoveries of spiked samples were in the range from 93.1% to 104.4% with a satisfactory precision (RSD lower than 7.4%). The chromatograms of the unspiked and spiked influent water sample are shown in Fig. 5. The obtained results indicate that the developed method is a  promising alternative for the rapid enrichment and determination of common phthalates in aqueous samples.

CONCLUSIONS

In this work, eight novel SUPRA solvents based on quaternary ammonium salt-perfluorinated compound mixtures were prepared and investigated for potential use as extractants in DLLME. The presented solvents were formed in situ during the analysis, exhibited higher than water densities and demonstrated good extraction efficiencies for phthalates. The  TDTMAB/HFIP solvent was successfully applied for the DLLME of parabens from wastewater samples prior to HPLC analysis. The obtained results demonstrate that the developed solvents may be considered as very promising alternative extractants for DLLME.

References

1. M. Sajid, Trends Analyt. Chem., 152, 116636 (2022).

2. M. Rezaee, Y. Assadi, M. R. M. Hosseini, E. Aghaee, F.  Ahmadi, S.  Berijani, J. Chromatogr. A., 1116, 1 (2006).

3. M. I. Leong, M. R. Fuh, S. D. Huang, J. Chromatogr. A, 1335, 2 (2014).

4. M. Saraji, M. K. Boroujeni, Anal. Bioanal. Chem., 406, 2027 (2014).

5. V. Vičkačkaitė, A. Padarauskas, Cent. Eur. J. Chem., 10(3), 652 (2012).

6. P. Makoś, E. Słupek, J. Gębicki, Microchem. J., 152, 104384 (2020).

7. M.  Moradi, Y.  Yamini, N.  Feizi, Trends Analyt. Chem., 138, 116231 (2021).

8. S. Rubio, Anal. Bioanal. Chem., 412, 6037 (2020).

9. A. Ballesteros-Gómez, S. Rubio, D. Pérez-Bendito, J. Chromatogr. A, 1216, 530 (2009).

10. J.  Grau, C.  Azorin, J.  L.  Benedé, A.  Chisvert, A. Salvador, J. Sep. Sci., 45, 210 (2022).

11. M. G. Khaledi, S. I. Jenkins, S. Liang, Langmuir, 29, 2458 (2013).

12. M. M. Nejati, M. G. Khaledi, Langmuir, 31, 5580 (2015).

13. S. I. Jenkins, C. M. Collins, M. G. Khaledi, Langmuir, 32, 2321 (2016).

14. J. Xu, M. Niu, Y. Xiao, Anal. Bioanal. Chem., 409, 1281 (2017).

15. J. Xu, Y. Li, C. Li, R. Zhang, Y. Xiao, Anal. Bioanal. Chem., 409, 4559 (2017).

16. J. Xu, X. Li, C. Li, J. Chen, Y. Xiao, Food Chem., 209, 122 (2018).

17. J.  Chen, W.  Deng, X.  Li, X.  Wang, Y.  Xiao, J. Chromatogr. A, 1591, 33 (2019).

18. J. Roy, N. Islam, J. Solut. Chem., 48, 758 (2019).

19. B. J. Van Middlesworth, J. G. Dorsey, J. Chromatogr. A, 1236, 77 (2012).

20. J.  Layne, T.  Farcas, I.  Rustamov, F.  Ahmed, J. Chromatogr. A, 913, 233 (2001).

21. S. Keunchkarian, M. Reta, L. Romero, C. Castells, J. Chromatogr. A, 1119, 20 (2006).

Vytautas Kavaliauskas, Vilma Olšauskaitė, Audrius Padarauskas

SUPRAMOLEKULINIAI TIRPIKLIAI KETVIRTINIO AMONIO DRUSKŲ IR PERFLUORINTŲ JUNGINIŲ PAGRINDU FTALATŲ DISPERSINEI SKYSČIŲ-SKYSČIŲ MIKROEKSTRAKCIJAI