Determination of aromatic amines in environmental water samples by deep eutectic solvent-based dispersive liquid-liquid microextraction followed by HPLC-UV

2.1 Chemical and reagents

4-Chloroaniline (4-CA, purity ≥ 99.0%), 3-nitroaniline (3-NA, purity ≥ 98.0%), 2-naphtylamine (2-NA, purity ≥ 99.0%), HPLC grade acetonitrile (ACN), bis(2-ethylhexyl) phosphate (BEHP, purity ≥ 97.0%), octanoic acid (OA, purity ≥ 99.0%), benzyltrimethylammonium chloride (BTEAC, purity ≥ 97.0%), phenol (Ph, purity ≥ 99.0%), ammonium acetate (purity ≥ 98.0%), and acetic acid (purity ≥ 99.7%) were purchased from Sigma-Aldrich (Steinhem, Germany). The chemical structure of the studied AAs is shown in Fig. 1a. A stock solution of the PAAs at concentration level of 2000 mg L⁻¹ was prepared by dissolution of proper amounts of each AAs in methanol. The solution kept in the refrigerator and stable at least for 6 month. A mix working solution of AAs at a concentration level of 40 mg L⁻¹ was also prepared in methanol by suitable diluting.

Close

Fig. 1

Chemical structure of PAAs (a). Suggested structure of the prepared DES (b).

Export to PPT

2.2 Apparatus

Chromatographic separation was performed by EuroChrom model high-performance liquid chromatography (HPLC) from Knauer Germany. A Capital chromatography column with specifications of C8: 250 mm × 4.6 mm, 5 μm (Broxburn, UK) was used. The injection volume and detection wavelength was 20 µL and 235 nm, respectively. The elution of the analytes was done under isocratic elution by using a mobile phase of 0.1 mol L⁻¹ ammonium acetate buffer (pH = 7.0) and ACN (65:35, %v/v) at a flow rate of 1.0 mL min⁻¹.

Fourier transform infrared (FTIR) spectra were measured with a PerkinElmer RXI spectrometer in the range of 400 cm⁻¹ to 4000 cm⁻¹ using the KBr method. Samples were centrifuged by an Universal 320R centrifuge from Hettich Zentrifugen (Tuttlingen, Germany). Samples were vortexed by a vortex-2 Genie vortex mixer from Scientific Industries (Bohemia, NY, USA).

2.3 Deep eutectic solvent preparation

Two hydrophobic DESs (BTEAC-octanoic acid and BEHP-phenol) were prepared by mixing 10 mmol of BTEAC and BEHP as the acceptor of hydrogen bonds (HBA) with 30 mmol of the octanoic acid and phenol as the donor of hydrogen bonds (HBD) in a round-bottom flask, respectively. The mixtures were stirred at 50 °C until the mixtures became clear and then the solvents kept at the ambient temperature.

2.4 Samples and analytical procedure

Three types of water samples were collected and analyzed. Lake water sample was collected from the Persian Gulf Martyrs Lake (Tehran, Iran). A fish pond water sample was collected from a fish farm from Karaj (Alborz, Iran). Tap water sample was collected from Standard Research Institute (Karaj, Iran). The samples were filtered by filter paper of 0.45 μm and then kept in the refrigerator until before analysis. pH of the samples was adjusted and the extraction procedure was carried out.

For the microextraction of AAs from the samples via DES-DLLME method, 12.0 mL of sample was transferred to the falcon centrifuge tube. The optimization condition was conducted by adding 1.0 g sodium chloride (10%) to the sample and pH adjusted (pH = 6). Then, mixture of (80.0 µL of DES (BHEP-Ph) + 100 µL of methanol) was added to the solution as expeditiously as possible to disperse properly. The solution was vortexed vigorously for 1.0 min to get the complete dispersion of the solvent. the solution was centrifuged for 5.0 min (5000 rpm) to separate the extraction solvent from the aqueous phase. After centrifugation, two-phase were separated and DES phase was collected at the top of the tube. After withdrawn of the aqueous phase, due to having high viscosity of DES, most of the solvent stuck to the inner wall of the tube. For this reason, to collect all extraction solvent and complete separation of aqueous phase, tube again centrifuged for 1.0 min. Finally, the DES phase (30 ± 2 µL) was diluted with 20 µL methanol and injected into HPLC.

3.1 Characterization of DES

FT-IR spectra of BEHP, Ph and its relevant DES (BEHP-Ph) were recorded to investigate structure of DES and its important functional groups. Based on the obtained results (Fig. 4), the appearance of characteristic peaks of each component of the solvent and the formation of a broad peak in the region of 3400 cm⁻¹ proves the formation of hydrogen bond between components and as a result, demonstrate the DES formation. Stretching vibrations at 3200 cm⁻¹, 2950 cm⁻¹, 2868 cm⁻¹, 1613 cm⁻¹, 1432 cm⁻¹ and 1241 cm⁻¹, and 1082 cm⁻¹ could be assigned to O-H, C-H aromatic, C-H aliphatic, C=C, and C-OH bonds of phenol (Stuart, 2004). Also, O-H, C-H aliphatic, P=O, P-O-C vibrations of BEHP are positioned at 3409 cm⁻¹, 2979 cm⁻¹, 1217 cm⁻¹, and 1072 cm⁻¹, respectively (Stuart, 2004).

Investigation of density, viscosity and melting point of the prepared DES showed that the physical properties of the solvent are favorable. The density and viscosity of the DES in 25 °C were 0.93 g mL⁻¹ and 180 cP, respectively. Also, the DES melting point investigation showed that it has a very low melting point and does not freeze in the refrigerator (4 °C).

3.2 Optimization of DES-Dllme

3.2.1

3.2.1 Investigation of type and composition of DES

To have an efficient extraction with pinpoint accuracy, the selection of proper HBD and HBA in the structure of the DES is crucial because with targeted selection, it is possible to have strong and different interactions in the solvent structure at the same time, which leads to very efficient extractions. Based on the structure of the aniline compounds, two compositions of bis(2-ethylhexyl) phosphate-phenol (BEHP-Ph) and benzyltriethylammonium chloride-octanoic acid (BTEAC-OA) were investigated and tested. On the one hand, BEHP is used as a strong ion-pairing agent in extracting basic polar compounds (Hansen et al., 2020; Altunay and Gürkan, 2019). Also, phenol can present strong π-π interactions with the studied compounds. On the other hand, aniline compounds can also have good ion-pair interactions with carboxylic acids (Torbati et al., 2018). In addition, the presence of a benzene ring in the BTEAC structure can also cause π-π interactions. The results showed that BEHP-Ph composition shows better and higher efficiency in the extraction of AAs. The average recoveries for the AAs were 65% and 80% for BTEAC-OA and BEHP-Ph, respectively.

To acquire a better result, the effectiveness of various molar ratio has been considered; hence the combination of BEHP to phenol molar ratio (1:1, 1:2, 1:3, 1:4) was prepared and investigated. The obtained results (Fig. 2a) show that as the molar ratio increases, the extraction efficiency should increase slightly and become almost constant. Therefore, the 1:2 molar ratio of BEHP-Ph was chosen for subsequent studies.

Close

Fig. 2

Effect of pH on the extraction of the PAAs (a), Effect of DES composition on extraction of PAAs (b).

Export to PPT

3.2.3

3.2.3 Influence of DES volume

The effect of different volumes of the DES (60 µL, 80 µL, 100, 120 µL) was investigated. By enhancing DES volume from 60 to 80 µL, extraction efficiency rose. However, further increases in the volume of the DES lead to decreasing signal as a result of the dilution effect (Faraji, 2019a, Faraji, 2019b). Therefore, based on Fig. 3a, 80 µL was selected as an optimum DES volume for subsequent studies.

Close

Fig. 3

Effect of DES volume on extraction efficiency of aromatic amines (a), Effect of salt concentration on extraction of aromatic amines (b).

Export to PPT

3.3 Analytical performance

Under optimal conditions, the performance of the developed method was evaluated by considering linear dynamic range (LDR), precision (intra-day and inter-day), the limit of detection (LOD), the limit of quantification (LOQ), and enrichment factor (EF). The figures of merit of the developed technique are presented in Table 1. The calibration curve was linear in the range of 0.5–200 μg L⁻¹, 0.3–200 μg L⁻¹, and 0.2–200 μg L⁻¹, for 4-CA, 3-NA, and 2-naphtyl amine, respectively (R² ≥ 0.99). Intra-day and inter-day RSD values are less than 6.0% at concentration levels of 5.0 and 25 µg L⁻¹. The LOD values based on S/N = 3 were 0.17, 0.1 and 0.07 µg L⁻¹ for 4-CA, 3-NA, 2-naphtylamine, respectively. Also, the LOQ values based on S/N = 10 were 0.5, 0.3 and 0.2 μg L⁻¹ for 4-CA, 3-NA, 2-naphtylamine, respectively. The enrichment factors (EFs) were calculated as the ratio between analyte concentrations in the acceptor phase (DES phase) to the initial concentration in the sample. EFs were 170, 180 and 190 for 4-CA, 3-NA, and 2-naphtyl amine, respectively.

3.4 Real sample analysis

The applicability of the developed method in real sample analysis was evaluated by analyzing some water samples under optimum conditions. The data are presented in Table 2. The concentration of the studied AAs in the water sample was below the LOD. Therefore, to verify the accuracy of the method, the samples were spiked at known concentration levels of PAAs. Good relative recoveries ranged from 90.0 to 98.0 with RSD less than 7.0% were obtained by analyzing three independent samples. Chromatogram of the lake water sample before and after the spike of level 1 (10, 5.0 and 5.0 µg L⁻¹ respect to 4-CA, 3-NA, 2-naphtylamine, respectively) and level 2 (20, 10.0 and 10.0 µg L⁻¹ respect to 4-CA, 3-NA, 2-naphtylamine, respectively) of mixture standard solution of the PAAs is shown in Fig. 4.

Close

Fig. 4

FT-IR spectrum of BEHP (blue), Ph (red) and BEHP-Ph (green) (a). Chromatograms of lake water before and after spike of level 1 (10, 5.0 and 5.0 µg L⁻¹ respect to 4-CA, 3-NA, 2-naphtylamine, respectively) and level 2 (20, 10.0 and 10.0 µg L⁻¹ respect to 4-CA, 3-NA, 2-naphtylamine, respectively) of the standard solutions of PAAs (b).

Export to PPT

3.5 Comparison of proposed technique with previously published techniques

The proposed method was compared with some of the methods that recently reported in the literatures for the extraction and determination of AAs in aqueous matrices (Table 3). As it can be seen, the linearity, detection limit, repeatability and enrichment factor of the proposed method are better or comparable to the LC-MS (Bie et al., 2017) and GC–MS (Ozkan et al., 2019, Torbati at al., 2018; Zhang and Duan, 2019) methods. In particular, the cost, time, simplicity, and repeatability of the proposed method are better and more appropriate than the SPE (Jalilian et al., 2017). The extraction solvent used in our proposed method, same as other DES based methods (Torbati et al., 2018; Werner, 2020), is safer than classical solvent based DLLME methods (Ozkan et al., 2019; Farajzadeh and Nouri, 2012, Han et al., 2013; Zou et al., 2010). However, extraction solvent separation in DES based methods is a bit difficult and usually an additional centrifugation step is required to complete separation of DES phase because of higher viscosity of DES solvents compared to the classical solvents that can stick to the walls of the vials. It should be noted that the analysis of AAs by GC-FID (Farajzadeh and Nouri, 2012, Han et al., 2013; Zou et al., 2010) or GC–MS (Ozkan et al., 2019; Torbati et al., 2018) compared to LC-based methods (Wang et al., 2020, Jalilian et al., 2017, Werner, 2020, Bie et al., 2017), requires a polar column or a derivation step which complicates the method and reduces its greenness. However, GC-based methods offer higher enrichment factors than LC-based methods due to the greater compatibility of DLLME with GC techniques (Rezaee et al., 2010). Accessibility, simplicity and cost of reagents and instrument, the proposed methods are better than MS-based methods, but sensitivity, reliability and selectivity of MS-based methods are better than the proposed method.