Ionic liquid coating for stir bar sorptive extraction and its application for extraction and nontargeted screening analysis via TD-GC–Orbitrap-HRMS of pollutants in river water
![<p>Journal of Chromatography A, Volume 1742, 2025: Fig. 3 Microscopy images (× 250) of a commercial PDMS-coated stir bar (a) and the prepared stir bars (number 7) using the [TESBPIM][NTf2] IL (b), and a photography of the IL-coated stir bar and uncoated stir bar (c).</p>](https://gcms.labrulez.com/labrulez-bucket-strapi-h3hsga3/Journal_of_Chromatography_A_Volume_1742_8_February_2025_465623_Fig_3_48415551fa_l.webp)
Journal of Chromatography A, Volume 1742, 2025: Fig. 3 Microscopy images (× 250) of a commercial PDMS-coated stir bar (a) and the prepared stir bars (number 7) using the [TESBPIM][NTf2] IL (b), and a photography of the IL-coated stir bar and uncoated stir bar (c).
This study introduces a novel imidazolium-based ionic liquid (IL) coating for stir bar sorptive extraction (SBSE) using a sol–gel method. The IL-coated stir bar demonstrated high mechanical and thermal stability, along with increased extraction efficiency across a wide polarity range of compounds. Compared to commercial PDMS stir bars, the IL coating significantly improved extraction yields, particularly in the presence of salt.
When applied to natural water samples from France, the IL-stir bar enabled the detection of over 1,000 compounds using TD-GC–Orbitrap-HRMS, with 334 compounds annotated after deconvolution. Notably, the IL-stir bar extracted five times more compounds than PDMS, including pollutants such as pesticides, personal care products, PCBs, and PAHs. These results highlight the IL-coated stir bar’s potential for comprehensive environmental analysis.
The original article
Ionic liquid coating for stir bar sorptive extraction and its application for extraction and nontargeted screening analysis via TD-GC–Orbitrap-HRMS of pollutants in river water
Amel Meziani, Ouassila Ferroukhi, Valerie Peulon-Agasse, Pascal Cardinael
Journal of Chromatography A, Volume 1742, 8 February 2025, 465623
https://doi.org/10.1016/j.chroma.2024.465623
licensed under CC-BY 4.0
Selected sections from the article follow. Formats and hyperlinks were adapted from the original.
1. Introduction
Water samples such as wastewaters are complex mixtures and may contain interferents that require extraction procedures to achieve low limits of detection (fixed at approximately µg/L by the European Union [12]). For a long time, liquid‒liquid extraction (LLE) was the most commonly used sample preparation technique. However, this method is time-consuming and requires the use of large volumes of organic solvents. As an alternative, solid-phase extraction (SPE) was introduced, and it allowed for considerably reduced solvent consumption. This sample preparation has been used to extract different pollutants, such as pesticides, polyaromatic hydrocarbons (PAHs) and PPCPs, from water samples [13]. Nevertheless, to achieve detection limits below ng/L, this technique requires sample volumes from 5 to 10 mL [14], the extract must be concentrated to a small volume (<1 mL), or a large-volume injection must be performed. However, the use organic solvents for elution step remains necessary. Miniaturized solvent-less extraction techniques, such as dispersive liquid‒liquid microextraction (DLLME) [15], solid phase microextraction (SPME) [16], microextraction-packed sorbent (MEPS) [17] and stir bar sorptive extraction (SBSE) [18], have recently been introduced, allowing extraction and preconcentration in a single step. The latter is based on sorption and partitioning between the sorption phase and aqueous phase [19]. These techniques have offered considerable advantages for trace analysis because they can be directly coupled to online analysis, ensuring the complete transfer of the extract to the analytical system.
The main limitation of SBSE is the lack of a variety of coatings in terms of polarity. Currently, only polydimethylsiloxane (PDMS), ethylene glycol (EG)-silicone and polyacrylate (PA) coatings are commercially available, which limits the extraction of solutes from water when the logarithm of the octanol–water partitioning coefficient (log Kow) is < 2 [22]. To overcome this limitation, different approaches, such as SBSE combined with a freeze concentration called ICECLES, have been proposed. In this case, SBSE was performed while the sample was frozen from the bottom of the vial to the top, which increased the analyte extraction yield [23]. Moreover, solvent-assisted SBSE (SA-SBSE) was proposed, where a solvent is used to modify the PDMS phase polarity to increase the recoveries of polar solutes [24].
In our study, the main goal was to develop a new ILs coating for SBSE using the sol‒gel method to improve the extraction yield of polar compounds while maintaining good extraction efficiency for apolar compounds. The yields obtained with these newly prepared stir bars will be compared to those obtained with commercial PDMS stir bars using a mixture of 11 target compounds with log Ko/w values between 0.65 and 7.21 to cover a wide range of compound polarities. The prepared IL-stir bar will also be tested on real water samples (Robec River, France) to extract emerging and/or persistent pollutants. A nontargeted screening TD-GC–Orbitrap-HRMS method was developed, and the pollutants were identified processed using Compound Discoverer software. The identified compounds were annotated according to the confidence scale proposed by Miller et al.[40].
2. Experimental
2.2. Preparation of IL-coated stir bars
Two imidazolium-based ionic liquids (1-butyl-3- (3-(triéthoxysilyl)propyl)-1H-imidazol-3-ium chloride [TESBPIM][Cl] and bis((trifluoromethyl)sulfonyl)amide 1-butyl-3-(3-(triethoxysilyl)propyl)-1H-imidazol-3-ium [TESBPIM][NTF2]) and a phosphonium-based ionic liquid (tributyl(3-(triéthoxysilyl)propyl) phosphonium chloride [P444TES] [Cl]) were synthesized according to previous laboratory work [41]. The three ILs were selected due to their high thermal stability (up to 325 °C) [42].
Noncoated glass-encased magnetic stir bars (10 mm length) obtained from Gerstel (Germany) were cleaned with water followed by washing with DCM and then dipped into a solution of NaOH (1 mol/L) for 8 h. Subsequently, the stir bars were cleaned with water, dipped again in a solution of HCl (0.1 mol/L) and then dried at 60 °C overnight. The sol‒gel process was used to prepare the IL coatings. The sol solution of the IL was composed of 200 mg of the IL, 150 µL of TEOS, 200 µL of 95 % TFA and 1 mL of DCM. The treated bare bars were immersed in the solution for 5 min, removed, and dried at room temperature. This procedure was repeated several times until a thick coating was formed. The coated stir bars were conditioned using a conditioning unit under a nitrogen atmosphere (Gerstel, TC 2, Germany), and the following temperature program was used from 30 °C to 200 °C and held for 2 h.
To evaluate the coating film homogeneity, microscopic images were obtained using a Nikon Eclipse LV100 microscope (maximum magnification: 1000 ×), coupled to a CCD camera connected to a computer. The data acquisition was performed via NIS-Elements D software (version 3.1).
DRIFT spectra were obtained using a PerkinElmer (Norwalk, CT, USA) Spectrum 100. The IL sample was placed in a 2 mm diameter disk. The spectra, which consisted of 256 scans, were acquired at a resolution of 4 cm−1 and over the range of 4000–600 cm−1.
2.3. Stir bar sorptive extraction procedure
Stirring extraction and TD were used in this work. Aqueous samples (10 mL) containing 11 target compounds (log Ko/w values between 0.65 and 7.21) were placed in 20 mL vials at room temperature. The prepared IL-stir bar was stirred in the sample mixture for 1 h. After extraction, the stir bar was dried and placed in a glass desorption tube. The thermodesorption was performed using a TDU 3.5 + 5 Gerstel thermal desorption unit (Mulheim, Germany) with the following temperature program: from 40 °C (held for 1 min) at 100 °C/min to 280 °C (held for 5 min) with 15.5 mL/min desorption flow. The same methodology was applied for commercial PDMS stir bars to compare the extraction yields.
The environmental water samples were collected from the Robec River (49.441515312762, 1.099119438792342 Rouen, France). The 500 mL water samples were filtered through a 0.45 µm nylon filter and then stored in a refrigerator at 4 °C. A volume of 10 mL of filtered water with 70 mg of NaCl was used for SBSE using the prepared IL-stir bar at room temperature for 1 h The same procedure was applied using a commercial PDMS stir bar once and without NaCl addition. The stir bars were thermally desorbed under the abovementioned conditions. Blank analysis of IL-stir bar and PDMS ones were performed before sample analysis highlighted similar noise.
2.4. GC–Q-Orbitrap analyses
Analyses were performed using a GC–Q-Orbitrap system (Q Exactive, Thermo Scientific, Bremen, Germany) comprising a Gerstel MPS (multi-purpose sampler) (Mülheim, Germany) autosampler, a trace 1310 GC with a programmable temperature vaporizer (PTV) injector, an electron ionization (EI) source, and a hybrid Q-Orbitrap mass spectrometer. The PTV cool injection system (CIS6) was used in splitless mode. The desorbed compounds were focused on a PDMS packed liner at 4 °C in the PTV inlet and programmed from 4 °C (held for 1 min) to 280 °C (held for 15 min) at 4 °C/s. The carrier gas was helium (99.999 %, Air liquid, Paris, France) at a constant flow rate of 1.0 mL/min. GC separations were performed using an Elite 5MS (60 m × 250 µm × 0.25 µm film thickness) (PerkinElmer, Waltham, Massachusetts, USA) column, and for the environmental water samples, a HP-5 MS UI (30 m × 250 µm × 0.25 µm film thickness) (Agilent Technologies, Santa Clara, USA) was used with the following temperature program: 40 °C (held for 2 min) to 320 °C (held for 5 min) at 10 °C/min. EI was performed at 70 eV with the source temperature set at 200 °C. Full-scan MS acquisition was performed in profile mode using a m/z range of 50–500. Nitrogen gas (Air liquid, Paris, France) was used for the C-Trap supply. The internal mass calibration (lock mass mode) was performed during the analyses using the following ions from column bleeding (m/z) (C3H9Si+, 73.04680; C3H9O2Si2+, 133.01356; C5H15O3Si3+, 207.03235; C7H21O4Si4+, 281.05114; C9H27O5Si+, 355.06993).
The data were reprocessed using QualBrowser 4.1, TraceFinder 4.1 and Compound Discoverer 3.3 software (Thermo Fisher Scientific, Waltham, MA, USA). The last two software programs were able to perform deconvolution and comparison with mass spectra and retention indices (RIs) collected from the NIST database (2.2, 2014) and the Orbitrap spectral database (Thermofisher, 2015). For the RI calculation, a standard n-alkane mixture from C6 to C28 was injected under the same chromatographic conditions.
3. Results and discussion
3.1. Preparation of the IL coating
To obtain a stable and long-lasting (with high thermal stability) coating of the IL on stir bars, sol–gel technology was chosen. This procedure involves several reactions starting with the hydrolysis of the monomer precursor, followed by the polycondensation of the hydroxylated compounds to form a tridimensional sol‒gel network. Several conditions for stir bar coating were tested, but for clarity, only some representative tests are given in Table 1 and discussed. Three different ILs were selected and used for sol‒gel coating, one with a phosphonium cation ([P444TES] [Cl]) and two with an imidazolium cation with two different anions ([TESBPIM][Cl] and [TESBPIM][NTf2]) (Fig. 1). The physical characteristics of the obtained coatings were evaluated by visual assessment at different steps of the process. The coatings prepared using [TESBPIM][Cl] and [P444TES] [Cl] showed good adhesion on the stir bar; however, after the conditioning step (240 °C and hold for 3 h), poor mechanical stability was observed as the coatings crumbled (numbers 1 and 2 in Table 1). A modification of their color was also observed, as shown in Fig. 2-(1). With [TESBPIM][NTf2], the physical characteristics of the coating showed a higher mechanical stability. However, after the conditioning step, cracks were observed in the coating layer (number 3 in Table 1), and a change of the color was also observed, as shown in Fig. 2-(1). This phenomenon was attributed to thermal degradation; thus, the temperature and duration of the conditioning step were reduced to 200 °C for 2 h.
Finally, the use of [TESBPIM][NTf2] with the new conditioning temperature (200 °C for 2 h) provided good mechanical stability of the coating (numbers 6 and 7 in Table 1). Fig. 3(c) shows the stir bars before and after the IL coating. The images obtained using an optical microscopy revealed the presence of gas bubbles in the coating. However, the coated film remained uniform, and no cracking was observed. The film thickness was measured at two locations of the stir bar (Fig. 3(b)) and was between 350 µm and 360 µm, demonstrating a uniform coating on the stir bar. Compared with a commercial PDMS coating (the measured film thickness was calculated to be approximately 430 µm (Fig. 3(a)), the film thickness of the prepared IL coating is satisfactory. Thus, the IL [TESBPIM][NTf2] stir bar (7), which has the best stability, was used for the extraction test in our following tests.
Journal of Chromatography A, Volume 1742, 8 February 2025, 465623: Fig. 3. Microscopy images (× 250) of a commercial PDMS-coated stir bar (a) and the prepared stir bars (number 7) using the [TESBPIM][NTf2] IL (b), and a photography of the IL-coated stir bar and uncoated stir bar (c).
The Fig. 4 shows the DRIFT spectrum of the IL [TESBPIM][NTf2] prepared coating. Peaks in the range from 2800 to 3000 cm-1 are attributed to the aliphatic CH of the butyl chain attached to the imidazolium ring. Peaks at 2881 and 2970 cm-1 are attributed to the symmetric stretching vibration of CH3 and asymmetric stretching vibration of CH3 respectively. The features observed above 3000 cm-1 are from the CH vibrational modes of the imidazolium ring [43]. Between 3300 and 3800 cm-1 there are main bands that are due to the symmetric and asymmetric stretching vibrations of H2O molecules and these bands tend to overlap the lower frequency bands [44].
Journal of Chromatography A, Volume 1742, 8 February 2025, 465623: Fig. 4. DRIFT spectrum of the [TESBPIM][NTf2] ionic liquid.
3.3. Extraction and analysis of emerging and/or persistent pollutants from natural freshwater
To test the performance of the prepared IL-stir bar, SBSE was performed on natural freshwater samples obtained from the Robec River (France) to determine emerging and persistent pollutants via nontargeted screening using GC–Orbitrap-HRMS instrument in full-scan mode.
After extraction and analysis, the compounds were identified by using Compound Discoverer software. This software allows to isolate compounds from background noise and separates coelutions via a defined workflow. The obtained mass spectra were compared with those of the NIST and GC-HRMS libraries, and different scores were calculated, such as RSI (reverse search index), ΔRI (difference in retention indices) and RHRMF (reverse high-resolution mass filter). A scale of identification confidence proposed by G. Miller et al.[40]. was used to annotate the different compounds on the basis of the scores calculated via Compound Discoverer software. Five levels of confidence in coumpound annotation are defined: level 5: unknown feature; level 4: possible chemical class or series; level 3: tentative candidate (RSI>600 and RHRMF>75 %); level 2: probable structure or closer isomer (RSI>600, RHRMF>75 % and ΔRI<50 and 1.5 %); and level 1: confirmed identification (RT, EI spectra using an in-house library). In our study, only level 2 and 3 annotated compounds were selected because no standards were injected.
A data table combining all the detected compounds with all the attributed scores was obtained. Overall, after filtering all the silylated compounds generated by the column and the stir bar bleedings from a blank analysis, 1040 compounds were extracted from the chromatograms using the PDMS and the IL-stir bars. The number of detected peaks extracted using the IL-stir bar with added salt was comparable to those obtained with a PDMS stir bar with and without added NaCl. The results are represented as a Venn diagram in Fig. 6.
Journal of Chromatography A Volume 1742, 8 February 2025, 465623: Fig. 6: Venn diagram of the compounds extracted from the Robec River water using the IL-stir bar with NaCl addition and a commercial PDMS stir bar with and without NaCl addition.
4. Conclusion
In the present study, the development of a new ionic liquid-based coating for SBSE using the sol‒gel method has led to significant improvements in the extraction and analysis of a wide range of pollutants. The [TESBPIM][Cl] ionic liquid allowed the best coating to offer a homogenous, crack-free surface with high thermal stability. The extraction was highly enhanced by the addition of NaCl, particularly for compounds with log Ko/w > 4, outperforming the commercial PDMS stir bar. This result could be due to structural modification of the IL through anion exchange, leading to an IL with different properties than the initial one. The application of the coated stir bar in a nontarget SBSE-TD-GC–Orbitrap-HRMS method successfully detected over 1072 compounds in the Robec River, including emerging and persistent pollutants such as PCBs, pesticides, personal care products and pharmaceuticals. The specific detection of >160 compounds and the increase in peak areas highlight the potential of the new SBSE IL-stir bar coating for comprehensive environmental monitoring. The high extraction efficiency, particularly under saline conditions, makes this IL-coated stir bar a valuable tool for analyzing contaminants in complex aquatic environments, including estuary water or seawater.
