Natural deep eutectic solvent-based dispersive liquid-liquid microextraction of pesticides in drinking waters combined with GC-MS/MS detection

This study aims to develop an environmentally friendly analytical method for detecting organochlorine and organophosphorus pesticides in drinking water. It combines dispersive liquid-liquid microextraction (DLLME) using a natural deep eutectic solvent (NADES) with GC-MS/MS analysis. The NADES, composed of DL-menthol and decanoic acid, is safe, biodegradable, and cost-effective.

Key experimental parameters were optimized using multivariate analysis to enhance extraction efficiency. The method demonstrated high sensitivity with LOQs as low as 0.2 ng/L and good recovery rates across various spiking levels in real water samples, proving its suitability for routine pesticide monitoring.

The original article

Natural deep eutectic solvent-based dispersive liquid-liquid microextraction of pesticides in drinking waters combined with GC-MS/MS detection

Аsya Hristozova, Lorena Vidal, Miguel Ángel Aguirre, Kiril Simitchiev, Antonio Canals  

Talanta, Volume 282, 2025, 126967

https://doi.org/10.1016/j.talanta.2024.126967

licensed under CC-BY 4.0

Selected sections from the article follow. Formats and hyperlinks were adapted from the original.

Organochlorine pesticides (OCPs) are resistant to environmental degradation and accumulate in soils and sediments, exposing humans to pesticide effects through the food chains, drinking water, and breathing [27]. The Stockholm Convention (2001) classifies certain OCPs as persistent organic pollutants (POPs), prohibiting their production and use [28]. Organophosphorus pesticides (OPPs) are less harmful than OCPs but have decreased usage due to their persistence and neurotoxicity [29]. The negative effects of pesticides on human health can be prevented by monitoring of the food, water, and air. Hence the monitoring of low-level pesticide residues in bottled drinking water is a global concern [[30], [31], [32]]. However, the monitoring itself requires and generates a huge amount of hazardous organic wastes with obvious harmful effects on the environment and humans. So, efforts have been made to lower the limits of detection of the target analytes by employing environmentally benign sample preparation processes such as stir bar extraction [33] or dispersive liquid-liquid microextraction [23]. In this view the development of new green solvents which are aimed to reduce or fully eliminate the hazardous wastes can be regarded as very worthful.

Therefore, the current research aims to develop a new and sensitive analytical method using a combination of dispersive liquid-liquid microextraction (DLLME) assisted by vortex and using an environmentally friendly extractant (i.e., NADES: DL-menthol and decanoic acid in molar ratio 2:1) for the preconcentration of organochlorine and organophosphorus pesticides followed by gas chromatography-tandem mass spectrometry (GC-MS/MS).

2. Materials and methods

2.2. Instrumentation, apparatus and software

Experiments were performed using a GC-MS/MS TSQ 9000 (Thermo Fisher Scientific, USA) with electron ionisation (EI) at 70 eV, equipped with a triple quadrupole mass filter and Programmable Temperature Vaporising (PTV) injector. The system was operated by Excalibur software 4.1. Extract with volume of 1 μL was injected manually with a 10 μL glass syringe. The study was conducted utilizing a PTV glass liner with Three Baffles, 1 mm ID, 2.75 mm OD, 120 mm Length (Thermo Fisher Scientific, USA). The PTV injector was used in split mode (split ratio 50:1), starting with 65 °C initial inlet temperature followed by gradient heating at 14.5 °C sec⁻¹ to 290 °C. GC column TG-SQC 15 m × 0.25 mm, 0.25 μm film thickness (Thermo Fisher Scientific, USA) was used for chromatographic separation. The oven temperature program was as follows: 120 °C – held for 1 min; increased at the rate of 40 °C min⁻¹ to 155 °C, rating 4 °C min⁻¹ to 187 °C, rating 1 °C min⁻¹ to 194 °C and rating 25 °C min⁻¹ to a final temperature of 290 °C and held for 10 min. The solvent delay time and total analysis time were 2.5 min and 31 min, respectively. Helium (purity 99.9999 %) at a flow rate of 1.2 mL min⁻¹ was used as a carrier gas. The transfer line and the ion source temperatures were set at 250 °C and 230 °C, respectively.

A Vortex Lab Dancer (IKA®, Germany), a centrifuge KA-1000 (Jiangsu Zhengji Instruments Co., Ltd., China), FT-IR VERTEX 70 Spectrometer (Bruker, Germany) and a pH meter with a combined glass electrode SensoDirect pH 110 (Lovibond® Water Testing, Germany) were used in the study.

The software Statistica® version 14.1 (TIBCO, Santa Clara, USA) was used in the optimisation procedure and graphical visualization of the results. AGREEprep software was applied for the evaluation of the greenness metric of the developed method.

3. Results and discussion

3.1. Assessment of the matrix effects in GC-MS/MS

To the best of our knowledge, the matrix effects provoked by NADES (MNT:DA) as a matrix medium analysed by GC-MS/MS have not been previously studied. The influence of MNT:DA on the instrumental sensitivity was evaluated by measurements of two sets of 3 standard solutions (with concentrations up to 2000 μg L⁻¹ in the final solution) prepared in acetonitrile as well as in NADES (matrix-matched medium). The slopes of the calibration curves were compared according to Equation (1). The standard deviations of the slopes were also assessed, and the propagated standard deviation of the corresponding slope ratios was further calculated.

Table 1 shows the derived slope ratio values, as well as the combined uncertainty. The chromatogram of the target analytes is presented on Fig. 2.

Talanta, Volume 282, 2025, 126967: Fig. 2. Reconstituted ion current chromatogram (RTIC) of each target analyte.

The sensitivity for all target analytes increased in the tested medium of MNT:DA compared to acetonitrile. The enhancement was in the range of 27–74 % (Table 1), indicating a favourable effect of the MNT:DA medium during the GC-MS/MS analysis. As a general trend, it can be seen that the longer the retention time of the target analyte, the higher sensitivity is achieved using NADES as a matrix solution. A possible explanation for this “matrix-induced response enchantment effect” could be a reduction of the thermal stress when the target analytes are passing through the GC liner [35,36]. Such favourable matrix impact was also seen when employing several types of liners for pesticide detection using a combination of cloud point extraction with Triton X-100 and GC-MS or GC-MS/MS [37].

3.2. Optimisation of dispersive liquid-liquid microextraction

Considering the simultaneous determination of 19 target analytes (organochlorine and organophosphorus pesticides) and the number of variables influencing the extraction step a multivariate study [38] was used for extraction variables optimisation. This optimisation approach follows the principles of Green Analytical Chemistry [39] to reduce significantly the number of experiments, consumption of samples, reagents, energy and time. An approach at two steps has been used for the multivariate optimisation of the experimental factors: a Plackett-Burman design, to identify between significant and non-significant factors as a screening method followed by a central composite design (CCD) to obtain the optimal values for the significant factors. Seven factors with potential effect on DLLME were evaluated by the Plackett-Burman design: (i) sample volume (6 and 10 mL); (ii) NADES volume (50 and 100 μL), (iii) sample pH (5 and 7); (iv) extraction time (1 and 3 min); (v) centrifugation speed (900 and 1300×g); (vi) centrifugation time (5 and 10 min); and (vii) ionic strength, NaCl concentration (0 and 3 %, w v⁻¹). Table S4 shows the constructed matrix of factors and their values used at each experimental run.

The results derived from the Plackett-Burman design for all studied analytes are presented as registered responses (i.e., peak areas) in Table S5 and as Pareto charts of the standardised effects of the factors on Figure S1. The factors showing a significant effect on the registered peak areas of all the analytes were the NADES and the sample volumes. The other five factors have no influence on the DLLME of the studied pesticides. Taking into account the effects from the Pareto charts the following values were set for further experiments: addition of NaCl is not necessary (negative effect), centrifugation speed at 900×g (negative effect) and extraction time 3 min (positive effect). Due to practical reasons, the values of the non-significant factors centrifugation time and pH are set to 5 min (due to the temperature increase with prolonged centrifugation time) and 7 for model solutions or native pH of the analysed drinking water, respectively.

Further optimisation of the NADES and sample volumes was performed by applying a central composite design (CCD). The factor values incorporated in the CCD were: (i) sample volume (5.2, 6.0, 8.0, 10.0, and 10.8 mL), and (ii) NADES volume (40, 50, 75, 100, and 110 μL). The constructed CCD matrix is given in Table S6 and the responses (i.e., peak areas) are shown in Table S7. The response surface diagrams obtained from CCD for alpha-HCH, Hexachlorobenzene, Chlorpyrifos, Aldrin, Heptachlor and p,p-DDE are shown in Fig. S2. It was found that the modelled functions for all studied pesticides have similar shapes – the registered peak area increases when the NADES volume is reduced and/or the sample volume is enlarged.

Aiming group extraction of the pesticides further a desirability function was calculated according to Equation (2) and response surface optimisation was applied (Fig. 3).

Talanta, Volume 282, 2025, 126967: Fig. 3. Desirability function response surface diagrams.

4. Conclusions

The NADES-based-DLLME methodology was successfully combined with GC-MS/MS and was applied for the analysis of organochlorine and organophosphorus pesticides in bottled drinking water by direct liquid injections of the NADES extractant phase into the gas chromatographic system. The last does not deteriorate the instrumental measurements in SRM mode allowing selective determination of the target analytes. The proposed analytical method is fast and simple and meets several conceptual green principles such as the miniaturisation of sample preparation, the use of a natural solvent, and the simultaneous quantification of 19 analytes, among others. Besides, the optimisation of the method was carried out by multivariate analysis, minimizing the consumption of sample, reagents and time. The figures of merit of the developed method were evaluated. It is important to highlight that the obtained absolute recoveries for most of the studied pesticides are close to 100 %; the gain in the value of the registered signals for the target analytes in the NADES phase is larger than the volume ratio of the initial and final solution due to positive matrix effect of the DES on the GC-MS/MS system. Finally, the developed method has been successfully applied to analyse drinking water samples, showing just a partial recovery for one out of nineteen studied pesticides (i.e., p,p-DDT).