Ultrasound Assisted-Homogeneous Liquid-Liquid Phase Microextraction based on Deep Eutectic Solvents and Ethyl Acetate for Preconcentration of Selected Organochlorine Pesticides in Water Samples

A novel and greener methodology for the simultaneous preconcentration of organochlorine pesticides in water samples based on ultrasound assistedhomogeneous liquid-liquid phase microextraction (UA-HLLME) has been developed. Gas chromatography-time of flight mass spectrometry was used for quantification of OCPS in water samples. In this method, choline-chloride-ethylene glycol deep eutectic solvent and ethyl acetate were used as the disperser solvent and extraction solvent, respectively. Univariate and multivariate approaches were used for optimization of the influential parameters that affect the extraction efficiency of the UA-HLLME method. Under the optimum conditions, enrichment factors ranging from 152 to 403 with acceptable recoveries of 85-100% were obtained. The dynamic linear ranges were obtained in the concentration range 0.015–1000 μg L−1 with correlation coefficients ranging 0.9952–0.9995. The limits of detection and quantification of the developed UAHLLME method were in the range 1.9-8.6 ng L−1 and 5.9-26 ng L−1, respectively. The intra-day and inter-day precision expressed in terms of relative standard deviation (%RSD) ranged from 2.1-4.5% and 3.9-7.3%, respectively. The developed method was successfully applied for the preconcentration and determination of the selected OCPs from 3 different river water samples. The developed procedure displayed simplicity, environmental friendliness, relatively high extraction efficiency, short analysis time and relatively low detection limits.


INTRODUCTION
Pesticides have remained in use for several years to control agricultural pests.Hence, their accumulation in the environment is due to their persistence properties [1].For this reason, high levels of pesticides are still being discovered in numerous environmental media, such as soil and groundwater [2].Due to unpleasant effects of pesticides on human and environmental health, it is important to identify and quantifying these contaminants in soil and groundwater matrices [1].The concentrations of these pollutants are usually at trace levels (especially in groundwater) or are trapped within complex matrices (such as soil), which then poses problems of detection and quantification.For this reason, sensitive, reliable and rapid techniques for accurate determination of pesticides in environmental matrices are required.
Chromatographic techniques such as liquid or gas chromatography coupled to different detectors are widely used for the separation and quantification of organochlorine pesticides in different matrices.However, direct analysis of pesticides using these techniques is not suitable.Therefore, sample preparation procedure prior to

Matong et al. / Preconcentration of Selected Organochlorine Pesticides
2 / 15 chromatographic quantification of pesticides is required.This is done in order to decrease the complexity of the matrix and increase the concentration of pesticides.Conventional sample preparation technique such as liquidliquid extraction (LLE) and solid phase extraction are the long-standing preconcentration and matrix separation method in analytical chemistry [3].However, their disadvantages include time-consuming, labor-intensive and need a large amount of toxic organic solvents (especially LLE) [4].Although SPE uses smaller volumes of potentially toxic solvents as compared to LLE, a significant amount of organic solvents, disposable cartridges, and discs with a special manifold are still required [5].Recently, liquid/solid phase microextraction techniques which are considered to be relatively green [5] have been developed for extraction of pesticides in different matrices.These include dispersive liquid-liquid microextraction (DLLME) [6,7], single-drop microextraction (SDME) [2] and hollow fiber liquid/solid-phase microextraction (HF-L/SPME) [8,9], among others.
In last decade, a novel sample pretreatment technique known as homogeneous liquid-liquid extraction (HLLE) was developed by Tavakoli et al. [10].The principle of this technique is similar to other liquid phase based techniques.The HLLE utilizes both the low and high-density solvents.Recently, the same research group has developed miniaturized homogeneous liquid-liquid extraction (MHLLE) and miniaturized counter current liquidliquid extraction using methanol (co-solvent) and low density solvents (extraction solvent) for determination of pesticides in soil samples [11,12].These methods combine high throughput analysis, low cost, and environmental sustainability, which is of great importance in analytical chemistry [12].In MHLLE the choice of suitable disperser or co-solvent is critical.This is to avoid the use of high disperser solvent which may lead to decreased partition coefficients of the analytes between sample and extraction solvent.Therefore, this can be avoided by using environmentally friendly solvents that have similar properties as frequently used solvents such as methanol.Deep eutectic solvents (DESs) have attracted the interest of many researchers as a green alternative solvent type in sample preparation applications.These solvents have attractive properties such as availability of materials at low cost, the ease of synthesis and low toxicity [14][15][16].They are obtained by mixing environmentally friendly (or naturally occurring) components that are compatible with each other mostly through hydrogen bonding [14][15][16].
The aim of this study was to develop and validate an analytical method based on ultrasound assistedhomogeneous liquid-liquid phase microextraction (UA-HLLME) for preconcentration of organochlorine pesticides in river water samples.The applicability of choline chloride-based deep eutectic solvents (DESs) and aprotic solvents was evaluated.The DESs were chosen because of their environmental friendliness.The DES acted only as a dispersion solvent and it assisted the dispersion of extraction solvent (ethyl acetate) within the aqueous solution.It should be noted that after centrifugation step, organic phase (ethyl acetate) was separated while DES remained in the aqueous solution due to its hydrophilic nature.Based on recent literature findings there are only a few publications on using DESs for extraction of organic pollutants [13][14][15].Low density solvents were selected because they can be withdrawn directly from the extraction using suitable micro-syringe.The most influential factors affecting the extraction efficiency of analytes were studied using the small central composite design.To the best of our knowledge, there are limited studies or no work at all, on the use of DESs together with ethyl acetate, for the preconcentration of pesticides in water samples.

Sample Collection and Preparation
River water samples were in Johannesburg South (South Africa) with the geographic coordinates (26°15′58″S 27°51′57″E) respectively.River water was taken from three different streams where there is domestic farming takes place and stored in dark glass containers and taken to the lab for pesticide analysis.It should be noted that the sample analysis was carried out on the same day.Before use, the samples were filtered through 0.45 μm cellulose acetate filters (Millipore HNWP, Bedford, MA, USA) in order to remove suspended particles.

Synthesis of DES
The synthesis of DES was carried out according to the procedure reported in the literature (16,17].Choline chloride-ethylene glycol, choline-oxalic acid, and choline-citric acid mixtures at appropriate molar ratios were synthesized and evaluated for extraction of OCPs in water samples.It should be noted that oxalic acid, citric acid, and ethylene glycol were used as a hydrogen bond donors.To describe the procedure briefly, appropriate amounts (in grams) of choline chloride and oxalic acid or citric acid or ethylene glycol were placed in a 50 mL round-bottom flask.The flask was heated on a temperature controlled hot plate stirrer at 80 °C.The mixture of the 2 components was continuously stirred using a magnetic stirrer bar until a homogeneous, colorless liquid formed (about 5-10 min).

Ultrasound Assisted-Homogeneous Liquid-Liquid Microextraction Procedure
The UA-HLLME was carried out as follows; 15 mL of an aqueous sample solution containing target analytes (50 µg L −1 ) was placed in a clean dry 25 mL glass sample tube (220 mm (H) × 17 mm (diameter)).An aliquot of 100-500 µL of DES as disperser solvent was added to the aqueous sample solution and a uniform solution was produced.Then, 20-100 µL of a low density solvent was introduced into the uniform solution leading to clustering of DES molecules and as a result, a cloudy solution was produced.So for the solution to entirely disperse the clustered DES droplets into the aqueous phase, the cloudy solution was sonicated for 5-20 min in an ultrasonic bath.At this step, the clustered DES droplets gradually fragmented into small droplets because of the ultrasonic wave involved by short-lived cavitation close to the interface of DES droplets [18].After centrifugation for 1-10 min at 4400 rpm, two clear phases were observed namely aqueous and organic phases.Then the water immiscible phase was extracted through a micro-syringe and transferred into clean screw-topped centrifuge tube.Furthermore, the acceptor phase (low density organic solvent phase) was diluted to the final volume of 0.2 mL (200 µL) hexane.Then, about 1 µL of the eluent was injected into the GC-TOFMS system for analysis.

RESULTS AND DISCUSSION
In UA-HLLME, the extraction efficiency is normally influenced by various parameters.These include including the type and the volume of disperser solvent and extraction solvent, extraction time, ionic strength (salt effect), extraction and centrifugation time [19].The effects of type of extraction solvent and disperser solvent on the extraction of OCPs were optimized using univariate (one-parameter-at-a-time) approach.Other factors were optimized using multivariate approach.The results obtained were expressed in terms of extraction efficiency or extraction recovery which was defined as the percentage of the moles of an analyte (n i ) that is transferred into the final organic phase (n f ).The percentage extraction efficiency was calculated according to the expression below.

Extraction efficiency (%EE) =
×     ×   � × 100, where C f is the amount of the analyte in the acceptor (organic phase), and can be obtained from experimental measurements; C i is the initial concentration.

Selection of disperser solvent
For the UA-HLLME method, the disperser solvent should be soluble with the acceptor phase (organic phase) as well as the donor phase (model and real samples).Therefore, suitability of choline chloride-based DES to be used as disperser solvent was investigated.The DESs investigated include choline chloride-ethylene glycol, cholineoxalic acid, and choline-citric acid.The preconcentration studies were carried out using 15 mL model aqueous samples containing 50 µg L −1 OCPs.Ethyl acetate was used as low density solvent.It can be seen from Figure 1 that higher extraction efficiencies for the studied OCPs were achieved when choline chloride-ethylene glycol was used as the disperser solvent compared to other DESs.This is because clear phase separation was observed when choline chloride-ethylene glycol is used as the disperser solvent.It is worth mentioning that when choline-oxalic acid and choline-citric acid were used, the water immiscibility layer was less distinct compared to choline chloride-ethylene glycol.This might be due to the presence of carboxylic groups which solubilized further the extraction solvent.Therefore, choline chloride-ethylene glycol was chosen a disperser solvent for further experiments.
In view of the results above, the effect of various molar ratios of choline chloride-ethylene glycol (DES1 1:1, DES2 1:2, DES 3 1:3, and DES 1:4) was investigated.The results obtained (Figure S1) revealed that quantitative extraction of OCPs was attained when the molar ratio was 1 choline: 1 ethylene glycol.Therefore, the 1 choline chloride: 1 ethylene glycol was selected for further investigations.The 1 choline chloride: 1 ethylene glycol resulted in a free flowing solvent which other were thicker (low viscosity).

Selection of extraction solvent
The type of extraction solvent is one of the very important factors in obtaining satisfactory extraction efficiency.This is because the physical-chemical properties of the solvents determine its extraction efficiency [20].In UA-HLLME, the extraction solvent should have a lower density than water, low water solubility and, compatible with analytical detection techniques, environmentally friendly and capable of extracting target analytes [11,[19][20][21][22]. Therefore, in this study, extraction efficiencies of THF and ethyl acetate for extraction of OCPs in river water were investigated using the UA-HLLME method.The study was carried out using a model sample solution of the analytes at a concentration level of 50 µg L −1 .Other parameters were fixed at 250 µL, 100 µL, 20 min, 10 min, 3% for DES volume, extraction solvent volume, extraction time, centrifugation time and ionic strength, respectively.The results obtained are presented in Figure 2. It can be seen from these results that relatively higher extraction efficiencies were achieved when using ethyl acetate compared to THF.In addition, only nine pesticides were extracted by THF.The findings were attributed to difficulties encountered in order to separate THF from the bulk solution.Ethyl acetate could be separated rapidly and completely and this was in accordance with other studies [11,22].Therefore, ethyl acetate was selected as an extraction solvent.

Optimization of the UA-HLLME Operation Parameters: Multivariate Approach
In order to obtain highest extraction efficiency, multivariate optimization of the UA-HLLME method was carried out using small central composite design (SCCD).In the latter, five variables, that is, the volume of extraction and disperser/emulsifier solvents, extraction time, ionic strength and were selected.STATISTICA software was used to estimate the optimum conditions for the performance of the UA-HLLME method.The low and high levels selected for the variables were as follows: DES volume (DESV) (100-500), ethyl acetate volume (EAV) (20-100), ionic strength (IS) (0-10%), extraction time (ET) (5-20 min), and centrifugation time (CT) (1-10 min).Typical design matrix and the response (extraction efficiency, %) are presented in Table S1.The analysis of variance (ANOVA) presented in terms of Pareto plot is shown in Figure S2.The ANOVA results revealed that all five factors were not significant at 95% confidence level.The response surfaces for factors against the analytical response and quadratic equations were used to estimate the optimum condition.Based to on the results obtained, the optimum conditions were; 500 µL of DES as a disperser solvent; 0% (w/v) salt concentration, 100 µL of ethyl acetate as the extraction solvent, 10 minute extraction time and 2 minutes centrifugation time.The performance of the developed method under optimum conditions was evaluated by analyzing triplicates of the model solution.This was done in order to validate the suitability of the estimated optimum conditions.The results obtained ranged from 95-99% and these results demonstrated that the selected conditions were suitable for preconcentration OCPs.

Analytical Figures of Merit
Under the optimum experimental conditions, the analytical performances of the developed UA-HLLME method preconcentration of OCPs were investigated.The calibration curves were obtained after a set of standard solutions (0 to 1000 µg L −1 ) was processed using the UA-HLLME method.The concentrations of the analytes in the eluent solutions (after extraction) were quantified by GC-TOFMS.The limits of detection and quantification were calculated using: LOD = 3×  and LOQ= 10×  , respectively, where Sd is the standard deviation of 12 replicate measurements at lower concentrations of calibration curves and b is the slope of each calibration curves [15].Dynamic linear ranges, correlation coefficients, LODs, LOQs and enrichment factors (defined as the as the ratio between the concentration of the analyte in the final phase (C f ) and the initial concentration of the analyte (C i ) ) [11] are presented in Table 1.The intraday (repeatability, n=15) and interday (reproducibility, n =7) precisions of the developed UA-HLLME method, expressed as the relative standard deviation (%RSD), was determined by carrying out at the 50 µg L − 1 concentration level and the results are presented in Table 1.
The accuracy of the developed method was validated by spiking double distilled deionized water with two concentration levels (1 and 5 µg L −1 ).The obtained results of recoveries are given in Table 2.According to the analytical results obtained, the percentage recoveries for 1.0 µg L −1 ranged from 95.0-99% and for 5.0 µg L −1 were between 96% and 99.8%.This demonstrated that developed UA-HLLME method had a potential of extracting and preconcentrating trace OCPs from real samples.

Comparison of the Proposed Methods with the Other Sample Preparation Techniques
A comparison of the developed method in terms of DLRs, LODs, EF, and RSDs with selected sample preparation reported in the literature for preconcentration of OCPs in different matrices is summarized in Table 3.It can be that LODs and DLRs of the UA-HLLME method were better than or comparable with other reported methods except for SPE-DLLME-GC-MS.The EFs were higher than those reported by Refs [12,23] and comparable those reported by [24].However, the EFs were lower than those reported by Refs [25][26][27][28].Eurasian J Anal Chem

Application to Real Water Samples
Real river water samples were analyzed by applying the developed UA-HLLME method.The target analytes were not detected in the river water samples (S1-3).These findings revealed that the OCPs were either not present in river water samples or they were below the detection limits of the developed technique.However, the real samples were then used to investigate the effect of the sample matrix.River water samples S1, S2 and S3 were spiked with the CRM at concentration levels of 1.0, 2.0 and 5 µg L −1 , respectively for method validation and the results are summarized in Table 4 and typical chromatograms are shown in Figures S4 and S5.It can be seen that the recoveries for the all the selected analytes were between 85.7% and 95 %.When compared to the results in Table 4, these findings suggest that real river water sample matrices have minor to medium interference effects on UA-HLLME.However, the overall results revealed that the developed method is applicable for extraction of trace pesticide residues in water samples.The results obtained in our study, are comparable with other studies reported by [19] in a similar analysis of water for this class of compounds.

CONCLUSION
A simple, rapid, environmentally friendly, inexpensive and sensitive analytical procedure based on UA-HLLME/GC-MS was developed for the quantification of organochlorine pesticides in river water sample.Factors affecting the developed method were optimized using univariate and multivariate approaches.The UA-HLLME method displayed relatively wide dynamic linear ranges, low LOD (ng L -1 ) and high preconcentration factors.In addition, UA-HLLME showed relatively good accuracy (in terms of recoveries) and precision (expressed in terms of %RSD).Moreover, developed method was successfully applied for the extraction and preconcentration of the selected OCPs in real water samples.

/ 15
Ultrasound assisted-homogeneous liquid-liquid phase microextraction based on deep eutectic solvents and ethyl acetate for preconcentration of selected organochlorine pesticides in water samples SUPPLEMENTARY DATA Matong et al. / Preconcentration of Selected Organochlorine Pesticides 4 / 15 Effect of Type of Disperser Solvent and Extraction Solvents: Univariate Approach

Figure S3 .
Figure S3.Typical response surface curves for optimization of ultrasound assisted-homogeneous liquid-liquid phase microextraction of pesticides

Table 1 .
Analytical figures of merit

Table 2 .
Percentage recoveries of OCPs from spiked double distilled deionized water

Table 3 .
Comparison of present method with reported liquid-liquid microextraction based methods for the preconcentration of pesticides in different matrices

Table 4 .
Analytical results from analysis of OCPs in spiked real river samples using UA-HLLME