Dispersive graphene-based silica coated magnetic nanoparticles as a new adsorbent for preconcentration of chlorinated pesticides from environmental water

Abstract

The present study describes the synthesis, characterization and application of new graphene-based silica coated magnetic nanoparticles (Fe₃O₄@SiO₂–G) for the simultaneous preconcentration of four chlorinated pesticides namely lindane, chlorpyrifos, hexaconazole and azaconazole from contaminated water. The newly synthesised adsorbent was characterized using FT-IR spectroscopy, field emission scanning electron microscopy (FESEM), energy dispersive spectroscopy (EDS) and X-ray diffraction (XRD). Chlorinated pesticide extraction efficiency of the Fe₃O₄@SiO₂–G was evaluated through magnetic solid phase extraction (MSPE) using gas chromatography with a micro electron capture detector (GC-μECD). Experimental parameters, i.e., desorption solvent, solvent volume, extraction time, desorption time, sample volume, adsorbent dosage and solution pH were optimized. Compared to commercial C18 sorbent and Fe₃O₄@SiO₂, the newly synthesized Fe₃O₄@SiO₂–G adsorbent showed a linear response (1–100 pg mL⁻¹), low limits of detection (0.12–0.28 pg mL⁻¹) and high adsorption capacity (13.04–18.69 mg g⁻¹) with a coefficient of determination (R²) of 0.999. Environmental water samples were used to assess the field applicability of the adsorbents. Excellent percent recovery (80.8–106.3%) was achieved for Fe₃O₄@SiO₂–G at pH 6.5. The results showed that the newly synthesized Fe₃O₄@SiO₂–G is an efficient adsorbent with good potential for the preconcentration of selected chlorinated pesticides from aqueous media.

---

1. Introduction

From the original discovery of pesticides to this day, a remarkable increase in the usage of pesticides has been observed. Recent literature surveys indicated that per annum worldwide consumption of pesticides has exceeded 2.5 million pounds.¹ The approximately 28% contribution of chlorinated pesticide to the total pesticide consumption makes them the largest class of pesticides. Generally chlorinated pesticides are used as triazine herbicides, fungicides and insecticides for crop protection.² The massive use of chlorinated pesticides is one of the most imperative issues of the modern era because due to their widespread usage these pesticides can be easily mixed with natural water.^1,3 Particularly water contamination resulting from chlorinated pesticides is becoming a hot issue for environmental researchers. From an environmental point of view the mixing of chlorinated pesticides as well as their degraded products into natural water is of great concern.⁴ Pesticides can be easily transferred to drinking water and become harmful to human health.⁵ The toxicity of chlorinated pesticides has been comprehensively reviewed; a literature survey showed that constant exposure to chlorinated pesticides can cause several life-threatening ailments, such as cancer, disorders of the reproductive and endocrine system.^6–8 The European Union has declared that the safe concentration levels for pesticides is 0.1 ng mL⁻¹, while a very high concentration of pesticides (i.e., 0.43 mg L⁻¹) has been reported.⁹ Consequently, due the exceeded limit as well as the well-known toxicity of chlorinated pesticides and their degraded products, precise determination and extraction is extremely necessary.¹⁰

Several techniques, including liquid–liquid extraction (LLE), liquid phase microextraction (LPME), stir bar sorptive extraction (SBSE), solid phase micro extraction (SPME), dispersive solid phase extraction (DSPE) and solid phase extraction (SPE), have been applied to remediate the pesticides and their degraded products from the aqueous environment.^11,12 Comparatively, SPE due to its simplicity, significant recovery, short extraction time, high enrichment factor and low cost is renowned as an advantageous and authentic method.¹³ A variety of materials such as electrospun modified silica,¹⁴ inorganic oxide nanoparticles/polyethylene,¹⁵ C18,¹⁶ TEOS-MTMOS,¹² multi wall carbon nanotubes,¹⁷ chitosan/polypyrrole¹⁸ and graphene have been used as extraction adsorbents.¹⁹ But due to the low adsorption efficiency and reusability, the use of conventional materials is limited. Consequently, exploitation of innovative highly selective pesticide sorbents which can be easily regenerated as well as effective use in aqueous media has currently become a focus of intensive research.^20–22 In this regard graphene due to its honeycomb-like structure deserves particular attention. In particular, the high surface area, significant adsorption capacity, variety of benzene rings and especially rich π–π electron arrangement makes graphene a suitable extractant for the extraction of benzene based pesticides.^23,24

A literature survey reveals that graphene-based materials are mostly used for pesticide decontamination.^25–27 However, graphene-based extraction reagents suffer from some drawbacks such as hydrophobicity as well as dispersive nature in aqueous media which makes the extraction process tedious and time consuming. These drawbacks can be abridged through the functionalization of graphene with appropriate molecular frameworks. In this respect, magnetic Fe₃O₄ nanoparticles have provided important avenues to prepare new, stable and efficient extraction reagents^28,29 as well as green sample preparation.³⁰ SiO₂ is cheap, environmentally friendly, chemically stable and highly dispersive in liquid due to it rich O–H groups. So, generally Fe₃O₄ nanoparticles are coated with different silane derivatives via the sol–gel method to increase the surface area porosity and effective binding sites.^31,32 Thus, in this study we report the synthesis, characterization and application of graphene-based silica-coated magnetic nanoparticles (Fe₃O₄@SiO₂–G) for the preconcentration of four chlorinated pesticides namely lindane, chlorpyrifos, hexaconazole and azaconazole from contaminated water. The chlorinated pesticide concentration was determined by GC-μECD.

2. Experimental

2.1 Materials

Lindane, chlorpyrifos, hexaconazole, azaconazole and propazine (internal standard) were from Riedel-de Haen (Hanover, Germany). Analytical grade reagents/chemicals were used. Graphite powder and tetraethoxysilane (TEOS) were purchased from Sigma Aldrich (St. Louis, MO, USA). Sulphuric acid (98%), nitric acid (65%), hydrochloric acid (37%), ethanol (97%) and HPLC grade methanol were from QReC (Selangor, Malaysia).

2.2 Instruments

FT-IR spectra (4000–400 cm⁻¹) were recorded on a 1600 series Perkin-Elmer FTIR Spectrometer (MA, USA) as KBr pellets. A D8-Advance X-ray diffractometer from Bruker (GmbH. Karlsruhe, Germany) was used for crystal analysis with Cu Kα radiation (λ = 1.54060 Å), high quality mode and 40 kV voltage. The size, morphology, structure and composition of the new magnetic materials were observed using a Carl Zeiss Model Supra 35-VP FESEM (Oberkochen, Germany) operated at 10.0 kV, magnification 50 000× and a working distance (WD) of 6.0 mm. An Agilent A7600 gas chromatograph equipped with a micro electron capture detector (GC-μECD) (Santa Clara, CA, USA) was used for the selected chlorinated pesticide analysis.

2.3 Gas chromatography conditions

GC-μECD with an HP-5MS column (5% phenyl methyl siloxane 325 °C: 30 m × 250 μm × 0.25 μm) was used for the identification of the selected chlorinated pesticides in the water samples. The optimum GC-μECD conditions are as follows: helium was used as back inlet carrier gas at a flow rate of 2.5 mL min⁻¹ and nitrogen was used a back detector make up gas at a flow rate of 5 mL min⁻¹. The back inlet port and detector temperatures were set at 280 °C (pressure set at 25 psi) and 300 °C, respectively. The gas chromatography temperature profile was set at 70–280 °C, starting at 70 °C (held for 1 min) ramp at 50 °C min⁻¹ to 190 °C (held for 1 min) and ramp at 30 °C min⁻¹ to 280 °C (held for 2 min). 1 μL of the extracted analyte in methanol was injected manually into the injection port under splitless mode. Triplicate injections were performed for each extract and the average peak area ratio reading was taken.

2.4 Preparation of Fe₃O₄ nanoparticles

Fe₃O₄ magnetic nanoparticles (MNPs) (Fig. 1A) were prepared via a modification of a previously reported procedure.³³ 0.5 g of (NH₄)₂Fe(SO₄)₂·6H₂O and 1 g FeCl₃·6H₂O were dissolved in 30 mL of deionized water and stirred vigorously for 5 min at 50 °C. Following the drop wise addition of 5 mL of ammonia solution (32%), the reaction mixture was further stirred at room temperature for 5 h. The resulting black Fe₃O₄ nanoparticles were collected using an external magnet and washed with excess deionized water and dried under vacuum at 80 °C for 24 h.

Fig. 1 Schematic for the synthesis of (A) magnetic Fe₃O₄ nanoparticles and (B) silica-coated magnetic nanoparticles.

2.5 Synthesis of Fe₃O₄@SiO₂

Tetraethyl orthosilicate (TEOS) coated magnetic nanoparticles (Fig. 1B) were prepared by a modification of the Stöber method.³⁴ SiO₂ coating of the Fe₃O₄ MNPs has been shown to enhance the lifetime as well as the binding sites of the MNPs.^31,32 0.5 g of the freshly prepared Fe₃O₄ were dispersed in 100 mL of water–ethanol solution (1 : 1) then 0.5 mL TEOS was added into the solution. Following the addition of 5 mL ammonia solution (25%) the reaction mixture was stirred for 1 h and then left at room temperature for 20 h. Finally, the resultant slightly darker brown Fe₃O₄@SiO₂ nanoparticles were collected and washed with deionized water and ethanol. The Fe₃O₄@SiO₂ was then dried under vacuum at 80 °C for 20 h.

2.6 Preparation of graphene oxide

2 g of natural graphite was soaked in 10 mL of concentrated HCl for 48 h. The mixture was then filtered and the filtrate was washed with excess of deionized water and dried under vacuum at 80 °C for 24 h. The dried natural graphite was suspended in a mixture of 120 mL of concentrated H₂SO₄–HNO₃ (2 : 1) and stirred at room temperature for 1 h. Potassium permanganate (3 g) was then added gradually into the reaction mixture and was further stirred at room temperature for 24 h. Finally, yellow-coloured graphene oxide (GO) was obtained by adding ice (500 g) and H₂O₂ (3 mL). The resultant product was washed with excess deionized water and then dried at 80 °C for 24 h under vacuum.

2.7 Synthesis of Fe₃O₄@SiO₂–graphene

Fe₃O₄@SiO₂–graphene MNPs (Fig. 2) were prepared as follows; 10 mg of freshly prepared Fe₃O₄@SiO₂ were dispersed into 50 mL of graphene oxide (GO) solution (10 mg mL⁻¹) followed by the addition of 2 mL of ammonia solution (32%) and 5 mL of hydrazine hydrate. The reaction mixture was stirred at room temperature for 48 h. The resulting Fe₃O₄@SiO₂–graphene MNPs were separated by an external magnet and washed with excess deionized water; then dried at 80 °C for 24 h.

Fig. 2 Schematic diagram of the preparation of graphene-based magnetic nanoparticles.

2.8 Water sample preparation

Four environmental water samples namely tap, river, lake and sea were used to assess the field application of the synthesized Fe₃O₄@SiO₂–graphene. Tap water was obtained from the laboratory, while river and lake water samples were taken from the Melana River and UTM Lake, Skudai, respectively. The sea water sample was taken from Danga Bay, Johor Bahru. Tap water was used directly without further treatment. River, lake and sea water samples were filtered using Whatman cellulose filter paper (125 mm diameter, 11 μm pore size and 180 μm thickness) for the removal of debris.

2.9 Extraction procedures for selected chlorinated pesticides

2.9.1 Optimization of MSPE parameters. Different parameters i.e., types of desorption solvent, solvent volume, adsorption time, desorption time, sample volume, adsorbent dosage and solution pH were optimized during the MSPE. Initially, 10 mL of sample volume, 10 mg of adsorbent and 20 min extraction time were used for extraction of the selected chlorinated pesticides. The chlorinated pesticides were eluted using 1 mL of different types of solvents.

2.9.2 Magnetic solid phase extraction. Fig. 3 shows the schematic diagram of the MSPE process for the extraction of selected chlorinated pesticides using the synthesized Fe₃O₄@SiO₂–G MNPs. The experiments were performed in Erlenmeyer flasks (100 mL) which contain 60 mg of the adsorbent and 50 mL of 1 ng mL⁻¹ solution of each chlorinated pesticide or adsorbed solution. For equilibrium studies the flasks were shaken on an orbital shaker at a constant speed of 250 rpm at 25 °C for 5 min. The adsorbent was then separated using an external magnet and the trapped analytes were desorbed using 3 mL of acetone. The eluted solution was evaporated by blowing over a gentle stream of nitrogen gas and reconstituted with 50 μL of propazine (100 ng mL⁻¹ in methanol) as an internal standard (IS). Finally, 1 μL of the reconstituted solution was injected into a GC-μECD to determine the extracted lindane, chlorpyrifos, hexaconazole and azaconazole. Triplicate injections were performed for each extract and an average peak area ratio reading was taken.

Fig. 3 Schematic representation of graphene-based MSPE procedure.

2.9.3 C18-SPE procedure. A commercial C18-SPE cartridge was optimized with important extraction parameters such as type of solvent, volume of solvent, sample volume and flow rate. Acetone, acetonitrile, methanol, ethyl acetate and n-hexane were used to optimize the solvents effect on C18 SPE. Comparatively, a high extraction efficiency was achieved using acetonitrile (0.5 mL to 5 mL). On the basis of the results, 2 mL of acetonitrile was selected as the optimized solvent volume. In order to improve the preconcentration factor, the sample volume was also optimized (1 mL to 50 mL). The maximum extraction efficiency was obtained with 10 mL of sample volume and thus was selected as the optimized sample volume. The best result was achieved at 0.5 mL min⁻¹ flow rate.

2.10 Adsorption capacity

Using the optimized conditions the adsorbent capacity (q_e) was investigated for residual chlorinated pesticides in water samples. The residual concentration of chlorinated pesticides was analyzed using GC-μECD by reconstituting the analyte with 100 μL methanol as described previously.³⁵ The adsorption capacity was calculated using eqn (1) as follows:where q_e is the adsorption capacity (mg g⁻¹), V is the initial volume of the sample (L) before pretreatment, m is the mass of applied adsorbent (g) for extraction, C₀ is the initial concentration (mg L⁻¹) and C_e is the residual concentration (mg L⁻¹) of analytes in the solution produced after extraction.

3. Result and discussion

3.1 Characterization

3.1.1 FT-IR spectra. The synthesis of Fe₃O₄@SiO₂, functionalization of graphene and immobilization of functionalized graphene onto the Fe₃O₄@SiO₂ were confirmed by FT-IR spectral analysis (Fig. 4A–C). Natural graphite does not contain any functional groups. However, following the functionalization process the resultant GO (Fig. 4A) shows some additional bands at 3372, 1720, 1625, 1436, 1168 and 1050 cm⁻¹ for of O–H, C=O, C=C, C–C, epoxy groups and C–O group stretching, respectively.²¹ The appearance of sharp bands (Fig. 4B) at 582 and 1100 cm⁻¹ corresponds to Fe–O and Si–O symmetric stretching, respectively.³⁵ The formation of Fe₃O₄@SiO₂–G MNPs was confirmed by the appearance as well as disappearance of characteristic bands. The IR spectra of GO (Fig. 4A) does not show a Si–O stretching band, however, following the immobilization with Fe₃O₄@SiO₂ the resultant material Fe₃O₄@SiO₂–G MNPs shows a band at 1100 cm⁻¹. Additionally, during the immobilization the Si–O group intensity also reduced and disappearance of characteristic bands at 1720 and 1168 cm⁻¹ for C=O and epoxy groups (Fig. 4C) respectively also offer proof for the formation of Fe₃O₄@SiO₂–G MNPs. Consequently appearance and disappearance of some characteristic peaks provide qualitative evidence which confirms the immobilization of GO onto Fe₃O₄@SiO₂ as well as formation of Fe₃O₄@SiO₂–G MNPs.

Fig. 4 FT-IR spectra of (A) GO, (B) Fe₃O₄@SiO₂ and (C) graphene-based magnetic nanocomposite.

3.1.2 Field emission scanning microscopy. The morphology of the newly synthesized magnetic graphene-based adsorbent (Fe₃O₄@SiO₂–G) was analyzed using field emission scanning microscopy (FESEM). The white cloud-like appearance (Fig. 5A) clearly shows that nanosize visible graphene sheets have been successfully immobilized with silica coated MNPs of Fe₃O₄@SiO₂. Furthermore, in order to examine the purity and elemental composition of the silica coated MNP immobilized material (Fe₃O₄@SiO₂–G), FESEM was coupled with energy dispersive spectroscopy (EDS). The EDS results (Fig. 5B) showed 50%, 17.91% and 5.50% of C, Fe and Si respectively, in the Fe₃O₄@SiO₂–G. As we know graphene oxide does not contain Si and Fe, and Fe₃O₄@SiO₂ does not contain C. But after the immobilization of Fe₃O₄@SiO₂ onto the modified graphene oxide, the presence of Si, Fe and C confirmed the successful immobilization.

Fig. 5 (A) Micrograph image and (B) EDS spectra of the graphene-based silica MNPs.

3.1.3 XRD. Fig. 6 shows the XRD signals of Fe₃O₄, Fe₃O₄@SiO₂, and Fe₃O₄@SiO₂–G. The Fe₃O₄ signals appeared as 2θ values of 30.14° (220), 35.51° (311), 43.15° (400), 53.44° (422), 57.09° (334), 62.68° (400) and 74.16° (533). In Fig. 6B, the broad peak at 20.44° revealed that following the coating with SiO₂ the intensity of Fe₃O₄ has decreased. The GO skeleton possesses significantly high diffraction signals at 2θ = 10.3° (001) and a weak one at 43.3° (101).^28,36,37 The disappearance of the high diffraction signal for GO and the appearance of a broad band signal at 26.1° in XRD spectra (Fig. 6C) may be attributed to the successful conversion of GO into graphene sheet through a reduction process in basic solution i.e., hydrazine hydrate and ammonia solution. The XRD results of the synthesized Fe₃O₄ MNPs completely matched the library template pattern number 01-071-6336. Furthermore, the quite small size of Fe₃O₄ particles can be predicted by the wider signals in the XRD spectra.³⁸

Fig. 6 XRD patterns of (A) Fe₃O₄ (B) Fe₃O₄@SiO₂ and (C) Fe₃O₄@SiO₂–G magnetic nanocomposite.

3.2 Optimization of different parameters

A batch-wise MSPE study was carried out to evaluate the extraction efficiency of the synthesized Fe₃O₄@SiO₂–G adsorbent for the selected chlorinated pesticides. Important parameters such as of desorption solvent, volume of solvent, extraction time, desorption time, sample volume and pH were optimized.

3.2.1 Types of desorption solvent. The effect of desorption solvent on MSPE performance has been investigated using seven organic solvents of different polarity. It is obvious from Fig. 7A that significant extraction efficiency was obtained with 3 mL acetone.

Fig. 7 Effect of (A) desorption solvents, (B) sample volume, (C) adsorbent dosage and (D) solution pH on chlorinated pesticide extraction efficiency.

3.2.2 Extraction and desorption time. The extraction and desorption times are the key parameters for an effective MSPE process. To optimize extraction time trials ranging from 1 to 90 min were used. The highest peak area ratio was observed within 3 min. For desorption time, time settings from 1 to 5 min were studied. The peak area was significantly highest in 1 minute. Thus, 3 min and 1 min were selected as the extraction and desorption times, respectively.

3.2.3 Sample volume. In order to obtain a high enrichment factor the sample volume (5 to 80 mL) was investigated (Fig. 7B). By increasing the volume (5–50 mL), the enrichment factor increased as well as the peak area ratio and it attains a maximum at 50 mL of sample volume. Beyond the 50 mL of sample volume there is no significant change in the peak area, consequently 50 mL was chosen as the optimized sample volume for further analysis.

3.2.4 Adsorbent dosage. The effect of adsorbent dosage on the percent extraction of chlorinated pesticides was assessed by changing the mass of adsorbent in the range of 10 to 120 mg (Fig. 7C). Increasing the mass of adsorbent, increased the % extraction of the chlorinated pesticides. The dosage study clarified that chlorinated pesticide extraction efficiency remarkably increased up to 60 mg of adsorbent. There after increasing the mass of adsorbent did not produce a significant improvement in % extraction. Consequently, all the experiments were performed with a fixed mass of adsorbent i.e., 60 mg.

3.2.5 Solution pH study. pH is also one of the key parameters and plays an important role during the adsorption process. It affects the aqueous chemistry as well as the dissociation of functional groups on the active sites of the adsorbent. Consequently, the effect of solution pH on the pesticide adsorption using the graphene-based magnetic adsorbent was examined at different pHs (i.e., 3.5, 5.0, 6.5, 8.0 and 10.0). It can be seen from Fig. 7D that the peak area of the isolated chlorinated pesticides in acidic conditions was higher as compared to basic media. The high extraction of chlorinated pesticides in a slightly acidic medium i.e., pH 6–7 can be explained on the basis of point zero charge (PZC) of graphene-based material. In an acidic medium (pH 5.0 to 6.5) the graphene-based adsorbent had a net positive surface charge.^39,40 Results showed that by decreasing the solution pH, extraction increases and it attains the maximum i.e., 88.5%, 92.9%, 100.9% and 96.9% for lindane, chlorpyrifos, hexaconazole and azaconazole pesticide, respectively at pH 6.5. In this case the probable mechanism is surface complexation by means of surface association between the acidic hydroxyl (⁺OH₂) groups of the graphene-based adsorbent and anionic sites of the chlorinated pesticides. While at low pH, the decrease in extraction efficiency may be due to the repulsion between the positive surface of the adsorbent and positive ions on the chlorinated pesticides. On the other hand at higher pH the deprotonation of the phenolic hydroxyl groups of the graphene moiety in the graphene-based adsorbent converts them into negatively charged phenoxide ions. Following the deprotonation as well as due to the abundance of OH⁻ ions, extraction of chlorinated pesticides decreases. The dramatic decrease in % extraction of chlorinated pesticides at higher pH (pH 8) is due to the fact that pesticides undergo hydrolysis rather than adsorption.^41,42

3.3 Comparison of extraction performance

Comparative studies were performed in order to compare the extraction performance of the newly synthesized Fe₃O₄@SiO₂–G adsorbent with Fe₃O₄@SiO₂ and the commercially available C18-SPE cartridge (100 mg) for the preconcentration of chlorinated pesticides. Fig. 8 shows that the high sensitivity of C18 is only for the non-polar chlorpyrifos (log K_O/W 4.7) and hexaconazole (log K_O/W 3.6) since it only provides non-covalent hydrophobic interactions. Due to SiO₂ porosity and polar hydroxide groups, Fe₃O₄@SiO₂ shows slightly higher efficiency for the moderately polar azaconazole (log K_O/W 2.7). The newly synthesised Fe₃O₄@SiO₂–G adsorbent showed a significantly higher extraction efficiency for both the non-polar as well as moderately polar chlorinated pesticides namely chlorpyrifos, hexaconazole and azaconazole. The higher extraction efficiency of the newly synthesized Fe₃O₄@SiO₂–G for non-polar as well as moderately polar pesticides can be explained by the fact that the newly synthesized adsorbent contains a benzene moiety which can form a strong π–π interaction with the benzene ring of the selected chlorpyrifos, hexaconazole and azaconazole pesticides. An H-bonding interaction was also observed (Fig. 9) and lindane affinity can be explained by hydrophobic and electrostatic interaction.

Fig. 8 Comparison of extraction performance (based on peak areas ratio) of three different sorbents for the chlorinated pesticides pre-concentration. Conditions for extractions: sample volume 10 mL, concentration of analytes 1 ng mL, EF 100, extraction time 5 min, desorption time 3 min and 3 mL acetonitrile as desorption solvent.

Fig. 9 Proposed mechanism for the adsorption of the chlorinated pesticides by Fe₃O₄@SiO₂–G.

3.4 Adsorption study

The adsorption capacity (q_e) of the newly prepared Fe₃O₄@SiO₂, Fe₃O₄@SiO₂–G adsorbents and C18-SPE sorbent for chlorinated pesticides was calculated using eqn (1). A plot of q_e versus C_e (residual concentration) as shown in (Fig. 10A) illustrates that the adsorption capacity of Fe₃O₄@SiO₂–G was increased by increasing the residual concentration until the adsorbent sites were saturated.

Fig. 10 (A) Experimental adsorption capacity of newly synthesized sorbent and (B) Langmuir linearity. Conditions: 50 mL sample in deionized water (0.1–7 μg mL⁻¹), extraction time 3 min, and adsorbent 60 mg.

3.5 Adsorption isotherms

The correlation between the amounts of solute adsorbed per unit amount of the adsorbent and concentration of adsorbate in bulk solution at a given temperature under equilibrium conditions can be analyzed by using adsorption isotherms. The analysis of isotherm data plays a significant role in predicting the maximum adsorption capacity (q_m) of the adsorbent. Consequently, three well known isotherm models namely Langmuir, Freundlich and Temkin isotherms were tested in the following way:⁴³

Temkin: q_e = B ln A + B ln C_e (4)

where C_e is the residual concentration of pesticides in the solution (mg L⁻¹), q_e is the experimental adsorption capacity (mg g⁻¹), q_m is the maximum adsorption capacity (mg g⁻¹), k is the Langmuir constant (L mg⁻¹), K_F [(mg g⁻¹)/(mg g⁻¹)^1/n] and n are the Freundlich isotherm constants. A is Temkin equilibrium constant (L g⁻¹) and B is the heat of sorption constant (J mol⁻¹).

Good linearity was obtained with the Langmuir isotherm model (Fig. 10B). Thus, adsorption of the selected chlorinated pesticides was well fitted by the Langmuir isotherm model as compared to the Freundlich isotherm and Temkin model due to the high value of coefficient of determination (R²). The q_m values (Table 1) showed that the newly synthesized Fe₃O₄@SiO₂–G adsorbent possesses a high adsorption capacity (q_m) for the selected chlorinated pesticides. However, the Fe₃O₄@SiO₂–G provided 16× and 30× higher adsorption capacity as compared to the Fe₃O₄@SiO₂ and C18-SPE, respectively.

Table 1 Maximum adsorption capacity (q_m) for C18-SPE, Fe₃O₄@SiO₂ NPs and Fe₃O₄@SiO₂–G

Model	Maximum adsorption capacity (qm)
	Lindane	Chlorpyrifos	Hexaconazole	Azaconazole
C18	0.82	0.73	0.63	0.64
Fe3O4@SiO2	1.58	1.82	2.1	1.18
Fe3O4@SiO2–G	13.04	16.58	18.69	15.35

---

3.6 Method validation

The analytical performance of the proposed magnetic graphene-based MSPE method was validated using different analytical parameters such as linearity, limit of detection (LOD), limit of quantification (LOQ), precision (RSD%), reusability and enrichment factor (EF). The linearity of the MSPE technique for chlorinated pesticide using Fe₃O₄@SiO₂–G, Fe₃O₄@SiO₂ and commercial SPE-C18 was examined at different concentration levels i.e., 0.001–0.1 ng mL⁻¹, 0.1–1 ng mL⁻¹ and 0.5–10 ng mL⁻¹, respectively. Good linearity with a high value of coefficient of determination (R²) i.e., 0.9991–0.9998 were obtained for the selected chlorinated pesticides (Table 2). The final concentration was calculated using a calibration graph method i.e., (A_x/A_is) versus (C₀/C_is), where A_x is the peak area after extraction, A_is is the peak area for the internal standard, C₀ is the initial concentration (ng mL⁻¹) and C_is is the internal standard concentration (ng mL⁻¹). The LODs (S/N = 3) obtained for all pesticides were between 0.1 and 0.3 pg mL⁻¹ (n = 3) with 1000 × EF. Table 2 shows that the LOD for Fe₃O₄@SiO₂–G is appreciably lower as compared to the Fe₃O₄@SiO₂ and commercial C18-SPE. The LOD obtained for Fe₃O₄@SiO₂–G is below the MRL i.e. 100 pg mL⁻¹ as set by US EPA or EU for each pesticide in drinking water.

Table 2 Statistical results of MSPE method, including limit of detection (LOD), limit of quantification (LOQ), linearity and enrichment factor (EF) using three different adsorbents

Validation	Sorbent	Analyte
		Lindane	Chlorpyrifos	Hexaconazole	Azaconazole
a Enrichment factor (EF) was calculated using (EF = Vaq/Vorg) equation.
Linearity (ng mL^−1)	C18	0.5–5	0.5–5	0.5–5	0.5–5
	Fe3O4@SiO2	0.1–1	0.1–1	0.1–1	0.1–1
	Fe3O4@SiO2–G	0.001–0.1	0.001–0.1	0.001–0.1	0.001–0.1
R^2	C18	0.9996	0.9997	0.9999	0.9997
	Fe3O4@SiO2	0.9998	0.9993	0.9994	0.9996
	Fe3O4@SiO2–G	0.9997	0.9995	0.9998	0.9994
LOD (pg mL^−1, n = 3)	C18	214	206	70	217
	Fe3O4@SiO2	16	18	30	25
	Fe3O4@SiO2–G	0.278	0.158	0.119	0.221
LOQ (pg mL^−1, n = 3)	C18	712.62	685.98	233.1	722.61
	Fe3O4@SiO2	53	60	99	83
	Fe3O4@SiO2–G	0.925	0.526	0.396	0.736
EF^a	C18	80	80	80	80
	Fe3O4@SiO2	1000	1000	1000	1000
	Fe3O4@SiO2–G	1000	1000	1000	1000

---

The repeatability and reproducibility of the proposed MSPE method were investigated using intraday and interday measurements. Relative standard deviation (%RSD) was assessed for extraction of 1 ng mL⁻¹ chlorinated pesticides followed by preconcentration thrice in a day (intraday) for five consecutive days (interday) (Table 3). The interday value looks a little bit larger but ANOVA showed no significant difference between the whole analysis, since, p > 0.05 and F_experimental < F_table at 95% confidence level. Thus repeatability and reproducibility values for proposed MSPE method are acceptable.

Table 3 Repeatability and reproducibility of MSPE method for chlorinated pesticide extraction (50 mL sample include 0.1 ng mL⁻¹ of each chlorinated pesticides)

Intra-day (n = 3)	MSPE (Fe3O4@SiO2–G)
	Lindane	Chlorpyrifos	Hexaconazole	Azaconazole
Day 1	4.34	7.21	4.69	6.95
Day 2	8.25	5.30	6.33	3.32
Day 3	5.52	1.38	1.57	6.26
Day 4	3.58	4.13	2.16	7.08
Day 5	4.60	1.25	1.01	1.37
Inter-day (n = 15)	8.47	9.72	4.94	8.06

---

Reusability of the newly synthesized adsorbent was studied for 40 continuous adsorption–desorption cycles. For regeneration, the adsorbent was washed sequentially with 2 mL acetone, 2 mL acetonitrile and 3 mL deionized water after each extraction process. The chlorpyrifos, hexaconazole, azaconazole and lindane were pre-concentrated at least 15–30 times without significant decrease in extraction efficiency.

3.7 Environmental water sample analysis

In order to assess the field application of the proposed MSPE method, chlorinated pesticides were isolated from environmental water samples i.e. tap, river, lake and sea water samples. Table 4 shows that tap water has a high percent recovery as compared to the sea water due to matrix interference. The results indicate that in spite of interference the graphene-based adsorbent was still capable of absorbing 0.50 pg mL⁻¹ of chlorinated pesticide from environmental water samples with good recovery 80.8–106.3% and lower RSD% (2.1–7.9%).

Table 4 Percentage recovery and precision (±RSD%) of chlorinated pesticides from different water samples for C18, Fe₃O₄@SiO₂ and Fe₃O₄@SiO₂–G

SPE sorbent	Sample	%Recovery^a (±RSD%, n = 3)
		Lindane	Chlorpyrifos	Hexaconazole	Azaconazole
a The numbers in the parenthesis shows the precision (%RSD, n = 3).
C18	Tap water	86.6 (1.8)	86.5 (3.6)	92.4 (3.3)	90.4 (1.7)
	River water	83.5 (4.1)	83.4 (10.2)	90.3 (2.4)	88.1 (6.6)
	Lake water	87.4 (4.9)	81.1 (5.2)	83.9 (3.7)	81.3 (2.9)
	Sea water	72.5 (2.9)	73.6 (4.3)	72.0 (13.3)	71.4 (12.4)
Fe3O4@SiO2	Tap water	95.2 (2.5)	103.7 (7.6)	102.1 (8.1)	101.2 (9.4)
	River water	92.0 (7.3)	98.6 (9.3)	99.9 (3.1)	98.0 (8.2)
	Lake water	97.4 (5.8)	101.6 (4.8)	97.2 (5.7)	97.9 (3.6)
	Sea water	88.4 (8.6)	87.9 (9.6)	83.4 (3.9)	94.7 (2.9)
Fe3O4@SiO2–G	Tap water	104.1 (2.1)	106.3 (2.1)	103.8 (3.7)	105.9 (6.5)
	River water	104.3 (5.1)	100.2 (2.5)	104.5 (4.9)	100.3 (2.9)
	Lake water	101.2 (2.0)	100.5 (7.9)	103.5 (5.3)	101.1 (4.9)
	Sea water	92.2 (3.9)	86.8 (7.1)	80.8 (3.6)	89.2 (7.7)

---

3.8 Chromatograms of the extracted chlorinated pesticides

GC-μECD was used for the analysis of chlorinated pesticides from water samples. Fig. 11A shows the chromatogram of the tap water sample. It was found that the tap water analysed did not contain the chlorinated pesticides of interest since expected peaks were not observe at expected retention times (t_R) of 5.9, 7.2, 7.9 and 8.2 min under similar GC-μECD conditions. Only a single sharp peak at 5.6 min, which belongs to internal standard (IS) propazine was observed. Fig. 11B shows the chromatogram of the spiked tap water sample for extracted chlorinated pesticides using the newly synthesised Fe₃O₄@SiO₂–G. The data revealed that the proposed MSPE method was suitable to analyze the chlorinated pesticides in water samples since sharp peaks were observed at 5.6, 5.9, 7.2, 7.9 and 8.2 min for the propazine (IS), lindane, chlorpyrifos, hexaconazole and azaconazole, respectively.

Fig. 11 GC-μECD chromatograms obtained using graphene-based MSPE from tap water samples for (A) unspiked, (B) spiked chlorinated pesticides (50 pg mL⁻¹). Peaks: (1) lindane; (2) chlorpyrifos; (3) hexaconazole; (4) azaconazole and propazine as an internal standard (IS).

3.9. Comparison with other results

For the validation of the study, the results obtained were compared with previously reported adsorbents from the literature. Comparison of LOD (Table 5) reveals that Fe₃O₄@SiO₂–G is 10–110× more sensitive compared to the previously reported adsorbents.

Table 5 Comparison of LOD of the current study with other recent MSPE sorbents

SPE sorbent	Analyte	LOD (pg mL^−1)	Method	Detector	Ref.
Fe3O4@SiO2–G	Lindane, chlorpyrifos, hexaconazole, azaconazole	0.12	MSPE	GC-μECD	This study
Pillararene–Fe3O4	Fusilazole, cyprodinil, pyrimethanil, trifumizole	500	MSPE	HPLC-UV	44
Graphene–Fe3O4	Chloroacetanilide herbicides	20.0	MSPE	GC-ECD	45
Graphene–Fe3O4	Herbicides	10.0	MSPE	GC-FID	46
Fe3O4@dioctadecyl dimethyl ammonium chloride@silica	Herbicides	70.0	MSPE	HPLC-UV	47

---

4. Conclusion

A new graphene-based silica coated magnetic adsorbent (Fe₃O₄@SiO₂–G) was successfully synthesized and characterized using FT-IR, FESEM, EDS and XRD. The newly synthesised Fe₃O₄@SiO₂–G was successfully applied for the preconcentration of four selected chlorinated pesticides from aqueous media. Outstanding adsorption efficiencies of 88.5%, 92.9%, 100.9% and 96.9% were achieved at pH 6.5 for lindane, chlorpyrifos, hexaconazole and azaconazole, respectively. The significant adsorption capacity (13.04–18.35 mg g⁻¹) and low LOD (0.12–0.28 pg mL⁻¹) with 1000× EF confirmed that the newly synthesized Fe₃O₄@SiO₂–G is a versatile adsorbent for the preconcentration of chlorinated pesticides as compared to Fe₃O₄@SiO₂ and C18-SPE. The field studies also supported the effectiveness of this new magnetic nanocomposite adsorbent which could be useful and has good potential for the extraction of selected pesticides from real water samples.

Acknowledgements

The authors would like to thank Universiti Teknologi Malaysia for facilitations and the Ministry of Education Malaysia for financial support (Q.J130000.2626.10J43). H. R. Nodeh also would like to thank UTM for the International Doctoral Fellowship (IDF) received.

References

1. U. S. McKnight, J. J. Rasmussen, B. Kronvang, P. J. Binning and P. L. Bjerg, Environ. Pollut., 2015, 200, 64–76.
2. Y. Gou, R. Eisert and J. Pawliszyn, J. Chromatogr. A, 2000, 873, 137–147.
3. W. Zhang, F. Jiang and J. Ou, Proc. Int. Acad. Ecol. Environ. Sci., 2011, 1, 125–144.
4. F. P. Carvalho, Environ. Sci. Policy, 2006, 9, 685–692.
5. J. M. Dabrowski, J. M. Shadung and V. Wepener, Environ. Int., 2014, 62, 31–40.
6. A. Bergman, J. J. Heindel, T. Kasten, K. A. Kidd, S. Jobling, M. Neira, R. T. Zoeller, G. Becher, P. Bjerregaard and R. Bornman, Environ. Health Perspect., 2013, 121, a104–106.
7. P. K. Seth, in Water and Health, Springer, 2014, pp. 375–384.
8. J. S. Burns, P. L. Williams, S. A. Korrick, R. Hauser, O. Sergeyev, B. Revich, T. Lam and M. M. Lee, Am. J. Epidemiol., 2014, 180, 909–919.
9. Rodal, ROADAL news, 2014, http://www.rodalenews.com/pesticides-apples.
10. L. Cai, J. Xing, L. Dong and C. Wu, J. Chromatogr. A, 2003, 1015, 11–21.
11. M. A. Farajzadeh, S. M. Sorouraddin and M. R. A. Mogaddam, Microchim. Acta, 2014, 1–23.
12. W. A. Wan Ibrahim, K. V. Veloo and M. M. Sanagi, J. Chromatogr. A, 2012, 1229, 55–62.
13. Y. Cai, G. Jiang, J. Liu and Q. Zhou, Anal. Chem., 2003, 75, 2517–2521.
14. H. Bagheri and A. Roostaie, J. Chromatogr. A, 2014, 1324, 11–20.
15. H. Bagheri and A. Roostaie, J. Chromatogr. A, 2015, 1375, 8–16.
16. X. Zhu, F. Tang, J. Yang and Y. Gao, J. Chin. Chem. Soc., 2014, 61, 227–232.
17. E. Sahmetlioglu, E. Yilmaz, E. Aktas and M. Soylak, Talanta, 2014, 119, 447–451.
18. H. Bagheri, A. Roostaie and M. Y. Baktash, Anal. Chim. Acta, 2014, 816, 1–7.
19. X. Luo, F. Zhang, S. Ji, B. Yang and X. Liang, Talanta, 2014, 120, 71–75.
20. Z. Shi, J. Hu, Q. Li, S. Zhang, Y. Liang and H. Zhang, J. Chromatogr. A, 2014, 1355, 219–227.
21. W. Guan, Z. Li, H. Zhang, H. Hong, N. Rebeyev, Y. Ye and Y. Ma, J. Chromatogr. A, 2013, 1286, 1–8.
22. Y. Ke, F. Zhu, F. Zeng, T. Luan, C. Su and G. Ouyang, J. Chromatogr. A, 2013, 1300, 187–192.
23. X. Chen, F. Meng, Z. Zhou, X. Tian, L. Shan, S. Zhu, X. Xu, M. Jiang, L. Wang and D. Hui, Nanoscale, 2014, 6, 8140–8148.
24. Z. Dong, D. Wang, X. Liu, X. Pei, L. Chen and J. Jin, J. Mater. Chem. A, 2014, 2, 5034–5040.
25. R. Sitko, B. Zawisza and E. Malicka, TrAC, Trends Anal. Chem., 2013, 51, 33–43.
26. Q. Han, Z. Wang, J. Xia, X. Zhang, H. Wang and M. Ding, J. Sep. Sci., 2014, 37, 99–105.
27. Q. Han, Z. Wang, J. Xia, L. Xia, S. Chen, X. Zhang and M. Ding, J. Sep. Sci., 2013, 36, 3586–3591.
28. M. Alvand and F. Shemirani, Microchim. Acta, 2014, 181, 181–188.
29. L. Liu, T. Feng, C. Wang, Q. Wu and Z. Wang, Microchim. Acta, 2014, 181, 1249–1255.
30. L. Wang, X. Xu, Z. Zhang, D. Zhang, X. Liu and L. Zhang, RSC Adv., 2015, 5, 22022–22030.
31. M. Sun, X. Ma, J. Wang, W. Wang, Q. Wu, C. Wang and Z. Wang, J. Sep. Sci., 2013, 36, 1478–1485.
32. N. Sun, C. Deng, Y. Li and X. Zhang, ACS Appl. Mater. Interfaces, 2014, 6, 11799–11804.
33. A. Demir, A. Baykal, H. Sözeri and R. Topkaya, Synth. Met., 2014, 187, 75–80.
34. C. Wu, H. He, H. Gao, G. Liu, R. Ma, Y. An and L. Shi, Sci. China: Chem., 2010, 53, 514–518.
35. Q. Tang, X. Wang, F. Yu, X. Qiao and Z. Xu, J. Sep. Sci., 2014, 37, 820–827.
36. Y. Tang, F. Huang, W. Zhao, Z. Liu and D. Wan, J. Mater. Chem., 2012, 22, 11257–11260.
37. S. Shubhda, J. Kiran, V. N. Singh, S. Sukhvir, N. Vijayan, D. Nita, G. Govind and T. D. Senguttuvan, Nanotechnology, 2012, 23, 205–501.
38. T. T. Baby and S. Ramaprabhu, Talanta, 2010, 80, 2016–2022.
39. J. Li, S. Zhang, C. Chen, G. Zhao, X. Yang, J. Li and X. Wang, ACS Appl. Mater. Interfaces, 2012, 4, 4991–5000.
40. X. Yu, S. Tong, M. Ge, L. Wu, J. Zuo, C. Cao and W. Song, Carbohydr. Polym., 2013, 92, 380–387.
41. B. Maddah and J. Shamsi, J. Chromatogr. A, 2012, 1256, 40–45.
42. A. M. Filho, F. N. dos Santos and P. A. d. P. Pereira, Microchem. J., 2010, 96, 139–145.
43. L. Li, L. Fan, H. Duan, X. Wang and C. Luo, RSC Adv., 2014, 4, 37114–37121.
44. M. M. Tian, D. X. Chen, Y. L. Sun, Y. W. Yang and Q. Jia, RSC Adv., 2013, 3, 22111–22119.
45. Z. Li, S. Bai, M. Hou, C. Wang and Z. Wang, Anal. Lett., 2013, 46, 1012–1024.
46. S. S. Bai, Z. Li, X. H. Zang, C. Wang and Z. Wang, Chin. J. Anal. Chem., 2013, 41, 1177–1182.
47. Z. He, D. Liu, R. Li, Z. Zhou and P. Wang, Anal. Chim. Acta, 2012, 747, 29–35.

---