Novel Functionalized Polythiophene-Coated Fe3O4 Nanoparticles for Magnetic Solid-Phase Extraction of Phthalates

Poly(phenyl-(4-(6-thiophen-3-yl-hexyloxy)-benzylidene)-amine) (P3TArH) was successfully synthesized and coated on the surface of Fe3O4 magnetic nanoparticles (MNPs). The nanocomposites were characterized by Fourier transform infra-red (FTIR), X-ray diffractometry (XRD), Brunauer-Emmett-Teller (BET) surface area analysis, analyzer transmission electron microscopy (TEM) and vibrating sample magnetometry (VSM). P3TArH-coated MNPs (MNP@P3TArH) showed higher capabilities for the extraction of commonly-used phthalates and were optimized for the magnetic-solid phase extraction (MSPE) of environmental samples. Separation and determination of the extracted phthalates, namely dimethyl phthalate (DMP), diethyl phthalate (DEP), dipropyl phthalate (DPP), dibutyl phthalate (DBP), butyl benzyl phthalate (BBP), dicyclohexyl phthalate (DCP), di-ethylhexyl phthalate (DEHP) and di-n-octyl phthalate (DNOP), were conducted by a gas chromatography-flame ionization detector (GC-FID). The best working conditions were as follows; sample at pH 7, 30 min extraction time, ethyl acetate as the elution solvent, 500-µL elution solvent volumes, 10 min desorption time, 10-mg adsorbent dosage, 20-mL sample loading volume and 15 g·L−1 concentration of NaCl. Under the optimized conditions, the analytical performances were determined with a linear range of 0.1–50 µg·L−1 and a limit of detection at 0.08–0.468 µg·L−1 for all of the analytes studied. The intra-day (n = 7) and inter-day (n = 3) relative standard deviations (RSD%) of three replicates were each demonstrated in the range of 3.7–4.9 and 3.0–5.0, respectively. The steadiness and reusability studies suggested that the MNP@P3TArH could be used up to five cycles. The proposed method was executed for the analysis of real water samples, namely commercial bottled mineral water and bottled fresh milk, whereby recoveries in the range of 68%–101% and RSD% lower than 7.7 were attained.


Introduction
Belonging to non-halogenated esters of phthalic acid, phthalates or phthalate esters are used as plasticizers for nitrocellulose, since it was first recognized in 1880, replacing camphor [1]. Nowadays, phthalates can be found in many different matrices in our environment and are widely utilized in the PVC industries as a plasticizer, from floors, hoses, cables (building materials), toys and medical appliances [2]. Other consumer-based products utilizing phthalates are as a component in inks, adhesive materials, lacquers, sealing and packing materials, materials for treating surfaces, solvents and fixing agents in fragrances, as well as additives in cosmetics [3][4][5]. They become emerging pollutants and harmful to humans, especially children, since they are not chemically bound in plastics

Standard, Reagents and Chemicals
Analytical grade ferric chloride, ferrous chloride, ammonia solution (25 wt %), thiophene, 4-hydroxybenzaldehyde, acetonitrile, potassium permanganate, 4-aminophenol, 3-bromothiophene, 1,6-dibromohexane, N-bromosuccinimide, acetic acid, sodium hydrogen bicarbonate, potassium iodide, potassium carbonate, tetrahydrofuran, methanol, hydrochloric acid, acetone and ethyl acetate were purchased from Merck (Darmstadt, Germany). Acetone was procured from Fisher Scientific (Loughborough, UK). Thiophene carboxaldehyde, polyvinyl alcohol and n-butyllithium (2.0 M in cyclohexane) were obtained from Sigma Aldrich (Milwaukee, WI, USA). Magnesium sulfate anhydrous, ethanol denatured and hexane were received from J. Kollins (Parkwood, Australia), while dimethyl sulfoxide-d 6 (DMSO-d 6 ) and phthalate esters were purchased from Acros Organics (Geel, Belgium). Ultrapure water was prepared by a model Aqua Max-Ultra ultra-pure water purification system (Zef Scientific Inc., San Diego, CA, USA). Stock solutions of 1000 mg¨L´1 of standards were prepared by dissolving appropriate amounts of compounds in methanol, which remain stable for three months if stored in a refrigerator at 4˝C. Working standard solutions were prepared daily by diluting the stock standard solution to the required concentrations.

Instruments
The Fourier transform infrared (FTIR) spectra were recorded on a Perkin-Elmer FTIR between 4000 and 400 cm´1, with a resolution of 2 cm´1. Structural elucidation was determined using 1 H NMR, JEOL 400 MHz. The pore diameter and surface area of Brunauer-Emmett-Teller (BET) analysis were determined from low-temperature nitrogen adsorption isotherms at 77.40 K using a Quantachrome Autosorb Automated Gas Sorption System (Quantachrome Instruments, Boynton Beach, FL, USA). X-ray powder diffraction (XRD) analysis was conducted with Panalytical model Empyrean (Panalytical, Almelo, Netherlands) at 40 kV and 35 mA using Cu Kα radiation (λ = 1.54059 Å). Morphological analyses of the synthesized products were conducted using transmission electron microscopy (TEM) analysis using an FEI Tecnai G2 spectra microscope (FEI, Hillsboro, OR, USA). The magnetic property was tested using a vibration sample magnetometer (VSM) Model 9600 (Quantum Design Inc., San Diego, CA, USA). Magnetization measurements were carried out in an external field of up to 15 kOe at room temperature.
Separation and detection of target analytes were performed by a Shimadzu 2010 gas chromatograph (Shimadzu, Kyoto, Japan) equipped with a split/splitless injector and a flame ionization detector (FID). A DB-5 Agilent fused-silica capillary column (Agilent, Santa Clara, CA, USA) (30 mˆ0.32 mm i.d.ˆ0.25 µm film thickness) was applied for separation of analytes. Helium (with 99.999% purity) was used as the carrier gas at a constant flow rate of 4 mL¨min´1. Chromatographic conditions were controlled as described; the temperatures of the injector and detector were set at 260 and 280˝C, respectively. The injection port was operated at splitless mode. Oven temperature was held at 150˝C for 1 min and increased to 280˝C at 8˝C¨min´1 for 3 min.

Polymerization of 3TArH and Thiophene Monomers on the Surface of MNPs
The preparation of MNP@PTh and MNP@P3TArH NPs involves two steps. Briefly, Fe3O4 has been prepared by the co-precipitation method [57]. FeCl3·6H2O (8.48 g, 30 mmol) and FeCl2·4H2O (2.25 g, 11.3 mmol) were dissolved in 400 mL deionized water under nitrogen atmosphere via vigorous stirring (1000 rpm) at 80 °C. Then, a 20-mL ammonia solution 25% (w/w) was added to the solution. The color of the bulk solution immediately changed from orange to black. After stirring the mixture for 5 min, the Fe3O4 NP precipitates were obtained via magnetic decantation and washed three times with deionized water. Finally, the Fe3O4 NPs were dried in a vacuum oven at 70 °C for 12 h.
The surface of Fe3O4 NPs was modified by being coated with the newly-designed modified thiophene monomers via oxidation polymerization with the generation of ferric cations on the Fe3O4 NPs' surface [54]. Fe3O4 NPs (1 mmol, 0.235 g) were discrete in polyvinyl alcohol (PVA) aqueous solution (0.001 M). Later, 3TArH (3.64 g, 10 mmol) was added into the mixed solution with vigorous stirring. Subsequently, 30 mL of HCl (0.5 M) solution were introduced into the mixture. Then, the products obtained were dried in a vacuum oven at 70 °C for 12 h. Experiments were repeated using freshly-distilled thiophene monomer (10 mmol, 0.84 g).

Solid Phase Extraction Optimization and Reusability Studies
Factors affecting the extraction efficiency of the proposed method, such as type of adsorbents, pH, extraction time, sample volume, elution solvent, elution solvent volume, desorption time, adsorbent dosage and effect of NaCl, were studied. All of the experiments were performed in triplicate, and the means of the results were used in plotting the optimization curves.
The reusability of the adsorbent was determined with optimized conditions for up to five cycles. The adsorbent was recycled after being washed with methanol and water and dried in vacuum at 70 °C for 12 h. Scheme 1. Synthesis pathway for (phenyl-(4-(6-thiophen-3-yl-hexyloxy)-benzylidene)-amine) (3TArH).

Polymerization of 3TArH and Thiophene Monomers on the Surface of MNPs
The preparation of MNP@PTh and MNP@P3TArH NPs involves two steps. Briefly, Fe 3 O 4 has been prepared by the co-precipitation method [57]. FeCl 3¨6 H 2 O (8.48 g, 30 mmol) and FeCl 2¨4 H 2 O (2.25 g, 11.3 mmol) were dissolved in 400 mL deionized water under nitrogen atmosphere via vigorous stirring (1000 rpm) at 80˝C. Then, a 20-mL ammonia solution 25% (w/w) was added to the solution. The color of the bulk solution immediately changed from orange to black. After stirring the mixture for 5 min, the Fe 3 O 4 NP precipitates were obtained via magnetic decantation and washed three times with deionized water. Finally, the Fe 3 O 4 NPs were dried in a vacuum oven at 70˝C for 12 h.
The surface of Fe 3 O 4 NPs was modified by being coated with the newly-designed modified thiophene monomers via oxidation polymerization with the generation of ferric cations on the Fe 3 O 4 NPs' surface [54]. Fe 3 O 4 NPs (1 mmol, 0.235 g) were discrete in polyvinyl alcohol (PVA) aqueous solution (0.001 M). Later, 3TArH (3.64 g, 10 mmol) was added into the mixed solution with vigorous stirring. Subsequently, 30 mL of HCl (0.5 M) solution were introduced into the mixture. Then, the products obtained were dried in a vacuum oven at 70˝C for 12 h. Experiments were repeated using freshly-distilled thiophene monomer (10 mmol, 0.84 g).

Solid Phase Extraction Optimization and Reusability Studies
Factors affecting the extraction efficiency of the proposed method, such as type of adsorbents, pH, extraction time, sample volume, elution solvent, elution solvent volume, desorption time, adsorbent dosage and effect of NaCl, were studied. All of the experiments were performed in triplicate, and the means of the results were used in plotting the optimization curves.
The reusability of the adsorbent was determined with optimized conditions for up to five cycles. The adsorbent was recycled after being washed with methanol and water and dried in vacuum at 70˝C for 12 h.

Analytical Performances and Real Sample Analysis
In order to evaluate the figures of merit of the proposed technique, linearity, the limit of detection (LOD), the limit of quantitation (LOQ) and repeatability were investigated under optimized conditions. The linearity was analyzed through the standard curves ranging from 0.1-50 µg¨L´1 by diluting appropriate amounts of phthalates stock solution (1000 mg¨L´1) with methanol and prepared in triplicate. The calibration curves were prepared using 10 spiking levels of analytes. For each level, three replicate experiments were performed.
To evaluate the reliability of the proposed method for the extraction of the plasticizers from real samples, two real samples were selected, spiked and subjected to the MSPE-GC-FID analysis. The two real samples were commercial bottled mineral water and bottled fresh milk. Figure 2 shows several additional peaks in the spectrum of nanocomposites, proportional to the MNP spectrum, which might be due to the surface functionalization. The strong absorption peaks in the range of~3400 cm´1 for MNP and all nanocomposites indicated the presence of OH vibration, while the peak at 530-632 cm´1 corresponds to Fe-O stretching modes [58]. The C-H aromatic stretching peak was observed for all nanocomposites, which falls at 3000 cm´1 for MNP@PTh and 2980 cm´1 for MNP@P3TArH. C-H sp 3 stretching (hexyl aliphatic side) occurred at 2934 cm´1 for MNP@P3TArH. Schiff base peaks (C=N) were observed at 1674 and 1685 cm´1 for MNP@P3TArH [59]. C=C aromatic symmetric and asymmetric absorption bands demonstrated in the range of 1573-1461 cm´1 occurred for both nanocomposites. Two absorption band peaks at 1250 and 1072 cm´1 indicated the presence of C-O in MNP@P3TArH. Hence, the FTIR study clearly revealed that the MNPs prepared have been successfully functionalized.

Analytical Performances and Real Sample Analysis
In order to evaluate the figures of merit of the proposed technique, linearity, the limit of detection (LOD), the limit of quantitation (LOQ) and repeatability were investigated under optimized conditions. The linearity was analyzed through the standard curves ranging from 0.1-50 µg·L −1 by diluting appropriate amounts of phthalates stock solution (1000 mg·L −1 ) with methanol and prepared in triplicate. The calibration curves were prepared using 10 spiking levels of analytes. For each level, three replicate experiments were performed.
To evaluate the reliability of the proposed method for the extraction of the plasticizers from real samples, two real samples were selected, spiked and subjected to the MSPE-GC-FID analysis. The two real samples were commercial bottled mineral water and bottled fresh milk. Figure 2 shows several additional peaks in the spectrum of nanocomposites, proportional to the MNP spectrum, which might be due to the surface functionalization. The strong absorption peaks in the range of ~3400 cm −1 for MNP and all nanocomposites indicated the presence of OH vibration, while the peak at 530-632 cm −1 corresponds to Fe-O stretching modes [58]. The C-H aromatic stretching peak was observed for all nanocomposites, which falls at 3000 cm −1 for MNP@PTh and 2980 cm −1 for MNP@P3TArH. C-H sp 3 stretching (hexyl aliphatic side) occurred at 2934 cm −1 for MNP@P3TArH. Schiff base peaks (C=N) were observed at 1674 and 1685 cm −1 for MNP@P3TArH [59]. C=C aromatic symmetric and asymmetric absorption bands demonstrated in the range of 1573-1461 cm −1 occurred for both nanocomposites. Two absorption band peaks at 1250 and 1072 cm −1 indicated the presence of C-O in MNP@P3TArH. Hence, the FTIR study clearly revealed that the MNPs prepared have been successfully functionalized.   (440)) [61]. This showed that the surface functionalization does not change the crystalline phase of MNPs [62].   (440)) [61]. This showed that the surface functionalization does not change the crystalline phase of MNPs [62]. The BET surface area is measured using the multipoint BET method, within the relative pressure (P/P0) range of 0.05-1. As described in ( Figure S5, Supplementary Material), the MNPs and all nanocomposites display an H3-type hysteresis loop, based on the Brunauer-Deming-Deming-Teller (BDDT) classification, demonstrating the existence of mesopores with pore diameters between 2 and 50 nm [63]. The pore size and BET surface area of MNPs and nanocomposites are tabulated in Table  1. The reduction in the pore size of nanocomposites is due to the addition of polymers on the surface. Meanwhile, escalation in the surface area could be because of the dispersity of particles that results from the enhancement of the spaces between them [64,65]. Morphological analysis of the synthesized products was performed using TEM techniques. As shown in Figure 4, TEM images of all materials demonstrated a sphere-shaped property. From the images, we could clearly observe the good dispersion of the functionalized nanoparticles (MNP@PTArH) in the TEM image. For instance, before polymerization, magnetic nanoparticles were highly agglomerated with each other. After polymerization of MNP with 3TArH, they showed lower agglomeration, and the nanocomposite became well dispersed. The dispersity of the nanocomposite influenced its surface area, as evidence by the BET result of MNP@P3TArH, which is higher compared to MNP@PTh and MNP, as tabulated in Table 1.  The BET surface area is measured using the multipoint BET method, within the relative pressure (P/P0) range of 0.05-1. As described in ( Figure S5, Supplementary Material), the MNPs and all nanocomposites display an H3-type hysteresis loop, based on the Brunauer-Deming-Deming-Teller (BDDT) classification, demonstrating the existence of mesopores with pore diameters between 2 and 50 nm [63]. The pore size and BET surface area of MNPs and nanocomposites are tabulated in Table 1. The reduction in the pore size of nanocomposites is due to the addition of polymers on the surface. Meanwhile, escalation in the surface area could be because of the dispersity of particles that results from the enhancement of the spaces between them [64,65]. Morphological analysis of the synthesized products was performed using TEM techniques. As shown in Figure 4, TEM images of all materials demonstrated a sphere-shaped property. From the images, we could clearly observe the good dispersion of the functionalized nanoparticles (MNP@PTArH) in the TEM image. For instance, before polymerization, magnetic nanoparticles were highly agglomerated with each other. After polymerization of MNP with 3TArH, they showed lower agglomeration, and the nanocomposite became well dispersed. The dispersity of the nanocomposite influenced its surface area, as evidence by the BET result of MNP@P3TArH, which is higher compared to MNP@PTh and MNP, as tabulated in Table 1. The BET surface area is measured using the multipoint BET method, within the relative pressure (P/P0) range of 0.05-1. As described in ( Figure S5, Supplementary Material), the MNPs and all nanocomposites display an H3-type hysteresis loop, based on the Brunauer-Deming-Deming-Teller (BDDT) classification, demonstrating the existence of mesopores with pore diameters between 2 and 50 nm [63]. The pore size and BET surface area of MNPs and nanocomposites are tabulated in Table  1. The reduction in the pore size of nanocomposites is due to the addition of polymers on the surface. Meanwhile, escalation in the surface area could be because of the dispersity of particles that results from the enhancement of the spaces between them [64,65]. Morphological analysis of the synthesized products was performed using TEM techniques. As shown in Figure 4, TEM images of all materials demonstrated a sphere-shaped property. From the images, we could clearly observe the good dispersion of the functionalized nanoparticles (MNP@PTArH) in the TEM image. For instance, before polymerization, magnetic nanoparticles were highly agglomerated with each other. After polymerization of MNP with 3TArH, they showed lower agglomeration, and the nanocomposite became well dispersed. The dispersity of the nanocomposite influenced its surface area, as evidence by the BET result of MNP@P3TArH, which is higher compared to MNP@PTh and MNP, as tabulated in Table 1.  The magnetic properties of the samples were recorded at room temperature with an external field of˘15 kOe. Important magnetic variables, such as saturation magnetization (M S ), were evaluated. The maximum saturation (M S ) of MNPs occurred at 69.2 emu¨g´1, respectively. After surface functionalization, the magnetization of MNP@PTh and MNP@P3TArH was reduced to 65.3 and 61.5 emu¨g´1 respectively. The magnetization decrease signified the presence of a dead magnetic layer on the surface of the nanocomposites [58]. Although the magnetization has declined, the value is still within the acceptable range, which suggests that it can be applied as the MSPE sorbent [66].

Type of Adsorbent
Hypothetically, the adsorption of phthalates is based on the hydrophobicity and π-π dispersion [67]. To prove that the structure architecture influences the adsorption studies of phthalates, three different types of sorbents, which are naked magnetic nanoparticles (MNP), MNP-PTh and MNP@P3TArH, were tested. As seen in Figure 5, MNP resulted in an insignificant peak area for all of the analytes studied. After the introduction of polythiophene derivatives on the surface of MNP, the peak area of phthalates increased. The presence of aliphatic and aromatic groups in the MNP@P3TArH enhances the dispersion of phthalates, which enhances the π-π dispersion and hydrophobic interaction. As evidenced, butyl benzyl phthalate (BBP) is more prone to the adsorbent with more aromatic sides, as in the MNP@PT3ArH, compared to the other adsorbents. Besides, the high surface area of MNP@P3TArH also contributes to the increase of extraction performance. Since the MNP@P3TArH has demonstrated the high peak area for all analytes studied, it was selected for further MSPE optimization.

Sample pH
To study the influence of the surface charge of adsorbent/adsorbate in the extraction process, experiments were performed under different pH conditions, ranging from pH 2-9. As shown from Figure 6a, the peak areas for phthalates increase when the pH rise from 2-7, but decline later from 8-9. At low pH, C=N, alkoxy in P3TArH was protonated, making the adsorbent surface positively charged. At pH < 7, phthalates hydrolyze to phthalic acid, thus making the carbonyl group nucleophilic, reacting with hydrogen ions in the aqueous solution, producing positive charges. Due to both the absorbate and adsorbent acquiring positive charges, the electrostatic repulsion occurred and retarded the adsorption performance [68]. At basic conditions, the surface adsorbent became negatively charged, while the adsorbate hydrolyzes to phthalate anions, reducing the extraction efficiency [69]. Thus, in neutral pH, the extraction increased due to the absence of electrostatic repulsion that disturbed the extraction capability. As the optimum performance was demonstrated at pH 7, this pH was selected for all of the experiments.

Extraction Time
It has been understood that prolonged extraction time might increase the recovery of analytes. Thus, the influence of extraction time on the recoveries of the analyte has been investigated. As demonstrated in Figure 6b, the peak area increased rapidly for the first 20 min, since more adsorption sites were available and phthalates could easily interact with these sites. After 30 min, the peak area was almost persistent; therefore, 30 min was sufficient to extract the maximum of the target analytes. In order to ensure that the extraction time was satisfactory, further experiments were carried out until 90 min, and they were found to be constant.

Desorption Studies
The elution solvent is one of the crucial parameters to be considered. In order to determine the best elution solvent, the solvent must be able to elute all of the analytes that were retained from the adsorbent in a small volume [70]. Six eluting solvents with dissimilar polarities, namely hexane, toluene, diethyl ether, acetonitrile, methanol and ethyl acetate, were studied.
As evidenced in Figure 7a, polar solvents (acetonitrile, methanol and ethyl acetate) were the best solvents, with high peak areas compared to non-polar solvents (hexane, toluene and diethyl ether), since phthalates contain a polar carbonyl group [71]. Among the polar solvents, ethyl acetate showed high solvent strength, since it gave the maximum peak area for the phthalates studied and was thus selected to be the eluent. The volume of ethyl acetate was tested from 0.1 mL-2.5 mL. As observed in Figure 7b, the peak area increased from 0.1 mL and remained constant after 0.5 mL. This showed that 0.5 mL may accommodate the maximum phthalates extracted from the sorbent. Further, desorption time was optimized to investigate the best time taken for the analytes to desorb from the sorbent ranging from 0-12 min. As revealed in Figure 7c, analytes were desorbed rapidly in the first 4 min and started to become linear after 10 min. This indicated that 10 min of time are sufficient to desorb back all of the analytes from the adsorbent. As for the case of BBP, desorption was found to be slower than other phthalates. This could be due to the presence of an additional aromatic ring in BBP, which makes it less polar to the eluent (ethyl acetate). After 6 min of desorption, most of the phthalates had reached near to equilibrium, whereas BBP was desorbed steeply after 6 min until it reached equilibrium at 10 min.

Mass of Adsorbent
Investigation of the adsorbent amount was executed in the range of 1-25 mg. As exposed in Figure 8a, the extraction peak area increased up to 10 mg, but decreased later with a further increase of the adsorbent. Increasing the adsorbent amount provides more active sites for the adsorption of target analytes. However, a high amount of adsorbent at a specific volume has weakened elution efficiency [30]. It is shown that this adsorbent only required a small amount of adsorbent to remove phthalates efficiently, which added the advantage of economic value. Therefore, for further experiments, the adsorbent amount of 10 mg was applied.

Sample Loading Volume
The effect of sample volume was investigated by the extraction of the phthalates ranging from 5-100 mL and shown in Figure 8b. Each sample was spiked with 10 mg¨L´1 analytes and 10 mg adsorbent. As can be seen, peak area increased until 20 mL and further decreased till 100 mL. A 20-mL volume of sample demonstrated the most efficient extraction. An increase in sample volume could lead to a high distribution of adsorbent to the aqueous phase, which lowered the amount of adsorbent in the volume unit sample solution, and the extraction became less effective [72]. Thus, a 20-mL sample volume was chosen as the optimized sample volume.

Effect of NaCl
Indeed, the addition of salt in the sample matrices effects the extraction efficiency. Thus, studies on the concentration of NaCl ranging from 0-25 g¨L´1 were conducted. As observed in Figure 8c, peak areas of the studied analytes increased from 0-15 g¨L´1, but decreased later from 20-25 g¨L´1. This can be due to the addition of salt, which increases the ionic strength and eventually decreases the solubility of the analytes in the media. However, as the concentration of salt increases, the diffusion rate of the analytes may reduce, since the solvation cage of the analytes is disturbed [51]. Since a 30 g¨L´1 NaCl concentration gave a high peak area for all analytes studied, it was chosen for subsequent experiments.

Reusability Studies
To investigate the probability of reusing and regenerating the sorbent, a reusability test was designed and implemented for Fe 3 O 4 @P3TArH, which was recycled after being washed with methanol and water and was dried in a vacuum at 70˝C for 12 h. From Figure 9, it could be surmised that after five repeated experiments, the adsorbent was still active. This may be due to some of the particles in the adsorbent accumulating due to the heat treatment after several cycles, which decreases the surface area.

Analytical Performances and Real Sample Analysis
The optimized method obtained for the extraction of phthalates using MNP@P3TARH involved the sample at pH 7, 30 min extraction time, ethyl acetate as the elution solvent, 500-µL elution solvent volumes, 10 min desorption time, 10 mg adsorbent dosage, 20-mL sample loading volume and a 15 g¨L´1 concentration of NaCl. In order to assess the validation of the proposed method, linearity, the limit of detection, the limit of quantitation and repeatability were performed under optimum conditions. Analytical performance figures of merits are tabulated in Table 2. Calibration curves obtained for the studied phthalates were linear over the range of 0.1-50 µg¨L´1 with R 2 more than 0.99. As per the U.S. EPA standard, the screening of phthalates in drinking water must be done at a concentration above 0.6 µg¨L´1 [4]. However, the LOD of our method lies within the range of 0.080-0.468, indicating the suitability of this method as an efficient phthalate detector.
Repeatability studies were conducted for inter-day (three consecutives replicates for three days) and intra-day (seven consecutives replicates on the same day). The results were expressed as relative standard deviations (RSD%). This method demonstrated good precision, since the RSD (%) values were in the range of 3%-5% [73]. Comparative studies on the analytical performance between the proposed methods with other developed methods are shown in Table 3. Obviously, the extraction of phthalates using MNP@P3TArH provides sensitivity and repeatability. To endorse the reliability of the method using MNP@P3TArH, it was applied to determine phthalates in the water from the mineral water bottle stored at room temperature and commercial fresh milk. Figure 10 shows the chromatogram of commercial fresh milk unspiked and spiked with phthalates. None of the targeted phthalates were found in the water samples under the optimized condition described. To evaluate the matrix effect, all of the samples were spiked with 50 µg¨L´1 of the phthalates studied. Recoveries and RSD (%) for all of the water samples were determined and are shown in Table 4. From the optimization procedures until the real sample analyses, DMP, DEP and DPP demonstrated lower recoveries; this may be due to the lower molecular weight of phthalates being more prone to aqueous solution than to the adsorbent [76]. From the chromatogram of mineral bottle stored at room temperature as shown in ( Figure S6, Supplementary Material), the recoveries obtained for water in the mineral bottle demonstrated higher values compared to the recovery for the milk sample. This might be caused by the matrix effect that holds the analyte in the milk sample to be higher compared to the water sample. RSD (%) values were found to be in the range of 1.3%-5.8%, which indicated a precise method.

DBP, DEHP
MNP@Zeolite-GC-FID 2.80-3. To endorse the reliability of the method using MNP@P3TArH, it was applied to determine phthalates in the water from the mineral water bottle stored at room temperature and commercial fresh milk. Figure 10 shows the chromatogram of commercial fresh milk unspiked and spiked with phthalates. None of the targeted phthalates were found in the water samples under the optimized condition described. To evaluate the matrix effect, all of the samples were spiked with 50 µg·L −1 of the phthalates studied. Recoveries and RSD (%) for all of the water samples were determined and are shown in Table 4. From the optimization procedures until the real sample analyses, DMP, DEP and DPP demonstrated lower recoveries; this may be due to the lower molecular weight of phthalates being more prone to aqueous solution than to the adsorbent [76]. From the chromatogram of mineral bottle stored at room temperature as shown in (Figure S6, Supplementary Material), the recoveries obtained for water in the mineral bottle demonstrated higher values compared to the recovery for the milk sample. This might be caused by the matrix effect that holds the analyte in the milk sample to be higher compared to the water sample. RSD (%) values were found to be in the range of 1.3%-5.8%, which indicated a precise method.

Conclusions
MNP@P3TArH has been successfully synthesized, characterized and utilized as a sorbent for the analysis of GC-FID in the determination of selected phthalates. The optimized conditions of MSPE were carefully selected as follows: sample at pH 7, 30 min extraction time, ethyl acetate as the elution solvent, 500-µL elution solvent volume, 10 min desorption time, 10 mg adsorbent dosage, 20-mL sample loading volume and a 15 g¨L´1 concentration of NaCl. The steadiness and reusability studies suggested that the MNP@P3TArH could be used up to five cycles without significantly impacting its extraction capacity. The adsorbent covers a wide range of phthalates with a dynamic linear range of 0.1-50 µg¨L´1 and a limit of detection at 0.08-0.468 µg¨L´1. The presence of new interfaces (π-π and hydrophobic interactions) among the sorbent and target analytes increased the adsorption capability. The application of MNP@P3TArH as the MSPE sorbent was successfully executed by the analysis of phthalate esters in the mineral water and commercial fresh milk.