Silicon dioxide–poly(dimethylsiloxane) with a bilayer structure, incorporating multi-walled carbon nanotubes, supported on stainless steel wire as a solid-phase microextraction fiber for the determination of trace phthalate esters in drinking water samples

Abstract

Silicon dioxide–poly(dimethylsiloxane) with a bilayer structure, incorporating multi-walled carbon nanotubes, supported on stainless steel wire were prepared by a sol–gel method. This was used as a novel solid-phase microextraction (SPME) fiber in the determination of four phthalate esters (i.e., dimethyl phthalate (DMP), diethyl phthalate (DEP), dibutyl phthalate (DBP), di-2-ethylhexyl phthalate (DEHP)) in drinking water samples. The reaction mechanism of the sol–gel coating process was discussed and confirmed by IR spectra and scanning electron microscopy (SEM). The SPME experimental conditions such as the desorption temperature and time, extraction temperature and time, agitation and the salt effect were optimized. Compared with commercial fibers (polyacrylate (PA), polydimethylsiloxane (PDMS) and divinylbenzene–carboxen–polydimethylsiloxane (DVB–CAR–PDMS)), the novel fiber exhibited better film forming ability, higher thermal stability, higher extraction efficiency, and longer life time. Under the optimized conditions, the detected signals showed good linearity with the concentrations of DMP and DBP in the range from 0.1 to 100 μg L⁻¹, and those of DEP and DEHP from 0.1 to 300 μg L⁻¹. Based on a signal to noise ratio (S/N) of 3, the detection limits of DMP, DEP, DBP and DEHP were estimated to be 0.01, 0.02, 0.01 and 0.02 μg L⁻¹, respectively. All calibration curves were found to have good linearity with the correlation coefficients (r²) which were greater than 0.998. The precision (RSD) of the system, measured by six repeated determinations of the analytes at 10 μg L⁻¹ were in the range of 5.67–12.2%. The recoveries of real samples ranged from 85.27% to 109.3%.

---

1. Introduction

Phthalate esters (PAEs) have a wide variety of applications in industrial, agricultural and domestic fields, but by far the most important is their use as plasticizers to improve the flexibility and workability of polymeric materials. But the excessive use of PAEs presents a potential risk to the environment, food products and human health.^1,2 It has been reported that some PAEs are suspected carcinogens that may affect the male reproductive system and that PAEs exercised acute or chronic toxicity toward aquatic organisms.^3,4 Hence, it is necessary to develop sensitive, rapid and simple analytical methods for the purpose of fast monitoring of PAEs in environmental samples and for accomplishing risk assessments.

The development of the analytical procedures to determine the presence of PAEs in environmental, biological, and human samples has received much attention in recent years.^5–10 The most widely used methods are gas chromatography (GC) techniques with mass spectrometry (MS) or flame ionization detection (FID). A sample pretreatment step prior to chromatographic analysis is usually necessary. Historically, liquid–liquid extraction (LLE) and solid-phase extraction (SPE) were often used for the extraction of PAEs from aqueous matrices.^11–13 However, LLE is generally time-consuming and labor-intensive, and requires large quantities of expensive, toxic, and environmentally unfriendly organic solvents. Although SPE needs less solvent, it is also a time-consuming multi-step process, and the enrichment efficiency of SPE is relatively low. Therefore, nowadays the simplification and miniaturization of the sample preparation methods is recommended, which has the advantage of using either no or very little amounts of toxic organic solvents.

Since the introduction of solid-phase microextraction (SPME) by Pawliszyn in 1989,¹⁴ SPME has been developed rapidly and it has been widely applied in sample analysis because it integrates sampling, extraction, concentration and sample introduction into a single solvent-free step. SPME is also portable, sensitive and convenient to couple with various analytical instruments, such as GC. Several studies employing direct SPME for extraction of PAEs from samples have been reported and SPME has proved to be an excellent method for determination of PAEs in environmental waters,¹⁵ cow’s milk,¹⁶ soil samples,¹⁷ and foods.¹⁸

The extraction efficiency of SPME is mainly determined by the distribution of analytes between the matrix and the fiber coating. Therefore, the development of fiber coatings with highly efficient extraction of the analytes is an important research direction in SPME. Sol–gel technology has been widely used to prepare fiber coatings because it can effectively create chemically bonded, porous, and highly cross-linked coatings on the fused-silica fiber surface.¹⁹ Sol–gel-based SPME fibers were first introduced by Malik et al.²⁰ to improve the performance of traditional fibers. They used methyltrimethoxysilane (MTMOS) as a precursor for preparing the chemically bonded PDMS fiber. Since then, sol–gel technology used to prepare SPME fibers has fast become a developing field.^21,22 Generally, methyltrimethoxysilane (MTMOS) or tetraethoxysilane (TEOS) is used as a precursor, and a compound containing a specific functional group is added into the sol–gel solution to prepare SPME fiber coatings for various needs.

Carbon nanotubes (CNTs), first found in 1991 by Iijima,²³ have a strong physical adsorption ability for hydrophobic organic pollutants and have been successfully used in coatings for SPME. Eshaghi et al.²⁴ reported a design of SPME fibers containing a CNT reinforced sol–gel which was protected by polypropylene hollow fibers (HF-SPME) for pre-concentration and determination of aromatic compounds in environmental waste water and human hair samples. Yan et al.²⁵ reported a SPME method using multi-walled carbon nanotubes (MWNTs) as the fiber coating for determination of polybrominated diphenyl ethers (PBDEs) in environmental samples. Zhang's group²⁶ used polymer-functionalized single-walled carbon nanotubes (SWNTs) as SPME coated fibers for determination of PBDEs in water samples. Rastkari et al.²⁷ used SWNTs as a SPME adsorbent for the determination of low-level concentrations of butyltin compounds in seawater. All the advantages of these methods are mainly due to the incorporation of CNTs, which enhances the π–π interaction with the analytes and increases the surface area of extraction in contact with the sample.

The choice of the coating for the extraction support material also greatly affects the efficiency of SPME. Fused-silica fiber has been applied widely as the coating for the extraction support material due to its characteristics of good toughness and the convenience of chemical modification in the sol–gel-based SPME technique. However, the lifetime of SPME fibers is frequently shortened by the breaking of the fragile silica fibers, no matter what coating was used. In order to overcome the above drawback, flexible and non-fragile stainless steel wire should be used to replace the fused-silica fiber, used as the support material of SPME fibers.^28–31 Unfortunately, the adhesion ability of the sol–gel-based extraction coatings on the pristine stainless steel wire surface is much lower than that on the fused-silica fiber surface, which greatly decreases the stability and the lifetime of the coating.

In this work, a novel SPME fiber supported on stainless steel wire was prepared by a sol–gel method and a SiO₂–PDMS-MWNTs coating, with a bilayer structure, was used as the extraction phase. In this strategy, the bilayer structure coating was designed to enhance the stability and durability of the novel fiber. To improve the adhesion of the PDMS-MWNTs coating on the stainless steel wire surface, a SiO₂ layer with a perfect Si–O network structure was coated on the surface of stainless steel wire via a sol–gel method prior to the coating of the PDMS-MWNTs. In addition, the sensitivity of the extraction phase could be increased due to the incorporation of MWNTs. The sol–gel coating process was confirmed by IR spectroscopy and SEM. Finally, the SiO₂–PDMS-MWNTs fiber was successfully applied for the determination of trace PAEs (DMP, DEP, DBP, and DEHP) in drinking water samples. Compared with commercial SPME fibers, the new coating exhibited better film stability, higher thermal stability, higher extraction efficiencies, and longer life times.

2. Experimental

2.1 Chemicals and materials

MWNTs (diam. 40–60 nm, length 5–15 μm) were purchased from the Shenzhen Nanotech Port Co., Ltd. (Shenzhen, China). Hydroxyl terminated silicone oil (TSO-OH) was purchased from Aldrich (Allentown, PA). Poly(methylhydrosiloxane) (PMHS) was purchased from the Chemical Plant of Wuhan University (Wuhan, China). The PAE standards of DMP, DEP, DBP, and DEHP were obtained from the Institute for Reference Materials of SEPA (Beijing, China). Stock standard solutions of PAEs were prepared in methanol at a concentration of 1 g L⁻¹ and stored in glass-stoppered bottles in the dark at 4 °C. Working solutions were prepared by dilution of the stock in the appropriate solvent. All the other chemicals were of analytical grade and used as received. Pure water was obtained from a Simplicity Personal Ultrapure Water System (Millipore, USA).

2.2 Instrumentation

Experiments were carried out on a Shimadzu GC-14C gas chromatograph equipped with a split–splitless injector and using the technique of flame ionization detection (FID) (Shimadzu, Tokyo, Japan). A Rtx-50 fused-silica column (30 m × 0.25 mm I.D.) with 50% phenyl and 50% methyl polysiloxane (film thickness 0.25 μm) (Restek, USA) was used. High purity nitrogen was used as carrier gas (1.0 mL min⁻¹) and make-up gas (30 mL min⁻¹). The instrument temperatures were as follows: injector temperature 280 °C; detector temperature 280 °C; initial oven temperature 80 °C for 1 min, increased to 160 °C (held for 3 min) at a rate of 20 °C min⁻¹, then increased to 280 °C at a rate of 10 °C min⁻¹, and held for 5 min. Hydrogen and air were used as carrier gases at 48 mL min⁻¹ and 450 mL min⁻¹, respectively. The inlet was operated in splitless mode. A Clarity Shimadzu was utilized to control the system and to acquire the analytical data.

2.3 SPME device

The SPME device was adapted from a used microsyringe. An inner needle of a 1 μL microsyringe, with the stainless steel wire present, was inserted into an outer needle of a 5 μL microsyringe. The wire pierced a piece of silicone rubber and was transferred into the hollow outer needle. The silicone rubber provided an effective seal during the movement of the wire. The outer needle was sealed at the top end of the wire to make sure that the lower side of the wire can be moved 2 cm out of the stainless steel needle for coating and extraction.

2.4 Preprocessing of MWNTs

Similar to the approach of Liu et al.,³² received MWNTs were further cut into short pipes by chemical oxidation in a mixture of concentrated sulfuric and nitric acid (3 : 1, v/v) under ultrasonication at 70 °C for 4 h. The reaction mixture was then diluted with water and allowed to stand overnight for precipitation. The supernatant was decanted, and the remains were diluted with pure water and filtered with a 0.22 μm pore diameter polytetrafluoroethylene (PTFE) membrane. Solid MWNTs were obtained by washing the remains on the PTFE filter with pure water until the filtrate pH became nearly neutral. The preprocessed-MWNTs were dried at 80 °C in a vacuum oven, and then powdered in a mortar.

2.5 Preparation of SPME fiber

Prior to coating, the stainless steel wire was cleaned with acetone and methanol in an ultrasonicator for 5 min, and then washed using distilled water, and finally air dried. The coating process was divided into two steps as shown in Fig. 1: first, the SiO₂ sol was prepared by vortexing a mixture of 200 μL of TEOS, 1 mL of ethyl alcohol, and 25 μL of 6% HCl for 2 min in a silanized glass tube. The treated wire was dipped vertically into the sol solution with a depth of 1.5 cm for 1 h and then taken out. This was repeated several times until the desired thickness of the coating was obtained. Afterwards, the wire was placed in a desiccator at room temperature for 1 h and then conditioned by placing it in a GC injector kept at a temperature of 100 °C for 3 h. The SiO₂ coating was obtained in this step. Second, 200 mg of preprocessed-MWNTs was put into 1 mL of ethyl alcohol and dispersed for 20 min by ultrasonic agitation, and then 200 μL of TEOS, 55 μL of TSO-OH, 25 μL of PMHS were added and mixed thoroughly for 20 min by ultrasonic agitation. 180 μL of TFA (95%) was sequentially added to the resulting solution with ultrasonic agitation for another 5 min. The SiO₂ fiber was dipped vertically into the sol solution to a depth of 1.5 cm and held for 1 h until a sol–gel coating was formed on the surface of the wire end. The coating process was repeated five times, and each time was performed in a freshly prepared sol solution. The sol–gel fiber was placed in a desiccator at room temperature for 24 h and the SiO₂–PDMS-MWNTs fiber was obtained. Prior to use, the SiO₂–PDMS-MWNTs fiber was conditioned in the injector of a GC (1 h at 100 °C and then 3 h at 280 °C) under 1 mL min⁻¹ of N₂. The remaining sol from the fiber coating processes was allowed to gel, was then ground, washed with distilled water and methanol, and conditioned under the same conditions as the fibers. The resulting xerogel was submitted to IR analysis.

Fig. 1 Schematic representation of the SiO₂–PDMS-MWNTs fiber preparation.

2.6 SPME procedure

A 20 mL glass vial was used as a sample container, and 15 mL of standard solution or sample extract was placed into the sample vial with a 1 cm spin bar. To prevent analyte evaporation, the vial was sealed with a septum. All SPME extraction was performed by direct immersion of the fiber in the sample for 30 min at 40 °C. The samples were continuously stirred at a constant speed (i.e. 300 rpm) with a magnetic stirrer. Finally, the fiber was introduced into the GC inlet and desorbed at 280 °C for 2 min.

3. Results and discussion

3.1 Optimization of the coating procedure

The coating procedure was optimized to obtain the better repeatability and higher sensitivity. To evaluate the effectiveness of SiO₂ coating on the fiber, PDMS-MWNTs coating was directly coated on the stainless steel wire using the same preparation process as the SiO₂–PDMS-MWNTs fiber and its extraction efficiency was investigated. The result shows that the PDMS-MWNTs coating (without SiO₂ coating between the stainless steel wire and the PDMS-MWNTs coating) is relatively thin, which makes the amount of analytes that it can extract relatively low. Moreover, its lifetime is significantly shorter than that of SiO₂–PDMS-MWNTs fiber.

To evaluate the effect of the MWNTs in the coating, a series of SiO₂–PDMS fibers that incorporated different amounts of MWNTs were prepared and their extraction amounts were investigated. The results show that the extraction efficiency increases with an increase of the amount of MWNTs (Fig. 2). The reason is mainly due to the incorporation of MWNTs, which enhances the π–π interaction with analytes and increases the surface area of extraction in contact with the sample. Unfortunately, when the amount of MWNTs is above 200 mg, the lifespan of the fiber obviously decreases due to the fracture of the fiber coating. Therefore, 200 mg of MWNTs is adopted in the following studies.

Fig. 2 Effect of the amount of MWNTs on extraction efficiency. Extraction conditions: sample, 15 mL of PAEs mixed standard solution (10.00 μg L⁻¹); no added salt; extraction time, 30 min; extraction temperature, 40 °C; stirring rate, 300 rpm; desorption time, 2 min; and desorption temperature, 280 °C. Peak identification: (■) DMP; (●) DEP; (▲) DBP; (▼) DEHP. Error bars show the standard deviation of the mean (n = 3).

3.2 Characterisation of SiO₂–PDMS-MWNTs

3.2.1 FTIR analysis. In order to verify the structure of the SiO₂–PDMS-MWNTs coating, FTIR spectra of the preprocessed-MWNTs, PDMS and PDMS-MWNTs were obtained. As shown in Fig. 3, the peaks at 2715 and 2807 cm⁻¹ are characteristic of C–H stretching, and the absorbance at 1629 cm⁻¹ was assigned to a C=C stretch vibration, implying that the MWNTs were successfully oxidized. Meanwhile, new adsorption peaks at 2964, 2174, 1260, and 1092 cm⁻¹ are found in the IR spectrum of PDMS-MWNTs (Fig. 3c), which can be identified as the characteristic peaks of PDMS, similar to those shown in Fig. 3b. This suggests that the polymer is grafted to the surfaces of the MWNTs.

Fig. 3 FTIR spectra. (a) preprocessed-MWNTs; (b) PDMS polymer; (c) PDMS-MWNTs coating.

3.2.2 SEM images. The surface of the coated fiber was characterized by SEM. It can be observed that a coating uniformly covers the fiber surface in Fig. 4, indicative of the success of the synthetic strategy. A lot of MWNTs are observed on the folds of the PDMS-MWNTs coating, meaning that the MWNTs are cross-linked and homogeneously distributed within the coating.

Fig. 4 SEM images of the SiO₂–PDMS-MWNTs fiber.

3.3 Lifespan of the coating

The coating's lifespan is important for practical application. The SiO₂–PDMS-MWNTs coating's lifespan was studied by monitoring the change of extraction peak areas of the PAEs after it had been used for 10, 50, 100, and 150 times. No obvious decline is observed (Fig. 5), indicating that the SiO₂–PDMS-MWNTs fiber is stable and can be used at least 150 times, this is obviously larger than the lifespan (50–100 times) of all commercial fibers. Such a long lifespan results from the presence of the SiO₂ part of the bilayer structured coating and also from the strong chemical bonding between the sol–gel generated organic–inorganic composite coatings.

Fig. 5 Lifespan profile of the SiO₂–PDMS-MWNTs fiber. Extraction conditions: the same as Fig. 2. Error bars show the standard deviation of the mean (n = 3).

3.4 Optimization of SPME procedures

3.4.1 Desorption time and temperature. To reach the highest sensitivity and to avoid carry over, desorption time and temperature were evaluated to ensure that the analytes were completely desorbed from the fiber. With the desorption temperature increasing from 200 to 280 °C, the peak areas of all the four analytes studied remain unchanged, but the tailing phenomenon appears with the peak when the desorption temperature is 200 °C. Therefore, the desorption temperature was set at 280 °C. The effect of desorption time on the chromatographic peak areas of the analytes was investigated at 0.5, 1, 2, 3, and 4 min under the desorption temperature of 280 °C (Fig. 6). It was found that the peak areas of the analytes slightly increase as desorption time increases from 0.5 to 2 min and then remain unchanged with a further increase in the desorption time to 4 min. The repeated desorptions with desorption time of 2 min indicates no carry over effect. Thus desorption time was set at 2 min in the subsequent experiments.

Fig. 6 Effect of desorption time on extraction efficiency. Extraction conditions: the same as Fig. 2, except desorption time.

3.4.2 Extraction temperature. For SPME, optimization of extraction temperature is generally more important. Increasing the temperature can significantly increase the analytes’ molecular thermodynamic movement. It is beneficial for the analytes to enter the fiber coating, resulting in an increase of the extraction amount on the fiber and a reduction of the equilibrium time. On the other hand, higher temperature increases the partition coefficients of the analytes in the fiber coating, thereby decreasing extraction efficiencies. The effect of temperature on extraction of the analytes was studied in the range 0–80 °C. Fig. 7 clearly shows that the peak area increases with temperature in the range of 0–40 °C and then decreases at higher temperatures. For further experiments, an extraction temperature of 40 °C was chosen.

Fig. 7 Effect of extraction temperature on extraction efficiency. Extraction conditions: the same as Fig. 2, except extraction temperature.

3.4.3 Extraction time. The extraction time is an important parameter for the extraction performance. To obtain extraction equilibrium in a reasonable analysis time, the extraction time was optimized in the range of 10–60 min. The results illustrated in Fig. 8 show that the peak areas of the analytes increase significantly in the range 10–30 min and then slightly decrease with prolonged extraction time. Therefore, extraction time of 30 min is adopted in the following studies.

Fig. 8 Effect of extraction time on extraction efficiency. Extraction conditions: the same as Fig. 2, except extraction time.

3.4.4 Stirring rate. In the extraction, an increase in the stirring rate reduced the thermodynamic equilibrium time and increased extraction efficiency since a suitable stirring rate enables a continuous exposure of the extraction surface to fresh aqueous sample. However, overstirring would make the solution vortex, resulting in an adverse effect on the immersion of the fiber. To evaluate the effect of sample stirring, the analytes were extracted in the fiber for 30 min at different stirring rates (0–600 rpm). The results shown in Fig. 9 reveal that the chromatographic peak areas of the analytes increase as the stirring rate increased from 0 to 300 rpm and then decrease at higher stirring rates. This means that the stirring speed of 300 rpm was suitable in this work.

Fig. 9 Effect of stirring rate on extraction efficiency. Extraction conditions: the same as Fig. 2, except stirring rate.

3.4.5 Salt effect. The influence of salt on the studied system was investigated by adding various amounts of NaCl in a series of concentrations (0, 5, 10, 15%, w/v) and the results are shown in Fig. 10. It was found that the ionic strength of sample solution had a negative effect on the extraction performance. The chromatographic peak areas of the analytes decrease with an increase in NaCl content in the tested range. To further study the effect of the addition of salt, the extraction was performed in the presence of different concentrations of other salts (such as CaCO₃, FeCl₃ and MgCl₂). It was found that the chromatographic peak areas for all the analytes decrease continuously when increasing salt concentration. So, in the following experiments, no salt was added to the samples.

Fig. 10 Effect of salt concentration on extraction efficiency. Extraction conditions: the same as Fig. 2, except salt concentration.

3.5 Comparison of the SPME efficiency of the SiO₂–PDMS-MWNTs fiber with commercial fibers

To evaluate the characterization of the newly developed SPME fiber, the extraction efficiencies of the SiO–PDMS-MWNTs fiber for PAEs were compared with those of commercial fibers: PA (85 μm polyacrylate), PDMS (100 μm polydimethylsiloxane) and DVB–CAR–PDMS (50/30 μm divinylbenzene–carboxen–polydimethylsiloxane). The results in Fig. 11 reveal that the SiO₂–PDMS-MWNTs fiber has a higher extraction efficiency for analytes than other commercial fibers. It is possibly because the carbon atoms in the walls of the carbon nanotubes possess mixed sp² and sp³ hybridization and therefore there is a highly delocalized conjugated π-electron system. This enhances the π–π interaction with the aromatic groups of the PAEs. On the other hand, the large surface area of the MWNTs with their curved surface also have a stronger binding affinity for hydrophobic molecules compared with that of a planar carbon surface.

Fig. 11 Comparison of extraction efficiency using the SiO₂–PDMS-MWNTs fiber and three commercial fibers. Extraction conditions: the same as Fig. 2. Error bars show the standard deviation of the mean (n = 3).

3.6 Quantitative analysis

Based on the optimal conditions, several experiments were carried out in order to determine analytical characteristics such as linear range, precision, reproducibility and detection limits of the method. As shown in Table 1, the calibration curves are linear from 0.1 to 100 μg L⁻¹ for DMP and DBP, and from 0.1 to 300 μg L⁻¹ for DEP and DEHP. Based on S/N = 3, the detection limits of DMP, DEP, DBP and DEHP are estimated to be 0.01, 0.02, 0.01 and 0.02 μg L⁻¹, respectively. All calibration curves were found to have good linearity with the correlation coefficients (r²) which are greater than 0.998.

Table 1 The linear equation, correlation coefficients, linear range and LOD of the method

Analyte	Regression equation^a	Correlation coefficient (r^2)	Linear range (μg L^−1)	LOD (μg L^−1)
a y: Peak area; x: concentration of analyte (μg L^−1).
DMP	y = 23.64x + 6.965	0.9989	0.1–100	0.01
DEP	y = 8.932x + 16.72	0.9991	0.1–300	0.02
DBP	y = 25.99x + 58.91	0.9993	0.1–100	0.01
DEHP	y = 11.99x + 49.53	0.9987	0.1–300	0.02

---

3.7 Sample analysis

The proposed method was used to quantify four PAEs in drinking water. The real drinking water samples were obtained from the Jiulong River of Zhangzhou, commercially bottled mineral water of a certain brand in the market and from tap water in a school. The SPME was operated under the optimum conditions. The detection results of PAEs in water samples are shown in Table 2. These results indicate that no PAEs compounds were detected in all three samples except for DBP and DEHP in the commercially bottled mineral water at concentrations of 5.26 and 8.47 μg L⁻¹. To further evaluate the credibility of the proposed SiO₂–PDMS-MWNTs fiber, these samples were analyzed by LLE according to ref. 11. Table 2 demonstrates that the results obtained by these two methods are in very good agreement, which indicates the reliability of this fiber for determination of PAEs in drinking water samples. Recoveries obtained at a 10 μg L⁻¹ spiking level are in the range of 79.62–109.3% for PAEs. The RSD from 5.67 to 12.2% indicates that the proposed method is feasible for the determination of PAEs in real drinking water samples. The chromatograms of PAEs in the commercially bottled mineral water sample before and after spiking with 10.00 μg L⁻¹ of PAEs mixed standard solution are shown in Fig. 12.

Table 2 Determination of PAEs in real samples (n = 6)

Sample analyte	Jiulong River of Zhangzhou				Commercially bottled mineral water of a certain brand				Tap water in a school
	This method^a (μg L^−1)	Recovery^b (%)	RSD (%)	LLE^c (μg L^−1)	This method (μg L^−1)	Recovery (%)	RSD (%)	LLE (μg L^−1)	This method (μg L^−1)	Recovery (%)	RSD (%)	LLE (μg L^−1)
a Determined with this method.b Determined with LLE according to ref. 11.c Recoveries determined with SiO2–PDMS-MWNTs fiber at spiked levels of 10 μg L^−1.d Not detected.
DMP	ND^d	95.26	6.03	ND	ND	85.91	8.82	ND	ND	104.2	7.21	ND
DEP	ND	85.27	8.42	ND	ND	79.62	11.4	ND	ND	97.26	8.24	ND
DBP	ND	107.8	7.25	ND	5.26	98.25	9.47	5.51	ND	89.71	5.67	ND
DEHP	ND	102.9	11.2	ND	8.47	109.3	12.2	8.92	ND	106.6	10.1	ND

---

Fig. 12 Chromatograms of PAEs obtained by the developed method for (a) the commercially bottled mineral water sample of a certain brand and (b) spiked with 10.00 μg L⁻¹ of PAEs mixed standard solution. Peak identification: (1) DMP; (2) DEP; (3) DBP; (4) DEHP. Extraction conditions: the same as Fig. 2.

4. Conclusion

In summary, we coated a stainless steel wire with a SiO₂–poly(dimethylsiloxane) multi-walled carbon nanotubes incorporated complex membrane (SiO₂–PDMS-MWNTs) to construct a SPME fiber. The novel SiO₂–PDMS-MWNTs fiber has been applied to determine trace PAEs in real drinking water samples. Compared with commercial SPME fibers such as 85 μm PA, 100 μm PDMS and 50/30 μm DVB–CAR–PDMS, this novel fiber showed higher extraction efficiency and selectivity for PAEs compounds. The main features of the fiber were its high extraction efficiency, low cost, good reproducibility, long lifespan, good thermal, chemical and mechanical stability. This proposed method possesses great potential in the analysis of trace compounds in real aqueous samples.

Acknowledgements

This research was financially supported by the National Nature Scientific Foundation of China (no. 21105088 and no. 21105123), the Natural Science Foundation of Fujian Province, China (no. 2013J01062) and the Program for Science and Technology Projects of the Education Department of Fujian Province, China (no. JA12206), which are gratefully acknowledged. Furthermore, we would like to extend our thanks to Dr Yidong Hou of Fuzhou University for his assistance with English.

References

1. G. D. Bella, M. Saitta, M. Pellegrino, F. Salvo and G. Dugo, J. Agric. Food Chem., 1999, 47, 1009–1012.
2. Y. Guo and K. Kannan, Environ. Sci. Technol., 2011, 45, 3788–3794.
3. W. Shi, X. X. Hu, F. X. Zhang, G. J. Hu, Y. Q. Hao, X. W. Zhang, H. L. Liu, S. Wei, X. R. Wang, J. P. Giesy and H. X. Yu, Environ. Sci. Technol., 2012, 46, 1811–1818.
4. Y. Kang, Y. B. Man, K. C. Cheung and M. H. Wong, Environ. Sci. Technol., 2012, 46, 8422–8430.
5. N. Casajuana and S. Lacorte, J. Agric. Food Chem., 2004, 52, 3702–3707.
6. H. Farahani, P. Norouzi, R. Dinarvand and M. R. Ganjali, J. Chromatogr., A, 2007, 1172, 105–112.
7. M. D. Carlo, A. Pepe, G. Sacchetti, D. Compagnone, D. Mastrocola and A. Cichelli, Food Chem., 2008, 111, 771–777.
8. E. Martendal and E. Carasek, J. Chromatogr., A, 2011, 1218, 1707–1714.
9. H. W. Sun, Y. L. Yang, H. Li, J. X. Zhang and N. Sun, J. Agric. Food Chem., 2012, 60, 5532–5539.
10. Y. P. Liu, S. H. Wang and L. Wang, J. Agric. Food Chem., 2013, 61, 1160–1164.
11. Y. Q. Cai, Y. Cai, Y. L. Shi, J. M. Liu, S. F. Mou and Y. Q. Lu, Microchim. Acta, 2007, 162, 73–79.
12. S. Jara, C. Lysebo, T. Greibrokk and E. Lundanes, Anal. Chim. Acta, 2000, 407, 165–171.
13. H. Y. Yan, X. L. Cheng and G. L. Yang, J. Agric. Food Chem., 2012, 60, 5524–5531.
14. R. P. Berladi and J. Pawliszyn, Water Pollut. Res. J. Can., 1989, 24, 179–191.
15. M. Polo, M. Llompart, C. G. Jares and R. Cela, J. Chromatogr., A, 2005, 1072, 63–72.
16. Y. L. Feng, J. Zhu and R. Sensenstein, Anal. Chim. Acta, 2005, 538, 41–48.
17. E. Martendala and E. Caraseka, J. Chromatogr. A, 2011, 1218, 1707–1714.
18. K. Holadova, G. Prokupkova, J. Hajslova and J. Poustka, Anal. Chim. Acta, 2007, 582, 24–33.
19. A. Kumar, Gaurav, A. K. Malik, D. K. Tewary and B. Singh, Anal. Chim. Acta, 2008, 610, 1–14.
20. S. L. Chong, D. Wang, J. D. Hayes, B. W. Wilhite and A. Malik, Anal. Chem., 1997, 69, 3889–3898.
21. Z. Wang, C. Xiao, C. Wu and H. Han, J. Chromatogr., A, 2000, 893, 157–168.
22. J. Yu, L. Dong, C. Wu, L. Wu and J. Xing, J. Chromatogr., A, 2002, 978, 37–48.
23. S. Iijima, Nature, 1991, 354, 56–58.
24. Z. Eshaghi, M. Ebrahimi and M. S. Hosseini, J. Chromatogr., A, 2011, 1218, 3400–3406.
25. J. X. Wang, D. Q. Jiang, Z. Y. Gu and X. P. Yan, J. Chromatogr., A, 2006, 1137, 8–14.
26. W. Y. Zhang, Y. Sun, C. Y. Wu, J. Xing and J. Y. Li, Anal. Chem., 2009, 81, 2912–2920.
27. N. Rastkari, R. Ahmadkhaniha, N. Samadi, A. Shafiee and M. Yunesian, Anal. Chim. Acta, 2010, 662, 90–96.
28. A. R. Chaves, G. C. Júnior and M. E. C. Queiroz, J. Chromatogr., B: Anal. Technol. Biomed. Life Sci., 2009, 877, 587–593.
29. H. Asadollahzadeh, E. Noroozian and S. Maghsoudi, Anal. Chim. Acta, 2010, 669, 32–38.
30. A. Mollahosseini and E. Noroozian, Anal. Chim. Acta, 2009, 638, 169–174.
31. A. Mohammadi, Y. Yamini and N. Alizadeh, J. Chromatogr., A, 2005, 1063, 1–8.
32. J. Liu, A. G. Rinzler, H. J. Dai, J. H. Hafner, R. K. Bradley, P. J. Boul, A. Lu, T. Iverson, K. Shelimov, C. B. Huffman, F. Rodriguez-Macias, Y. S. Shon, T. R. Lee, D. T. Colbert and R. E. Smalley, Science, 1998, 280, 1253–1256.

---