Polymer-Functionalized Single-Walled Carbon Nanotubes as a Novel

Mar 18, 2009 - Single-walled carbon nanotubes (SWNTs) were function- alized with a ... Shenzhen Entry-Exit Inspection and Quarantine Bureau. (1) Davis...
0 downloads 0 Views 945KB Size
Download PDF
Anal. Chem. 2009, 81, 2912–2920

Polymer-Functionalized Single-Walled Carbon Nanotubes as a Novel Sol-Gel Solid-Phase Microextraction Coated Fiber for Determination of Polybrominated Diphenyl Ethers in Water Samples with Gas Chromatography-Electron Capture Detection Weiya Zhang,†,‡ Yin Sun,† Caiying Wu,*,† Jun Xing,† and Jianying Li‡ College of Chemistry and Molecular Science, Wuhan University, Wuhan, 430072, and Shenzhen Entry-Exit Inspection and Quarantine Bureau, Shenzhen, 518045, China Single-walled carbon nanotubes (SWNTs) were functionalized with a hydroxyl-terminated silicone oil (TSO-OH). It is synthesized by the reactions of carbonyl chloride groups on the surface of SWNTs and hydroxyl groups of silicone oil (TSO-OH). The functionalized product SWNTsTSO-OH was first used as precursor and selective stationary phase to prepare the sol-gel derived poly(SWNTsTSO-OH) solid-phase microextraction (SPME) fiber for determination of polybrominated diphenyl ethers (PBDEs) in water samples. The possible major reaction of the sol-gel coating process was discussed and confirmed by IR spectra, Raman spectroscopy, and scanning electron microscopy. Some parameters of SPME fiber for the determination of PBDEs were investigated by headspace SPME/gas chromatography with electron-capture detection (HS-SPME/GC-ECD). Compared with the commercial SPME fiber, the new coated fiber showed higher extraction efficiency to PBDEs, better thermal stability (over 340 °C), and longer life span (over 200 times). All of these advantages are mainly due to the incorporation of SWNTs, which enhanced the π-π interaction with PBDEs and increased the surface area of extraction in contact with the sample. Moreover, the sol-gel coating technology additionally provided the porous structure of the 3-D silica network and the strong chemical binding provided which also will improve the extraction efficiency. Under optimized conditions, the method detection limits for seven PBDEs were 0.08-0.8 ng/L (S/N ) 3) and the precision (RSD, n ) 5) was 2.2-7.5% at the 50 ng/L level. The linearity of the developed method is in the range of 5-500 ng/L with coefficients of correlation greater than 0.995. The developed method was successfully applied for the analysis of trace PBDEs in reservoir water and wastewater samples. The recoveries obtained at spiking 50 ng/L were between 74% and 109% (n ) 5) for PBDEs in water samples. Polybrominated diphenyl ethers (PBDEs) have been widely used in the last years as additives in computer plastics, furniture, foams, textile, and other product due to their good flame-retardant properties.1,2 PBDEs are additive flame-retardants which blend with the material, eventually they can be released into the * Corresponding author. Phone: +86-27-68761367. Fax: +86-27-87647617. E-mail: [email protected]. † College of Chemistry and Molecular Science. ‡ Shenzhen Entry-Exit Inspection and Quarantine Bureau. (1) Davis, J. J.; Green, M. L. H.; Hill, H. A. O.; Leung, Y. C.; Sadler, P. J.; Sloan, J.; Xavier, A. V.; Tsang, S. C. Inorg. Chim. Acta 1998, 272, 261–266. (2) Watanabe, I.; Sakai, S. Environ. Int. 2003, 29, 665–682.

2912

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

environment by the utilization of old electronic instruments, burnt out-of-order of finished products containing PBDEs, and other ways. The increasing level of PBDEs has been detected in the global environment, including water,3 sediment,4 fish,5 human blood,6 breast milk,7 and other biota. Several epidemiological studies have shown that PBDEs are posing health risks, such as endocrine disruption, adverse neurobehavioral effects, reproductive toxicants, and probable carcinogens. In addition, the combustion of these compounds could generate highly toxic photoproducts.8-10 These PBDEs have 209 congeners. The low molecular weight congeners, tri to hexa-BDEs, are dominant in the PBDE profiles of environmental matrixes. They can be almost completely absorbed and be highly bioaccumulated, but it is difficult to eliminate them.8,11 In order to evaluate the global distribution, movement, and fate of PBDE, it is crucial to develop a simple, fast, and sensitive method to monitor trace PBDEs in the environment. Several articles and reviews have been published in recent years regarding the development of the analytical procedures to determine the presence of PBDEs in environmental, biological, and human samples.3,12-22 In these procedures, the most widely used methods were GC or HPLC techniques with MS or electron (3) Polo, M.; Gomez-Noya, G.; Quintana, J. B.; Llompart, M.; Garcia-Jares, C.; Cela, R. Anal. Chem. 2004, 76, 1054–1062. (4) Allchin, C. R.; Law, R. J.; Morris, S. Environ. Pollut. 1999, 105, 197–207. (5) Jansson, B.; Andersson, R.; Asplund, L.; Litzen, K.; Nylund, K.; Sellstrom, U.; Uvemo, U. B.; Wahlberg, C.; Wideqvist, U.; Odsjo¨, T.; Olsson, M. Environ. Toxicol. Chem. 1993, 12, 1163–1174. (6) Wehler, E. K.; Hovander, L.; Bergman, Å. Organohalogen Compd. 1997, 33, 420–425. (7) Darnerud, P. O.; Atuma, S.; Aune, M.; Cnattingius, S.; Wernroth, M. L.; Wicklund-Glynn, A. Organohalogen Compd. 1998, 35, 411–414. (8) McDonald, T. Chemosphere 2002, 46, 745–755. (9) Palm, A.; Cousins, I. T.; Mackay, D.; Tysklind, M.; Metcalfe, C.; Alaee, M. Environ. Pollut. 2002, 117, 195–213. (10) Environmental Health Criteria 192. Flame Retardants: A General Introduction;World Health Organization (WHO): Geneva, Switzerland, 1994. (11) Younes, M. The Second International Workshop on Brominated Flame Retardants; Stockholm, Sweden, 2001. (12) Hyotylainen, T.; Hartonen, K. TrAC, Trends Anal. Chem. 2002, 21, 13–29. (13) Covaci, A.; Voorspoels, S.; Ramos, L.; Neels, H.; Blust, R. J. Chromatogr., A 2007, 1153, 145–171. (14) Eljarrat, E.; de la Cal, A.; Barcelo, D. Anal. Bioanal. Chem. 2004, 378, 610–614. (15) Salgado-Petinal, C.; Garcia-Chao, M.; Llompart, M.; Garcia-Jares, C.; Cela, R. Anal. Bioanal. Chem. 2006, 385, 637–644. (16) Zhou, J. J.; Yang, F. X.; Cha, D. M.; Zeng, Z. R.; Xu, Y. Talanta 2007, 73, 870–877. 10.1021/ac802123s CCC: $40.75  2009 American Chemical Society Published on Web 03/18/2009

capture detector (ECD) as the detector. But their sensitivity and selectivity are usually insufficient for direct determination of these contaminants at a very low concentration level especially in the complex matrix. Therefore, a sample pretreatment step prior to chromatographic analysis is usually necessary. Generally, such sample treatments include a number of steps for exhaustive extraction, preconcentration of the target compounds, followed by purification and fractionation before the final chromatographic separation. Liquid-liquid extraction (LLE) and solid-phase extraction (SPE) are usually used as sample pretreatment method for the determination of PBDEs in water samples with GC-ECD or MS. LLE usually requires a large volume of samples and solvents. SPE needs less solvent but still needs multisteps, plus the enrichment efficiency of SPE is relatively low. Solid-phase microextraction (SPME) is a fast, simple, solventless, low-cost, and more efficient extraction technique. It can integrate the extraction preconcentration and sample introduction into one step when coupled with GC. Polo et al. 3 first applied SPME for determination of PBDEs in tap water and wastewater samples with a commercial PDMS fiber. The method obtained high sensitivity with GC/MS/ MS. Linear responses were observed in the 0.2-500 ng/L range, with recoveries >87% and RSD 50%, specific surface area > 400 m2/g) was purchased from Shenzhen Nanotech Port (Shenzhen, China). The fused-silica fiber (140 µm, o.d.) with a protective polyimide coating was obtained from Academy of Post and Telecommunication, Wuhan, China. The commercial SPME fibers used included 100 µm poly(dimethylsiloxane) (PDMS) and 65 µm poly(dimethylsiloxane)divinylbenzene (PDMS/DVB) and were purchased from Supelco (Bellefonte, PA). The seven PBDEs standards (50 mg/L in isooctane for each) were purchased from AccuStandard (New Haven, CT) and stored in amber bottles in the refrigerator at -20 °C. They were 3,3′,4tribromodiphenylether (BDE-35), 2,2′,4,4′-tetrabromodiphenylether (BDE-47), 3,3′,4,4′-tetrabromodiphenylether (BDE-77), 2,2′,4,4′,5pentabromodiphenylether (BDE-99), 2,2′,4,4′,6-pentabromodiphenylether (BDE-100), 2,2′,4,4′,5,5′-hexabromodiphenylether (BDE153), and 2,2′,4,4′,5,6′-hexabromodiphenylether (BDE-154). Tetraethoxysilane(TEOS, 99%), trifluoroacetic acid (99%), and hydroxylterminated silicone oil (TSO-OH) were purchased from Aldrich (Allentown, PA). Poly(methylhydrosiloxane)(PMHS) was purchased from the Chemical Plant of Wuhan University. Doubly deionizedwater (DDW, 18.2 MΩ cm-1) was obtained from a WaterPro water purification system (Labconco, Kansas City, MO). All solvents used were of analytical grade from Guangzhou Chemicals (Guangzhou, China). Instrumentation. The SPME devices for manual sampling were obtained from Supelco and were also prepared by modificaAnalytical Chemistry, Vol. 81, No. 8, April 15, 2009

2913

Figure 1. The procedure for fabricating SWNTs-TSO-OH.

tion of a commercial SPME fiber holder and assembly. SPME-GC experiments were carried out on an Agilent HP6890 GC system equipped with a capillary splitless injector system and an electron capture detection system. A 30 m × 0.32 mm i.d. × 0.25 µm HP-5 coating fused-silica capillary column was used. High-purity nitrogen was used as the carrier gas at a flow rate of 1.6 mL/min. The instrumental temperature were as following: injector temperature 300 °C; detector temperature 320 °C; initial oven temperature 110 °C, increased to 250 °C at a rate of 30 °C/min, and then increased to 275 °C at a rate of 5 °C/min, and then increased to 300 °C at a rate of 40 °C/min, held for 8 min. The total time required for one GC run was 18.3 min. The inlet was operated in splitless mode. A HP Chemstation (Agilent Technologies, DE) was utilized to control the system and to acquire the analytical data. The 20 mL sample vials were purchased from Supelco. In order to mix the various ingredients in solution thoroughly, a model KQ-50DE ultrasonator (Kunshan Ultrasonator Instrument Corporation, Kunshan, China) was used. A centrifuge model TGL16C (Shanghai Anting Instrument Factory, Shanghai, China) was used to separate the sol solution from the precipitate during fiber preparation. A magnetic stirrer DF-101B (Leqing, China) was employed for stirring the sample during extraction. A microscope BA22031 (Guangdian, Chongqing, China) was used for the evaluation of fiber thickness. IR spectra were done on IR instrument of model FTIR-360 (Nicolet). Raman spectra were recorded using an excitation wavelength of 632.8 nm on a laser confocal Raman microspectroscope (LabRAM HR 800 UV) from Jobin Yvon Horiba (France). The SEM micrographs of the SWNTs-TSO-OH coating on the surface of the SPME fiber were obtained on a scanning electron microscope (Quanta 200, FEI, Holland) at 30.0 kV. Synthesis of SWNTs-TSO-OH. A typical procedure for fabricating polymer-functionalized SWNTs is described in Figure 1. The preparation of SWNTs-COOH and SWNTs-(COCl)m was referred to in the literature.24,25 Raw SWNTs were first refluxed in 2.6 M nitric acid for 45 h. Then acid-purified SWNTs were oxidized by a 3:1 (v:v) mixture of sulfuric acid (98%)/nitric acid (70%) for 24 h to SWNTs-COOH. SWNTs-COOH were stirred in SOCl2 [containing 1 mL of dimethylformamide (DMF)] at 70 °C for 24 h to yield (SWNT-COCl)m. After this, 30 mg of SWNTs-(COCl)m was thoroughly mixed with TSO-OH (500 mg) in a flask, heated to 75 °C, and vigorously stirred under nitrogen protection for 48 h. The reaction mixture was dissolved in roomtemperature dichloromethane. The as-prepared solution was centrifuged. Then a 0.2 µm PTFE filter was used to remove (24) Liu, J.; Rinzler, A. G.; Hafner, J. H.; Bradley, R. K.; Boul, P. J.; Lu, A.; Iverson, T.; Shelimov, K.; Huffman, C. B.; Rodriguez-Macias, F.; Colbert, D. T.; Smalley, R. E. Science 1998, 280, 1253–1256. (25) Chen, J.; Hamon, M. A.; Hu, H.; Chen, Y.; Rao, A. M.; Eklund, P. C.; Haddon, R. C. Science 1998, 282, 95–98.

2914

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

any residual insoluble materials from the mixture. Finally, ethanol was added to the mixture to produce a black precipitate. This product, which was SWNTs-TSO-OH, was washed several times for further purification. SPME Fiber Preparation. Prior to sol-gel coating, the protective polyimide layer on a 1 cm tip of a 10 cm piece of fusedsilica was removed by dipping it into acetone for 3 h. Then the fiber was dipped into 1 M NaOH solution for 1 h in order to expose the maximum number of silanol groups on the surface of the fiber, and then the fiber was cleaned with water. Subsequently, it was placed in 0.1 M HCl solution for 30 min to neutralize the excess NaOH, cleaned again, and dried. Care needs to be taken during these experiments due to the fused-silica becoming fragile after removing the polyimide coating. The sol-gel solution was prepared as follows: 90 mg of SWNTs-TSO-OH was dissolved in 300 µL of dichloromethane 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. A volume of 180 µL of TFA (95% water solution) was sequentially added to the resulting solution with ultrasonic agitation for another 5 min. The mixture was then centrifuged at 8000 rpm for 3 min, and the top clear sol solution was used for fiber coating. The treated fiber was dipped vertically into the sol solution and held for 5 min until a sol-gel coating was formed on the bare outer surface of the fiber end. For each fiber, this coating process was repeated several times and each time for 2 min in the same sol solution until the thickness of the coating desired was obtained. The fiber was placed in a desiccator at room temperature for 24 h. The fiber was initially conditioned by placing it in a GC injector kept at a temperature of 100 °C with a gentle N2 flow for 1 h, then conditioned again at 220-340 °C for another 2 h. The final thickness of the fiber was 68 µm. The length of each coated fiber is 1 cm. The color of the coated fiber is a nut-brown. Preparation of Working Standards and Samples. Concentrations of 0.5, 1, 5, 10, 50, 100, 200, and 500 µg/L of PBDEs solutions were prepared by diluting the seven PBDEs standards (50 mg/L in isooctane for each) with hexane. For the above PBDEs hexane solutions, 10 µL of each was diluted with 10 mL of doubly deionized water, respectively. The final concentration of PBDEs working solution is 0.5, 1, 5, 10, 50, 100, 200, and 500 ng/L PBDEs.26 All PBDEs solutions were stored at -20 °C before use. One reservoir water sample and two wastewater samples were collected from a water inspection institution in Shenzhen, Guangdong province, China. Three water samples were stored in amber glass containers and maintained in the dark at 4 °C before analysis. (26) Zeng, Z. R.; Qiu, W. L.; Huang, Z. F. Anal. Chem. 2001, 73, 2429–2436.

Figure 2. FT-Infrared spectra of (a) TSO-OH and (b) SWNTs-TSO-OH.

Solid-Phase Microextraction. A volume of 10 mL of sample solution was placed in a 20 mL glass vial with a stir bar (PTFE) and headspace cap. The vial was then immersed in a thermostatic water bath at 90 °C. During extraction, the fiber was exposed to the headspace over the water for 30 min. After HS-extraction, the SPME fiber was inserted into the GC injector at 300 °C for 4 min. Before sampling, the fiber was cleaned in another injector port at 300 °C for 10 min to eliminate any carry-over of analytes from the previous extraction. SPME-GC blanks were performed before each SPME extraction. RESULTS AND DISCUSSION Characteristics of SWNTs-TSO-OH. The structure of SWNTsTSO-OH is shown in Figure 1, and it can be seen that the formation of SWNTs-TSO-OH is through SWNTs-(COCl)m chemical bonding to TSO-OH with the ester group (-COO-). Therefore, FT-IR spectra (Figure 2) were done first to verify the existence of the ester group. Compared with the TSO-OH (Figure 2a), the SWNTs-TSO-OH showed an absorbance peak at

about 1701.6 cm-1 (Figure 2b) due to the carboxyl stretch of the ester group. It indicated the polymer-functionalized reaction of SWNTs was successful. In order to verify the structure of SWNTs-TSO-OH further, Raman spectra (Figure 3) were conducted. Figure 3a illustrates the Raman spectra of SWNTs. The peak at 211 cm-1 in the Raman breathing mode region (dashed circle 1) is the characteristic peak of SWNTs, the peaks at 1330 and 1590 cm-1 in dashed circle 2 are peaks D and G of the SWNTs, respectively. The bigger the ratio peaks G to D showed, the better orderly the structure of carbon.27 The Raman spectrum of TSO-OH is presented in Figure 3c. The peaks at 160 and 190 cm-1 (dashed circle 3) are the feature peaks of TSO-OH; the peaks at 490 and 710 cm-1 (dashed circle 4) are absorbance peaks due to the group of Si-O-Si. Figure 3b gives the Raman spectra of the finished product SWNTs-TSO-OH. It has the homologous peak (211 cm-1) in Figure 3a in dashed circle 1, which showed SWNTsTSO-OH preserves the SWNT structure. The ratio of peak G to (27) Li, B.; Lian, Y. F.; Shi, Z. J. Chem. J. Chin. Univ. 2000, 21, 1633–1635.

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

2915

Figure 3. Raman spectra of (a) SWNTs, (b) SWNTs-TSO-OH, and (c) TSO-OH.

D is large in Figure 3b, and it showed that the damage caused on the surface structure of SWNTs from the preparation of SWNTsCOOH is minimal and the structure of the carbon is preserved. The peaks in dashed circles 3 and 4 of Figure 3b are the same as in Figure 3c. It showed SWNTs-TSO-OH and TSO-OH has the common feature peaks, which further confirmed SWNTs were incorporated into the polymer structure of TSO-OH triumphantly. Coating Ingredients and Major Reaction of the Coating Process. SWNTs-TSO-OH, TEOS, TSO-OH, PMHS, and 95% trifluoroacetic acid are the principal ingredients of the sol-gel coating solution in this work. SWNTs-TSO-OH is used as the selective stationary phases for compounds and TEOS is used as a precursor. TSO-OH, as sol-gel active, is added into this system to lengthen the silica network and help uniformly spread the stationary phase on the fiber. PMHS is used as the deactivation reagent. 95% Trifluoroacetic acid is the catalyst of the sol-gel process. Five major reaction processes occur during the sol-gel formation: (1) tetraethoxysilane groups in the monomer can be catalytically hydrolyzed to silanol groups; (2) condensation of the hydrolyzed products with SWNTs-TSO-OH; (3) polycondensation of the condensation products into a three-dimensional sol-gel network and chemical bonding of OH-TSO and SWNTs-TSO-OH to the evolving sol-gel network; (4) chemical anchoring of the evolving sol-gel polymer to the surface of the fiber to create a surface-bonded polymeric coating; (5) deactivation of the surface bonded sol-gel coating. The structure of the coated fiber is represented in Figure 4. Surface Structure of the Coating. The morphological structure of the SWNTs-TSO-OH coating can be assessed from Figure

Figure 4. The structure of the sol-gel SWNTs-TSO-OH coating. 2916

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

5. The surface of the fiber was homogeneous and dense (Figure 5a), so good reproducibility could be achieved for the preparation of the fiber. The coating possesses a porous structure (Figure 5b), and it should significantly increase the available surface area on the fiber as well as improve the kinetics of the extraction process, therefore, enhance extraction efficiency. Thermal Stability of the Coating. The thermal stability is the crucial characters of coating in practical application. The SWNTs-TSO-OH fiber’s thermal stability was investigated by performing extraction after the fiber’s being conditioned for 1 h at 280, 300, 320, and 340 °C, respectively. The results (Figure 6) indicated that SWNTs-TSO-OH coating can withstand higher temperature up to 340 °C without loss of extraction efficiency (peak area). Such a high operating temperature achieved is due to the special nature of SWNTs and the strong chemical bonding provided by sol-gel technology. Enhanced thermal stability allowed the use of higher injection port temperatures for efficient desorption of semivolatile analytes. Thus, an extended range of analytes can be handled by the SPME-GC method. Lifespan of the Coating. A coating’s lifespan is important for practical application. The SWNTs-TSO-OH coating’s lifespan was studied by monitoring the change of extraction peak areas of seven PBDEs after it had been used for 50, 100, 150, and 200 adsorption/ desorption times. No obvious decline was observed (Figure 7), which indicated that SWNTs-TSO-OH fiber was stable and can be used least 200 times, while all commercial fibers can only be used in the range of 50-100 times. Such a long life span also benefits from the strong chemical bonding between the sol-gel generated organic-inorganic composite coating and the silica fiber surface. Optimization of HS-SPME Procedures. According to the literature,3 HS-SPME offered higher sensitivity than direct-SPME at high temperature for the determination of PBDEs in water. Furthermore, HS-SPME can minimize disturbing the sample matrix, prolonging the life of the fiber. HS-SPME method is based on the multiphase (coating/headspace/aqueous) equilibration principle. In order to achieve the best extraction efficiency of the new coating for seven PBDEs (Table 1), several parameters, including salt addition, extraction temperature, extraction time, and desorption time, were investigated. Salt Addition. Salting can increase or decrease the amount extracted, depending on the compound, the type of fiber, and salt concentration. Here, the salt addition of NaCl on the HS-SPME of PBDEs in water was investigated with 10 mL of water in a 20 mL vial containing seven PBDEs at 0.2 ng/mL. The amounts of PBDEs extracted from a saturated sodium chloride solution were compared with that from a solution with no sodium chloride added. An obvious increase in the efficiency of extraction (Figure

Figure 6. Thermal stability of SWNTs-TSO-OH coated fiber (68 µm). PBDEs at 0.2 ng/mL. Conditions: saturated out with NaCl; extraction time, 30 min; extraction temperature, 90 °C; desorption time, 4 min. Error bars show the standard deviation of the mean (n ) 3). Peaks: no. 1, BDE-35; no. 2, BDE-47; no. 3, BDE-77; no. 4, BDE-100; no. 5, BDE-99; no. 6, BDE-154; no. 7, BDE-153.

Figure 7. Lifetime profile of sol-gel SWNTs-TSO-OH coated fiber. See Figure 6 for conditions and peak identification. Error bars show the standard deviation of the mean (n ) 3).

8) presented as the peak area was observed for all seven PBDEs when adding saturated NaCl in the spiked solution. This is due to the salting-out effect which is also commonly found in the extraction process involving hydrophobic interaction. The method of salting out with NaCl was selected for the subsequent experiments. Extraction Temperature. For HS-SPME, optimization of extraction temperature is generally more important, especially for semivolatile compounds such as PBDEs. Since these compounds usually have large Kow values [octanol/water partition coefficient, i.e., hydrophobic coefficient], and small KH values [Henry’s constant], which leads to high extraction sensitivity (as a PDMS fiber) but long equilibration time. For analysis of aqueous samples, the equilibration time in the HS-SPME is controlled by both Kow and KH.28 Increasing the temperature can signifi-

cantly increase the analysis evaporation rates (raise KH of analyte) and mass transfer at the water-headspace interface (distribution velocity). It is apt for the analytes to enter the headspace gas phase and fiber coating, resulting in the increasing of the extraction amount on the fiber and the reducing of the equilibrium time. Therefore, at higher temperature, the HS-SPME extraction process can be completed without reaching equilibrium and still provide excellent sensitivity, if Kow values of the analytes are large. On the other hand, higher temperature increases the extraction efficiency only in nonequilibrium situations. If the system were in equilibrium, higher temperatures would not give higher responses since adsorption is an exothermic process.29 On the basis of Figure 9, to provide a higher extraction efficiency and more convenient manipulation in this study, the optimum extraction temperature was set at 90 °C. Extraction Time. In order to further evaluate the extraction ability of the new fiber for PBDEs, the extraction times of these analytes were investigated from 10-60 min at 90 °C (Figure 10). The responses of all the seven PBDEs increase with the longer exposure time in the headspace. The equilibrium was not achieved even after 60 min sampling time. It indicated that the new coating

(28) Zhang, Z. Y.; Pawliszyn, J. Anal. Chem. 1993, 65, 1843–1852.

(29) Liompart, M.; Li, K.; Fingas, M. Anal. Chem. 1998, 70, 2510–2515.

Figure 5. Scanning electron micrographs of SWNTs-TSO-OH coated fiber at different magnifications: (a) 400-fold and (b) 2 000fold.

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

2917

Table 1. Structures of PBDEs

fiber showed remarkable extraction capacity toward these analytes. On the other hand, the sol-gel SWNTs-TSO-OH coating is a porous polymer-based solid extracting phases. They extract analytes via adsorption rather than absorption (as PDMS), while adsorption is a competitive process and the best extraction efficiency is not at the point of equilibrium. Thus SPME analysis with the use of solid coatings is usually working at nonequilibrium. As long as the exposure time of the fiber is kept exactly constant, the results are reliable.30 Therefore, 30 min was fixed as the working condition based on the sensitivity and time efficiency.

Figure 8. Effect of NaCl added to the sample matrix on the extraction. See Figure 6 for conditions and peak identification. Error bars show the standard deviation of the mean (n ) 3). 2918

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

Desorption Time. The desorption time had a significant effect on the responses of the target compounds, especially for compounds of high boiling point. In this work, the effect of desorption time on the HS-SPME of PBDEs was investigated at 1, 2, 3, 4, and 5min under the desorption temperature of 300 °C. The chromatographic area of all the seven PBDEs reached the equilibrium after 4 min. Thereby, the desorption time was set at 4 min.

Figure 9. Influence of the extraction temperature on the response of 0.2 ng/mL PBDEs in a water sample. Conditions: saturated out with NaCl; extraction time, 30 min; desorption time, 4 min. See Figure 6 for peak identification. Error bars show the standard deviation of the mean (n ) 3).

Figure 12. Chromatograms obtained by the developed method for standard solution containing 0.05 ng/mL of each PBDE. Conditions: sample volume, 10 mL; saturated out with NaCl; extraction time, 30 min; extraction temperature, 90 °C; desorption temperature, 300 °C; desorption time, 4 min. See Figure 6 for peak identification. Figure 10. Effect of the extraction time on the response of 0.2 ng/ mL PBDEs in a water sample. Conditions: saturated out with NaCl; extraction temperature, 90 °C; desorption time, 4 min. See Figure 6 for peak identification. Error bars show the standard deviation of the mean (n ) 3).

Table 2. Linearity Range, Repeatability, and Detection Limits of the Proposed Method and Compared with Other Methods

no. no. no. no. no. no. no.

1, 2, 3, 4, 5, 6, 7,

BDE-35 BDE-47 BDE-77 BDE-100 BDE-99 BDE-154 BDE-153

a

LODs (ng/L, S/N ) 3)

linearity range (ng/L)

RSD (%,n ) 5) (peak area)

this work

5-500 5-500 5-500 5-500 5-500 5-500 5-500

2.2 6.4 4.7 7.2 4.4 7.0 7.5

0.08 0.2 0.2 0.2 0.2 0.4 0.8

Wang et al.17

Tian et al.18

3.6

8 1.1

4.9 5.9 7.7 8.6

1.9 2.4 7.2 16.0

a The concentration of the standard solution was 50 ng/L for each compound.

Figure 11. Comparison of the extraction efficiency of the sol-gel SWNTs-TSO-OH coated fiber and two commercial fibers for PBDEs. See Figure 6 for conditions and peak identification. Error bars show the standard deviation of the mean (n ) 3).

Comparison with Commercial Fiber. The seven PBDEs are slightly polar and benzenoid compounds. PDMS is a nonpolar coating which was usually applied to analyze PBDEs in much of the literature. PDMS/DVB fiber is a middle-polar coating which has affinity for benzenoid compounds. Therefore, PDMS and PDMS/DVB fiber was selected for comparing the extraction effect with the SWNTs-TSO-OH fiber. The results are shown in Figure 11. It revealed that SWNTs-TSO-OH has superior extraction efficiency for PBDEs than commercial PDMS and PDMS/DVB. It is possibly due to carbon atoms in the walls of carbon nanotubes possessing the effect of mixed hybridization of sp2 and sp3; therefore, there is a highly delocalized conjugate system of the π-electron,31 which enhances the π-π interaction with the aromatic PBDEs. On the other hand, the large surface area of carbon nanotubes as their curly surface also has stronger (30) Ai, J. Anal. Chem. 1997, 69, 1230–1236. (31) Yuan, W. K.; Wu, H.; Jiang, Z. Y.; Xu, S. W. Chin. J. Org. Chem. 2006, 26, 1508–1517.

binding affinity for hydrophobic molecules compared with the aplanar carbon surface.32 Furthermore, a large sorption surface on the outside and in the internal pores of the CNTs and the porous structure of the 3-D sol-gel silica network of the novel SPME coating also favor the extraction efficiency to some extent. All these indicate that SWNTs have a strong physical adsorption ability to hydrophobic PBDEs. Quantitative Calibration and Reproducibility. The characteristics of the quantitative calibration curves are given in Table 2. The linearity concentration range of the developed analytical methods for all seven PBDEs was from 5 to 500 ng/L. All of the correlation coefficients are greater than 0.995. Under the optimal conditions, the reproducibility of one particular fiber and the reproducibility between different fibers were studied. The one fiber reproducibility (RSD) was evaluated by determining the chromatographic peak areas (shown in Table 2) ranged from 2.2 to 7.5%. The fiber-to-fiber reproducibility (RSD) of the seven PBDEs at 50 ng/L extracted with three same thickness SWNTsTSO-OH coated fibers prepared in the same way ranged from 2.7 to 8.8%. The results of the proposed method in this study were compared with the data obtained by other researchers using MWNTs for SPME-GC-ECD(Table 2).17,18 The proposed method showed higher sensitivity than the method using MWNTs fiber coating. The sensitivity improved 5-29 times. The peaks of the (32) Saridara, C.; Brukh, R.; Iqbal, Z.; Mitra, S. Anal. Chem. 2005, 77, 1183– 1187.

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

2919

Table 3. Analytical Results for the Determination of PBDEs in Real Samples

s1a

s2b

s3c

a

real sample concentration (ng/L), mean ± SD (n ) 3) average recoveryd (% (RSD %, n ) 5)) real sample concentration (ng/L), mean ± SD (n ) 3) average recoveryd (% (RSD %, n ) 5)) real sample concentration (ng/L), mean ± SD (n ) 3) average recoveryd (% (RSD %, n ) 5))

no. 1, BDE-35

no. 2, BDE-47

no. 3, BDE-77

no. 4, BDE-100

no. 5, BDE-99

no. 6, BDE-154

no. 7, BDE-153

2.0 ± 0.06

1.3 ± 0.04

1.8 ± 0.04

2.2 ± 0.08

2.8 ± 0.1

4.0 ± 0.15

2.8 ± 0.14

107 (1.8)

109 (1.9)

103 (1.5)

93 (2.7)

103 (2.4)

81 (1.8)

85 (2.7)

17.8 ± 0.32

51.7 ± 0.98

1.5 ± 0.04

2.1 ± 0.07

2.6 ± 0.09

2.6 ± 0.11

105 (2.3)

93 (1.7)

95 (3.2)

86 (1.7)

85 (3.1)

78 (1.4)

84 (2.2)

4.1 ± 0.1

86.8 ± 1.7

1.7 ± 0.08

1.7 ± 0.09

2.2 ± 0.11

4.1 ± 0.18

2.1 ± 0.1

92 (4.7)

74 (4.0)

78 (4.6)

75 (5.6)

83 (6.8)

77 (2.4)

75 (1.6)

e

s1, reservoir water sample. b s2, wastewater sample. c s3, wastewater sample. d Recoveries obtained at 50 ng/L spiking level. e Not detected.

13 shows a typical chromatogram of a real wastewater sample with the sol-gel SWNTs-TSO-OH fiber.

Figure 13. Chromatograms obtained by the developed method for one wastewater sample. Conditions: sample volume, 10 mL; saturated out with NaCl; extraction time, 30 min; extraction temperature, 90 °C; desorption temperature, 300 °C; desorption time, 4 min. See Figure 6 for peak identification.

proposed method are all sharp (Figure 12.). Application to Real Samples. The developed method was successfully applied for the analysis of trace PBDEs in reservoir water and wastewater samples. The results are given in Table 3. The PBDEs except BDE-153 (is not detected in sample s2) were almost detected in all three samples, ranging from 1.3 to 86.8 ng/ L. Recoveries obtained at 50 ng/L spiking level were in the range of 74-109% for PBDEs. The recovery of the reservoir water sample is slightly higher than wastewater samples. The recovery of the reservoir water sample was in the range of 81-109%; the recoveries of two wastewater samples were the range of 78-105% and 74-92%, respectively. The precision for the determination of three samples for all seven PBDEs was below 7% (RSD). Figure

2920

Analytical Chemistry, Vol. 81, No. 8, April 15, 2009

CONCLUSIONS A novel sol-gel SWNTs-TSO-OH coated SPME fiber had been prepared and applied to determine seven PBDEs in water samples. Compared with the commercial SPME fibers, e.g., PDMS and PDMS/DVB, the novel porous sol-gel SWNTs-TSO-OH fiber exhibited high sensitivity and selectivity for PBDEs compounds, higher thermal stability (to 340 °C), and long service life (more than 200 times). HS-SPME coupled with GC-ECD using the sol-gel SWNTsTSO-OH fiber was successfully applied to the determination of PBDEs in reservoir water and wastewater samples. The detection limit is in the subnanogram per liter range (0.08-0.8 ng/L) and has good linearity in a wide range of concentrations as well as good precision between 2.2% and 7.5%. The recoveries for three real wastewater samples were 81-109%, 78-105%, and 74-92%, respectively. Such a method offers a simple, rapid, sensitive, and very efficient tool for determination of trace PBDEs in water samples. ACKNOWLEDGMENT The authors kindly thank Fei Niu, Dr. Xiaohua Wang, and Dr. Hua Tong of Wuhan University for their technical assistance in the IR spectra experiments, Raman spectra experiments, and SEM experiments. Received for review October 8, 2008. Accepted February 9, 2009. AC802123S