Separation Science and Technology ISSN: 0149-6395 (Print) 1520-5754 (Online) Journal homepage: http://www.tandfonline.com/loi/lsst20 Activated Carbon Nanofiber Produced from Electrospun PAN Nanofiber as a Solid Phase Extraction Sorbent for the Preconcentration of Organophosphorus Pesticides Bozorgmehr Maddah, Mostafa Soltaninezhad, Korosh Adib & Mahdi Hasanzadeh To cite this article: Bozorgmehr Maddah, Mostafa Soltaninezhad, Korosh Adib & Mahdi Hasanzadeh (2016): Activated Carbon Nanofiber Produced from Electrospun PAN Nanofiber as a Solid Phase Extraction Sorbent for the Preconcentration of Organophosphorus Pesticides, Separation Science and Technology, DOI: 10.1080/01496395.2016.1221432 To link to this article: http://dx.doi.org/10.1080/01496395.2016.1221432 Copyright © Taylor & Francis Group, LLC Accepted author version posted online: 23 Aug 2016. Published online: 23 Aug 2016. Submit your article to this journal View related articles View Crossmark data Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=lsst20 Download by: [163.172.152.231] Date: 23 August 2016, At: 19:37 t rip Activated Carbon Nanofiber Produced from Electrospun PAN Nanofiber as a Solid Phase Extraction Sorbent for the Preconcentration of Organophosphorus Pesticides Bozorgmehr Maddaha, Mostafa Soltaninezhada, Korosh Adiba, Mahdi Hasanzadehb,* Department of Chemistry, Imam Hossein University, Tehran, Iran us c a b Advanced Materials and Nanotechnology Research Center, Faculty of Engineering, Imam Hossein University, Tehran, Iran * Corresponding author. Tel.: +98 21 33516875; fax: +98 2133516875. an E-mail addresses: m_hasanzadeh@aut.ac.ir, hasanzadeh_mahdi@yahoo.com (M. Hasanzadeh). M Abstract Activated carbon nanofibers (CNFs) were derived from electrospun nanofibers and subsequent ed heat treatment. They were characterized by scanning electron microscopy, and Fourier transform infrared spectroscopy. The applicability of activated CNFs for preconcentration and pt determination of organophosphorus pesticides (OPPs) were investigated by high performance ce liquid chromatography with diode array detector (HPLC-DAD). Some important parameters influencing the extraction efficiency, such as amount of sorbent, pH, flow rate, and amount of Ac salt were investigated by response surface method (RSM). The obtained results showed that this analytical method will be useful for the analysis of OPPs in tap water with high precision and accuracy. Keywords: carbon nanofiber; electrospinning; preconcentration; solid phase extraction; experimental design 1 Introduction Carbon nanofibers (CNFs) have been derived from electrospun polymer nanofibers, such as rip t polyacrylonitrile (PAN) by the processes of stabilization, carbonization and graphitization. Since CNFs derived from electrospun PAN have a higher specific surface area, higher loading us c capacity, and high mechanical flexibility, they are good candidates for applications in filters, scaffolds and fuel cells for bio-medical, energy applications, and also as a sorbents used in solid phase extraction (SPE) [1-5]. The SPE is a common preconcentration procedure for extraction an and determination of low concentration of pesticides. Although this process is a powerful technique, it suffers from some limitations such as excessively time-consuming and labor M intensive [6-8]. ed Organophosphorus pesticides (OPPs) such as ethion, fenitrothion, and diazinon, are neurotoxic and are released into the environment from manufacturing, transportation, and agriculture pt applications. Due to their low cost and broad spectrum activity, they are widely used for agricultural purposes [9-11]. When they absorbed by human organisms, because of acetyl- ce cholinesterase deactivation, they become very toxic. Therefore, developing novel analytical Ac methods capable of performing rapid detection of these compounds in the environment becomes necessary and is of special interest. Nowadays, several numerical methods are widely used for either modeling or optimizing the complex and non-linear processes generally found in nanoscience and technology.12-16 Response surface methodology (RSM) is a combination of mathematical and statistical techniques used to 2 evaluate the relationship between a set of controllable experimental factors and observed results. The main goal of RSM is to optimize the response, which is influenced by several independent variables, with minimum number of experiments [17, 18]. Therefore, the application of RSM in rip t extraction will be helpful in an effort to find and optimize an important parameter. This paper will focus on the fabrication of activated CNFs and their application as an efficient us c sorbent for preconcentration and determination of organophosphorus pesticides (OPPs) in water samples by HPLC with diode array detection (DAD) following a discussion on the several important parameters influencing the extraction efficiency (recovery percent of OPPs). an According to the literature, there has not been a previous report regarding the utilization of RSM for the predicting recovery percent of OPPs. Hence, in this work, a study has been conducted to M investigate the effects of four parameters including amount of sorbent, solution pH, flow rate, ed and amount of salt on recovery percent of OPPs. For this purpose, response surface model based on the central composite design (CCD) was employed to investigate quantitatively the effect of pt variables on the response. ce Experimental Ac Materials Polyacrylonitrile (PAN, Mw=80,000) was purchased from Polyacryle Co. (Iran). N-N, dimethylformamide (DMF), hydrochloric acid, sodium hydroxide, acetonitrile, methanol, ethanol, potassium hydroxide, and sodium chloride were obtained from Merck Co. (Germany). 3 All OPPs (ethion, fenitrothion, and diazinon) were obtained from Sigma-Aldrich Co. (Germany) and were used without further purification. All OPPs and reagents used in this study were of analytical grade. rip t Stock standard solution (300 µg mL-1) of ethion, fenitrothion, and diazinon were prepared in methanol and stored at 4 °C. Working standard solutions were prepared daily by diluting with an Electrospinning and preparation of CNF us c deionized water. The polymer solutions (8-16 wt.%) were prepared by dissolving PAN powder in DMF via M magnetic stirrer for 24 h at 50 °C. These polymer solutions were used for electrospinning. The electrospinning device used in this experiment was produced by Fanavaran Nano-meghyas Co. ed (Iran). To produce electrospun nanofibers, the polymer solutions were placed into a 5 mL plastic syringe with an 18 gauge (diameter=0.12 mm) needle tip as a nozzle for electrospinning. A pt syringe pump fed polymer solution to the needle tip. The metallic needle was connected to a positive high voltage and the collector (aluminum foil) was connected to the ground. The ce grounded collector was located at different distances (10-15 cm). All electrospinning Ac experiments were carried out at room temperature. The electrospun nanofibers were picked-up from the collector and were stabilized by heating in air from room temperature to 280 °C at a rate of 1 °C/min and holding for 4 h under air flow. Then the stabilized nanofibers were carbonized in a tubular high-temperature furnace (EX120030L, Exciton) at a heating rate of 30 °C/min up to 800 °C for 1 h in a N2 atmosphere. The 4 obtained carbonized carbon nanofiber has been activated with the addition of KOH and has been placed in a furnace under N2 flow up to 850 °C. The activated product was washed using distilled water, filtered and then dried at room temperature. rip t Measurement and Characterization electron microscope (SEM) (Hitachi S-4160, Japan). us c The morphology of the gold-sputtered electrospun nanofibers were observed by scanning The infrared spectra of samples were recorded on a Perkin Elmer Spectrum 100 in the wave an number range of 500-4000 cm−1; a nominal resolution for all spectra was 4 cm−1. M Analyses were performed using a Knauer HPLC system with an EA4300F SMART LINE pump and a S-260 UV detector (Knauer, Germany). The analytical column was C18 column (250 mm ed × 4.6 mm, 10 μm) perfect sill Target ODS-3. The working conditions of HPLC was isocratic; mobile phase was acetonitrile:water; (85:15); and flow rate was 1.1 mL min-1. Injection volume pt was 20 μL and the wavelength of the Diode-array detector (DAD) was fixed at 220 nm for the ce residual analysis of ethion and fenitrothion and also 247 nm for the analysis of diazinon. The pH of solutions was measured with a Metrohm 781 pH/Ion meter (Herisau, Switzerland) supplied Ac with a combined electrode. Sampling River water sample were collected in 1 L amber glass bottles (without any further treatment), and cooled in refrigerator. Prior to extraction, each sample was filtered through a 0.45 µm membrane 5 filter and then was used for extraction [19]. The spiked with concentration level of 3 ng mL-1 of each OPPs were extracted using the optimized procedure and then analyzed using HPLC. rip t Extraction Procedure The packed fiber solid phase extraction (PF-SPE) device is a rectangle made of 5 mg PAN us c nanofibers into the tip of a needle syringe using a fine steel rod (about 0.5 mm diameter). Prior to a preconcentration step, the PF-SPE column was activated with 200 mL of ethanol and then 200 mL of water. Then, the water sample (1 mL) was loaded through the sorbent by the pressure of an air forced by gas tight plastic syringe (5 mL), with the flow carefully controlled in a slow dropwise manner. After the analytes were eluted through the sorbent, the content of the column was M washed with 100 mL of water, and the sorbed targets were desorbed by 1.0 mL of methanol. ed Finally, 20 mL of elution solvent were analyzed by HPLC system. pt Experimental design and optimization In this study, response surface methodology (RSM) was used in the experimental design and ce optimization. The effects of four independent analytical parameters, such as amount of sorbent (mg), solution pH, flow rate (mL min-1), and amount of salt (mg) in the extraction of analytes Ac have been considered using RSM based on a central composite design (CCD). The experiment was performed for at least three levels (coded as -1, 0, and +1) of each factor to fit a quadratic model. The coded values were calculated according to the following equation: 6 Hi and Li (1) refer to the high and low levels of the variables i (i= 1,2,3,4), respectively. rip where, -[ Hi + Li ] / 2 [ Hi - Li ] / 2 t i Xi = 0 + i =1 4 3 i Xi +  2 ii X i +  i =1 i =1 4  ij Xi X j j=2 an 4 Y= us c The behavior of the system is explained by the following quadratic polynomial equation: where, Y is the predicted response, X i and X j are the independent variables, i , ii , ij 0 is constant are coefficients estimated from the regression.17,18 The RSM was M coefficient, and (2) applied to the experimental data using statistical software, Design-expert (Version 8.0.3, Stat- ed Ease, Minneapolis, MN, 2010). The experimental parameters and their levels are given in Table pt 1. ce Results and Discussion Ac Structure and properties In this study, the electrospun nanofibers with different diameters were prepared by changing the electrospinning parameters to determine the optimum conditions for improved structure and morphology which is required for carbonization process. It is found that PAN nanofibers electrospun at conditions of 14.4 wt.% polymer concentration, 12.5 cm of tip-to-collector 7 distance, and 15 kV of the applied voltage had suitable structure and morphology for production of CNFs. Further PAN nanofibers characteristics, which provided under different electrospinning conditions, are provided in the Table A-1 (Appendix). Figure 1 shows SEM micrographs and rip t fiber diameter distribution of as-spun PAN and PAN-derived CNFs. It can be seen that CNFs have a smooth surface morphology and uniform diameter along their length, which could us c enhance their absorbance capability. Considering the average fiber diameter (AFD) in Figure 1 reveals that the CNFs is much thinner than the PAN nanofibers, which is attributed to the shrinkage of PAN nanofibers and weight loss caused during the carbonization. an ourier Transform Infrared (FTIR) spectra of the as-spun PAN nanofibers, and those carbonized at M 800 °C are shown in Figure 2. PAN nanofibers had a characteristic vibration at 2241-2243 cm-1 due to the C≡N Nitrile group. ed The vibrations of the aliphatic CH groups (CH, CH2, and CH3) were observed at 2870-2931 cm-1, 1450-1460 cm-1, 1350-1380 cm-1, and 1220-1270 cm-1. The strong bands were observed at 1737 pt cm-1 and 1684 cm-1 correspond to C=O stretching and the amide group, respectively.3,20 ce The FTIR spectrum of heat-treated PAN nanofibers is characterized by the absence of the 22412243 cm-1 peak intensity (C≡N nitrile group), the absence of the intensity of the aliphatic CH Ac groups as well as the amide group. The band around 1588 cm-1 is due to a mix of C=N, C=C, and N–H groups, which is attributed to the cyclization and cross-linking and therefore preparing the chemical structure for subsequent high temperature carbonization [3, 20]. 8 Analysis of variance The experimental results concerning recovery percent of ethion (Y1), fenitrothion (Y2), and t diazinon (Y3) using four parameters are shown in Table 2. As illustrated, the values of recovery rip percent of diazinon are more than other pesticides that could be attributed to the presence of two pyrimidine nitrogen atoms. According to the FTIR spectra, a mix of C=N, C=C, and N–H groups us c in CNFs (band around 1588 cm-1) can react with two pyrimidine nitrogen atoms in diazinon resulting in a strong adsorption and thus high values of recovery percent. However, further an investigation is needed to understand any specificity of the CNFs to this analyte. The results of the second-order response surface model in the form of analysis of variance M (ANOVA) for ethion and fenitrothion recovery percent are shown in Table 3. A similar trend was also observed for diazinon recovery percent. In this work, statistical conclusions were ed obtained at 95% confidence level. Using 5% significance level, the factor is considered pt significant if the p-value is less than 0.05. By linear regression analysis of Equation (2), the predicted response functions (Y1 and Y2) were Ac ce obtained and given as Y1 = +23.88 − 20.88X 2 − 4.63X3 (3) −21.53X1 X 4 + 5.34 X 2 X 4 − 20.75X12 +17.97 X 22 Y2 = +11.34 −17.63X 2 −19.48X1 X 4 − 7.17 X12 + 4.85X 22 9 (4) As can be seen in Table 3 the prob > F-values for both the responses Y1 and Y2 are lower than 0.05 indicating that quadratic models were significant and have a good agreement with experimental data. The coefficient of determination (R2) that was found to be close to 1 (0.96 for rip t Y1 and 0.95 for Y2) also revealed a high correlation between observed and predicted values. This can be also seen in Figure 3 by comparing the actual values against the predicted responses us c by the models for recovery percent of ethion and fenitrothion. The “adequate precision” measures the signal to noise ratio. The ratio greater than 4 is desirable and demonstrating that models are significant. The adequate precision for Y1 and Y2 were 16.8 and 13.3, respectively an which indicate both models were significant. M Figure 3 about here ed Effects of significant parameters In this study, the influence of all the single factors and interactions of two factors on recovery pt percent of ethion was investigated. The main effect plots which describe the influence of single Figure 4 about here Ac ce factors on recovery percent of ethion are shown in Figure 4. It is illustrated that increasing the amount of sorbent cause an increase followed by a decrease in recovery percent of ethion (Figure 4a). It was found that recovery percent increased by increasing the amount of sorbent (activated CNFs mass) to 15 mg of packing quantity and then it decreased with increasing amount of sorbent. Furthermore, maximum ethion recovery occurred at solution pH=4, as shown in Figure 4b. It is obvious that increasing the solution pH cause a 10 decrease in recovery percent of ethion. The same behavior were also observed for fenitrothion and diazinon. The reason may be related to the degradation of OPPs in alkaline medium [21, 22]. The influence of flow rate on recovery percent of ethion is shown in Figure 4c. Higher ethion rip t recovery was found in low flow rate (10 mL min-1). It is obvious that CNFs have very high surface areas resulting in a high extraction capacity. However, the interaction between CNFs and us c analytes could not be complete within a short time. Therefore, increasing the flow rate, due to the limited rate of adsorption of analytes on the surface of CNFs, leads to a decrease in recovery percent of OPPs. From the Figure 4d, it is observed that increasing the amount of salt cause an an increase in recovery of ethion and the maximum ethion recovery was found to be at 200 mg. Increasing trends of recovery percent were also found for fenitrothion and diazinon. This effect M can be attributed to decrease of analyte solubility in water samples with increasing ionic strength. Thus, the partitioning from the aqueous solution to the acceptor is improved and therefore the ed recovery percent increases [8]. The same result was found to be correct by Maddah et al. [8] when they studied the application of electrospun polystyrene nanofiber as an efficient sorbent for pt preconcentration and determination of organophosphorus pesticides in water. Figure 5 about here Ac ce The response surface and contour plots for ethion recovery are illustrated in Figure 5. Figure 5a shows the effect of sorbent and salt amount (mg) on recovery percent of ethion. This figure presents the response surface and the contour plots as an estimate of recovery (%) as a function of two parameters: amount of sorbent and amount of salt (solution pH= 7, flow rate= 20 mL min-1). It can be seen that ethion recovery increases as the amount of salt is increased at low 11 sorbent values. For a 13 mg sorbent, increasing the salt content from 0 to 150 mg leads to an increase in ethion recovery from 2 to 20%. However, at higher sorbent values, the decrease in ethion recovery with an increase in salt content is observed. Further investigation shows that at rip t any given salt content, increasing the amount of sorbent leads to an increase followed by decrease in recovery percent of analyte. The result indicated that the optimum amount of sorbent us c for the extraction of the investigated pesticide was about 15 mg. The response surface and the contour plots in Figure 5b represented the recovery percent of ethion at different solution pH and amount of salt. It is obvious that at any given salt content, the recovery percent of ethion an increases with decreasing the solution pH. It can be seen that maximum recovery percent of ethion (60%) was achieved at lowest solution pH as well as salt content. By visualizing these M results it can be said that salt content played a small role in recovery percent of ethion in combination with solution pH. However, it was the key component in combination with amount ed of sorbent. Three-dimensional surface plots reveals that the interactions between salt content and pt amount of sorbent and also between solution pH and salt content were statistically significant. Figure 6 shows the combined effect of varying salt content and sorbent value on recovery ce percent of fenitrothion under solution pH and flow rate of 7 and 20 mL min-1, respectively. It can be seen that at higher sorbent values, increasing the salt content leads to decrease in recovery. On Ac the other hand, at lower sorbent values, the fenitrothion recovery increases with increasing salt content. In fact, maximum recovery percent of fenitrothion (30%) was obtained at the lowest sorbent value (10 mg) and highest salt content (200 mg). Moreover, fenitrothion recovery percent increased, when salt content was at low value and amount of sorbent was increased from 10 to 20 mg. 12 Figure 6 about here Optimization rip t Response surface methodology has been used successfully to optimize the parameters affecting the recovery percent. The optimal conditions of the analytical parameters for efficient extraction us c of analytes were established from the quadratic form of the RSM. Independent variables, namely, the amount of sorbent (mg), solution pH, flow rate (mL min-1), and amount of salt (mg) were set in the range and dependent variables (recovery percent of ethion, fenitrothion, and diazinon) was an fixed at maximum. The optimum conditions in the tested range for maximum recovery percent of OPPs were: amount of sorbent= 12.38 mg, solution pH= 4, flow rate= 21.44 mL min-1, and M amount of salt= 199.89 mg. The obtained value of desirability (1) shows that the estimated function may represent the experimental model and desired conditions. In order to confirm the ed predictive ability of the RSM model for response, a further experiment was carried out according to the optimized conditions. It was observed that the experimental value obtained was in good pt agreement with the value predicted from the model, with relatively small error between the ce predicted and the actual values. Ac Application to real sample analysis The present method was applied to the preconcentration and determination of ethion, fenitrothion, and diazinon in Rodsar river water samples. The recoveries of the spiked ethion, fenitrothion, and diazinon were determined and the results shown in Table 4. 13 Figure 7 shows the HPLC chromatograms obtained following preconcentration of 2.0 mL of unspiked (Figure 7a) and spiked (at 3 ng mL-1 for each analyte; Figure 7b) water sample on PFSPE under optimized conditions. The retention times were 5.1 min for ethion, 5.9 min for obtained at 105%, 98%, and 102%, respectively. us c Figure 7 about here rip t diazinon and 7.1 min for fenitrothion. The recoveries of ethion, fenitrothion, and diazinon were an Comparison of PF-SPE with other methods A comparison of the important features of the proposed method with the other previously method M in the literature [23-26] was carried out for further demonstration of the superiority of this method. LOD: limits of detection ed a pt Table 5 gives information on methodologies for determination of OPPs in water samples. A comparison of this method with other previously method demonstrates the feasibility of using ce PF-SPE-HPLC method and its reliability for analysis of OPPs in real water sample. Ac Conclusion In this study, the fabrication of activated CNFs form electrospun PAN nanofibers and its applicability as sorbent for solid phase extraction of organophosphorus pesticides at trace levels in water samples were investigated. The obtained results showed good precision, linear dynamic 14 range, and was a convenient, safe, and simple method for the determination of trace quantities of ethion, fenitrothion and diazinon in water samples with satisfactory results. Moreover, the impact of four analytical parameters for efficient extraction of analytes, including the amount of sorbent rip t (mg), pH of the sample, flow rate (mL min-1), and amount of salt (mg) in the extraction of analytes was quantitatively determined by response surface model based on the CCD technique. us c Acknowledgements The authors gratefully acknowledge the Department of Chemistry of Imam Hossein University an and Iran Nanotechnology Initiative Council (INIC) for partial financial support of this work. M Appendix ed Table A-1. SEM micrographs and fiber diameter distribution of PAN nanofibers electrospun n (wt.%) Ac . Concentratio 1 Voltage 16 Distanc (kV) e (cm) 20 10 ce No pt under different conditions. Fiber diameter AFD StdFD distribution (nm) (nm) 294 31 SEM micrographs 15 16 10 15 3 8 20 15 4 8 10 10 5 9.6 398 18 100 19 48 9 123 31 ce pt ed M an us c rip t 2 12.5 Ac 15 16 14.4 15 12.5 7 12 12 12.5 8 12 18 12.5 9 ce 108 21 146 28 95 27 98 18 pt ed M an us c rip t 6 Ac 12 15 11 17 12 15 14 11 12 15 12.5 94 22 98 12 M an us c rip t 10 ed References: pt [1]. Inagaki, M.; Yang, Y.; Kang, F. (2012) Carbon nanofibers prepared via electrospinning. Adv. Mater., 24, 2547-2566. ce [2]. Yoon, S.-H.; Lim, S.; Song, Y.; Ota, Y.; Qiao, W.; Tanaka, A.; Mochida, I. (2004) KOH Ac activation of carbon nanofibers. Carbon, 42, 1723-1729. [3]. Arshad, S. N.; Naraghi, M.; Chasiotis, I. (2011) Strong carbon nanofibers from electrospun polyacrylonitrile. Carbon, 49, 1710-1719. 18 [4]. Wang, G.; Pan, C.; Wang, L.; Dong, Q.; Yu, C.; Zhao, Z.; Qiu, J. (2012) Activated carbon nanofiber webs made by electrospinning for capacitive deionization. Electrochimica Acta, 69, 65-70. rip t [5]. Hammel, E.; Tang, X.; Trampert, M.; Schmitt, T.; Mauthner, K.; Eder, A.; Pötschke, P. (2004) Carbon nanofibers for composite applications. Carbon, 42, 1153-1158. us c [6]. Zulin, Z.; Huasheng, H.; Xinhong, W.; Jianqing, L.; Weiqi, C.; Li, X. (2002) Determination and Load of Organophosphorus and Organochlorine Pesticides at Water from Jiulong River an Estuary. China Mar. Pollut. Bull., 45, 397–402. [7]. Ahmadi, F.; Rajabi, M.; Faizi, F.; Rahimi-Nasrabadi, M.; Maddah, B. (2014) Magnetic solid- M phase extraction of Zineb by C18-functionalised paramagnetic nanoparticles and determination by first-derivative spectrophotometry. Int. J. Environ. Anal. Chem., 94, 1123-1138. ed [8]. Maddah, B.; Javadi, S. S.; Mirzaei, A.; Rahimi-Nasrabadi, M. (2015) Application of Electrospun Polystyrene Nanofibers as Solid Phase Extraction Sorbent for the Preconcentration pt of Diazinon and Fenitrothion in Environmental Waters. J. Liq. Chromatogr. R. T., 38, 208-214. ce [9]. Prieto, A. (1999) Analysing organophosphorus pesticides in wines using graphitized carbon black extraction cartridges. Food Addit. Contam, 16, 57-61. Ac [10]. Slobodnik, J.; Öztezkizan, Ö.; Lingeman, H.; Brinkman, U. A. T. (1996) Solid-phase extraction of polar pesticides from environmental water samples on graphitized carbon and Empore-activated carbon disks and on-line coupling to octadecyl-bonded silica analytical columns. J. Chromatogr. A., 750, 227-238. 19 [11]. Wang, X.; Tang, Q.; Wang, Q.; Qiao, X.; and Xu, Z. (2014) Study of a molecularly imprinted solid‐phase extraction coupled with high‐performance liquid chromatography for simultaneous determination of trace trichlorfon and monocrotophos residues in vegetables. J. Sci. rip t Food. Agr., 94, 1409-1415. [12] Moghri, M.; Shamaee, H.; Shahrajabian, H.; and Ghannadzadeh, A. (2015) The effect of us c different parameters on mechanical properties of PA-6/clay nanocomposite through genetic algorithm and response surface methods. Int. Nano Lett., 5, 133-140. [13] Pourmortazavi, S. M.; Rahimi-Nasrabadi, M.; Fazli, Y.; Mohammad-Zadeh, M. (2015) M nano-plates. Appl. Phys. A., 119, 929-936. an Statistical optimization of synthesis procedure and characterization of europium (III) molybdate [14] Honary, S.; Ebrahimi, P.; Rad, H. A.; and Asgari, M. (2013) Optimization of preparation of ed chitosan-coated iron oxide nanoparticles for biomedical applications by chemometrics approaches. Int. Nano Lett., 3, 1-5. pt [15] Rahimi-Nasrabadi, M.; Pourmortazavi, S. M.; Ganjali, M. R. (2015) Facile synthesis ce optimization and structure characterization of zinc tungstate nanoparticles. Mater. Manuf. Process., 30, 34-40. Ac [16] Shariati-Rad, M.; Irandoust, M.; Amri, S.; Feyzi, M; and Ja’fari, F. (2014) Magnetic solid phase adsorption, preconcentration and determination of methyl orange in water samples using silica coated magnetic nanoparticles and central composite design. Int. Nano Lett., 4, 91-101. 20 [17]. Myers, R. H.; Montgomery, D. C.; Anderson-cook, C. M. (2009) Response Surface Methodology: Process and Product Optimization using Designed Experiments, 3rd ed.; John Wiley and Sons: USA. rip t [18]. Gu, S. Y.; Ren, J.; Vancso, G. J. (2005) Process optimization and empirical modeling for electrospun polyacrylonitrile (PAN) nanofiber precursor of carbon nanofibers. Eur. Polym. J., us c 41, 2559-2568. [19] Larki, A.; Rahimi Nasrabadi, M.; Pourreza, N. (2015) UV-Vis spectrophotometric determination of trinitrotoluene (TNT) with trioctylmethylammonium chloride as ion pair an assisted and disperser agent after dispersive liquid–liquid microextraction. Forensic Sci. Int., M 251, 77-82. [20]. Pavia, D. L.; Lampman, G. M.; Kriz, G. S.; Vyvyan, J. R. (2001) Introduction to ed Spectroscopy, 4th ed.; Brooks/Cole; Cengage Learning: Independence, KY. [21] Konrad, J. G. ; Chesters , G.; Armstrong, D. E. (1969) Soil degradation of malathion, a pt phosphorodithioate insecticide. Solid Sci. Soc. Am. Proc., 33, 259-262. ce [22]. Han, X.; Balakrishnan V. k.; Buncel, E. (2007) Alkaline Degradation of the Organophosphorus Pesticide Fenitrothion as Mediated by Cationic C12, C14, C16, and C18 Ac Surfactants. Langmuir, 23, 6519-6525. [23] El-Kabbany, S.; Rashed, M.M.; Zayed, M.A. (2000) Monitoring of the pesticide levels in some water supplies and agricultural land, in El-Haram, Giza (A.R.E.). J. Hazard. Mater., 72, 11-21. 21 [24] Lambropoulou, D. A.; Albanis, T. A. (2001) Optimization of headspace solid-phase microextraction conditions for the determination of organophosphorus insecticides in natural waters. J. Chromatogr. A., 922, 243-255. rip t [25] Liang, P.; Guo, L.; Liu, Y.; Liu, S. Zhang, T. (2005) Application of liquid-phase microextraction for the determination of phoxim in water samples by high performance liquid us c chromatography with diode array detector. Microchem. J., 80, 19-23. [26] Li, J.; Zhao, X.; Shi, Y.; Cai, Y.; Mou, S.; Jiang, G. (2008) Mixed hemimicelles solid-phase extraction based on cetyltrimethylammonium bromide-coated nano-magnets Fe3O4 for the an determination of chlorophenols in environmental water samples coupled with liquid Ac ce pt ed M chromatography/spectrophotometry detection. J. Chromatogr. A., 1180, 24-31. 22 Ac ce pt ed M an us c rip t Figure 1. SEM micrographs and fiber diameter distribution of (a) PAN nanofiber and (b) PANderived CNFs. 23 Ac ce pt ed M an us c rip t Figure 2. FTIR spectra of (a) PAN and (b) PAN-derived CNFs. 24 Ac ce pt ed M an us c rip t Figure 3. Comparison between the actual and predicted value of (a) ethion and (b) fenitrothion recovery. 25 Ac ce pt ed M an us c rip t Figure 4. Main effect plots of factors on ethion recovery: (a) amount of sorbent, (b) solution pH, (c) flow rate, and (d) amount of salt. 26 Ac ce pt ed M an us c rip t Figure 5. Response surface and contour plots showing the effect of: (a) salt and sorbent amount, and (b) salt amount and solution pH, on ethion recovery. 27 Ac ce pt ed M an us c rip t Figure 6. Response surface and contour plots showing the effect of salt and sorbent amount on fenitrothion recovery. 28 Ac ce pt ed M an us c rip t Figure 7. HPLC chromatograms of Roudsar river water sample (a) unspiked and (b) spiked with 3 ng mL-1 ethion, fenitrothion, and diazinon under optimized conditions. 29 Table 1. Design of experiment (factors and levels). Actual values Coded flow rate, solution pH, / mg / amount of salt, 2 10 4 0 15 7 1 20 10 10 Ac ce pt ed M an -1 30 / mg rip (mL min-1) us c 1 4 t amount of sorbent, values 3 0 20 100 30 200 Table 2. Experimental design and results of the central composite design. Variables diazinon ethion Sorbent / solution flow rate / salt / mg pH (mL min-1) mg fenitrothion t Run Responses recovery / recovery / % 20 100 2 15 4 20 100 3 20 10 30 0 4 15 7 30 5 10 7 6 15 7 20 8 15 68.41 66.80 37.82 65.37 10.42 4.13 70.42 100 18.46 7.21 78.74 20 100 2.66 10.62 65.87 10 20 100 25.04 2.57 52.47 10 10 0 17.55 4.50 44.37 20 100 23.21 12.50 61.79 ed pt ce 9 9.40 7 21.47 us c 7 an 15 % M % 1 rip recovery / 20 4 30 200 9.81 8.26 7.15 10 4 30 0 17.95 5.10 68.06 11 15 7 20 100 24.20 6.66 59.74 12 15 7 10 100 22.31 14.07 65.73 13 15 7 20 100 22.48 13.13 58.81 Ac 10 31 10 4 10 0 28.76 2.11 64.78 15 15 7 20 0 12.68 6.18 58.66 16 15 7 20 200 28.03 13.43 65.55 17 20 4 10 200 21.22 18 10 10 30 200 20.95 19 10 10 10 200 20 20 7 20 100 21 15 7 20 rip t 14 65.80 11.27 63.94 34.03 19.60 76.71 11.75 5.74 37.49 an us c 9.20 11.48 67.24 Ac ce pt ed M 100 32 33.99 Table 3. ANOVA analysis for responses Y1 [ethion recovery] and Y2 [fenitrothion recovery]. Sum of Degree of Prob > F- Mean F-value Source freedom square Model 3163.93 14 226.00 Residual 122.23 6 Lack of Fit 18.78 2 Pure error 103.45 4 Total 3286.17 20 value t squares rip For Y1 us c 11.09 0.0037 20.37 0.36 0.7163 8.15 0.0083 2.23 0.2233 an 9.39 M 25.86 ed R 2 = 0.9628 Adeq Precision = ce pt 16.814 Ac For Y2 Model 1116.66 14 79.76 Residual 58.71 6 9.79 Lack of Fit 30.97 2 15.49 33 Pure error Total 27.74 4 6.94 1175.38 20 rip t R 2 = 0.9500 Adeq Precision = Ac ce pt ed M an us c 13.333 34 Table 4. Results of determination and recoveries of Rodsar river water samples spiked at concentration level of 3 ng mL-1 of each OPPs. Initial Found / ng mL-1 Recovery percent / % ethion Not detected 3.15 105 fenitrothion Not detected 2.94 diazinon Not detected us c rip t OPPs Ac ce pt ed M an 3.06 35 98 102 Table 5. Comparison of the current PF-SPE method with the other previously method for determination of OPPs in water samples. Final determination Recovery LODa Ref. technique rip percent / % t Method SPE GC-MS 8.9–19 µg L-1 HS-SPME GC-FTD/MS 10–40 ng L-1 80–120 [24] LPME HPLC-DAD 10 µg L-1 92.3–96.7 [25] Magnetic SPE HPLC-UV 0.11-0.15 µg L-1 83-98 [26] PF-SPE HPLC-DAD 0.09-0.22 µg L-1 37-78 This work us c an M ed pt ce Ac 36 - [23]