Automated Electric-Field-Driven Membrane Extraction System

Article pubs.acs.org/ac

Automated Electric-Field-Driven Membrane Extraction System Coupled to Liquid Chromatography−Mass Spectrometry Hong Heng See*,†,‡,§ and Peter C. Hauser*,† †

Department of Chemistry, University of Basel, Spitalstrasse 51, 4056 Basel, Switzerland Ibnu Sina Institute for Fundamental Science Studies, Universiti Teknologi Malaysia, 81310 UTM Skudai, Johor, Malaysia

‡

ABSTRACT: An automated analyte electroextraction and preconcentration system, which was used as the front end for a liquid chromatography− electrospray mass spectrometry instrument, is described. The extraction was based on driving the anionic analytes across a polymer inclusion membrane by application of a potential of 200 V to the cell. Five milliliters (5 mL) of sample were passed through a flow-through cell at a flow rate of 0.2 mL/min containing a membrane 20 μm thick. This consisted of 75% cellulose triacetate as base polymer, 12.5% of tris(2-ethylhexyl)phosphate as plasticizer, and 12.5% of Aliquat 336 as cationic carrier. The target analytes were enriched in 20 μL of a stagnant acceptor solution prior to online LC/MS analysis. The performance of the system was demonstrated for the determination of chlorinated phenoxyacetic acid herbicides in spiked river water. Enrichment factors of ∼200 were achieved with recoveries of typically 99% and precision values of typically 5%. The limit of detection (LOD) values were found to be between 0.03 ng/mL to 0.08 ng/mL.

F

with extraction efficiencies up to 99% was demonstrated. The membranes demonstrated a significantly improved mechanical robustness and easier handling, in comparison to our previous experience with the SLMs based on impregnated porous polypropylene,13−15 and since the cellulose acetate membranes are completely dry, they can be stored and do not have to be prepared immediately before use. When employed in flowthrough cells, they displayed a robust pressure tolerance. This work was followed by more-fundamental studies on the effect of the membrane composition and on the nature of the transported species, i.e., organic and inorganic anions and cations, on the extraction efficiency.16,17 Further progress was made by the introduction of flow cells to optimize the separation.16,18 In these systems, sample solution was passed continuously through the donor chamber with the help of a peristaltic or syringe pump, while the acceptor solution was stagnant. The thin layer arrangement that has been employed promotes efficient extraction and the enrichment factor can be easily varied by adjustment of the volume of the sample passed through. For the PIM-based electro-driven preconcentration cells reported so far, the acceptor solution had to be removed manually after completion of the extraction and for quantification was then injected into a capillary electrophoresis or ion chromatography instrument in a conventional way.12,16−18 The automated system reported herein makes use

requently, analyte extraction is required prior to quantification in order to achieve the required detection limits and/or for matrix elimination. This can be achieved by solvent extraction and is currently often performed with solid phase extraction (SPE) cartridges. In recent years, a range of advanced miniaturized methods have been proposed in order to be able to deal with small sample volumes, to speed up or simplify the procedure, or to reduce the amount of consumables.1−7 One interesting reported variation has been the technique termed “electromembrane extraction” (EME), which is an electrodialytic method that is based on supported liquid membranes (SLMs).8,9 The latter consist of a porous material that is impregnated with a water-immiscible organic solvent, such as octanol, through which the ions to be extracted migrate upon application of an electric field. Frequently, porous hollow fibers made from polypropylene have been employed. The fibers are dipped into the sample solution and an acceptor solution is placed inside the fiber. Electrodes placed into the two solutions allow the application of the extraction voltage. The main advantage of these methods, compared to conventional extraction techniques, is the high transfer rate of this single-step extraction procedure, since the analyte transport process is not limited by slow passive diffusion. Because the SLMs have limited mechanical stability,10,11 the use of polymer inclusion membranes (PIMs) for electric-fielddriven membrane extraction was recently investigated by us.12 The membranes were based on a homogeneous nonporous material consisting of cellulose triacetate as a base polymer, onitrophenyl octyl ether (NPOE) as a plasticizer, and Aliquat 336 as an anion carrier. The extraction of selected alkylsulfonates and anionic herbicides present in water samples © 2014 American Chemical Society

Received: April 29, 2014 Accepted: August 2, 2014 Published: August 2, 2014 8665

dx.doi.org/10.1021/ac5015589 | Anal. Chem. 2014, 86, 8665−8670

Analytical Chemistry

Article

of a miniaturized flow through extraction cell. For liquid handling, a sequential-injection analysis (SIA) manifold,19 based on a reversible syringe pump, a selection valve, and a holding coil, was employed. This was directly coupled to a HPLC-MS instrument, so that the EME-SIA system functions as a sophisticated online sample processing front end. The use of EME prior to quantification by HPLC has been reported (see, for example, refs 20−24) and Pedersen-Bjergaard and coworkers recently have also described similar online extraction systems based on SLMs.25,26 The utility of the integrated system proposed herein was investigated by the demonstration of the determination of chlorinated phenoxyacetic acids (CPAs) in environmental water samples down to ∼0.1 ng/mL. These herbicides are widely used for the selective control of broadleaf weeds in crops; however, because of their toxicity to humans and animals, monitoring of residues and contamination is necessary. Currently, the U.S. Environmental Protection Agency (USEPA) has set the maximum tolerated contaminant level of 2,4-dichlorophenoxyacetic acid (2,4-D) in drinking water to 70 μg/L.27 The European Union (EU) has set the drinking water standard for any herbicide/pesticide at 0.1 μg/L, regardless of its toxicology.

Modification of the counterion in the membranes was achieved by immersing the prepared membranes for 24 h in 20 mL of a stirred 2.0 M NaClO4 solution. The thickness of the membrane containing 2.5 mg of Aliquat was determined with a digital micrometer (MDC-1, Mitutoyo Corporation, Kawasaki, Japan) to be ∼20 μm. Extraction System. The assembly was based on a two-way syringe pump (Cavro XLP 6000) fitted with a 10-mL syringe and a nine-port channel selection valve (Cavro Smart Valve), both purchased from Tecan (Crailsheim, Germany). All fluid connections to the selector valve were made with Teflon PFA tubing that had an inner diameter (ID) of 0.02 in. and an outer diameter (OD) of 1/16 in. (IDEX, Oak Harbor, WA, USA). The extraction cell was made of two poly(methyl methacrylate) (PMMA) blocks with dimensions of 20 mm × 35 mm × 10 mm and with an engraved flow channel with a length of 20 mm, a width of 4 mm, and a depth of 0.25 mm. The membrane (30 mm length × 6 mm width) was placed between the two halves and these were held together with four M3 screws. The exposed membrane area was 80 mm2. Appropriate holes were cut into the two blocks to attach tubings with 1/4 in. (1/4-28-UNF) fittings and platinum wires (0.5 mm diameter) as electrodes. The extraction voltage was applied using a small high-voltage module (CA05N, EMCO, Sutter Creek, CA, USA). The operation of pump, selection valve, high-voltage module, and LC sample injection valve were controlled from a computer via an Arduino Mega microcontroller board (RS Components, Wädenswil, Switzerland) programmed using the Arduino integrated development environment. The RS232-driver (MAX232) to interface between the microcontroller and the pump, as well as the valve, was obtained from Maxim Integrated (San Jose, CA, USA). The operational amplifier for the current monitor was a TL071 (Texas Instruments, Dallas, TX, USA). The 19-in. rack and chassis were obtained from Schroff (Straubenhardt, Germany). HPLC-MS Quantification. Quantitation was carried out on a system consisting of a Merck Hitachi LaChrom L-7100 pump (Darmstadt, Germany) equipped with an injector fitted with a 10-μL sample loop and connected via an electrospray interface to a Thermo Finnigan LCQ-Deca mass spectrometer (Thermo, San Jose, CA, USA). Chromatographic separation was performed with a 3 μm Nucleodur EC 150/2 C-18 column (15 cm × 2.0 mm ID) (Macherey-Nagel, Oensingen, Switzerland) thermostated at 60 °C (the relatively high temperature was employed to reduce the viscosity and thus allow a higher flow rate for faster separation), using a L-7200 column oven (Merck Hitachi, Darmstadt, Germany). The mobile phase consisted of methanol (solution A) and 5 mM aqueous ammonium acetate (solution B). Gradient elution at a total flow rate of 0.5 mL/min was as follows: 20% A linear to 45% A in 8 min, followed by an increase to 80% A at 10 min. The mobile phase was readjusted to the initial conditions by returning the fraction of solution A to 20% in 1 min. The equilibrium time before the next injection was 9 min and the total run time was 28 min. The ESI-MS analyses of phenoxyacetic acid herbicides (extract obtained after electrodriven extraction) were carried out using selected ion monitoring (SIM) in the negative ion mode. To establish the appropriate SIM conditions standard solutions were injected directly into the mass spectrometer. The peak areas of the ions at m/z 185.3, 219.2, 219.0, and 253.0 for 4-CPA, 3,4-DA, 2,4-D, and 2,4,5-T, respectively, were determined.

■

EXPERIMENTAL SECTION Chemicals, Reagents, and Sample Preparation. 4Chlorophenoxyacetic acid (4-CPA), 3,4-dichlorophenoxyacetic acid (3,4-D), 2,4-dichlorophenoxyacetic acid (2,4-D), 2,4,5trichlorophenoxyacetic acid (2,4,5-T), sodium hydroxide, cellulose triacetate (CTA), tris(2-ethylhexyl)phosphate (TEHP), sodium perchlorate (NaClO4), and Selectophoregrade dichloromethane were obtained from Fluka (Buchs, Switzerland). Aliquat 336 (a lipophilic quaternary amine salt with a mixture of C8 and C10 chains with C8 predominating) was purchased from Aldrich (Milwaukee, WI, USA). Ultrapure deionized (DI) water was produced on a Barnstead Nanopure water purification system (Thermo Scientific, Reinach, Switzerland). All other reagents were of analytical grade and used without any further purification. Stock solutions of 4-CPA, 3,4D, 2,4-D, and 2,4,5-T at a concentration of 1 mg/mL were prepared in methanol and were stored in a refrigerator. Standard solutions for studying the extraction performance contained the target analytes at a concentration of 5 ng/mL and contained a background of 100 μM NaClO4. River water was collected from the Rhine River, Basel, Switzerland in Teflon bottles precleaned with acetone and filtered through 0.45 μm membrane filters and stored at −4 °C until analysis. To this, a background of 100 μM NaClO4 was also added (0.5 μL of 1 M NaClO4 was added to 5 mL of river water). Analysis of the blank river water showed no contamination by the analyte species. The target analytes were then spiked into the blank river water sample at trace level. The octanol−water partitition coefficients (log P) were calculated with the ACD/Chemsketch software package from Advanced Chemistry Development, Inc. (http://www.acdlabs.com, Toronto, Ontario, Canada). Membrane Preparation. The membranes were prepared by casting a solution of 15 mg of CTA as the base polymer, 2.5 mg of tris(2-ethylhexyl)phosphate (TEHP) as the plasticizer, and varying amounts of Aliquat 336 (i.e., 2.5, 5, 7.5, or 10 mg, corresponding to concentrations of 12.5%, 22.2%, 30%, and 36.4%, respectively) as a cationic carrier in 2 mL of dichloromethane (DCM). The solutions were poured and spread evenly into a 8-cm-diameter glass Petri dish and the solvent was allowed to evaporate gradually overnight. 8666

dx.doi.org/10.1021/ac5015589 | Anal. Chem. 2014, 86, 8665−8670

Analytical Chemistry

Article

Figure 1. Schematic diagram of the automated electro-driven extraction system coupled to the HPLC-MS instrument. The miniature extraction cell has dimensions of 20 mm × 35 mm × 20 mm, and fluid handling is based on a standard SIA manifold.

■

RESULTS AND DISCUSSION System Design and Operation. A schematic diagram of the integrated system is shown in Figure 1. The central part is the miniature thin layer flow-through cell used for extraction, which had been adapted from earlier designs.16,17 The two flat channels have a depth of 0.25 mm and were designed thus for efficient transport. Both chambers have a volume of 20 μL. The solution on the acceptor side was stagnant during the extraction, and the volume of 20 μL was well matched to the volume of the loop of the injection valve of the HPLC system of 10 μL. The fluidic handling was carried out with a two-way syringe pump and a multiport valve arranged in a standard sequential injection analysis (SIA) manifold. The total volumes of sample and acceptor solution passed through the donor and acceptor cell channels were fully controlled by the programmable syringe pump. The fluidic and electronic parts were assembled into a standard 19-in. rack with a height of 12 in. All fluidic components, including pump, valves, connecting tubings, liquid containers, and extraction cell, were fixed onto the front panel, which allowed easy access for manipulations. A photograph of the system assembled in its rack is shown in Figure 2. A block diagram of the electronic control circuitry employed is shown in Figure 3. This was based on the Arduino electronics prototyping platform (www.arduino.cc). This open-source initiative provides standardized microcontroller printed circuit

Figure 3. Diagram of the control circuitry based on the Arduino opensource microcontroller platform. The input of the operational amplifier in the current-to-voltage convertor configuration, employed for monitoring of the cell current, is at ground potential.

boards, together with a streamlined integrated development environment (IDE) for software production. The latter hides much of the inherent complexity from the user, and thus enables relative newcomers to build electronic control systems previously beyond reach. We would like to point out that the Arduino platform greatly simplified the construction of the control circuitry and programming of the system reported here. Indeed, as Pearce has pointed out, there is a growing opensource hardware movement in the scientific community which can greatly benefit from such devices.28 The use of the board took care of most of the required electronic hardware, and only a few extra items were necessary. The syringe pump and multiposition valve were controlled with code sent through a serial port that was connected to the Arduino board via a level shifting interface (MAX232). The high-voltage module was controlled via a pulse width modulation (PWM) output and a low pass filter. Also provided was a trigger output to the HPLC system. During operation, the software ran on the microcontroller and the PC attached via a USB interface acted as a terminal for command-line-based user interaction with the program. Note that the circuitry also allowed measurement of the cell current with the help of the operational amplifier in the current follower configuration attached to the cell. This is not essential, but it does give feedback to the operator if the extraction is proceeding properly. A typical sequence of the operations, which starts with the aspiration of the sample into the manifold and ends with the injection into the HPLC system, is outlined in Table 1. Note that both sides of the extraction cell are connected to the multiposition valve and the SIA system is used to pump

Figure 2. Photograph of the extraction system showing the syringe pump, multiposition valve, and extraction cell mounted together with liquid containers on the front panel of a standard 19-in. rack (note that the connection to the HPLC-MS system is not shown here). 8667

dx.doi.org/10.1021/ac5015589 | Anal. Chem. 2014, 86, 8665−8670

Analytical Chemistry

Article

of donor and acceptor solution and the membrane, had to be considered. First of all, the compounds must be present in a charged form for electrokinetic transport. The herbicides are weak acids and are present in the anionic form at neutral pH. Nevertheless, the effect of the pH value of the donor solution was tested in the pH range of 5−9. It was found that, at pH >6, consistently high extraction efficiencies were obtained. Hence, the aqueous solutions with measured pH values in the range of 6−7.5 were used directly without addition of a pH buffer. A homogeneous plasticized cellulose acetate material including Aliquat 336 as a carrier was employed as a separation membrane. The membranes should be as thin as possible for fast transport while maintaining sufficient mechanical strength. A thickness of 20 μm had been found to be suitable.12 It had previously been found for octanesulfonate as a model for anionic analytes that the concentration of the carrier had a strong effect on the rate of extraction, with an optimum at 20%.12 This was later adopted for the extraction of the herbicide glyphosate.18 However, further studies with inorganic anions16 indicated that the optimum membrane composition may also depend on the nature of the species transported, and a moredetailed investigation17 with organic model analytes showed that the nature and concentration of the plasticizer also has a significant effect on the extraction efficiency. TEHP was used as a plasticizer for the membranes employed in the current project, since it gave the best results with the organic model analytes used in this previous study.17 Because of the number of parameters, a fully comprehensive optimization of the membrane composition for each new analyte species is not readily possible and was not carried out in this case, because it was not the main aim of the current project. Nevertheless, the effect of the carrier concentration on the extraction efficiency of the chlorinated phenoxyacetic acid herbicides was investigated and the results for a given set of extraction conditions are given in Figure 5 and confirm the dependence on the carrier content, as well as the effect of the nature of the species transported, also for these target compounds. There is a maximum in the extraction efficiency for the carrier concentration, which is most likely due to trapping of the analytes in the membrane for the high carrier contents as observed previously.17 The pattern for the transported species shows a clear correlation with their lipophilicities (log P values: 4-CPA, 2.0; 2,4-D, 2.6; 3,4-D 2.7; 2,4,5-T, 3.2). All solutions employed in the extraction contained 100 μM of NaClO4, to provide a background conductivity. The lipophilic perchlorate anion had previously been found to give better results than other inorganic background ions, in terms of extraction efficiency of organic model analyes,12 and the relatively low concentration was used as higher concentrations caused excessive currents, leading to instability of the system, as evidenced by the current monitor. This is at least partly due to excessive bubble formation at the electrode (due to water electrolysis) and perhaps also to Joule heating effects. Other important operating parameters are the applied voltage and the flow rate. The applied voltageand, hence, the field strength should be as high as possible for a high rate of electrophoretic transport through the membrane as these parameters are nearly proportionally related, and no transport is observed without an applied voltage.18 However, high voltages also lead to high currents. It was found that 200 V could be applied with the present system without leading to instability. The flow rate of the donor solution has an optimum: a flow rate that is too low will lead to excessively long extraction times, while a flow rate

Table 1. Outline of Operation Sequence step

operation

1 2 3 4 5 6

picking up acceptor solution dispensing acceptor solution to cell picking up sample solution starting the dispensing of sample to cell turning the extraction voltage on extraction while dispensing sample through cell (with repeated pickup of sample) turning the extraction voltage off transfer of the acceptor solution to HPLC injection valve triggering injection and data acquistion

7 8 9

position of selection valve 4 8 3 7 7 7 7 8 8

solutions through both cell halves. The electrolyte solutions are, of course, conducting electricity and therefore the fluid manifold is exposed to the voltage applied for extraction. In order to avoid complications, the acceptor side, which was connected to the HPLC injection valve, was always grounded while the extraction voltage was applied to the donor side. This means that the multiposition valve was subjected to the voltage. However, all wetted parts were made from polymeric materials so that the fluidic paths were insulated from the metallic materials. A slight current leakage was observed between solutions at different ports even for closed ports, but this was negligible and did not impair the operation of the extraction cell. The Arduino software also ensured that the extraction voltage was never turned on when the corresponding port was open. Operating Conditions. A chromatogram for the extract of a standard mixture of the four chlorinated phenoxyacetic acids, 4-CPA, 3,4-DA, 2,4-D, and 2,4,5-T, as was obtained after optimization of conditions, is given in Figure 4. In order to get the best extraction efficiency, several parameters concerning the analytes and the extraction process, as well as the composition

Figure 4. Chromatogram for the acceptor solution after electro-driven extraction of spiked river water (with an added background of 100 μM NaClO4) containing CPAs at 0.5 ng/mL. Membrane: 15 mg CTA, 2.5 mg TEHP, 2.5 mg Aliquat 336. Sample volume = 5 mL. Flow rate = 0.2 mL/min. Acceptor phase = 20 μL of 100 μM NaClO4. Applied voltage = 200 V. 8668

dx.doi.org/10.1021/ac5015589 | Anal. Chem. 2014, 86, 8665−8670

Analytical Chemistry

Article

solution) corresponded to 201-, 213-, 226-, and 235-fold for 4CPA, 3,4-D, 2,4-D, and 2,4,5-T, respectively; the maximum possible factor, as determined by the volumes of the donor and acceptor solutions, was 250. The intraday and interday variabilities for analyte peak intensities were found to be 8% and 10%, respectively (RSD). Spiked River Water. The potential of the automated system for the determination of CPAs in environmental water samples was examined by analyzing spiked river water. The method reproducibility was evaluated by determining the concentrations of the analytes spiked at three different concentrations (0.5, 50, and 200 ng/mL). The concentrations were obtained by comparison with the calibration curves for peak areas obtained with aqueous standards also subjected to the extraction, which compensates for the different degrees of extraction efficiency for the different analyte species. The results are summarized in Table 3, are certainly acceptable, and demonstrate that the extraction efficiencies are stable and not affected by the ionic background of the river water. Table 3. Recoveries and Reproducibilities for the Phenoxyacetic Acid Herbicides Extracted from Spiked River Water Samplesa

Figure 5. Effect of the concentration of Aliquat 336 in the membrane on the extraction efficiency for CPAs. Donor phase: 5 ng/mL CPAs in 5 mL 100 μM NaClO4 (pH 7). Acceptor phase: 20 μL of 0.1 mM NaClO4. Sample volume = 5 mL. Flow rate = 0.2 mL/min. Applied voltage = 200 V. Each data point represents the mean value of five measurements.

compound

that is too high is not efficient, in terms of the required sample volume. A rate of 0.2 mL/min was found to be optimum for the current setup. A single membrane could be used repeatedly, and a test for carry over, which was conducted by doing an extraction experiment with a blank following the extraction of a standard, was found to be negative. Method Validation. A series of experiments to determine linearity, limits of detection (LODs), limits of quantification (LOQs), repeatability, and analyte enrichment factor was performed. The results are given in Table 2. Calibration curves were acquired for standard solutions at eight concentration levels in the range from 0.1 ng/mL to 200 ng/mL, using optimized extraction conditions. The curves of peak area (in terms of counts per second, cts/s) versus analyte concentration (ng/mL) were found to be linear for this range with good correlation coefficients of at least 0.9994. The LOD values for the CPAs, which are in the range of 0.03−0.08 ng/mL, are well below the Maximum Contaminant Level (MCL) allowable for the contaminants in drinking water set by the USEPA (70 μg/L for 2,4-D) and EU (0.1 μg/L). The analyte enrichment factors also given in the table were determined for three significantly different concentrations (0.5, 50, and 200 ng/mL). They were found to be identical within the precision of the method with the relative standard deviation (RSD) being lower than 7%. The enrichment factors for the given conditions (5 mL of donor

amount added (ng/ mL)

amount found (ng/ reproducibility, RSD mL) (%)

4-CPA

200 50 0.50

199 49.4 0.49

5.5 4.8 5.2

3,4-D

200 50 0.50

199 49.6 0.49

4.9 5.5 5.1

2,4-D

200 50 0.50

199 49.5 0.49

5.8 5.2 5.3

2,4,5-T

200 50 0.50

199 49.7 0.49

5.5 5.2 4.8

a

Eight consecutive extractions were carried out for each concentration employing the same membrane throughout.

Finally, a test was carried out to evaluate if it is possible to achieve higher enrichment factors by increasing the sample-toacceptor volume ratio. Three river water samples of 5, 10, and 25 mL were spiked with 0.5, 0.25, and 0.1 ng/mL, respectively, of the CPAs. Note that the total amount of the CPAs was identical in each experiment and the concentration expected in the acceptor solution therefore was supposed to be the same in each case. The results in form of the chromatograms are shown in Figure 6. Clearly the peaks for all CPAs studied are

Table 2. Method Validation compound 4-CPA 3,4-D 2,4-D 2,4,5-T a

linearity range (ng/mL) correlation coefficient, r 0.1−200 0.1−200 0.1−200 0.1−200

0.9995 0.9996 0.9995 0.9994

LODa (ng/mL) 0.08 0.05 0.07 0.03

LOQb (ng/mL) repeatability (%), n = 10 for 0.5 ng/mL 0.28 0.17 0.23 0.11

4.7 5.2 4.8 5.1

enrichment factor 235 226 213 201

LOD for 3× signal-to-noise-ratio. bLOQ for 10× signal-to-noise-ratio. 8669

dx.doi.org/10.1021/ac5015589 | Anal. Chem. 2014, 86, 8665−8670

Analytical Chemistry

Article

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank the Swiss National Science Foundation (Grant No. 200020-137676/1) for financial support. The authors also would like to thank S. Krähenbühl for the mass spectrometer.

■

Figure 6. Chromatograms for CPAs enriched in the acceptor phase on extraction of spiked river water samples (with an added background of 100 μM NaClO4) at different concentrations and varying volumes, but identical amounts: (a) 5 mL of 0.5 ng/mL CPAs, (b) 10 mL of 0.25 ng/mL CPAs, and (c) 25 mL of 0.1 ng/mL CPAs each in water sample containing 0.1 mM NaClO4. Acceptor phase: 100 μM NaClO4 in 20 μL. Membrane: 15 mg CTA, 2.5 mg TEHP, 2.5 mg Aliquat 336. Flow rate = 0.2 mL/min. Applied voltage = 200 V.

comparable. Hence, the introduction of larger sample volumes into the extraction cell to increase the enrichment factor of the method is thus possible. Nevertheless, there was a drawback, because introducing larger sample volumes during the extraction will prolong the extraction time. Approximately 125 min was required to pass 25 mL of sample through the extraction cell. This may not always be suitable.

■

CONCLUSION The automated electro-driven extraction technique could be readily connected to the LC-MS system and provides a straightforward and robust means to extend the limit of detection (LOD). The single-step method requires a minimum of sample manipulation and has the advantage of not requiring any organic solvents (other than for the casting of the membrane). It should be applicable to other charged species of both polarities, such as organic acids, other pesticides, or different classes of drugs. Future work may also concern further optimizations of the speed of extraction, and will need to address a study of the lifetime of the membranes as well as the effect of the matrices of more-contaminated samples, such as wastewater.

■

REFERENCES

(1) Kocúrová, L.; Balogh, I. S.; Andruch, V. Microchem. J. 2013, 110, 599−607. (2) Hu, B.; He, M.; Chen, B. B.; Xia, L. B. Spectrochim. Acta, Part B 2013, 86, 14−30. (3) Miró, M.; Hansen, E. H. Anal. Chim. Acta 2013, 782, 1−11. (4) Kokosa, J. M. TRACTrends Anal. Chem. 2013, 43, 2−13. (5) Silva, E. A. S.; Risticevic, S.; Pawliszyn, J. TRACTrends Anal. Chem. 2013, 43, 24−36. (6) Huang, Z. Z.; Lee, H. K. TRACTrends Anal. Chem. 2012, 39, 228−244. (7) Lee, J. Y.; Lee, H. K.; Rasmussen, K. E.; Pedersen-Bjergaard, S. Anal. Chim. Acta 2008, 624, 253−268. (8) Pedersen-Bjergaard, S.; Rasmussen, K. E. J. Chromatogr. A 2006, 1109, 183−190. (9) Gjelstad, A.; Pedersen-Bjergaard, S. Anal. Methods 2013, 5, 4549− 4557. (10) Balchen, M.; Jensen, H.; Reubsaet, L.; Pedersen-Bjergaard, S. J. Sep. Sci. 2010, 33, 1665−1672. (11) de Gyves, J.; Rodríguez de San Miguel, E. Ind. Eng. Chem. Res. 1999, 38, 2182−2202. (12) See, H. H.; Hauser, P. C. Anal. Chem. 2011, 83, 7507−7513. (13) See, H. H.; Hauser, P. C.; Sanagi, M. M.; Ibrahim, W. A. W. J. Chromatogr. A 2010, 1217, 5832−5838. (14) Xu, L.; Hauser, P. C.; Lee, H. K. J. Chromatogr. A 2008, 1214, 17−22. (15) Sanagi, M. M.; See, H. H.; Ibrahim, W. A. W.; Naim, A. A. J. Chromatogr. A 2007, 1152, 215−219. (16) Schmidt-Marzinkowski, J.; See, H. H.; Hauser, P. C. Electroanalysis 2013, 25, 1879−1886. (17) See, H. H.; Hauser, P. C. J. Membr. Sci. 2014, 450, 147−152. (18) See, H. H.; Stratz, S.; Hauser, P. C. J. Chromatogr. A 2013, 1300, 79−84. (19) Cerdà, V.; Avivar, J.; Cerdà, A. Pure Appl. Chem. 2012, 84, 1983−1998. (20) Seip, K. F.; Gjelstad, A.; Pedersen-Bjergaard, S. J. Chromatogr. A 2013, 1308, 37−44. (21) Ramos-Payán, M.; Villar-Navarro, M.; Fernández-Torres, R.; Callejón-Mochón, M.; Bello-López, M. A. Anal. Bioanal. Chem. 2013, 405, 2575−2584. (22) Davarani, S. S. H.; Pourahadi, A.; Nojavan, S.; Banitaba, M. H.; Nasiri-Aghdam, M. Anal. Chim. Acta 2012, 722, 55−62. (23) Alhooshani, K.; Basheer, C.; Kaur, J.; Gjelstad, A.; Rasmussen, K. E.; Pedersen-Bjergaard, S.; Lee, H. K. Talanta 2011, 86, 109−113. (24) Lee, J.; Khalilian, F.; Bagheri, H.; Lee, H. K. J. Chromatogr. A 2009, 1216, 7687−7693. (25) Li, B.; Petersen, N. J.; Payan, M. D. R.; Hansen, S. H.; PedersenBjergaard, S. Talanta 2014, 120, 224−229. (26) Dugstad, H. B.; Petersen, N. J.; Jensen, H.; Gabel-Jensen, C.; Hansen, S. H.; Pedersen-Bjergaard, S. Anal. Bioanal. Chem. 2014, 406, 421−429. (27) National Primary Drinking Water Regulations; Office of Water, U.S. Environmental Protection Agency (USEPA): Washington, DC, May 2009. (28) Pearce, J. M. Science 2012, 337, 1303−1304.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *Tel.: +41 61-267-1003. Fax: +41 61-267-1013. E-mail: Peter. [email protected]. Present Address §

Australian Centre for Research on Separation Science (ACROSS), School of Physical Sciences Chemistry, University of Tasmania, Hobart, Tasmania 7001, Australia. 8670

dx.doi.org/10.1021/ac5015589 | Anal. Chem. 2014, 86, 8665−8670