Electro-driven extraction across a polymer inclusion membrane in a flow-through cell

Electro-driven extraction across a polymer inclusion membrane in a flow-through cell

Journal of Chromatography A, 1300 (2013) 79–84 Contents lists available at SciVerse ScienceDirect Journal of Chromatography A journal homepage: www...

567KB Sizes 0 Downloads 29 Views

Journal of Chromatography A, 1300 (2013) 79–84

Contents lists available at SciVerse ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Electro-driven extraction across a polymer inclusion membrane in a flow-through cell Hong Heng See a,b,∗ , Simone Stratz a , Peter C. Hauser a,∗∗ a b

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

a r t i c l e

i n f o

Article history: Available online 22 January 2013 Keywords: Electro-membrane extraction Polymer inclusion membrane Capillary electrophoresis Contactless conductivity detection Glyphosate

a b s t r a c t A flow-through arrangement for electrodriven extraction across a polymer inclusion membrane was developed. Sample introduction into the donor chamber was continuous, while the acceptor solution was stagnant. By adjustment of the total volume of the donor solution pumped through the cell the best compromise between enrichment factor and extraction time can be set. The enriched extract was analyzed by capillary electrophoresis with contactless conductivity detection. Membranes of 20 ␮m thickness were employed which consisted of 60% cellulose triacetate as base polymer, 20% o-nitrophenyl octyl ether as plasticizer, and 20% Aliquat 336. By passing through 10 mL of sample at a flow rate of 1 mL/min the model analytes glyphosate (a common herbicide) and its major metabolite aminomethylphosphonic acid could be transported from the aqueous donor solution to the aqueous acceptor solution with efficiencies >87% in 10 min at an applied voltage of 1500 V. Enrichment factors of 87 and 95 and limits of detection down to 43 and 64 pg/mL were obtained for glyphosate and aminomethylphosphonic acid, respectively. The intra- and interday reproducibilities for the extraction of the two compounds from spiked river water were about 6 and 7% respectively when new membranes were used for each experiment. For consecutive extractions of batches of river water with a single piece of membrane a deterioration of recovery by about 16% (after 20 runs) was noted, an effect not observed with purely aqueous standards. © 2013 Elsevier B.V. All rights reserved.

1. Introduction The use of electric fields for sample pretreatment in order to achieve pre-concentration of charged analytes or a matrix clean-up has seen considerable recent attention. Compared to conventional extraction techniques the methods are faster as the transport processes are not limited by passive diffusion. Back-extraction (into an aqueous solution) is also not normally needed. An introduction to the methods can be found in several recent review articles [1–4]. Prominent among these techniques has been the method termed electro-membrane extraction (EME) [5–9]. This approach has been based on supported liquid membranes (SLM) usually employing porous hollow fibers made from polypropylene. The material is impregnated with an organic solvent such as 1-heptanol, 1-octanol, o-nitrophenyl octyl ether (NPOE) or ethyl nitrobenzene. Extraction of either anionic or cationic species is then achieved by applying

∗ Corresponding author at: Department of Chemistry, University of Basel, Spitalstrasse 51, 4056 Basel, Switzerland. Tel.: +41 61 267 1053; fax: +41 61 267 1013. ∗∗ Corresponding author. Tel.: +41 61 267 1003; fax: +41 61 267 1013. E-mail addresses: [email protected], [email protected] (H.H. See), [email protected] (P.C. Hauser). 0021-9673/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chroma.2013.01.062

a voltage of the appropriate polarity between the sample solution and the acceptor solution behind the membrane. However, a weakness of the SLMs is the limited mechanical stability of the solvent soaked membranes. Several researchers have indeed reported problems such as leaching of the organic liquid phase into sample and receiving solutions during the electrodriven extraction [10–14]. For this reason, a variation of the EME approach, which is based on a homogeneous, non-porous polymer inclusion membrane (PIM), rather than a supported liquid membrane, was recently developed by us [15]. The plasticized polymer membrane, which consisted of 60% cellulose triacetate as base polymer, 20% NPOE as plasticizer, and 20% Aliquat 336 as cationic carrier and had a thickness of 20 ␮m, demonstrated a significantly improved mechanical robustness and easier handling compared to our previous experience with the SLMs based on porous polypropylene [8,16]. In this arrangement the solvent is the plasticizer for the polymer rather than being suspended in pores. The membranes are dry and can be stored readily without deterioration. This robust material, which had been reported previously for use in waste water treatment [17,18], was proven to be suitable for extraction of lipophilic organic analytes using an applied voltage as driving force [15]. The strong mechanical structure of the PIM allowed repeated analyses without requiring membrane reconditioning steps.

80

H.H. See et al. / J. Chromatogr. A 1300 (2013) 79–84

The extraction cell used in the previous report consisted of two PTFE compartments clamped together with a flat membrane sandwiched between them [15]. The donor and receiving sections had volumes of 20 mL and 750 ␮L, respectively. This fixed volume ratio between the two chambers limited the maximum possible enrichment factor to 26. Such a restriction can be removed by employing a flow through cell for continuous extraction from a sample stream. The enrichment factor can then be adjusted by variation of the total volume of sample pumped through the cell. Recently, Petersen et al. indeed demonstrated flow-cell based EME employing SLMs for the determination of selected basic drugs [19–21]. Herein a flow-through cell for electro-driven extraction based on a solid PIM is described. It was designed with a large ratio of membrane surface to sample volume in order to facilitate efficient analyte transport. As model analytes glyphosate (GLYP) and aminomethylphosphonic acid (AMPA) were used. The widely used herbicide glyphosate and its degradation product are of importance in environmental analysis, and were chosen as model analytes in order to allow a comparison with earlier studies on miniature extraction methods conducted in our research groups [15,16]. For quantification of the poorly UV-absorbing species, capillary electrophoresis with capacitively coupled contactless conductivity detection (CE-C4 D) was employed. More details on this simple and versatile method for the determination of ionic species can be found in recent review articles [22–28]. 2. Experimental 2.1. Chemicals and reagents Aminomethylphosphonic acid (AMPA), Aliquat-336 (a commercial product mainly consisting of the lipophilic quaternary amine salt methyltrioctylammonium chloride) and sodium perchlorate (NaClO4 ) were obtained from Aldrich (Buchs, Switzerland). Glyphosate (GLYP), 3-(N-morpholino)propanesulfonic acid (MOPS), histidine (His), hexadecyltrimethylammonium bromide (CTAB), cellulose triacetate (CTA), o-nitrophenyl octyl ether (NPOE) and dichloromethane were purchased from Fluka (Buchs, Switzerland). Ultrapure deionized water was produced on a NanoPure water purification system (Barnstead, IA, USA). All other reagents were of analytical grade and used without any further purification. 2.2. Membrane preparation The polymer inclusion membranes (PIMs) were prepared by casting a solution of CTA (15 mg) as base polymer, NPOE (5 mg) as plasticizer and Aliquat-336 (5 mg) as cationic carrier in 4 mL dichloromethane. The solution was spread evenly into an 8 cm diameter glass petri-dish and the solvent was allowed to gradually evaporate overnight. All membranes tested were modified by replacing chloride by perchlorate as counter-ion by immersing the membrane for 24 h in 20 mL of a stirred 2.0 M NaClO4 solution. The membrane thickness was measured with a digital micrometer (MDC-1, Mitutoyo Corporation, Kawasaki, Japan) and was determined to be approximately 20 ␮m. 2.3. Solutions Stock solutions of GLYP and AMPA at a concentration of 1 mg/mL were prepared in deionized water and stored in the refrigerator. River water was collected from the Rhine in Basel, Switzerland. An analysis of this sample did not show the presence of GLYP and AMPA before adding these compounds to validate the method. Donor test solutions and spiked river water samples were obtained by appropriate dilution of the stock solution and NaClO4 was added to these

to obtain a final concentration of 1 mM NaClO4 in order to assure a constant background conductivity. All samples had a pH in the range between 6 and 7. The acceptor solution also contained 1 mM NaClO4 in all cases. 2.4. Electro-driven extraction system A high voltage DC power supply with negative voltage output in the range from 0 to 1500 V was obtained from Advance Hivolt (GM12-1K5N, Woburn, MA, USA). The peristaltic pump was obtained from Gilson (Minipuls 3, Villiers le Bel, France). The operational amplifier used for recording the current was a TL071 (Texas Instruments, Austin, TX, USA) and the data acquisition system an e-corder (eDAQ, Denistone East, NSW, Australia). 2.5. CE-C4 D analysis The CE experiments were performed on a PrinCE 500 2-lift system (Prince Technologies, Emmen, The Netherlands) equipped with a contactless conductivity detector system (eDAQ). The excitation frequency of the detector was set to 300 kHz and the amplitude was fixed at 100%. A bare fused silica capillary of 50 ␮m I.D. and 365 ␮m O.D. (Polymicro Technologies, Phoenix, AZ, USA) with total and effective lengths of 55 and 45 cm respectively was employed. The new capillary was conditioned by flushing with 1 M sodium hydroxide (NaOH) for 15 min and water for 5 min. The pretreated capillary was then rinsed with the running buffer for 30 min. After each analysis run, the capillary was rinsed for 3 min with the running buffer to maintain the reproducibility of the analysis. The running buffer consisted of 12 mM His, 8 mM MOPS and 50 ␮M CTAB (pH 6.3). Large volume sample stacking (LVSS) was performed using hydrodynamic injection for 100 s at 50 mbar [29]. The separation voltage was −30 kV. 2.6. Calculations The extraction recovery (R) and enrichment factor (EF) were calculated according to the following equations: R=

Va · Ca,final na = nd Vd · Cd,initial

EF =

Ca,final V =R× d Cd,initial Va

(1) (2)

The recovery is the ratio between the number of moles of the analyte collected in the acceptor solution (na ) and the number of moles initially present in the donor solution (nd ), which can be obtained from the acceptor and donor volumes (Va and Vd ) and the final and initial concentrations (Ca ,final and Cd ,initial ) of the two solutions. The enrichment factor is the ratio of the concentrations in the two solutions (and can also be obtained from the recovery and the volume ratio). 3. Results and discussion 3.1. Cell design The continuous electro-driven extraction setup is shown schematically in Fig. 1. The extraction cell was based on two polytetrafluoroethylene (PTFE) blocks each having the dimensions of 1 cm × 3.5 cm × 3.5 cm (width × height × depth). The donor and receiving compartments had a round cross-section of 2 cm diameter with depths of 1.6 mm and 0.3 mm and volumes of 500 ␮L and 100 ␮L respectively (Fig. 1). Appropriate holes allowed access for tubings and electrodes. The membrane used was cut in a square with each dimension being 2.5 cm. To assemble the extraction

H.H. See et al. / J. Chromatogr. A 1300 (2013) 79–84

81

and it had been found that 1 mM of KClO4 was best in terms of transfer efficiency due to the relative lipophilic nature of the perchlorate anion [15]. In preliminary tests (results not shown) it was also found that the membrane should be as thin as possible in order to obtain fast transfer rates, and thus a membrane with a thickness of 20 ␮m was employed corresponding to the lower limit of mechanical viability. 3.2.1. Effect of the pH of the donor phase As the molecular structures of GLYP and AMPA are highly pHdependent, the pH of the sample solution was expected to be an important factor influencing the extraction performance. The structures of GLYP and AMPA are given in Fig. 2 together with the ionization processes and pKa -values [30]. The effect of the pH value was first investigated in the range of 2–9 by adjusting the sample solution with 1.0 M HClO4 or NaOH solutions. An aliquot of 10 mL of a donor solution containing GLYP and AMPA at 500 ng/mL as well as 1 mM NaClO4 was passed through the cell using a peristaltic pump at a flow rate of 1 mL/min. The receiving solution was stagnant and consisted of 100 ␮L of 1 mM NaClO4 solution. The extraction voltage was fixed at 1000 V. As shown in Fig. 3a, a significant improvement of extraction recovery for GLYP was observed when the sample pH was increased from 2 to 6. GLYP exists in a zwitterionic state with a net charge of zero when present in the sample with pH 2. Hence, the rate of transport of GLYP was unsatisfactory at this point. When the donor pH was increased to 3 and 4, the carboxylic group of GLYP was deprotonated and the molecule possessed a net charge of −1. A significant increase in recovery was observed as GLYP started to electromigrate across the membrane. Optimum recovery was obtained for GLYP when the sample pH was increased to 6 and remained constant up to pH 9. A donor pH higher than pH 9 was found to result in non-reproducible recovery (results not shown). This might be due to a deterioration of the membrane due to hydrolysis of CTA in the alkaline medium [31]. On the other hand, the rate of transport of AMPA was found unsatisfactory with a donor pH in the range from 2 to 5 due to its existence in the zwitterion state. The recovery of AMPA was found to increase significantly at pH 6 and reached its optimum at pH 7. At this pH-value AMPA has a net charge of −1. The recovery remained constant when the pH was further increased to 8 and 9. Hence, a donor pH in the range of 6–7 (which does not require any adjustment of the pH in aqueous test solutions or river water) was employed for the subsequent experiments.

Fig. 1. Schematic illustration of the continuous electro-driven extraction setup.

cell, the two compartments were screw-clamped together with the membrane sandwiched between them. The membrane surface area exposed to the solutions was approximately 3.14 cm2 . Two platinum wires (0.5 mm diameter) were used as electrodes. Negative voltages were applied at the donor side. Note that the elevated voltages used in this project require special care to avoid accidental exposure, such as a safety cage fitted with a microswitch to interrupt the high voltage on opening. The electrode in the acceptor solution was at ground potential but connected to the input of an operational amplifier in the current-to-voltage convertor configuration to allow monitoring of the current passing through the cell. The sample solutions were introduced into the donor chamber by using a peristaltic pump at the desired flow rate. Besides the flow-through feature, this new cell has a much improved ratio of membrane surface area to donor volume of approximately 6.30 cm2 /mL compared to the previous design (approximately 0.06 cm2 /mL [15]) in order to expedite the analyte transport and speed up the overall process. 3.2. Optimization of operating conditions It had been shown in previous work that the anionic model analytes glyphosate (GLYP) and its metabolite aminomethylphosphonic acid (AMPA) could be transported by electromigration across a plasticized cellulose triacetate based membrane [15]. Its composition had been comprehensively optimized and the best mixture consisting of 60% cellulose triacetate, 20% o-nitrophenyl octyl ether (NPOE) as plasticizer, and 20% Aliquat 336 as cationic carrier in the perchlorate form was adopted again. A background electrolyte was used in both, donor as well as acceptor, solutions

O HO

HO H2N+

OH P

O

pK1 0.8

O

HO H2N+

HO

OP

O

pK2

O

-O

2.2

3.2.2. Effect of extraction voltage The effect of the applied voltage on the transport efficiency for GLYP and AMPA was then studied. The applied voltage was varied in the range from 0 to 1500 V and the results are shown in Fig. 3b. A strong dependence on the voltage is evident. First of all, the two target compounds are not detectable in the receiving solution if

HO H 2N +

OP

O

pK3

O

5.4

-O

O-

-

O

H2N+

P

O

pK4

O

10.2

-O

-O

HN

Glyphosate (GLYP)

HO H3N+

OH P

O

pK1 1.8

HO H3N+

OP

O

pK2

-

5.4

H3N+

O

OP

O

pK3 10.0

O-

-

O

H2N

P

O

Aminomethylphoshonic acid (AMPA) Fig. 2. Structures of glyphosate (GLYP) and aminomethylphosphonic acid (AMPA) with ionization processes and pKa -values.

OP

O

82

H.H. See et al. / J. Chromatogr. A 1300 (2013) 79–84

Fig. 3. Influence of (a) donor pH, (b) applied voltage, and (c) donor flow rate on the extraction recovery. Donor solution: 10 mL of 1 mM NaClO4 containing GLYP and AMPA at 500 ng/mL; acceptor solution: 100 ␮L of 1 mM NaClO4 ; extraction time: 10 min. Error bars represent standard deviation of results, n = 3.

no voltage is applied across the two compartments. This clearly shows that the transport of analytes is forced by the applied potential difference across the membrane and any passive diffusion is negligible. The recovery was almost linearly related to voltage up to 1000 V where a recovery of about 80% is achieved. For higher voltages (1250 and 1500 V) a saturation is apparent, with AMPA reaching almost 100% at 1500 V, but not GLYP which showed a maximum of about 82%. Note that for these high applied voltages relatively high currents of approximately 1.4 mA were obtained (not due only to the transport of the target ions [15]), which cause bubble formation because of water electrolysis at the electrodes. In order to avoid the disruption of current flow during extraction, the electrode in the donor compartment was placed at the flow outlet. In this case, excessive bubbles generated are immediately removed from the cell. On the receiving side an opening was provided next to the electrode to release any excessive bubbles and avoid any possible pressure increment in the receiving chamber. With this design, the current remained constant in all experiments and satisfactory reproducibility was achieved. Based on the results obtained, the applied voltage was fixed at 1500 V for subsequent experiments. Note that, due to the electrolysis of water, in the stagnant acceptor chamber the pH-value was found to decrease from 6–7 to about 3–4 (as tested with an indicator strip) during a run, but this was not found to cause any adverse effects. 3.2.3. Effect of the flow rate of the donor solution To evaluate the effect of the sample flow rate on the extraction recovery, a test solution containing 500 ng/mL of GLYP and AMPA in 1 mM of NaClO4 was passed through the extraction cell at different flow rates (0.5–3.0 mL/min). The extraction time was fixed at 10 min which meant that different total volumes, and different amounts of the target compounds, were pumped through the cell in the experiments. The effect of the sample flow rate on the extraction recovery of the analytes is illustrated in Fig. 3c. It can been noted that the recoveries increased with the increase of flow rate from 0.5 to 1.0 mL/min to maxima of 88% for GLYP and 96% for AMPA. At this optimum 10 mL of sample solution was processed in 10 min and enrichment factors of 88 and 96 for GLYP and AMPA respectively could be achieved. A further increase of flow rate from 1.5 to 3.0 mL/min resulted in gradually decreased recoveries. As the flow rate increased, the residence time of GLYP and AMPA at the sample/membrane interface decreased and presumably there was insufficient time for complete analyte permeation. When 30 mL of the test solution was passed through the cell at 3 mL/min, the sample-to-acceptor volume ratio was 300

corresponding to the theoretical maximum enrichment factor. However, as the recoveries for GLYP and AMPA were only 25 and 30%, the enrichment factors achieved were only approximately 75- and 90-fold respectively. Thus, the highest enrichment factors should only be possible when large sample volumes are introduced at relatively low flow rates which result in high extraction recovery. To prove this, a test volume of 20 mL was introduced into the flow-through cell at an optimum flow rate of 1 mL/min. Recoveries of GLYP and AMPA were found to remain constant and the enrichment factors were therefore indeed found to increase to 176 and 192-fold respectively (results not shown). In practice, one would generally use the flow rate with the highest recovery (1 mL/min) and then choose the sample volume which just yields the minimum required enrichment factor in order to keep the analysis time relatively short. 3.3. Method validation A series of experiments to determine linearity, limits of detection (LOD), repeatabilities, and recoveries was performed and the results are given in Table 1. Note that new membranes were used for each experiment. Calibration curves were acquired by using optimized extraction conditions for spiked standard solutions at eight concentration levels in the range from 0.1 to 500 ng/mL. The curves of peak area (mV s) versus analyte concentration (ng/mL) were found to be linear for this range with good correlation coefficients (regression equations: y = 3.8941x + 0.3291 for GLYP, and y = 3.5321x + 0.1767 for AMPA). The limits of detection (40 and 60 pg/mLfor GLYP and AMPA respectively) were found to be significantly lower than those obtained by EME with the cellulose acetate membrane employing the static-mode extraction cell (0.8 ng/mL and 1.5 ng/mL respectively)[15]. A comparison can also be made with the recently reported supported liquid membrane tip extraction (SLMTE) approach for these compounds employing a passive, diffusion based sampling arrangement [16]. Although the LODs for GLYP and AMPA using SLMTE were better (5 pg/mL and 40 pg/mL respectively), the extraction time required was 6 times longer (60 min) than the time required in the present study. The detection limits thus obtained are well below the legislated maximum concentration limits for drinking water set by the US Environment Protection Agency (0.7 ␮g/mL) and European Union (0.1 ␮g/L). The extraction recoveries also given in the table were determined for three significantly different concentrations (0.5, 250, and 500 ng/mL). They were found to be identical within the precision of the method and, importantly, are thus independent of

H.H. See et al. / J. Chromatogr. A 1300 (2013) 79–84

83

Table 1 Performance data for the continuous flow electro-driven extraction of GLYP and AMPA. Analytes

Concentration range (ng/mL)

Correlation coefficient (r)a

LOD (pg/mL)b

Recovery (%)c

Reproducibility (%)d Intra-day

Inter-day (3 days)

GLYP AMPA

0.1–500 0.1–500

0.9987 0.9981

43 64

86–88 94–96

3.5–5.5 3.7–5.4

5.5–7.2 5.4–7.6

a

From peak areas (mV s) for 8 concentrations. Calculated from signal-to-noise = 3. c For 3 concentration (0.5, 250 and 500 ng/mL). d From peak areas for eleven determinations each for 3 concentrations (0.5, 250 and 500 ng/mL). Extraction conditions: 10 mL donor solution containing 1 mM NaClO4 pumped at 1 mL/min; 100 ␮L of acceptor solution containing 1 mM NaClO4 ; extraction time, 10 min; applied voltage, 1500 V. b

the analyte level. The enrichment factors for the given conditions (10 mL of donor solution) corresponded to 87 and 95 for GLYP and AMPA respectively; the maximum possible factor as determined by the volumes of the donor and acceptor solutions was 100. The intra- and inter-day variabilities for peak areas were found to be in the single digit percentage range and are certainly acceptable. 3.4. Spiked river water In order to examine the potential of applying the new approach for the determination of GLYP and AMPA in environmental samples, the extraction recovery and method reproducibility (eleven consecutive experiments, new membranes were used each time) were evaluated by spiking the analytes at three different concentration (0.5, 250 and 500 ng/mL) into river water. The recovery results obtained were found to be 85–87% for GLYP and 93–95% for AMPA, respectively. These results were thus almost identical to the ones determined for the aqueous standards reported in the previous section. Similarly, the intra- and interday variabilities were also found to be comparable with values of 4.8–6.2% (RSD) and 5.3–7.1% respectively. For illustration, electropherograms for a blank river

Fig. 4. Electropherograms for (a) blank river water sample, (b) river water sample with GLYP and AMPA added at 100 ng/mL, (c) spiked river water sample after electrodriven extraction and (d) receiving solution after electro-driven extraction. PIM: CTA 60%, NPOE 20%, Aliquat-336 20%; flow rate: 1 mL/min; applied voltage: 1500 V; extraction time: 10 min.

Fig. 5. Variation of extraction recovery in continuous extraction of GLYP and AMPA in (A) unfiltered and (B) filtered river water samples. Sample: spiked river water containing GLYP and AMPA at 1 ng/mL. Other conditions as for Fig. 4.

84

H.H. See et al. / J. Chromatogr. A 1300 (2013) 79–84

water sample, a river water sample spiked with 100 ng/mL of the two compounds before and after extraction, as well as for the acceptor solution are shown in Fig. 4. The depletion of GLYP and AMPA from the donor solution is clearly evident as well as the significant up-concentration in the acceptor solution. The enrichment factors represented in the figure correspond to 88 for GLYP and 96 for AMPA. To evaluate the stability of the membrane, twenty batches of spiked river water containing 1 ng/mL of GLYP and AMPA were introduced consecutively into the cell for extraction. The enriched receiving solutions were collected at each 10 min interval and subjected to analysis. The acceptor compartment was refilled with 100 ␮L of fresh receiving solution after each collection. The same piece of membrane was used for all runs. Two sets of experiments were carried out, one with unfiltered, one with filtered water (0.45 ␮m polypropylene syringe filters). As can be seen from Fig. 5A and B a deterioration of the membrane is apparent for both runs, albeit for the filtered water the results were more stable. For the first 10 analyses the relative standard deviations for GLYP and AMPA were within 1.3 and 1.4, respectively. The recoveries then gradually decreased and a loss of recovery of approximately 16% was noted for the twentieth extraction in comparison to the first extraction. Similar results were observed with another batch of the membrane, but the deterioration was not present when working with aqueous standards (see the previous section) rather than river water. 4. Conclusions It could be demonstrated that the flow cell allows indeed adjustment of the enrichment factor by variation of the volume of the sample passed through. The expectation in this regard could therefore be fulfilled. For efficient extraction an optimum flow rate was found to exist and the time needed for an extraction is then determined by the enrichment factor required for an application at hand, and hence the sample volume that needs to be passed through the cell at this flow rate. The limits in terms of maximum volume of donor solution and lowest possible analyte concentration have not yet been explored. In comparison to the batch-wise extraction cell reported previously [15], it was found that the cell was about 7 times more efficient in terms of the product of enrichment factor and extraction time. This must be due to geometrical factors alone, as the membrane was identical, and further improvements in this regard should be possible.

Acknowledgement The authors are grateful for financial support by the Swiss National Science Foundation through Grant no. 200020-137676/1 and the Ministry of Higher Education Malaysia (MOHE) for financial support through Research University Grants (GUP). References [1] G. Morales-Cid, S. Cádenas, B.M. Simonet, M. Valcárcel, TrAC Trends Anal. Chem. 29 (2010) 158. ˇ ˇ A. Slampová, P. Boˇcek, Electrophoresis 31 (2010) 768. [2] P. Kubán, [3] C.J. Collins, D.W.M. Arrigan, Anal. Bioanal. Chem. 393 (2009) 835. [4] S. Pedersen-Bjergaard, K.E. Rasmussen, TrAC Trends Anal. Chem. 27 (2008) 934. [5] S. Pedersen-Bjergaard, K.E. Rasmussen, J. Chromatogr. A 1109 (2006) 183. [6] I.J.O. Kjelsen, A. Gjelstad, K.E. Rasmussen, S. Pedersen-Bjergaard, J. Chromatogr. A 1180 (2008) 1. [7] C. Basheer, S.H. Tan, H.K. Lee, J. Chromatogr. A 1213 (2008) 14. [8] L. Xu, P.C. Hauser, H.K. Lee, J. Chromatogr. A 1214 (2008) 17. ˇ L. Strieglerová, P. Gebauer, P. Boˇcek, Electrophoresis 32 (2011) [9] P. Kubán, 1025. [10] P.R. Danesi, L. Reichley-Yinger, P.G. Rikert, J. Membr. Sci. 31 (1987) 177. [11] A.J.B. Kemperman, D. Bargeman, T. van den Boomgaard, H. Strathmann, Sep. Sci. Technol. 31 (1996) 2733. [12] A.M. Sastre, A. Kumar, J.P. Shukla, R.K. Singh, Sep. Purif. Methods 27 (1998) 213. [13] J. de Gyves, E.R. de San Miguel, Ind. Eng. Chem. Res. 38 (1999) 2182. [14] M. Balchen, H. Jensen, L. Reubsaet, S. Pedersen-Bjergaard, J. Sep. Sci. 33 (2010) 1665. [15] H.H. See, P.C. Hauser, Anal. Chem. 83 (2011) 7507. [16] H.H. See, P.C. Hauser, M.M. Sanagi, W.A.W. Ibrahim, J. Chromatogr. A 1217 (2010) 5832. [17] L.D. Nghiem, P. Mornane, I.D. Potter, J.M. Perera, R.W. Cattrall, S.D. Kolev, J. Membr. Sci. 281 (2006) 7. [18] M. O‘Rourke, R.W. Cattrall, S.D. Kolev, I.D. Potter, Solvent Extr. Res. Dev. Jpn. 16 (2009) 1. [19] N.J. Petersen, H. Jensen, S.H. Hansen, S.T. Foss, D. Snakenborg, S. PedersenBjergaard, Microfluid. Nanofluid. 9 (2010) 881. [20] N.J. Petersen, S.T. Foss, H. Jensen, S.H. Hansen, C. Skonberg, D. Snakenborg, J.P. Kutter, S. Pedersen-Bjergaard, Anal. Chem. 83 (2011) 44. [21] N.J. Petersen, J.S. Pedersen, N.N. Poulsen, H. Jensen, C. Skonberg, S.H. Hansen, S. Pedersen-Bjergaard, Analyst 137 (2012) 3321. [22] W.K.T. Coltro, R.S. Lima, T.P. Segato, E. Carrilho, D.P. de Jesus, C.L. do Lago, J.A.F. da Silva, Anal. Methods 4 (2012) 25. [23] T.D. Mai, P.C. Hauser, Chem. Rec. 12 (2012) 106. ˇ P.C. Hauser, Electrophoresis 32 (2011) 30. [24] P. Kubán, [25] A.A. Elbashir, H.Y. Aboul-Enein, Biomed. Chromatogr. 24 (2010) 1038. [26] M. Trojanowicz, Anal. Chim. Acta 653 (2009) 36. ˇ P.C. Hauser, Anal. Chim. Acta 607 (2008) 15. [27] P. Kubán, [28] F.M. Matysik, Microchim. Acta 160 (2008) 1. [29] H.H. See, P.C. Hauser, W.A.W. Ibrahim, M.M. Sanagi, Electrophoresis 31 (2010) 575. [30] M.G. Cikalo, D.M. Goodall, W. Matthews, J. Chromatogr. A 745 (1996) 189. [31] J.S. Gardner, J.O. Walker, J.D. Lamb, J. Membr. Sci. 229 (2004), 87.