On-chip electromembrane extraction of acidic drugs

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This is the peer reviewed version of the following article: Roman-Hidalgo, C., Santigosa-Murillo, E., Ramos-Payán, M., Petersen, N.J., Kutter, J.P. and Pedersen-Bjergaard, S. (2019), On-chip electromembrane extraction of acidic drugs. ELECTROPHORESIS, 40: 2514-2521, which has been published in final form at https://doi.org/10.1002/elps.201900024. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions. This article may not be enhanced, enriched or otherwise transformed into a derivative work, without express permission from Wiley or by statutory rights under applicable legislation. Copyright notices must not be removed, obscured or modified. The article must be linked to Wiley’s version of record on Wiley Online Library and any embedding, framing or otherwise making available the article or pages thereof by third parties from platforms, services and websites other than Wiley Online Library must be prohibited." 1 On-chip electromembrane extraction of acidic drugs Cristina Roman Hidalgoa, Elia Santigosa Murillob, María Ramos Payána,*, Nickolaj J. Petersenc, Jörg P. Kutterc, Stig Pedersen-Bjergaardc,d,* aDepartment of Analytical Chemistry, Faculty of Chemistry, University of Seville, c/Prof. García González s/n, 41012, Seville, Spain bDepartment of Analytical Chemistry, Universitat Autónoma de Barcelona, 08193 Bellaterra, Barcelona, Spain cDepartment of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, 2100 Copenhagen, Denmark dSchool of Pharmacy, University of Oslo, P.O. Box 1068 Blindern, 0316 Oslo, Norway Abstract In the present work, a new supported liquid membrane (SLM) has been developed for onchip electromembrane extraction of acidic drugs combined with HPLC or CE, providing significantly higher stability than those reported up to date. The target analytes are five widely used non-steroidal anti-inflammatory drugs (NSAIDs): ibuprofen (IBU), diclofenac (DIC), naproxen (NAX), ketoprofen (KTP) and salicylic acid (SAL). Two different microchip devices were used, both consisted basically of two poly(methyl methacrylate) (PMMA) plates with individual channels for acceptor and sample solutions, respectively, and a 25 m thick porous polypropylene membrane impregnated with the organic solvent in between. The SLM consisting of a mixture of 1-undecanol and 2-nitrophenyl octyl ether (NPOE) in a ratio 1:3 was found to be the most suitable liquid membrane for the extraction of these acidic drugs under dynamic conditions. It showed a long-term stability of at least 8 hours, a low system current around 20 A, and recoveries over 94% for the target analytes. NPOE was included in the SLM to significantly decrease the extraction current compared to pure 1-undecanol, while the extraction properties was 2 almost unaffected. Moreover, it has been successfully applied to the determination of the target analytes in human urine samples, providing high extraction efficiency. 1. Introduction Sample preparation is a time consuming step in many analytical procedures. For this reason, substantial research has been devoted to the development of new sample preparation techniques in recent years. In this research, focus has been on analysis time, analyte pre-concentration, sample clean-up, automation, and miniaturization. The introduction of solid-phase microextraction (SPME) initiated this development [1], and was, among others, followed by single drop microextraction (SDME) [2], hollow-fiber liquid-phase microextraction (HF-LPME) [3], and electromembrane extraction (EME) [4]. The latter technique was introduced in 2006 and proposed the use of an external electrical field to promote the extraction of charged analytes from the sample to an acceptor solution across a SLM. EME is interesting because mass transfer is rapid due to the electrical field, and because pre-concentration and sample clean-up is achieved. In addition, EME provides high selectivity based on the direction and the magnitude of the electrical field, and based on the chemical composition of the SLM. Because the acceptor solution is aqueous, it can be injected and analyzed directly by high performance liquid chromatography (HPLC), liquid chromatography–mass spectrometry (LC–MS), or capillary electrophoresis (CE). The SLM is a microliter volume (3-15 µL) of an organic solvent immobilized in the pores of a polymeric membrane support, and the amount of organic solvent used per sample is reduced to a minimum compared to the large volumes used in traditional techniques. Thus, EME can be considered as a green chemistry approach to analytical sample preparation [5-7]. 3 EME has been developed into the 96-well format for high-throughput sample preparation, but can also be performed in microfluidic devices. Thus, in 2010, Petersen et al. demonstrated for the first time down-scaling of EME to a chip system, which was termed on-chip EME [8]. This system was operated with a stagnant acceptor solution and a dynamic (i.e., flowing) sample, and new sample was continuously delivered to the membrane. Due to the latter, the system provided high extraction efficiency. Later, a double-flow system was developed, where both acceptor solution and sample were pumped constantly into the chip device [9]. This further improved efficiency, and on-chip EME shows potential as a future sample preparation technique. The composition of the SLM plays an important role in the extraction efficiency, and selection of a proper solvent is critical [10]. Based on current understanding, the ideal organic solvent has to be water-immiscible and with a certain hydrophobicity. Immiscibility with water is important to avoid leakage of the SLM during extraction (samples and acceptor solutions are both aqueous), and hydrophobicity is important in order to impregnate completely the membrane support. In addition, hydrophobicity is also important to avoid excessive flux of matrix ions across the SLM. This reduces current (and conductivity) in the system and ensures system stability. The organic solvent should preferably be of low viscosity to facilitate fast mass transfer of the analytes, and of low volatility to avoid evaporation. Finally, the organic solvent should facilitate strong hydrogen binding and dipole-dipole interactions with the analytes, to facilitate their transfer as ionic species into the SLM. The latter can to some extent be predicted by the Kamlet and Taft solvatochromic parameters ,  and π* [11, 12]. Thus, according to literature, EME of acidic analytes requires solvents as SLM with high hydrogen bonding acidity (α) and dipolarity–polarizability (π*), and low hydrogen bonding basicity (). Therefore long-chain alcohols seem to be the most suitable organic solvents for acidic 4 analytes. For EME of basic analytes, high hydrogen bonding basicity () is required, and 2-nitrophenyl octyl ether (NPOE), either as pure solvent or mixed with additives such as di-(2-ethylhexyl) phosphate (DEHP) or tris-(2-ethylhexyl) phosphate (TEHP) is a highly suitable solvent [13, 14]. In most papers related to EME of acidic substances, 1-heptanol or 1-octanol were found to be the most efficient SLMs [10, 14-19]. To our knowledge, only one single paper is available in the literature describing on-chip EME of acidic drugs [20], where simultaneous extraction of acidic and basic drugs was proposed, achieving the best extraction efficiency with 1-octanol as SLM for the acidic analytes. 1-Octanol is optimal in terms of hydrogen bonding acidity (α), hydrogen bonding basicity (), and dipolarity– polarizability (π*) [21]. However, due to slight water solubility (1.2 g/L) it may not be stable under flow conditions and, in our experience, its use as organic solvent in microfluidic chip devices should be, therefore, avoided. Although a considerable number of research papers have been published on EME of acidic analytes, still no highly stable SLMs have been identified. Therefore, the aim of this work was to develop a stable SLM under flow conditions in a microchip device for EME of acidic drugs. For this purpose, several organic solvents have been tested in two different experimental set-ups for on-chip EME, with primary focus on extraction efficiency, stability, and extraction current. Five non-steroidal antiinflammatory drugs (NSAIDs) were used as target analytes (ibuprofen, naproxen, ketoprofen, salicylic acid, and diclofenac), and optimization of the experimental conditions has been carried out. Moreover, on-chip EME has been also evaluated with human urine samples. 2. Experimental 2.1. Chemicals and sample solutions 5 All chemicals were of analytical-reagent grade. Salicylic acid (SAL), ketoprofen (KTP), naproxen (NAX), diclofenac (DIC) and ibuprofen (IBU) were purchased from Fluka– Sigma–Aldrich. 1-octanol, 1-nonanol, 1-decanol, 1-undecanol, 1-dodecanol, 1,1,2,2fluorooctanol, 2-nitrophenyl octyl ether (NPOE), Aliquat®336, sodium hydroxide, hydrochloric acid, ammonia and methanol were purchased from Fluka (Buchs, Switzerland). Isobutanol, butanol, 1-pentanol and 1-heptanol were purchased from FlukaSigma-Aldrich (Madrid, Spain). Aqueous working solutions of NSAIDs were daily prepared by adequate dilutions from methanolic (IBU, KTP, NAX) and aqueous (SAL, DIC) 400 mg L-1 stock solutions stored at 4°C. 2.2. Capillary electrophoresis Capillary electrophoresis (CE) was performed with an Agilent Technologies HP3D CE instrument (Agilent Technologies, Waldbronn, Germany) equipped with a UV-DAD detector. The wavelengths used were 220 nm for IBU, 225 nm for NAX and 253 for KTP. The running buffer was 30 mM NaAc:ACN (75:25, v/v) pH 5 (adjusted with HAc). Separations were performed at 25 kV in a 50 μm i.d. fused-silica capillary (TSP050375, Polymicro Technologies, Phoenix, AZ) with an effective length of 56 cm. The temperature of the capillary was set to 20ºC. Hydrodynamic injections were performed by applying 45 mbar for 8 seconds. Daily, before use, the capillary was successively rinsed with 0.1 M NaOH, water, and the running buffer for 5 min each. Between runs, it was also conditioned with 0.1 M NaOH (2 min), water (3 min), and the running buffer (5 min). The acceptor solution was spiked with 30 µg mL-1 benzoic acid serving as an internal standard (IS) to correct for possible evaporative loss of the acceptor solution. 2.3. Liquid chromatography 6 The Agilent 1100 series HPLC system consisted of a G1312A binary pump. The injector was an G1313A autosampler allowing an injection volume of 5 µL. Separations were carried out at 25°C using a Purospher® STAR RP-18e LiChroCART® 3 m (75 mm x 4.0 mm i.d.) (VWR, Barcelona, Spain) HPLC column proceeded by a guard column Kromasil 100 Å, C18, 5 m (20 mm x 4.6 mm i.d.) (Scharlab S.L., Barcelona, Spain). The mobile phase consisted of 0.1% formic acid (pH 2.6) (component A) and methanol (component B) at a flow rate of 0.5 mL min-1. An initial elution gradient was programmed from 35% to 25% A for 2 min, then an isocratic mode for 5 min and finally a gradient mode from 25% A to 0% A. The final condition was kept for 1.5 min, followed by 4 min re-equilibration. The wavelengths used for DAD were 235, 255, 230, 280 and 225 nm for SAL, KTP, NAX, DIC and IBU, respectively. Wavelengths were slightly different in CE and HPLC, due to instrumental differences and because the chemical conditions in CE buffer and the HPLC mobile phase were different, which affected the UV absorption maxima. The chromatogram was completed in 9 min and the retention time was 3.2, 4.7, 5.45, 8.05 and 8.5 min for SAL, KTP, NAX, DIC and IBU, respectively. 2.4. Calculation of recovery and enrichment factor Recovery was defined as the fraction of analyte i in the sample solution that was transferred to the acceptor phase, and was calculated, for the individual analytes i, according to the following equation (Eq. 1): Ri (%) = 𝑛𝑎𝑖𝑜𝑢𝑡𝑙𝑒𝑡 𝑛𝑠𝑖𝑖𝑛𝑡𝑙𝑒𝑡 · 100% = 𝑉 𝑎 · 𝐶𝑎𝑖𝑜𝑢𝑡𝑙𝑒𝑡 𝑉 𝑠 · 𝐶𝑠𝑖𝑖𝑛𝑡𝑙𝑒𝑡 · 100% (Eq. 1) where 𝑛𝑎𝑖𝑜𝑢𝑡𝑙𝑒𝑡 is the amount of analyte, i, transferred to the acceptor phase and 𝑛𝑠𝑖𝑖𝑛𝑡𝑙𝑒𝑡 is the amount of analyte, i, originally present in the volume, Vs , of sample processed through the chip. Va is the volume of the acceptor having the concentration of the analyte i, 𝐶𝑎𝑖𝑜𝑢𝑡𝑙𝑒𝑡, at the outlet of the acceptor channel, whereas 𝐶𝑠𝑖𝑖𝑛𝑡𝑙𝑒𝑡 is the initial sample concentration of the analyte i at the sample inlet reservoir. 7 Enrichment factor (EFi) for the analyte i was calculated according to the following equation (Eq. 2): EFi = 𝐶𝑎𝑖𝑜𝑢𝑡𝑙𝑒𝑡 𝐶𝑠𝑖𝑖𝑛𝑡𝑙𝑒𝑡 = Ri · 𝑉 𝑠 𝑉 𝑎 (Eq. 2) The concentration, 𝐶𝑎𝑖𝑜𝑢𝑡𝑙𝑒𝑡, of analyte in the acceptor solution was estimated by CE or HPLC with UV-detection using external calibration. 2.5. On-chip EME Two set-ups have been used in this study. The two on-chip EME devices are shown in Figure 1. One of the chip devices used (Figure 1A) was composed of two poly(methyl methacrylate) (PMMA) (53 × 53 × 2.1 mm3) plates, containing five individual channels for extraction. In both plates, the 6 mm long sample and acceptor channels with a depth of 50 μm and a width of 2.00 mm were milled on a CNC micro-milling machine (Folken M3400 E CNC mini mill, Folken Industries, Glendale, CA). At both ends of the channels, 1.6 mm i.d. holes were drilled through the plate to serve as inlet and outlet for the sample and acceptor solution. Above the sample channels, a porous polypropylene membrane (support) was located (covering the whole PMMA plate), with 25 μm thickness, 55% porosity, and 0.21 x 0.05 μm pores (Celgard 2500 microporous membrane; Celgard, Charlotte, NC). The second PMMA plate was placed above the polypropylene membrane and aligned to match the positions of the sample and acceptor channels, and the whole assembly was fixed by solvent-assisted bonding with ethanol and cured in a 70°C oven following the specific instructions described previously by Petersen et al. [22]. Small platinum wires (0.076 mm i.d.; Sigma-Aldrich) were inserted into the outlet of the sample and the inlet of the acceptor channels, respectively, and connected to an EL302T Triple power supply (DC) (Thurlby-Thandar Instruments LTD, Cambs, UK). 8 The second poly(methyl methacrylate (PMMA) device used (Figure 1B) consisted of two symmetrical plates with six holes of 6 mm for assembling and four holes of 1.4 mm diameter for in/outlets Teflon tubes. The microfluidic device contained two channels (one for acceptor solution and another for sample solution). The geometry of each channel was 23×3×0.12 mm and the channels were separated by the same polypropylene membrane as discussed above. Finally, two electrodes (100 µm i.d.) were located in each channel. The main difference of both chips is related with the assembly mode as described above and the possibility of replacing the flat membrane in the same channel (chip B). For chip A, 5 channels were built simultaneously in the same device to be able to test different SLMs without the need to manufacture a new one, since the assembly requires a longer time. Both devices were connected to syringe pumps to deliver both the sample and acceptor solution into the channels of the chip, and they were used in parallel to study the stability of the SLM in the EME of acidic drugs. Once the extraction was completed, the acceptor solution was collected using a micropipette and transferred to a vial for analysis by capillary electrophoresis in case of set-up A or by liquid chromatography for set-up B. 3. Results and discussion 3.1. Selection of the support liquid membrane (SLM) Selectivity and efficiency of the EME system are highly dependent on the chemical properties of the SLM [10]. As indicated above, long-chain alcohols are the most suitable organic solvents for EME of acidic analytes due to their high hydrogen bonding acidity (). This allows deprotonated acidic drugs, with hydrogen bonding basicity and high dipole moment, to be transferred into the SLM as a result of molecular interactions [14]. The composition of the SLM also determines the electric current level in the system, originating from transfer of analyte ions, back-ground electrolyte ions, and sample matrix ions. Generally, high and unstable currents usually provide higher standard deviation and 15 compatibility with very small volumes of biological fluids, (d) potential for dynamic extraction, and (e) the green nature of the concept. From author point of view, examples of highly specialized applications may be related to micro-physiological systems or organ-on-chip technologies. 1-Octanol (C8 alcohol) provided high extraction efficiency, but suffered from poor long term stability and high current. C9-C11 alcohols were more stable, provided less current, and provided similar recoveries as 1-octanol. C12 alcohol was also tested, but this was less efficient in terms of extraction efficiency. Thus, as pure liquid, 1-undecanol (C11) turned out to be the optimal selection for acidic drug substances. However, mixing 1-undecanol with 2-nitrophenyl octyl ether (NPOE) was found to improve performance further. Thus, the addition of NPOE reduced current, while extraction efficiencies remained unaffected as long as the 1-undecanol/NPOE ratio was not more than 1:3. The 1-undecanol:NPOE SLM was highly stable for long-term operation, and initial evaluation of on-chip EME combined with capillary electrophoresis or high-performance liquid chromatography indicated that acidic drug substances can be extracted and measured reliably from biological fluids by such a concept in the future. Acknowledgements CRH is grateful to University of Sevilla for personal funding through the V Plan Propio de Investigación de la Universidad de Sevilla. MRP is grateful to the program “Juan de la Cierva-Incorporación” (Grant number JCI-2015-26647) and the Project TEC200679367-C2-1-R from the “Dirección General de Investigación y Gestión del PlanNacional de I + D + I”. The authors would also like to acknowledge Frederik André Hansen from the Department of Pharmacy, University of Copenhagen, for fabricating the chips design A used in this work. Conflict of interest 16 The authors declare no conflict of interests. References [1] Arthur, C.L., Pawliszyn, J., Anal. Chem. 1990, 62, 2145-2148. [2] Jeannot, M. A., Cantwell, F. F., Anal. Chem. 1996, 68, 2236–2240. [3] Pedersen-Bjergaard, S., Rasmussen, K.E., Anal. Chem. 1999,71, 2650-2656. [4] Pedersen-Bjergaard, S., Rasmussen, K.E., J. Chromatogr. A. 2006, 1119, 183–190. [5] Pedersen-Bjergaard, S., Huang C., Gjelstad, A., J. Pharm. Anal. 2017, 7, 141–147. [6] Huang C., Chen, Z., Gjelstad, A., Pedersen-Bjergaard, S., Shen, X., TrAC. 2017, 95, 47-56. [7] Ocaña-González, J. A., Fernández-Torres, R., Bello-López, M. A., Ramos-Payán, M., Anal. Chim. Acta. 2016, 905, 8-23. [8] Petersen, N.J., Jensen, H., Hansen, S.H., Foss, S.T., Snakenborg, D., PedersenBjergaard, S., Microfluid. Nanofluid. 2010, 9, 881-888. [9] Petersen, N.J., Pedersen, J.S., Poulsen, N.N., Jensen, H., Skonberg, C., Hansen, S.H., Pedersen-Bjergaard, S., Analyst. 2012, 137, 3321-3327. [10] Huang C., Gjelstad, A., Pedersen-Bjergaard, S., Rev. Anal. Chem. 2016, 35, 169183. [11] Taft,R.W., Kamlet, M.J., J. Am. Chem. Soc. 1976, 98, 2886–2894. [12] De Juan, A., Fonrodona, G., Casassas, E., TrAC. 1997, 16, 52–62. [13] Yamini, Y., Seidi, S., Rezazadeh, M., Anal. Chim. Acta. 2014, 814, 1-22. [14] Huang, C., Gjelstad, A., Seip, K.F., Jensen, H., Pedersen-Bjergaard, S., J. Chromatogr. A. 2015, 1425, 81-87. [15] Balchen, M., Rasmussen, K.E., Gjelstad, A., Pedersen-Bjergaard, S., J. Chromatogr. A. 2007, 1152, 220-225. 17 [16] Ramos-Payán, M., Bello-López, M.A., Fernández-Torres, R., Villar-Navarro, M., Callejón-Mochón, M., Talanta, 2011, 85, 394-399. [17] Koruni, M.H., Tabani, H., Gharari, H., Fakhari, A.R., J. Chromatogr. A. 2014, 1361, 95–99. [18] Román-Hidalgo, C., Martín-Valero, M.J., Fernández-Torres, R., Callejón-Mochón, M., Bello-López, M.A., Talanta, 2016, 162, 32–37. [19] Román-Hidalgo, C., Martín-Valero, M.J., Fernández-Torres, R., Bello-López, M.A., Talanta, 2018, 179, 601–606. [20] Asl, Y. A., Yamini, Y., Seidi, S., Rezazadeh, M., Anal. Chim. Acta, 2016, 937, 6168. [21] Cerón-Carrasco, J. P., Jacquemin, D., Laurence, C., Planchat, A., Reichardt, C., Sraïdi, K., J. Phys. Chem. B , 2014, 118, 4605-4614. [22] Petersen N.J., Jensen H., Pedersen-Bjergaard S., Methods in Molecular Biology, 2015, vol 1274. Humana Press, New York, NY. [23] Seip, K. F., Faizi, M., Vergel, C., Gjelstad, A., Pedersen-Bjergaard, S., Anal. Bioanal. Chem. 2014, 406, 2151–2161. [24] Huang, C., Jensen, H., Seip, K. F., Gjelstad, A., Pedersen-Bjergaard, S., J. Sep. Sci. 2016, 39, 188–197. [25] Huang C., Gjelstad, A., Pedersen-Bjergaard, S., Anal. Chim. Acta, 2015, 853, 328– 334. [26] Restan, M. S., Jensen, H., Shen, X., Huang, C., Martinsen, Ø. G., Kubáň, P., Gjelstad, A., Pedersen-Bjergaard, S., Anal. Chim. Acta. 2017, 984, 116-123. [27] Miller, J.N., Miller, J.C., Statistics and Chemometrics for Analytical Chemistry, 4th ed., Prentice Hall, London, 2000. 18 Table 1 Table 1. Obtained recoveries (%) for each NSAID and achieved current with different supported liquid membranes (SLM) in both chip devices (%RSD). Chip A IBU DIC NAX KTP SAL Current (A) 1-octanol longterm stability (8h, n=15)a 61 (19) * 56 (19) 57 (20) * 30-60 Undecanol:NPOE 1:3 (v/v) longterm stability (8h, n=23)a 51 (9) * 45 (9) 47 (9) * 11-24 Undecanol:NPOE 1:3 (v/v) (n=3)b 105 (1) * 98 (4) 104 (1) * 14-23 Chip B 1-octanol longterm stability (1.5h, n=8) 93 (3) 76 (4) 95 (5) 92 (4) 33 (6) 12-100 Decanol longterm stability (4h, n=34) 98 (1) 94 (1) 100 (2) 99 (2) 18 (2) 10-58 aSample flow rate = acceptor flow rate: 3 L min-1. bSample flow rate: 0.5 L min-1; acceptor flow rate: 3 L min-1. *DIC and SAL were not studied with the set-up A. 19 Table 2 Table 2. Calibration data, Limit of Detection (LOD), Limit of Quantitation (LOQ) and recoveries for the target analytes in the optimal conditions with both chip devices (A and B). Set-up IBU DIC NAX KTP SAL Linear range (g mL-1) A 16-60 * 18-60 18-60 * B 1-10 0.23-10 0.26-10 0.3-10 0.26-10 Linearity (R2) A 0.9963 * 0.9956 0.9957 * B 0.9986 0.9990 0.9993 0.9991 0.9984 LOD (g mL1) A 5 * 5 5 * B 0.3 0.07 0.08 0.1 0.08 LOQ (g mL1) A 16 * 18 18 * B 1.0 0.23 0.26 0.3 0.26 Recovery (%) (%RSD, n=3) Aa 105 (1) * 98 (4) 104 (1) * Bb 98 (1) 94 (1) 100 (2) 99 (2) 18 (2) Current (A) A 14-23 (SLM: undecanol:NPOE 1:3 v/v) B 10-58 (SLM: decanol) Stability A 8h at least (n=23) B 4h at least (n=34) aExtraction conditions: 10V, 5 min, sample flow rate= 0.5 L min-1, acceptor flow rate= 3 L min-1, sample and acceptor composition at pH 10 (100 mM NH4Cl/NH3 buffer). bExtraction conditions: 20V, 7 min, sample flow rate= acceptor flow rate= 1 L min-1, sample pH= 10.5, acceptor pH= 11.5. *DIC and SAL were not studied with the set-up A. 20 Table 3 Table 3. Recoveries of the target NSAIDs using EME+CE/HPLC from non-diluted spiked urine samples (%RSD, n=3). Spiked level (g mL-1) IBU DIC NAX KTP SAL Urine 1a 40 97 (6) * 95 (5) 94 (6) * Urine 2b 0.75 93 (1) 87 (1) 92 (2) 93 (2) 14 (2) Urine 3b 0.75β 92 (1) 89 (1) 100 (1) 94 (2) 17 (2) aSet-up A; SLM: undecanol:NPOE 1:3 (v/v); Sample = spiked female urine in 100 mM NH4Cl/NH3 buffer pH 10; sample flow rate = 0.5 L min-1; Acceptor phase = 100 mM NH4Cl/NH3 buffer pH 10; acceptor flow rate = 3 L min-1; Extraction time = 5 min; Voltage = 10 V. bSet-up B; SLM: decanol; Sample = spiked female (2) and male (3) urine in NaOH at pH 10.5; sample flow rate = 1 L min-1; Acceptor phase = NaOH pH 11.5; acceptor flow rate = 1 L min-1; Extraction time = 7 min; Voltage = 20 V. *DIC and SAL were not studied with the set-up A. βSpiked level for IBU was 1.75 (µg mL-1). Figure 1 Figure 1 Schematic illustration of the two on-chip EME systems used. 21 Figure 2 22 Figure 2. Recovery (Ri) versus sample flow rate. SLM: undecanol:NPOE 1:3 (v/v); Sample = 40 g mL-1 of ibuprofen (IBU), naproxen (NAX) and ketoprofen (KTP) in 100 mM NH4Cl/NH3 buffer pH 10; acceptor flow rate = 3 L min-1 (ammonium chloride/ammonia buffer pH 10); Extraction time = 5 min; Voltage = 10 V. The error bars reflect the overall standard deviation, SD (n = 3). Figure 3 0 20 40 60 80 100 120 140 160 00,5 1 1,5 2 2,5 3 3,5 4 Recovery (%) Sample flow (L min-1) IBU NAX KTP 23 Figure 3. Enrichment factor (EFi) and recovery (R, %) versus acceptor flow rate. SLM: undecanol:NPOE 1:3 (v/v); Sample = 40 g ml-1 of ibuprofen (IBU), naproxen (NAX) and ketoprofen (KTP) in 100 mM NH4Cl/NH3 buffer pH 10; sample flow rate = 3 L min-1; Acceptor phase = 100 mM NH4Cl/NH3 buffer pH 10; Extraction time = 5 min; Voltage = 10 V. The error bars reflect the overall standard deviation, SD (n = 3). 0 10 20 30 40 50 60 70 0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 1,6 1,8 2,0 0,7 1,2 1,7 2,2 2,7 Recovery (%) Enrichment factor (EF) Acceptor flow (µL min-1) EF IBU EF NAX EF KTP %R IBU %R NAX %R KTP 24 Figure 4 Figure 4. Recovery (Ri) versus voltage. SLM: undecanol:NPOE 1:3 (v/v); Sample = 40 g ml-1 of ibuprofen (IBU), naproxen (NAX) and ketoprofen (KTP) in 100 mM NH4Cl/NH3 buffer pH 10; Sample/acceptor flow rates = 3L min-1; Acceptor phase = 100 mM NH4Cl/NH3 buffer pH 10; Extraction time = 5 min; Voltage = 10 V. The error bars reflect the overall standard deviation, SD (n = 3). 0 10 20 30 40 50 60 70 80 90 0 5 10 15 20 25 30 Recovery (%) Voltage (V) IBU NAX KTP