A flow-through potentiometric sensor for an integrated microdialysis system

Sensors and Actuators B 103 (2004) 350–355

A flow-through potentiometric sensor for an integrated microdialysis system D.G. Pijanowska a,b,∗ , A.J. Sprenkels a , H. van der Linden a , W. Olthuis a , P. Bergveld a , A. van den Berg a b

a MESA+ Research Institute, University of Twente, P.O. Box 217, 7500 AE, Enschede, The Netherlands Institute of Biocybernetics and Biomedical Engineering, Polish Academy Sciences, ul. Trojdena 4, 02-109 Warsaw, Poland

Available online 8 June 2004

Abstract In this paper, the performance of a flow-through potentiometric sensor based on a semi-permeable dialysis tubing implemented in silicon, is presented. The sensor is designed as part of a lab-on-a-chip system and has been successfully incorporated into an integrated microdialysis system. Results of a potassium sensor, operating stand-alone, and as part of the integrated system are presented. © 2004 Elsevier B.V. All rights reserved. Keywords: Potentiometric sensor; Flow-through system; Potassium sensor; Microdialysis; Lab-on-a-chip; ␮TAS.

1. Introduction Microdialysis is a sampling technique based on a sizeselective diffusion of the analyte through a semi-permeable membrane. Molecules smaller than the pore size may permeate into the perfusion medium following the concentration gradient, while molecules larger than the membrane cut-off (e.g. proteins), are excluded. Mass transport through the dialysis membrane is driven by the concentration gradient of a particular compound between the perfusate and the extracellular fluid and depends on the flow velocity of the perfusate inside the microdialysis probe. Ungerstedt was one of the pioneers in the field of microdialysis [1–3] and made the technique more versatile through the microdialysis probes he designed and introduced [4]. The technique has been established as one of the major research tools in brain research to measure for instance extracellular neurotransmitter concentrations [5,6]. Dialysates are protein-free samples requiring no pretreatment prior to analysis. They may be analysed by different analytical techniques such as liquid chromatography coupled with UV detection [7–13] or with electrochemical sensors [14–18], capillary electrophoresis, fluorometry [19] or radioactivity tests, among them radioimmunoassays [20,21]. The analytical techniques are rather sophisticated

∗ Corresponding author. E-mail address: [email protected] (D.G. Pijanowska).

0925-4005/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2004.04.108

and expensive. An alternative for these analytical techniques is the use of (bio)-chemical sensors for direct measurement of the concentration of desired species in the dialysate. The possibility of miniaturisation of the components of (bio)-chemical analysis systems by adopting silicon micromachining technology has brought about a rapid development in the design of micro total analytical systems (␮TASs) [22]. Advantages of those systems are minimisation of sample volume and diminishing of the total assay time. Although several miniaturised microdialysis systems based an electrochemical sensors have been designed [18,23–26], they are still equipped with externally connected microdialysis probes. In recent years, the very first integration of both, the microdialysis probe and biosensors has been presented [27,28]. In this paper an approach consisting in utilisation of an electrochemical sensor array in an integrated microdialysisbased system is presented. The system consists of a microdialysis probe, a sensor array and a calibration facility [27]. A generic flow-trough potentiometric microsensor based on a semi-permeable tubing made in Perspex® has been described previously by Böhm et al. [29]. As proposed in this paper the microchannel itself is an integral part of the sensor geometry and is formed by a tubular semi-permeable membrane. The major advantage of this geometry is the fact that commercially available ion-selective cocktails can be applied, yielding sensors for different analytes. The sensor based on this concept is designed in the form of an integral component of our lab-on-a-chip system.

D.G. Pijanowska et al. / Sensors and Actuators B 103 (2004) 350–355

351

The sensor fabrication and the performance of the potassium microsensor operating stand-alone, and as part of an integrated microdialysis system are presented.

2. Experimental 2.1. Materials The dialysis tube of regenerated cellulose with a MW cut-off of 20 kD was adapted from an artificial kidney Filtral® 6, AN69 HF, Hospal France. Epoxy resin (Hysol® , Dexter, USA) was used to fix the dialysis tube. The sensors windows were closed with a piece of Pyrex fixed with UV curable glue (Loctite 350). To form the electrodes, a conducting silver−silver chloride paste (Electrodag® 6088 SS, Acheson Colloiden B.V., The Netherlands) was used. The components of the hydrogel are: acrylamide AAm, N,N-methylene bisacrylamide (crosslinker) and 2,2-dimethoxy-2-phenylacetophenone DMPAP (photoinitiator) obtained from Fluka, and DMSO. 2.2. Sensor fabrication Originally, the sensors were designed to be filled with two liquids: an internal electrolyte and the ionophore cocktail. However, it was found that due to the specific configuration of the sensor, it was difficult to create a well-defined interface between the two liquids. To solve this problem an UV sensitive hydrogel based on polyacrylamide was applied for entrapment of the internal electrolyte. A cross section of the potentiometric flow-trough microsensor is shown in Fig. 1. The estimated swelling capability of the hydrogel, containing 3.6% crosslinker versus AAm, was about 45%. The photographs in Fig. 2 show the pAAm hydrogel in a dry and in a swollen state.

Fig. 2. The pAAm hydrogel in a dry state (a), and in a swollen state (b).

Before use the hydrogel was soaked in the internal electrolyte for at least 24 h. Prior to placing the hydrogel into the sensor cavity, it was cut to the proper square size. Then the hydrogel was placed in the sensor cavity and dried for about 15–30 min. Afterwards the cavity was closed with a piece of Pyrex and fixed with the UV curable adhesive. Next the cavity was filled with the internal electrolyte to allow the hydrogel to swell again for at least 1 h. Finally the excess of internal electrolyte in the surrounding of the dialysis tube was removed and replaced with an ionophore (valinomycin) cocktail (60,031 obtained from Fluka). To form a reference electrode, the sensor cavity located next to the ion-selective one was completely filled with pAAm hydrogel saturated with 3 M KCl. The sensor and the reference electrode have been integrated in a microdialysis system constructed in a generic technology of a glass–silicon bonded sandwich as has been described previously [27]. The system consists of a microdialysis probe (loop or cannula type), a sensor array and a calibration facility (Fig. 3). 2.3. Measurements

Fig. 1. Layout of the flow-trough potentiometric microsensor implemented in silicon.

The stand-alone sensors were operating at a constant continuous flow rate of 2 ml/h driven by a peristaltic pump (P-1, Pharmacia Biotech, Sweden). The concentration of the potassium ions was changed by a standard addition method using 0.1 M and 1 M KCl aqueous solutions. As a background electrolyte 100 mM NaCl was used. Sampling was performed using the integrated microdialysis probe under continuous flow at a flow rate of 2 ␮l/min driven by a microdialysis pump (CMA 102, CMA Microdilysis AB, Sweden) for 60 s. Two different microdialysis probe styles a cannula type and a loop type were tested. The cannula type microdialysis probe is constructed as a double-lumen system with an inner glass tube and outer dialysis tubing.

352

D.G. Pijanowska et al. / Sensors and Actuators B 103 (2004) 350–355

Fig. 3. Layout of an integrated microdialysis based system implemented in silicon.

The loop type probe consists of a single piece of dialysis tubing leaving the system and entering the system in two different channels (Fig. 3). Calibration of the sensors was performed with built-in electrochemically actuated dual dosing pumps [30]. These pumps were actuated with a current pulse of 50 ␮A for 36 s. Oxygen and hydrogen, resulting from the electrolysis of water taking place in the electrochemical cells of the pumps, push out calibrants from the reservoirs into the main channel. These calibration plugs were transported by the carrier solution to the sensor array. The potential of the detecting electrode was measured against the integrated reference electrode, using a homemade high-impedance instrumentation amplifier. Before the actual measurements, the sensors were allowed to stabilise by flushing with a 50 mM KCl solution for about 2 h.

3. Results and discussion The potassium microsensor was tested as a stand-alone sensor, and as part of an integrated microdialysis system. The results representing the dynamic response of the sensor and calibration curves for the continuous flow at a constant flow rate of 2 ml/h are shown in Figs. 4a and 4b, respectively. The sensitivity of the sensors varied from 50 to 55 mV/dec. The sensitivity depends on the time of the sensor conditioning (stabilisation) before measurements. The minimal time for the sensor conditioning is 2 h. The activities of the cations in the aqueous solution were calculated according to the Debye–Huckel approximation [31]. The potentiometric selectivity coefficient for the sensor for potassium over sodium (Kpot = 161.5) was estimated according to the fixed interference method (FIM) [32] by increasing the concentration of the activity of the primary ion in the solution in steps of about 0.7 decade.

The sensor has been successfully integrated in a microdialysis system [33]. A dynamic response and a calibration curve for the sensor operating in the microdialysis system for 60 s sampling through a microdialysis probe at a flow rate 2 ␮l/min are shown in Figs. 5 and 6(upper curve), respectively. The recovery of a component from the sample depends on the dialysis membrane cut-off and length, the flow rate of the perfusate, and the diffusion coefficient of the compound. Sampling was performed at a very low flow rate, low enough to obtain homogeneity and 100% recovery rate (RR) for potassium in the plug formed via the microdialysis probe. This results from mass transport conditions, namely a diffusion time for potassium ions (diffusion coefficient for potassium ions is 1.9565 × 10−5 cm2 /s) for the diffusion path through the dialysis membrane (lm = 50 ␮m) and the perfusate layer (lp ) in the microdialysis probe. In the case of the cannula type microdialysis probe lp = 25 ␮m so that the diffusion time for potassium ions yields ca. 2.9 s, while for the loop type microdialysis probe lp = 100 ␮m, resulting in a diffusion time 11.5 s. Concluding, for both the cannula and loop type microdialysis probes, the diffusion time is shorter than the residence time of the plug in the probes, which is about 8.8 s for the cannula and 18.8 s for the loop type probe for effective probe lengths of 20 mm. Since under experimental conditions, the samples formed in the microdialysis probes were homogenous, the FIA theory [34] is applicable for an adequate description of the experiment. For the longer sampling time, the volume of the homogenous plug formed via the microdialysis probe is larger so that the travel time of the plug through the sensor area was longer, resulting in the higher output signal peak. Calibration of the sensors was performed with two built-in electrochemically-actuated dosing systems. A typical response of the system to calibrants (5 and 50 mM KCl) and to samples of different concentration of the potassium ions

D.G. Pijanowska et al. / Sensors and Actuators B 103 (2004) 350–355 0,06

31 mM 160 mM

0,04

Potential [V]

353

83 mM

0,02 0 -0,02 0.1 mM

-0,04

0.6 mM

-0,06

31 mM

-0,08

9.5 mM

-0,1

0

400

800

1200 1600 2000 2400 2800 3200

(a)

3600 4000

time [s] 0.08

Potential [V]

0.04 0 -0.04 -0.08 -0.12 -6

(b)

-5

-4

-3

log aK+

-2

-1

0

Fig. 4. Time response (a) and calibration curve (b) for the potassium sensor for continuous flow. Background electrolyte was 100 mM NaCl.

sampled via that integrated microdialysis probe (sampling time 60 s) is shown in Fig. 5. In contrast to the calibration peaks (Cal 1 and Cal 2) obtained for very long calibration plugs, the sample peaks were obtained for short plugs. For very large plug volumes, there is a part of the plug, which has not undergone dilution and therefore, over a certain time span the sample passing the sensor yields the maximum peak height. The time span decreases with the shorter plugs so that the output signal peaks may become smaller.

This does not affect the calculated sensor sensitivity for different fixed volumes of the plugs. However, this may result in a shift of the low detection limit towards higher values as can be seen in Fig. 6. Fig. 7 shows calibration curves corresponding to the microsensor response to calibration plugs generated by a build-in microdosing system (a two point calibration) and for 60 s sampling via a microdialysis probe (Fig. 5). The discrepancy for the slope calculated for the two-point calibration (Fig. 7, lower curve) and that calculated for

0,1 500 mM

0,08 100 mM

Potential [V]

0,06 0,04

0.1 mM

Cal 2

5 mM

50 mM

0,02 0 0.01 mM

-0,02 Cal 1

1 mM

10 mM

-0,04 0

400

800

1200

1600

2000

2400

2800

3200

3600

4000

time [s]

Fig. 5. Response of the potassium potentiometric sensor to: calibrants 5 and 50 mM KCl (Cal 1 and Cal 2, respectively) and samples of concentration from 0.01 to 500 mM KCl. Background electrolyte was 100 mM NaCl. Sampling time 60 s.

354

D.G. Pijanowska et al. / Sensors and Actuators B 103 (2004) 350–355

The sensors have been successfully integrated into a chip resulting in lab-on-a-chip type system consisting of several functional blocks: microdialysis probe, calibration facility and sensor array. The estimated inaccuracy of the calibration method of the potassium sensor operating in the integrated microdialysisbased system was 1.2%.

0.12

Potential [V]

0.08 0.04 0 -0.04 -0.08 -0.12 -6

-5

-4

-3

log aK+

-2

-1

Acknowledgements

0

Fig. 6. Calibration curve for the potassium sensor for continuous flow – long plugs – (lower curve) and also the calibration curve for the sensor as integrated in the microdialysis system – short plugs (60 s sampling) – (upper curve). Background electrolyte was 100 mM NaCl. 0.1

This project is financially supported by the Dutch foundation for Fundamental Research on Matter (FOM) and the Royal Dutch Academy of Science (KNAW).

References

0.08

Potential [V]

0.06 0.04 0.02

60 s sampling

50 mM

0 -0.02

5 mM

-0.04 -0.06

calibration

-0.08

-6

-5

-4

-3

-2

-1

0

log aK+ Fig. 7. Calibration curve for the potassium sensor for a two-point calibration – long plugs – (lower curve) and also the calibration curve for the sensor as integrated in the microdialysis system – short plugs (60 s sampling) – (upper curve). Background electrolyte was 100 mM NaCl.

the calibration curve obtained with the 60 s sampling (upper curve) was 1.2%. Owing to mass transport parameters under the experimental conditions (RR = 100% and plug homogeneity), this value may be used for an estimation of the inaccuracy of the calibration method based on the use of the dual pump electrochemically actuated microdosing system applied in the integrated microdialysis system.

4. Conclusions The potentiometric sensor fabrication and the performance of the potassium microsensor operating stand-alone, and as part of an integrated microdialysis system were presented. A flow-trough potassium sensitive potentiometric microsensor based on semi-permeable tubing has been implemented in silicon and shows a near Nernstian response and a good selectivity over sodium.

[1] U. Ungerstedt, A. Hallstrom, In vivo microdialysis – a new approach to the analysis of neurotransmitters in the brain, Life Sci. 41 (1981) 861–864. [2] U. Ungerstedt, in: C.A.Marsden (Ed.), Measurement of Neurotransmitter Release by Intracranial Dialysis, Measurement of Neurotransmitter Release In Vivo, Wiley, 1984. [3] U. Ungerstedt, Microdialysis – a new bioanalytical sampling technique, Curr. Sep. 7 (1986) 2. [4] U. Ungerstedt, Dialysis probe, US Patent No. 4,694,832 (1987). [5] P.F. Morrison, P.M. Bungay, J.K. Hsiao, I.N. Dykstra, R.L. Dedrick, Quantitative microdialysis, in: T.E. Robinson, J.B. Justice, Jr. (Eds.), Technique in the Behavioural and Neural Sciences in Microdialysis in the Neurosciences, vol. 7, Elsevier, Amsterdam, 1991, pp. 47–80. [6] J. Kehr, A survay on quantitative microdialysis: theoretical models and practical implications, J. Neurosc. Methods 48 (1993) 251–261. [7] S. Menacherry, W. Hubert, J.B. Justice Jr, In vivo calibartion of microdialysis probes for exogenous compounds, Anal. Chem. 64 (1992) 577–583. [8] K. van Belle, T. Dzeka, S. Sarre, G. Ebinger, Y. Michotte, In vitro and in vivo microdialysis calibration for the measurement of carbamazepine and its metabolites in rat brain tissue using the internal reference technique, J. Neurosc. Methods 49 (1993) 167–173. [9] Y. Hashimoto, T. Murakami, C. Kumasa, Y. Higashi, N. Yata, M. Takano, In vivo calibration of microdialysis probe by use of endogenous glucose as an internal recovery marker: measurement of skin distribution of tranilast in rats, J. Pharm. Pharmacol. 50 (1998) 621–626. [10] N. Borg, L. Stahle, Recovery as a function of the osmolality of the perfusion medium in microdialysis experiments, Anal. Chim. Acta 379 (1999) 319–325. [11] Y. Song, C.E. Lunte, Calibration methods for microdialysis sampling in vivo: muscle and adipose tissue, Anal. Chim. Acta 400 (1999) 143–152. [12] K.E. Hepper, M.I. Davies, The microdialysis shunt probe: the influence of bile salts on recovery of caffeine and phenylphthalein gluccuronide and use of in vitro recovery data for calibration in vivo, Curr. Sep. 18 (1999) 55–59. [13] R. Chen, J.A. Stenken, An in vitro hydroxyl radical generation assay for microdialysis sampling calibration, Anal. Biochem. 306 (2002) 40–49. [14] D. Moscone, M. Pasini, M. Mascini, Subcutaneous microdialysis probe coupled with glucose biosensorfor in vivo continuous monitoring, Talanta 39 (8) (1992) 1039–1044. [15] E.F. Pfeiffer, C. Meyerhoff, F. Bischof, F.C. Keck, W. Kerner, On line continuous monitoring of subcutaneous tissue glucose is feasible by

D.G. Pijanowska et al. / Sensors and Actuators B 103 (2004) 350–355

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

combining portable glucosensor with microdialysis, Horm. Metab. Res. 25 (1993) 121–124. T. Laurell, Microdialysis implemented in the design of a system for continuous glucose monitoring, Sens. Actuators B Chem. 13–14 (1993) 323–326. G. Yadid, K. Pacak, I.J. Kopin, D.S. Goldstain, Modified microdialysis probe for sampling extracellular fluid and administering drugs in vivo, Am. J. Physiol. (Regul. Integr. Comp. Physiol.) 34 (1993) R1201–R1211. J.A. Stenken, E.M. Topp, M.Z. Southard, C.E. Lunte, Examination of microdialysis sampling in a well-characterized hydrodynamic system, Anal. Chem. 65 (1993) 2324–2328. J. de Boer, J. Korf, H. Plijter-Groendijk, In vivo monitoring of lactate and glucose with microdialysis and enzyme reactors in intensive care medicine, Int. J. Artif. Organs 17 (3) (1994) 163–170. L. Strindberg, P. Lönnroth, Validation of an endogenous reference technique for calibration for the calibration of microdialysis catheters, Scand. J. Clin. Lab. Invest. 60 (2000) 205–212. M. Rizell, P. Naredi, P. Lindner, K. Hellstrand, P.-A. Jansson, Validation of the internal reference technique for microdialysis measurements of interstitial histamine in the rat, Life Sci. 65 (1999) 1765–1771. A. Manz, N. Graber, H.M. Widmer, Miniaturized total chemical analysis systems: a novel concept for chemical sensing, Sens. Actuators B Chem. 1 (1990) 244–248. M.G. Boutelle, L.K. Fellows, C. Cook, Enzyme packed bed system for the on-line measurements of glucose, glutamate, and lactate in brain microdialysate, Anal. Chem. 64 (1992) 1790–1794. R. Steinkuhl, C. Sundermeier, H. Hinkers, C. Dumschat, K. Cammann, M. Knoll, Microdialysis system for continuous glucose monitoring, Sens. Actuators B Chem. 33 (1996) 19–24. E. Dempsey, D. Diamond, M.R. Smyth, G. Urban, G. Jobst, I. Moser, E. Verpoorte, A. Manz, H.M. Widmer, K. Rabenstein, R. Freaney,

[26]

[27]

[28]

[29]

[30]

[31]

[32] [33]

[34]

355

Design and development of a miniaturised total chemical analysis system for on-line lactate and glucose monitoring in biological samples, Anal. Chim. Acta 346 (1997) 341–349. J. Perdomo, H. Hinkers, C. Sundermeier, W. Seifert, O.M. Morell, M. Knoll, Miniaturized real-time monitoring system for l-lactate and glucose using microfabricated multi-enzyme sensors, Biosens. Bioelectron. 15 (2000) 515–522. A. Sprenkels, W. Olthuis, P. Bergveld, On the integration of a microdialysis-based ␮TAS with calibration facility on a silicon–glass sandwich, in: Proceedings of the First Annual International IEEE-EMBS Special Topic Conference on Microtechnologies in Medicine and Biology, Lyon, France, 12–14 October 2000, pp. 326– 330. P.S. Petrou, I. Moser, G. Jobst, BioMEMS device with integrated microdialysis probe and biosensor array, Biosens. Bioelectron. 17 (2002) 859–865. S. Böhm, W. Olthuis, P. Bergveld, A flow-through micro sensor array based on semi-permeable dialysis tubing, in: Proceedings of the SPIE Conference on Industrial and Environmental Sensing, Boston, USA, 19–20 September 1999, pp. 24–30. A.J. Sprenkels, D.G. Pijanowska, J.G. Bomer, S. Herber, K.M. van Delft, W. Olthuis, P. Bergveld, A closed-loop calibration system for a microdialysis-based lab-on-a-chip, in: Proceedings of the MicroTAS, Nara, Japan, 4–7 November 2002. Y. Umezawa, K. Umezawa, H. Sato, Selectivity coefficients for pot ion-selective electrodes: recommended methods for reporting KAB values, Pure Appl. Chem. 67 (1995) 507–518. K. Umezawa, Y. Umezawa, CRC Handbook of Ion-Selective Electrodes: Selectivity Coefficients, CRC Press, 1990, pp. 3–9. D.G. Pijanowska, A.J. Sprenkels, J. Bomer, W. Olthuis, P. Bergveld, An integrated microdialysis-based system, in: Proceedings of the microTAS, Nara, Japan, 4–7 November 2002, pp. 34–36. J. Ruzicka, E.H. Hansen, Flow Injection Analysis, Wiley, NY, 1981.