Analytica Chimica Acta

Transcription

1 Analytica Chimica Acta 687 (2011) 7 11 Contents lists available at ScienceDirect Analytica Chimica Acta journal homepage: Integrated electrochemical transistor as a fast recoverable gas sensor Ulrich Lange a, Vladimir M. Mirsky b, a University of Regensburg, Institute of Analytical Chemistry, Chemo- and Biosensors, D Regensburg, Germany b Lausitz University of Applied Sciences, BCV-Nanobiotechnology, Grossenhainer str. 57, Senftenberg, Germany article info abstract Article history: Received 13 October 2010 Received in revised form 1 December 2010 Accepted 4 December 2010 Available online 11 December 2010 Keywords: Gas sensor Conducting polymer Electrochemical transistor Six-point resistance Measurement Sensor recovery A new design of conductometric chemical sensors based on conducting polymers as chemosensitive elements was suggested. The sensor includes six electrodes. Four inner electrodes coated by chemosensitive polymer are used for simultaneous two- and four-point resistance measurements thus providing information on the bulk polymer resistance and on the resistance of the polymer/electrode contacts. Two outer electrodes wired to inner electrodes by polymeric electrolyte are used for electrical control of redox state of the chemosensitive polymer. The outer electrodes are connected to potentiostat as reference and counter electrodes. It allows us to control redox state of the inner (working) electrodes. This new measurement configuration, resembling chemosensitive electrochemical transistors, provides an internal test of the sensor integrity and an electrically driven sensor regeneration. It was tested as a sensor for the detection of nitrogen dioxide. Polythiophene or polyaniline was used as receptors. Cyclic voltammograms of these polymers on the sensor surface measured in air atmosphere were very similar to that measured in aqueous electrolyte. A control of conductivity of these chemosensitive polymers by electrical potential applied vs. incorporated reference electrode was demonstrated. This effect was used for the regeneration of the chemosensitive material after exposure to nitrogen dioxide: in comparison to usual chemiresistors displaying an irreversible behavior in such test even in the time scale of hours, a completely reversible sensor regeneration within few minutes was observed Elsevier B.V. All rights reserved. 1. Introduction A usual problem in the development of chemosensitive materials for chemical sensors is that an optimization of the receptor affinity required to get a higher sensitivity and selectivity, is corresponded by a decrease in the efficiency of the sensor recovery. It is caused by the fact that the binding constant is the ratio of association and dissociation kinetic constants; therefore this ratio can be increased by the decrease of dissociation kinetic constant. A very slow dissociation was observed for many types of chemical sensors including gas sensors based on conducting polymers. These materials possess many important features such as a high sensitivity, an ability to work at room temperature, a compatibility with organic flexible electronics and a low price [1]. However the sensor recovery after exposure to analytes is very slow [2], leading to a decrease of the sensor signal in each next operation cycle. Perhaps, namely this problem is the main reason why conducting polymers are not used so far in commercial chemical sensors and sensor arrays. Many attempts to speed up this process by heating [3] or UV-light exposure [4] were performed. However these approaches are poor-compatible with flexible organic electronics, require exact Corresponding author. Tel.: ; fax: address: vmirsky@hs-lausitz.de (V.M. Mirsky). balancing of temperature extension of contacting materials and can lead to a destruction of the polymers or the polymer/metal contacts. In this paper we present a new concept based on electrochemical control of the redox state of chemosensitive materials through integrated auxiliary and reference electrodes and describe an example of its application for the detection of nitrogen dioxide. The sensor configuration suggested in the current work is an advanced electrochemical transistor (ECT). This term was introduced by White et al. [5 7] for a configuration including two electrodes (source and drain electrodes) and an additional (gate) electrode(s). Later this approach was developed in different groups [8 16]. In contrast to ChemFETs, in which the gate electrode is separated from the channel by an insulating layer, and the channel (source drain) resistance is controlled by chemical processes on the gate through electrostatic effects, ECTs have a different action mechanism. The source drain resistance in ECTs is controlled directly by the electrochemical oxidation/reduction of a conducting polymer placed between these electrodes and acting as the channel. The configuration presented in the current communication has an additional feature in comparison with known ECT: it is based on four-point configuration to measure resistance and two additional electrodes (gate electrodes in terminology of FETs) which are used to control redox state of the chemosensitive polymer are integrated into the same chip. Therefore, the new sensor combines the advantages of separate measurements of bulk and contact resistances /$ see front matter 2010 Elsevier B.V. All rights reserved. doi: /j.aca

2 8 U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) 7 11 with optional features to perform analytical measurements and electrochemical control of sensor layer. 2. Experimental 2.1. Materials Thiophene, boron trifluoride diethyl etherate (BFEE), aniline, poly(2-acrylamido-2-methyl-1-propane-sulfonic acid) (PAMPSA), ethylene glycol, LiClO 4, NaCl, AgNO 3 and acetonitrile were obtained from Sigma Aldrich, EDTA and sorbitol were from Serva. 300 ppm NO 2 in synthetic air was obtained from Linde. All solutions were prepared with deionized water additionally purified by Millipore Milli-Q system. The sensor chip was fabricated by sputtering of 150 nm thick gold structures on a Ti/W adhesion layer on a glass wafer. The measurement electrodes used for analysis of bulk and contact resistances have a shape of four parallel strips of 35 m width separated by 8 m gaps. The counter and reference electrodes were formed as 300 m strips around working electrode and separated by a 150 m gap, the gap between reference and measurement electrodes was also 150 m Instrumentation The technology of simultaneous 2- and 4-point resistance measurements (S 24 ) was described earlier [17,18]. This technology provides an internal integrity test. Even a slight desorption of chemosensitive material from the measurement electrodes (which can be induced by chemical treatment or by different physical factors) leads to a strong increase of the ratio of resistances measured by 2- and 4-point techniques [19]. Two additional electrodes (reference and auxiliary electrodes) connected to potentiostat Radiometer-PGP201 were used to control the redox state of chemosensitive polymers. A home-made gas mixing device with computer control was used for the gas sensor measurements Procedures Prior to modification, the electrodes were cleaned with acetone, ethanol, water, by treatment with piranha solution (freshly prepared mixture of 30% H 2 O 2 :conc. H 2 SO 4, 1:3 (v/v). Caution: this solution reacts violently with most organic materials and must be handled with extreme care) and rinsed by water. The silver electrode was formed by the galvanostatic deposition of silver at 10 A on the corresponding gold electrode from a solution containing 10 mm AgNO 3, 20 mm EDTA, 120 mm NH 3 and 80 mm NaOH; it was then calibrated as a pseudo-reference electrode by ferrocene in acetonitrile, a value of V vs. NHE was obtained. Silver/silver chloride reference electrode was formed by a short dipping of the silver modified electrode into solution of 1 mg FeCl 3 in 1 ml ethanol. Polythiophene and polyaniline were selected as well characterized chemosensitive conducting polymers to develop and test the new 6-point measurement concept. Therefore thin films of these polymers were electropolymerized over the four resistance measurement electrodes and on the auxiliary electrode. Polythiophene was electropolymerized from a 0.2 M solution of thiophene in 90% boron trifluoride diethyl etherate (BFEE)/10% acetonitrile at +1.3 V vs. Ag/Ag + (10 mm). The amounts of the polymer deposited on the measurements and auxiliary electrodes correspond to the polymerization charges of 2.5 mc and 7.5 mc. The BFEE based electrolyte decreases oxidation potential of thiophene [20], allowing to make electropolymerization at low potential and to avoid overoxidative damage. Then the chip was covered by dropcoating a thin layer of a poly(2-acrylamido-2-methyl-1-propane-sulfonic acid) (PAMPSA) electrolyte on its surface. The electrolyte was prepared by mixing 500 mg of a 15% solution of PAMPSA in water, 75 mg ethylene glycol, 25 mg sorbitol, 16 mg LiClO 4 and 0.5 ml water. Ethylene glycol and sorbitol were added to the polymer solution in water to obtain a polymer gel electrolyte with a sufficient high conductivity at normal humidity [21]. Before measurements the electrolyte layer was dried in air for few hours. Alternatively, a layer-by-layer deposition of six double layers of 0.1% (w/w) polystyrenesulfonate and 0.1% (w/w) polyallylaminhydrochloride solutions containing 0.1 M NaCl was performed in some experiments to obtain a polymeric electrolyte. Polyaniline (PANI) based electrochemical transistors were formed by the electropolymerization of aniline from a 5 M solution of aniline in 1 M HCl at V vs. Ag/AgCl and subsequent coating the chip with the PAMPSA electrolyte containing 0.1 M NaCl instead of LiClO 4. The polymerization charges were 1 mc for the measurement electrodes and 3 mc for the counter electrode. Characterisation of sensor properties without analyte was performed in air over a saturated calcium chloride providing a constant humidity of about 30% [22]. The sensor response was measured in a flow through cell of 10 ml volume at flow rate of 120 ml min ppm nitrogen dioxide in nitrogen was diluted with synthetic air and humidified by passing through the headspace of a water containing flask. During analyte measurements the electrodes were disconnected from potentiostat. Sensor regeneration was performed by flushing the sensor with synthetic air for 5 min and subsequent application of 50 mv potential vs. the silver pseudo reference electrode for 30 s. To achieve a very fast regeneration, the reduction was carried out until the current decreased about 50% lower than the initial baseline drain current; further disconnecting of the measurement electrodes led to an increase of the drain current till its initial value. The measurements were performed at room temperature (22 C). 3. Results and discussion The suggested approach was tested with different chemosensitive redox-active polymers including polyaniline and polythiophene, but the main part of the work was performed with a well known [1,2] combination of sensor material and analyte: polythiophene and nitrogen dioxide. Polythiophene has a high oxidation potential and is therefore stable in its reduced non-conducting form. An exposure to gaseous nitrogen dioxide in ppm concentration scale leads to oxidation of this polymer corresponded by a strong increase in its conductance. The design and wiring of the solid state electrochemical chemotransistor with integrated control electrodes (gate electrodes) as well as a photo of this electrode before deposition of the conducting polymer are presented in Fig. 1. The four inner electrodes (further referred as measurement electrodes) of the sensor chip were used for resistance measurements by S24 technique. The potential difference applied between measurement electrodes was 10 mv or less, therefore variations of the redox state of the polymer between the four measurement electrodes are negligible. The two outer electrodes (reference and auxiliary electrodes) were connected through a potentiostat with one of the four measurement electrodes for an electrochemical control of the chemosensitive polymer. The combination of the four measurement electrodes and two external (reference and auxiliary) electrodes leads to a new, six-electrode measurement configuration in which each electrode has its own function. The deposition of chemosensitive polymers onto measurement electrodes was performed by electropolymerization. The (quasi)reference electrode was formed electrochemically by the deposition of a silver layer (for polythiophene based sensors with chloride-free gel electrolytes used for outer coating) or silver and silver chloride layers (for PANI based sensors with chloride-

3 U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) A I / µa E / V vs. Ag 3 B 2 I / µa E / V vs. Ag / AgCl Fig. 2. Cyclic voltammogram of polythiophene (A) and polyaniline (B) measured in air. The electrolyte: poly(2-acrylamido-2-methyl-1-propane-sulfonic acid) (PAMPSA) containing 0.1 M LiClO 4 (A) or 0.1 M NaCl (B). (Quasi)reference electrode: Ag (A) or Ag/AgCl (B). Scan rate: 10 mv s 1. The measurements were performed at 30% humidity. Fig. 1. Design and wiring of the chemical sensor with electrical affinity control and a photo of the sensor chip before the deposition of electroactive polymer on the measurement and auxiliary electrodes. The four inner electrodes connected with potentiostat as working electrodes, serve as measurement electrodes for simultaneous 2- and 4-point measurements. The next outer electrode is the reference electrode and the outer surrounding electrode is the auxiliary electrode. The four gold strips of the measurement electrodes have a width of 35 m and are separated by an 8 m gap. containing gel electrolytes) on the corresponding gold strip. The auxiliary electrodes in the current work were electrochemically coated by the same conducting polymer as the measurement electrodes, however other types of redox-active compounds can be also used for this coating. A function of this coating is to provide a reservoir of electric charges for the oxidation or reduction of the chemosensitive redox-active polymer on the measurement electrodes. This coating minimizes a possible polarization of the interface between auxiliary electrode and solid electrolyte. The total amount of conducting polymer on the auxiliary electrode was few times higher than that on the measurement electrodes, therefore a strong oxidation of polymer on the measurement electrodes leads only to a small reduction of the polymer on the auxiliary electrode. To provide an electrical connection between measurement and control electrodes, all electrodes were coated by an inert polymeric electrolyte. Cyclic voltammograms of polythiophene (Fig. 2A) and polyaniline (Fig. 2B) films demonstrate successful electrochemical control of the polymer film in gas atmosphere. Due to reduced electroactivity of polythiophene in aqueous media, the oxidation/reduction peaks are not clearly visible in the voltammogram, however the measurements of the polymer resistance display a potential dependent switch for about four orders of magnitude. These results confirmed that the redox state of the polymer can be controlled electrically by electrodes connected through gel electrolyte. The voltammograms measured in air are similar to that measured in

4 10 U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) I D / µa E / V vs. Ag / AgCl Fig. 4. Influence of the gate potential on the drain current of a polyaniline based sensor measured at constant drain voltage of 10 mv. deposited onto measurement electrodes by an electrical potential applied between the measurement and control electrodes. The difference in sensing performance between a usual polythiophene chemiresistor (the same setup as the electrochemical transistor, but without gel electrolyte) and the new sensor is illus- Fig. 3. Current voltage characteristics of the polythiophene basedsensor at different redox states controlled by potential difference between measurement and reference electrodes (gate potential) V G (A) and an influence of the gate potential on the drain current measured at constant drain voltage of 10 mv (B). liquid aqueous electrolytes. Some shift in the oxidation/reduction peaks of polyaniline in comparison with the literature [23] and our data obtained in liquid electrolytes can be referred to partial drying of the gel and to an influence of the gel polymer on the activity of chloride ions. The current voltage characteristics of the sensor were measured at three different potentials relative to the reference electrode, corresponding to different redox states of polythiophene by changing the drain potential between the four measurement electrodes from 1 to 50 mv and measuring the resulting current (Fig. 3A). The gate potential +0.7 V corresponds to the oxidized state, the potential 0.1 V corresponds to the reduced state, while the potential +0.4 V corresponds to some intermediate state. In all cases the characteristic was linear till at least 50 mv (Fig. 3), therefore the drain potential is not influencing the polymer resistance in this potential interval. The dependence of the current through measurement electrodes (this is an analogue of the drain current in FET) on gate voltage (i.e. a potential difference between measurement electrodes and integrated reference electrode) at a drain voltage of 10 mv is presented in Fig. 3B. Upon switching the polymer from its reduced state ( 0.1 V) to its oxidized state (+ V) the conductance changed for over four orders of magnitude (Fig. 3B). A similar behavior and the same slope of the voltage dependence of the polymer conductance were observed in aqueous solution of 0.1 M LiClO 4. The influence of the gate potential on the drain current of a PANI based electrochemical transistor at a drain voltage of 10 mv is shown in Fig. 4. The results confirm the direct electrochemical control of the redox states of electroactive polymers Fig. 5. Current response of chemiresistor (A) as well as of a new sensor with electrical affinity control (B) to different concentrations of nitrogen dioxide in synthetic air. The dashed line represents the concentration pulses of nitrogen dioxide. The pulse duration was 5 min. Regeneration time was 10 min in the case of the electrochemical transistor and 65 min in the case of the chemiresistor. To emphasize different kinetics for spontaneous and electrically driven sensor recovery, the time scales in the panels (A and B) are different.

5 U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) trated in Fig. 5. Polythiophene film is oxidized upon exposure to nitrogen dioxide; this leads to an increase in the film conductivity and a corresponding increase in the drain current. Both sensors based on chemiresistor and 6-electrode configuration exhibited a high sensitivity towards nitrogen dioxide. However the recovery of the chemiresistor performed by perfusion with synthetic air was not complete even after 1 h (Fig. 5A). The new sensor exploiting electrical control of the redox state of chemosensitive polymers was electrically regenerated within a few minutes (Fig. 5B). The concentration dependence of the sensor signal being measured till 100 ppm is almost linear. 4. Conclusion We demonstrated an application of a new 6-point measurement configuration for gas sensors. Electrochemical control of the redox state of the polymer allows a fast regeneration of the sensors thus providing a way to overcome one of the main drawbacks of gas sensors based on conducting polymers. The suggested type of integrated electrochemical transistor can be applied also to other types of chemical sensors and biosensors based on conducting redox active materials thus providing a great enhancement of sensors performance in sense of reversibility, reproducibility and signal stability. A problem of selectivity known for such sensors [1,2] could be overcome by combining different sensors in an array. However, the selectivity depends on the type of particular chemosensitive material and not on the measurement approach, which is out of the scope of this communication. Acknowledgment References [1] U. Lange, N.V. Roznyatovskaya, V.M. Mirsky, Anal. Chim. Acta 614 (2008) [2] H. Bai, G. Shi, Sensors 7 (2007) [3] J. Kong, N.R. Franklin, C. Zhou, M.G. Chapline, S. Peng, K. Cho, H. Dai, Science 287 (2000) [4] V. Dua, S.P. Surwade, S. Ammu, X. Zhang, S. Jain, S.K. Manohar, Macromolecules 42 (2009) [5] H.S. White, G.P. Kittlesen, M.S. Wrighton, J. Am. Chem. Soc. 106 (1984) [6] J.W. Thackeray, M.S. Wrighton, J. Phys. Chem. 90 (1986) [7] S. Chao, M.S. Wrighton, J. Am. Chem. Soc. 109 (1987) [8] P. Bartlett, Y. Astier, Chem. Commun. (2000) [9] P. Bartlett, P. Birkin, J. Wang, F. Palmisano, G. De Benedetto, Anal. Chem. 70 (1998) [10] Y. Astier, P. Bartlett, Bioelectrochemistry 64 (2004) [11] J. Mabeck, G. Malliaras, Anal. Bioanal. Chem. 384 (2006) [12] M. Nikolou, G.G. Malliaras, Chem. Rec. 8 (2008) [13] D.J. Macaya, M. Nikolou, S. Takamatsu, J.T. Mabeck, R.M. Owens, G.G. Malliaras, Sens. Actuators B 123 (2007) [14] P. Svensson, D. Nilsson, R. Forchheimer, M. Berggren, Appl. Phys. Lett. 93 (2008) [15] Z. Mousavi, A. Ekholm, J. Bobacka, A. Ivaska, Electroanalysis 21 (2009) [16] D. Nilsson, T. Kugler, P.O. Svensson, M. Berggren, Sens. Actuators B 86 (2002) [17] Q. Hao, V. Kulikov, V.M. Mirsky, Sens. Actuators B 94 (2003) [18] U. Lange, V.M. Mirsky, J. Electroanal. Chem. 622 (2008) [19] V. Kulikov, V.M. Mirsky, T.L. Delaney, D. Donoval, A.W. Koch, O.S. Wolfbeis, Meas. Sci. Technol. 16 (2005) [20] G.Q. Shi, C. Li, Y.Q. Liang, Adv. Mater. 11 (1999) [21] M. Hamedi, R. Forchheimer, O. Inganas, Nat. Mater. 6 (2007) [22] P.W. Winston, D.H. Bates, Ecology 41 (1960) [23] W. Huang, B. Humphrey, A. MacDiarmid, J. Chem. Soc. Faraday Trans. 82 (1986) U.L. was supported by DFG Graduate College GRK.