Chemiresistors based on conducting polymers: A review on measurement techniques

Analytica Chimica Acta 687 (2011) 105–113

Contents lists available at ScienceDirect

Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca

Review

Chemiresistors based on conducting polymers: A review on measurement techniques Ulrich Lange a , Vladimir M. Mirsky b,∗ a b

University of Regensburg, Institute of Analytical Chemistry, Chemo- and Biosensors, 93040 Regensburg, Germany Lausitz University of Applied Sciences, BCV-Nanobiotechnology, 01968 Senftenberg, Germany

a r t i c l e

i n f o

Article history: Received 19 September 2010 Received in revised form 11 November 2010 Accepted 12 November 2010 Available online 19 November 2010 Keywords: Chemiresistor Chemotransistor Conducting polymer Chemosensor Contact resistance s24-Technique

a b s t r a c t This review covers the development of measurement configurations for chemiresistors based on conducting polymers. The simplest chemiresistors are based on application of a two-electrode technique. Artifacts caused by contact resistance can be overcome by application of a four-electrode technique. Simultaneous application of the two- and four-electrode measurement configurations provides an internal control of sensor integrity. An incorporation of two additional electrodes controlling the redox state of chemosensitive polymers and connecting to the measurement electrodes through liquid or (quasi)solid electrolyte results in a six-electrode technique; an electrically driven regeneration of such sensors allows one to perform fast and completely reversible measurements. © 2010 Elsevier B.V. All rights reserved.

Ulrich Lange studied chemistry at the University of Regensburg from 2002–2007. His master thesis was focused on in-situ resistance measurements of thin conducting polymer films. The work was awarded with the Siemens VDO Award. His Ph.D. work which he did in the same university under the supervision of Prof. V.M. Mirsky and finished in 2010 was focused on new measurement techniques and new composite nanomaterials in chemiresistors and electrochemical transistors based sensors.

1. Introduction The most fascinating property of conducting polymers is their intrinsic conductivity and the ability to switch this conductivity over 10 orders of magnitude. The conductivity is not present in

∗ Corresponding author. Tel.: +49 357385917. E-mail address: [email protected] (V.M. Mirsky). 0003-2670/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2010.11.030

Vladimir M. Mirsky is a professor at Lausitz University of Applied Sciences, Germany. He graduated from Moscow Medical University in biophysics and went on to study physical chemistry and electrochemistry at the Frumkin Institute of Electrochemistry, obtaining his Ph.D. in 1986. He subsequently held an AvHumboldt Research Fellowship and a research position at CNRS prior to joining the Institute of Analytical Chemistry, Chemical Sensors and Biosensors at Regensburg University in 1995. After habilitation he became a professor of nanobiotechnology and moved to Lausitz. He is an editor of three books. His work has led to about 20 patents and patent applications as well as to about 120 peer-reviewed scientific papers.

their neutral (uncharged) state, but results from the formation of charge carriers upon oxidizing (p-doping) or reducing (n-doping) their conjugated backbone. Such a formation of charge carriers is accompanied by changes in optical and electrical properties and by changes of the dopant (counterions) concentration in the polymer. These changes are used in chemo- and biosensors [1–5]. There are several reasons to apply conducting polymers in chemo- and biosensors. All of them possess an intrinsic affinity towards redox active species and many to acidic or basic gases and

106

U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) 105–113

Fig. 1. Main configurations used for the analysis of the resistance of conducting polymers. (A) Two point configuration without fixation of polymer potential. (B) Typical configuration used in electrochemical experiments. (C) Two point configuration with fixation of polymer potential. (D) “Classical” four-point technique with a current source. (E) s24-configuration providing simultaneous two- and four-point measurements without fixation of polymer potential. (F) s24-configuration with fixation of the electrode potential.

solvent vapours. They can be modified with receptors to obtain a specific interaction with the analyte. Being immobilized on conducting polymers, such receptors provide an important advantage in comparison to monomer based receptors: the CP-wire provides a collective system response leading to high signal amplification in comparison to single molecular receptors [6]. Zhou and Swager demonstrated that a conjugated polymeric receptor for methylviologen shows a 65 times signal amplification in comparison to the monomer based receptor [7,8]. The amplification depends on the molecular weight. The application of conducting polymers in chemo- and biosensors can be realized with a number of different transducing techniques, allowing one to choose the most appropriate one for a particular sensor design. A detailed discussion of these aspects is given in [1]. Conductometric transducing of the sensor response is probably the most common method in chemo- and biosensors based on conducting polymers. There are several advantages of this transducing technique in chemosensors based on conducting polymer (such devices are named chemiresistors). (i) Small perturbations anywhere along the polymer chain can alter the conductance of the whole chain. Therefore, this approach provides a higher sensitivity than other techniques based on the modification of integral volume properties of the polymer, as, for example, electrochemical or colorimetric techniques [9]. (ii) Conductometric sensors

can be realized with a simple setup which nevertheless allows high precision measurements. (iii) Conductometric chemosensitive measurements can be realized even on nanowires of CP [10], therefore this technique is perfectly compatible with the actual trend for miniaturisation of analytical devices. (iv) Single chemiresistors can be easy combined into sensor arrays. (v) Using RFID technology, such sensors can also be adapted for non-contact measurements [11]. There are several devices for measuring the conductometric response of such sensors. The simplest and most often used is a chemiresistor (Fig. 1A, D and E). In the more common two point technique (Fig. 1A) the conducting polymer is deposited between two (typically interdigitated) electrodes separated by a narrow gap. The conductivity is measured by applying a constant current or voltage (dc or ac) between these electrodes and measuring the resulting voltage or current. The less used four-point measurement technique measures the conductance of the bulk polymer layer without the influence of the potential drop on the polymer–metal contacts (Fig. 1D). This technique was modified recently by combining the two- and four-point techniques for simultaneous measurements (Fig. 1E) [12,13] and was named as s24. Another possibility is the use of organic field effect transistors as sensors [4,14,15]. Here the current between the source and drain electrodes is controlled by the gate voltage. However these devices are out of the scope of this review. Based on the similarity

U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) 105–113

of the measurement configuration, Wrighton et al. [16–19] introduced the term “electrochemical transistor” for a device based on CP in which the redox state of this polymer is controlled by applying a voltage between the working electrodes (source and drain) and a reference electrode (gate) (see Section 4 for details) (Fig. 1C and F). The measurements of electrical current between two or more electrodes placed on the same solid support is certainly the most practicable configuration, however the low conductivity of some CP at typical conditions of bioanalytical applications (neutral pH, modest oxidation potential) may complicate this approach. In such cases, measurements of the resistance between an electrode coated by a conducting polymer and an electrode in the solution are used (Fig. 1B). These measurements can be realized by two- or three-electrode circuits which are common in electrochemistry. An advantage of such measurements is the about 100 times higher ratio of the electrode area to the layer thickness (or an effective distance between electrodes placed on solid support), resulting in an about 100 times lower absolute resistance than for the measurements of lateral resistance. All the configurations shown in Fig. 1 can be used in dc or ac mode and for impedance spectroscopy. In this review the different measurement configurations will be discussed in detail. Starting from the simplest two-electrode configuration, an evolution towards a four- and finally to a sixelectrode configuration will be discussed.

2. Two-electrode configuration In most chemiresistors conductance changes are measured by the two point technique. A small dc probe voltage (∼5–100 mV) or a constant current is applied between two electrodes separated by a small gap and the resulting current/voltage drop is measured. Such microelectrodes are mainly fabricated using photolithography. About 200 nm thick gold or platinum films are sputtered on oxidized silicon or glass wafers using a very thin adhesion layer. The gap between the electrodes is usually between 1.5 ␮m and 100 ␮m, however narrower gaps are favoured, as a smaller amount of polymer is needed to cover the gaps, resulting in an increased sensitivity. If the polymer is grown electrochemically over the gaps, larger gaps usually lead to thicker polymer films and therefore in a loss of sensitivity [20,21]. Pre-treatment of the electrode support with, for example, hydrophobic silanes can improve the lateral growth of the polymer over the gaps [22–25], whereas pretreatment of gold electrodes with thiol modified monomers can improve the contact between the electrode and the polymer film and therefore lower the influence of the contact resistance [12]. To achieve a longer gap length at limited electrode area, interdigitated electrodes are commonly used. An application of a constant potential difference or current to measure conductance changes in sensors based on conducting polymer leads to some problems. It may induce irreversible or slowreversible changes in the polymer, which can be avoided by using of ac technique or by alteration of polarity of dc pulses [3]. Such changes may be a result of polarisation effects in the polymer or at a high potential difference are due to overoxidation of the polymer in the vicinity of one contact. In addition, passing a high current through CP microelectrodes leads to a self-heating of the polymer, which makes its conductance sensitive to such parameters as airflow [26]. The thermal response time was found to be in the order of 1 ms. Therefore the probe power should not exceed a certain limit. Another reason to keep the voltage probe as small as possible, is the non-linearity of the current voltage dependence. In contrast, the signal to noise (S/N) ratio increases with increasing current. As a consequence, a compromise between a sufficiently high current to provide a suitable S/N ratio and the need to keep self-heating

107

effects negligible has to be found [26]. Ac measurements have some advantages over dc measurements. It was reported that the current noise during conductivity measurements displays flicker noise behaviour and decreases as 1/f (where f is frequency) with frequency increase. This noise behaviour was explained by a contribution of the contact resistance between single polymer grains to the overall resistance of the polymer film. This intergrain resistance is shunted by a capacitance bypassing the resistance at higher frequencies [26]. Furthermore, using ac measurements, one can make complete impedance analysis or monitor simultaneously resistive and capacitive changes which enhance the sensor selectivity [27,28]. A sensor device interrogating the response at different frequencies was reported. Measurements at few selected frequencies are much faster than impedance spectroscopy [29]. Using multiple electrodes and multiplexing between the electrodes allows increasing the measurement throughput. Slater et al. used such multiple electrodes to design a multilayered conducting polymer gas sensor [30]. Based on the measurements of electrical conductance between electrodes with different spacing, the authors proposed that it is possible to relate the conductance to different zones of the polymer. Hence the conductance between closely spaced electrodes is provided mainly by a thin polymer layer close to the electrode substrate. The thickness of this layer estimated from combinatorial experiments for polyaniline [13,31] corresponds to the polymerisation charge of about 2.5 C m−2 , while maximal relative sensitivity to gaseous HCl is observed at about 0.5 C m−2 . Thinner films were reported to be more sensitive towards vapours than thicker ones [21,32]. A higher polymer grow in lateral direction in comparison with transversal direction suggests that the polymer chains are oriented mainly parallel to the support surface. It can lead to electrical anisotropy and to higher conductivity in the lateral direction. If equilibrium between polymer and analyte is not reached, the analyte concentration decreases through the film thickness, therefore the polymer near the electrode support is less affected by analyte interaction in thicker films. However, because of the anisotropic structure, the polymer layer near support is mainly responsible for the resistance changes. This effect is expected to be especially strong if an interaction of polymer with analyte leads to an increase of the polymer resistance. The resistance measured by the two-point technique includes the bulk polymer resistance and the resistance between the contacts and the polymer. If the contact resistance is high and shows in comparison to the polymer resistance only small changes upon analyte interaction, it can limit the sensitivity of the system. Most synthesis techniques leads to the formation of micrometer or submicrometer thick layers of conducting polymers, therefore the method of Cox and Strack [33] based on the variation of the ratio of the contact area to the material thickness can hardly be applied to distinguish bulk and contact resistances of conducting polymers. Principally, it can be done by impedance spectroscopy: by measuring a wide frequency spectra one can separate the resistance of the polymer and the resistance between the polymer and the metal contacts [34,35]. However these measurements are relatively slow as a wide frequency range has to be covered. Additionally, the results of impedance spectroscopy are influenced by the selection of equivalent circuit used for the data analysis. More easily this problem can be solved by measurement techniques based on four-electrode configuration.

3. Four-electrode configuration In the four-electrode configuration whose invention relates to Sir W. Thompson (Lord Kelvin), the conductivity is usually measured by applying a constant current between the two outer electrodes and by measuring the potential difference between the

108

U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) 105–113

Fig. 2. Evolution of the electrode geometry from linear to interdigitated four contacts (Reprinted with permission from [12]. Copyright 2003 Elsevier). Below: Array of 96 of such electrodes on a wafer (Reprinted with permission from [36]. Copyright 2004 Wiley-VCH).

inner electrodes. This potential difference is measured by a highimpedance voltmeter, and therefore is not influenced by ohmic potential drop on the contact resistance of the inner electrodes. Usually, four parallel metallic strips are used as contacts for the four electrode configuration. To get a more effective use of the sensor surface, interdigitated electrodes for four-point measurements have been designed. Fig. 2 shows a “mental evolution” of the electrode geometry [12]. Such electrodes were used as single sensors and as arrays consisting of 96 such electrodes (Fig. 2), allowing a combinatorial screening of sensing materials [13,36–38]. If the contact resistance is so high that it limits the dynamic range of the sensor signal, the application of a four-electrode configuration results in a higher sensitivity [12–20]. On the other hand four-electrode measurements alone provide no information on the resistance of polymer/metal contacts. However, many processes lead to detaching of the polymer or to formation of low conducting contacts due to modification of polymer band structure. This results in a higher contact resistance. For example, a detachment of polymer induced by usual procedure of chemical immobilization of proteins was reported [13]. On the other hand, the contact resistance can contain additional analytical information. One of these possibilities is based on measuring the time delay between two- and four-point resis-

tance signals caused by diffusion through polymer layer (Fig. 3). This approach was reported in [39] for detection of gaseous HCl by polydialkoxybipyrrols using interdigitated gold electrodes, a ratio of two- to four-point resistances was used as an analytical signal. HCl leads to decrease of both bulk- and contact resistances. The delay caused by analyte diffusion through chemosensitive polymer leads to a peak of the kinetics of this ratio, a width of this peak depends monotonously on the analyte concentration [39]. One can imagine a further development of this technology using a chromatography-like determination of analytes in complex mixtures. Simultaneous two and four-point resistance measurements (s24) (Fig. 1E), introduced by our group, provide a possibility to measure both four- and two-point resistance and by subtraction of both values to determine a contact resistance [12,13,40]. The ratio of the resistances measured by two- (R2 ) and four (R4 ) techniques provides valuable information on the quality of metal/polymer contacts which can be used to control immobilization of biomolecules [13] and for other applications. An extraction of quantitative data from such measurements is based on the following idea. The twopoint technique gives the value R2 which is the sum of bulk resistance and two contact resistances, while the four-point technique measures the bulk resistance between some mean points

analyte

diffusion

}

response of bulk resistance time delay

analyte concentration

response of contact resistance Fig. 3. Principle of analytical application of simultaneous two- and four-point detection [39]. The signal caused by contact resistance is delayed relative to the signal caused by bulk polymer resistance. The delay depends on the diffusion through the polymer layer and therefore is a function of analyte concentration.

U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) 105–113

4. Configuration for the in situ conductivity measurements and electrochemical transistors

s24 impedance

10000

R / R0

1000

100

10

1 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1

0.0

0.1

0.2

E / V vs. SCE Fig. 4. Comparison of the contact resistance between polypyrrole and gold measured by simultaneous two- and four-point techniques (s24 approach) and by impedance spectroscopy (Reprinted with permission from [40]. Copyright 2008 Elsevier).

of the inner electrode R4 . The contact resistance can therefore be calculated as: RC = R2 − ˛ · R4 ,

109

(1)

where ˛ is the geometrical factor. One can expect that for electrodes with a geometry drawn in Fig. 2, coated by thin homogeneous polymer films whose thickness is much less than the width, the value of the geometric factor ˛ is about 3 [40]. This was confirmed by many hundreds of experimental observations performed by combinatorial technique: the minimum value of the ratio R2 /R4 is 3.1 ± 0.1 for various coatings (by polyaniline and by binary copolymers of polyaniline derivates with polyaniline) [13,37]. If the electrode geometry is the same for the all studied polymers, one can apply Eq. (1) to calculate contact resistance from the values of R2 and R4 postulating some definitive value of the factor ˛ (for our electrodes ˛ = 3). It was demonstrated that small variations of ˛ do not lead to qualitative changes of the behaviour of the contact resistance. This approach was applied to study the dependence of the contact resistance between polypyrrole and gold electrodes at different electrode potentials. The results were compared with impedance spectroscopy (Fig. 4). A good confidence for a four order range of the resistance changes was observed [40]. Several other approaches to get quantitative evaluation of the contact resistance were described. Similar to the s24 technique other techniques to measure the contact resistance are discussed in literature. This value can be obtained by switching between a fourand a three electrode configurations [41]. Potential measurements between source and sensing electrodes and linear extrapolation of the voltage drop allow calculating the voltage drop at the source and drain contacts [42]. An investigation of the resistance dependence on the distance between two electrodes and determination of the contact resistance by subsequent extrapolation of this dependence to zero distance was reported [43,44]. The technique also provides an evaluation of the film homogeneity, but requires a number of measurements to get a single measurement of the contact resistance and changes in the contact resistance in a single sensor can be therefore not monitored. The s24-technique is important for also such conductometric sensors in which an analytical signal is caused only by bulk polymer properties. In this case the value of the contact resistance or the ratio of the resistances measured by two- and four-point techniques provides valuable information on the sensor integrity: even a slight desorption of chemosensitive polymer leads to a high increase of the contact resistance. Therefore, the s24 technique can be used for internal quality control of sensor devices.

A number of applications require the control of the polymer redox state. This can be done electrically, by fixation of the polymer potential relative to a reference electrode. Such measurements are often referred as in situ resistance measurements and the measurement configurations are often called as electrochemical transistors. This designation was accepted in literature and is therefore used here, however it should not be confused with semiconductor or organic transistors operating in a different way. The first application of electrochemical transistors for conducting polymers was described in 1984 by White et al. [16]. They used a symmetric configuration consisting of three gold electrodes 3 ␮m wide separated by a 1.4 ␮m gap placed on insulation support. The middle electrode used to control the redox-state of the polymer was connected as a working electrode with a potentiostat. External auxiliary and reference electrodes were placed in the electrolyte solution. Due to analogy to field effect transistors, the middle gold microelectrode was named gate and the two outer electrodes source and drain (Fig. 5, left). This configuration was further simplified by excluding the middle electrode and by control of the redox state of the polymer through the source (or drain) electrode (Fig. 5, right). This allowed to get a narrower gap between the source and drain electrode and therefore a faster response time. However, at high polymer resistivity this asymmetric configuration may lead to deviation of the polymer redox state near the second electrode and therefore to inhomogeneous polymer properties between the source and drain. A further simplification of the configuration of the electrochemical transistor can be performed by replacement of the potentiostat with reference and auxiliary electrodes just by a reference electrode. However in this case the ohmic drop at low polymer resistance can lead to deviations from the applied potential. A potential difference between source and drain electrodes required for conductivity measurements can be achieved by using a bipotentiostat [45]. Using a bipotentiostat, the potentials of the source and drain electrode (W1 and W2) are controlled simultaneously by keeping a constant small potential difference (e.g. 5 mV). The currents through drain (IW1 ) and source (IW2 ) electrodes consist of a faradaic components caused by the redox reactions in the polymer and an ohmic component caused by the potential difference between the drain and source electrodes. Assuming that the magnitudes of the faradaic currents (iF ) at both electrodes are equal, one can calculate the ohmic current through the polymer film as I = (IW1 − IW2 )/2 [45]. Another approach is based on combination of usual two- or four-electrode configurations for resistance measurement with a reference electrode connected directly or through a potentiostat. However if a constant voltage is applied between the source and drain electrodes, the measured current contains not only the current between the two electrodes but also the current between polymer layer and auxiliary or reference electrode in the electrolyte [46]. To overcome this problem, low frequency (<1 Hz) voltage pulses [40,46] or triangle voltage waveforms of alternating polarity are applied [47]. In these cases different methods to calculate the polymer resistance were used. Wrighton and coworkers calculate it from the slope of the I–V characteristic at zero voltage [47]. If pulses instead of triangle waveforms are used, the current is usually measured at the end of each pulse, when transient effects are minimal. In [46] the current is calculated from three successive measurements, by subtracting the negative pulse current from the average current of the two surrounding positive pulses. Averaging of the positive and negative pulse yields almost the same results. The elimination of the current crossing the electrochemical cell is particularly important if the current between source

110

U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) 105–113

Fig. 5. Symmetric (left) (Reprinted with permission from [16]. Copyright 1984 American Chemical Society) and asymmetric (right) (Reprinted with permission from [9]. Copyright 2007 Chemical Society of Japan) configurations of electrochemical transistors.

and drain is very small, for example, if the polymer has a high resistance. To use such electrochemical transistors in sensor applications, a detailed characterisation of factors influencing their in situ resistance should be done first. Such in situ resistance measurements have been widely used to study thin conducting polymer films [16,17,25,46–61]. This approach was also combined with other techniques, like cyclic voltammetry [51,53,54,56,58–61] or electron spin resonance [46,54]. However in most studies two-point resistance measurements were used so far. An application of simultaneous two- and four-point measurements demonstrated that the data obtained by two-electrode configuration can contain a significant error due to the contact resistance, especially at high polymer conductance [40] (Fig. 6). It has to be noted that in such measurements the exact values of the conductivity are often less important than its change as a function of applied potential. The conductance changes in such experiments can be few orders of magnitude, therefore a logarithmic scale is often used [61]. Measurements of the in situ conductance during polymer growth provide a simple possibility to stop polymerisation at a desired polymer resistance [20,25]. This may be important for a reproducible sensor fabrication. Additionally, one can obtain the minimum polymerisation charge to cover the gaps between two electrodes. As the sensitivity is influenced by the amount of polyR4

10000

RC

R/Ω

1000 100 10 1 0.1 -0.6

-0.4

-0.2

0.0

0.2

0.4

E / V vs. SCE Fig. 6. Contact (RC ) and bulk (R4 ) (measured by four-electrode techniques) resistances of electropolymerised polypyrrole measured in configuration of electrochemical transistor at different potentials relative to saturated calomel electrode (Reprinted with permission from [40]. Copyright 2008 Elsevier).

mer deposited between the electrodes, this technique can be used to optimize sensor sensitivity. The possibility to control the oxidation state of conducting polymers is important in conducting polymer based conductometric chemo- and biosensors. In such sensors which respond to pH [24,25], ions [62,63] or other analytes which bind to receptors linked to the conducting polymer backbone [64], the electrochemical control of the redox state provides a higher reliability and reproducibility of the sensor response because spontaneous variations of the redox potential can be excluded. Furthermore in sensors which measure changes in redox potential due to interaction with any redox-active analytes (NADH, hydrazine, different redox active gases, etc.) the electrochemical control provides a fast regeneration and conditioning of the sensor before the measurement [65–67]. A modification of conducting polymers with enzymes which interact directly with the polymer or release/uptake any compounds effecting the polymer resistance during their enzymatic cycle is an approach to design conductometric enzymatic biosensors [25,68–71]. To enhance the sensitivity of such devices, a collector electrode can be connected to the source electrode. If the potential at the collector electrode changes, these changes will be transferred to the conducting polymer film bridging source and drain electrodes resulting in a change of its conductivity [47,72,73]. Another approach is based on using of conducting polymers as transducer. In this case the polymer is deposited between source and drain while the gate electrode is coated by a receptor layer. In this case potential changes on the gate electrode influence the conducting polymer thus changing its conductivity. Such configurations were used for detection of hydrogen peroxide; a platinum gate electrode in combination with a PEDOT–PSS film between source and drain electrodes was used [74,75]. In a simple realization, external reference or reference/auxiliary electrodes can be used in electrochemical transistors; however these electrodes can also be implanted on the microchip surface. This is especially important in solid state devices. Solid state electrochemical transistors were first reported by Chao and Wrighton [76,77]. They deposited a solid electrolyte over an array consisting of eight microelectrodes. Two of these electrodes were used as counter electrode, the other six were connected by a conducting polymer film. A spot of silver glue was used as reference electrode. The chip was coated by polyvinylalcohol and suggested as a humidity sensor [76].

U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) 105–113

The influence of faradaic and non-faradaic processes at the gate electrode on the performance of electrochemical transistors was evaluated [78]. The charge needed for oxidation (doping) of a conducting polymer can be compensated by discharging of the double layer at the gate electrode (non-faradaic process) or by a reduction process at the gate electrode (faradaic process). However, the relatively small capacitance of the ionic double layer can limit the charge of the polymer oxidation. Faradaic processes have much higher pseudocapacitance, this leads to a higher sensitivity of the transistor current to the changes of its gate potential. To enable a faradaic process on the gate electrode, it can also be covered by a conducting polymer. A high surface ratio between gate and work electrodes decreases the switching time. The switching time and the charge needed to switch the device also depend strongly on the amount of polymer necessary to bridge the gap between source and drain electrode. A switching time of 0.1 ms was reported for PANI using a gap of 50–100 nm. This device was switched by an electric charge below 1 nC. For a similar device with 1.5 ␮m gap, about 100 times higher charge was required [79]. Devices with very small gaps were also used in sensors. By modifying the polyaniline with glucose oxidase, a biosensor for glucose designed for in vivo applications was fabricated [80]. Electrochemical transistors with conducting polymers were also fabricated on flexible substrates [81,82]. In this case PEDOT/PSS is used as contact, channel and gate material. Fast switching rates at low humidity can be achieved using a very hygroscopic solid electrolyte consisting of PSS, ethyleneglycol, sorbitol and LiClO4 [83]. Nilsson et al. suggested several measurement configurations for such transistors. The simplest one is a three electrode electrochemical transistor which has a gate electrode referenced to the source electrode. A solid electrolyte covers the channel between the source and drain electrode and the gate electrode [82,84]. A modification of this configuration can be achieved if the gate electrode is not referenced to the source but to a second gate electrode which is in contact with the channel [81,82]. This is similar to the configuration of White et al. (Fig. 5) with a middle electrode as gate electrode between the source and drain electrodes. The four terminal electrochemical transistor was tested as a humidity sensor [81]. The conductance change of the channel upon applying a gate voltage of 1.2 V depends strongly on the humidity dependent conductance of the solid electrolyte (Nafion). The same principle was used for sensing of the ionic strength of an electrolyte however a configuration using two three terminal electrochemical transistors, in which one served as reference, was used [85]. In a third configuration (Fig. 7) both gate electrodes are not in contact with the channel [82]. By applying a gate voltage between these electrodes, an electrochemical reaction is also induced in the areas c and d. If area a is oxidized, the area b is reduced leading to an oxidation of area d and a reduction of area c. This configuration was called dynamic electrochemical transistor

111

Fig. 7. Configuration of the dynamic electrochemical transistor. Reprinted with permission from [82]. Copyright 2002 Wiley-VCH.

because it switches back to its initial state by switching off the gate voltage. The initial motivation to develop the configuration of electrochemical transistor was the requirement to control the redox-state of conducting polymers thus fixing its physical and chemical properties at the state corresponding to the optimal chemosensitivtiy. However it also provides a possibility for electrochemical regeneration and conditioning. This approach was recently extended to gas sensors in our group: electrical switching between redox states was used for a fast regeneration of the sensor material after an interaction with a gaseous analyte [86]. 5. Integrated electrochemical transistors for non-conducting liquids and gaseous phases: from four- to six-electrode configuration A combination of the four-point configuration for simultaneous measurements of polymer conductivity and contact resistance with configuration of electrochemical transistor to control redox state of chemosensitive materials leads to the six-point configuration. For a convenient application in chemo- and biosensors, the counter and reference electrode together with the four electrodes for resistance measurements were installed on one chip (Fig. 8). The reference electrode gold strip was further modified by Ag or Ag/AgCl, whereas the counter electrode and the four electrodes for resistance measurements were coated by polyaniline or polythiophene. Fig. 9 demonstrates the potential dependent two- and four-point resistance of polyaniline measured with that configuration in 1 M HCl. The suggested configuration provides a number of advantages. The device combines the advantages of the four-point configuration for resistive measurements with that of electrochemical transistors. By coating the chip with a gel- or solid electrolyte it can be used as a solid-state sensing device. No external electrodes

Fig. 8. Electrodes designed for six-point measurements with interdigitated (left) and linear (right) four stripes working electrodes for simultaneous two- and four-point measurements. The two surrounding stripes are used as reference and counter electrode.

112

U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) 105–113

40 10

30

4

3

10 10

2

I / µA

R/Ω

20 10

0 -10

10

1

A -20 0.0

0.2

0.4

0.6

0.8

E / V vs. Ag / AgCl Fig. 9. Potential dependent two- (open circles) and four- (filled squares) point resistance and cyclic voltammogram of polyaniline in 1 M HCl measured by six-electrode configuration.

are needed for the measurements; this makes the device very convenient and technological and is especially important for further miniaturisation. Coating of the chip with a gel electrolyte allows one to use it also as a gas sensor. Such an electrochemical transistor based on polythiophene was used as a gas sensor for nitrogen dioxide. The electrochemical control of the sensor film allowed a fast regeneration of the sensor after analyte exposure [86]. 6. Conclusion Measurement configurations used for the fabrication of conductometric chemosensors based on conducting polymers passed a long way from simple two-electrode circuits to sophisticated multielectrode systems. The four-electrode configuration provides a separated analysis of bulk and contact resistances; both values can be applied in chemosensors independently. The configuration of electrochemical transistors provides a control of the polymer redox state and can be realized in liquid as well as in solid electrolytes. Finally, a six-electrode configuration combining advantages of four-point configuration and electro-chemical transistor was suggested. The described principles were developed and tested with conducting polymers, but they are quite general and can be transferred to other types of conductometric chemosensors. Further one can also expect a combination of numerous from each other electrically insulated six-electrode sensors into sensor arrays for the formation of artificial noses and tongues. Acknowledgement U.L. was supported by DFG fellowship. References [1] U. Lange, N. Roznyatovskaya, V.M. Mirsky, Anal. Chim. Acta 614 (2008) 1–26. [2] S. Brahim, A.M. Wilson, D. Narinesingh, E. Iwuoha, A. Guiseppi-Elie, Microchim. Acta 143 (2003) 123–137. [3] A. Guiseppi-Elie, G. Wallace, T. Matsue, Handbook of Conducting Polymers, 1998. [4] J. Mabeck, G. Malliaras, Anal. Bioanal. Chem. 384 (2006) 343–353. [5] S.W. Thomas, G.D. Joly, T.M. Swager, Chem. Rev. 107 (2007) 1339–1386. [6] T.M. Swager, Acc. Chem. Res. 31 (1998) 201–207. [7] Q. Zhou, T.M. Swager, J. Am. Chem. Soc. 117 (1995) 7017–7018. [8] Q. Zhou, T.M. Swager, J. Am. Chem. Soc. 117 (1995) 12593–12602. [9] K. Sugiyasu, T.M. Swager, Bull. Chem. Soc. Jpn. 80 (2007) 2074–2083. [10] A. Wanekaya, W. Chen, N. Myung, A. Mulchandani, Electroanalysis 18 (2006) 533–550. [11] R.A. Potyrailo, W.G. Morris, Anal. Chem. 79 (2007) 45–51. [12] Q. Hao, V. Kulikov, V.M. Mirsky, Sens. Actuators B 94 (2003) 352–357.

[13] V. Kulikov, V.M. Mirsky, T.L. Delaney, D. Donoval, A.W. Koch, O.S. Wolfbeis, Meas. Sci. Technol. 16 (2005) 95–99. [14] C. Bartic, G. Borghs, Anal. Bioanal. Chem. 384 (2006) 354–365. [15] J. Janata, M. Josowicz, Nat. Mater. 2 (2003) 19–24. [16] H.S. White, G.P. Kittlesen, M.S. Wrighton, J. Am. Chem. Soc. 106 (1984) 5375–5377. [17] J.W. Thackeray, H.S. White, M.S. Wrighton, J. Phys. Chem. 89 (1985) 5133–5140. [18] D. Ofer, R. Crooks, M. Wrighton, J. Am. Chem. Soc. 112 (1990) 7869–7879. [19] M. Wrighton, J. Thackeray, M. Natan, D. Smith, G. Lane, D. Belanger, W. Albery, Philos. Trans. R. Soc. Lond. B: Biol. Sci. 316 (1987) 13–30. [20] A. Partridge, P. Harris, M. Andrews, Analyst 121 (1996) 1349–1353. [21] E. Stussi, R. Stella, D. De Rossi, Sens. Actuators B 43 (1997) 180–185. [22] T. Matsue, M. Nishizawa, I. Uchida, Nippon Kagaku Kaishi (1995) 493–501. [23] M. Nishizawa, Y. Miwa, T. Matsue, I. Uchida, J. Electroanal. Chem. 371 (1994) 273–275. [24] M. Nishizawa, T. Matsue, I. Uchida, Sens. Actuators B 13 (1993) 53–56. [25] M. Nishizawa, T. Matsue, I. Uchida, Anal. Chem. 64 (1992) 2642–2644. [26] P. Harris, W. Arnold, M. Andrews, A. Partridge, Sens. Actuators B 42 (1997) 177–184. [27] F. Musio, M.C. Ferrara, Sens. Actuators B 41 (1997) 97–103. [28] M. Amrani, A. Peter, K.C. Payne, Persaud, Sens. Actuators B 33 (1996) 137–141. [29] M. Amrani, R. Dowdeswell, P. Payne, K. Persaud, Sens. Actuators B 47 (1998) 118–124. [30] J. Slater, J. Paynter, E. Watt, Analyst 118 (1993) 379–384. [31] R.A. Potyrailo, V.M. Mirsky, Chem. Rev. 108 (2008) 770–813. [32] P.N. Bartlett, S.K. Ling-Chung, Sens. Actuators 19 (1989) 141–150. [33] R. Cox, H. Strack, Solid-State Electron. 10 (1967), 1213–1214, IN7–IN8, 1215–1218. [34] C. Deslouis, M. Musiani, B. Tribollet, M.V. Vorotyntsev, J. Electrochem. Soc. 142 (1995) 1902–1908. [35] X. Ren, P.G. Pickup, J. Electroanal. Chem. 420 (1997) 251–257. [36] V.M. Mirsky, V. Kulikov, Q. Hao, O.S. Wolfbeis, Macromol. Rapid Commun. 25 (2004) 253–258. [37] V. Kulikov, V.M. Mirsky, Meas. Sci. Technol. 15 (2004) 49–54. [38] V.M. Mirsky, Combinatorial Methods for Chemical and Biological Sensors, Springer, New York, 2009, pp. 315–330. [39] M. Krondak, G. Broncova, S. Anikin, A. Merz, V.M. Mirsky, J. Solid State Electrochem. 10 (2006) 185–191. [40] U. Lange, V.M. Mirsky, J. Electroanal. Chem. 622 (2008) 246–251. [41] S.G. Haupt, D.R. Riley, J. Zhao, J.T. McDevitt, J. Phys. Chem. 97 (1993) 7796–7799. [42] P.V. Pesavento, R.J. Chesterfield, C.R. Newman, C. Frisbie, J. Appl. Phys. 96 (2004) 7312–7324. [43] Q.J. Cai, M.B. Chan-Park, J. Zhang, Y. Gan, C.M. Li, T.P. Chen, B.S. Ong, Org. Electron. 9 (2008) 14–20. [44] P.V. Necliudov, M.S. Shur, D.J. Gundlach, T.N. Jackson, Solid-State Electron. 47 (2003) 259–262. [45] M. Nishizawa, I. Uchida, Electrochim. Acta 44 (1999) 3629–3637. [46] J. Kruszka, M. Nechtschein, C. Santier, Rev. Sci. Instrum. 62 (1991) 695–699. [47] E.W. Paul, A.J. Ricco, M.S. Wrighton, J. Phys. Chem. 89 (1985) 1441–1447. [48] M. Gholamian, T. Kumar, A. Contractor, Proc. Indian Acad. Sci. Chem. Sci. 97 (1986) 457–464. [49] L. Groenendaal, G. Zotti, F. Jonas, Synth. Met. 118 (2001) 105–109. [50] R. Holze, J. Lippe, Synth. Met. 38 (1990) 99–105. [51] M.C. Morvant, J.R. Reynolds, Synth. Met. 92 (1998) 57–61. [52] G. Schiavon, S. Sitran, G. Zotti, Synth. Met. 32 (1989) 209–217. [53] G. Zotti, G. Schiavon, Synth. Met. 39 (1990) 183–190. [54] G. Zotti, Synth. Met. 97 (1998) 267–272. [55] E. Csahok, E. Vieil, G. Inzelt, J. Electroanal. Chem. 482 (2000) 168–177. [56] L. Olmedo, I. Chanteloube, A. Germain, M. Petit, E. Genies, Synth. Met. 30 (1989) 159–172. [57] H. Huang, P. Pickup, Acta Polym. 48 (1997) 455–457. [58] H. John, R. Bauer, P. Espindola, P. Sonar, J. Heinze, K. Müllen, Angew. Chem. Int. Ed. 44 (2005) 2447–2451. [59] D. Wei, P. Espindola, T. Lindfors, C. Kvarnstroem, J. Heinze, A. Ivaska, J. Electroanal. Chem. 602 (2007) 203–209. [60] E. Sezer, M. Skompska, J. Heinze, Electrochim. Acta 53 (2008) 4958–4968. [61] J. Heinze, B.A. Frontana-Uribe, S. Ludwigs, Chem. Rev. 110 (2010) 4724–4771. [62] Z. Mousavi, A. Ekholm, J. Bobacka, A. Ivaska, Electroanalysis 21 (2009) 472–479. [63] D.A. Bernards, G.G. Malliaras, G.E. Toombes, S.M. Gruner, Appl. Phys. Lett. 89 (2006) 053505. [64] B. Fabre, L. Taillebois, Chem. Commun. (2003) 2982–2983. [65] P.N. Bartlett, J.H. Wang, E.N.K. Wallace, Chem. Commun. (1996) 359–360. [66] P.N. Bartlett, P.R. Birkin, Anal. Chem. 66 (1994) 1552–1559. [67] U. Lange, V.M. Mirsky, Electrochim. Acta (2010), doi:10.1016/j.electacta. 2010.08.092. [68] P. Bartlett, Y. Astier, Chem. Commun. (2000) 105–112. [69] P. Bartlett, P. Birkin, J. Wang, F. Palmisano, G. De Benedetto, Anal. Chem. 70 (1998) 3685–3694. [70] D.T. Hoa, T.N.S. Kumar, N.S. Punekar, R.S. Srinivasa, R. Lal, A.Q. Contractor, Anal. Chem. 64 (1992) 2645–2646. [71] H. Sangodkar, S. Sukeerthi, R.S. Srinivasa, R. Lal, A.Q. Contractor, Anal. Chem. 68 (1996) 779–783. [72] Y. Astier, P. Bartlett, Bioelectrochemistry 64 (2004) 15–22. [73] Y. Astier, P.N. Bartlett, Bioelectrochemistry 64 (2004) 53–59.

U. Lange, V.M. Mirsky / Analytica Chimica Acta 687 (2011) 105–113 [74] Z. Zhu, J.T. Mabeck, C. Zhu, N.C. Cady, C.A. Batt, G.G. Malliaras, Chem. Commun. (2004) 1556–1557. [75] D.A. Bernards, D.J. Macaya, M. Nikolou, J.A. DeFranco, S. Takamatsu, G.G. Malliaras, J. Mater. Chem. 18 (2008) 116–120. [76] S. Chao, M.S. Wrighton, J. Am. Chem. Soc. 109 (1987) 6627–6631. [77] S. Chao, M. Wrighton, J. Am. Chem. Soc. 109 (1987) 2197–2199. [78] F. Lin, M.C. Lonergan, Appl. Phys. Lett. 88 (2006) 133507–133513. [79] E.T. Jones, O.M. Chyan, M.S. Wrighton, J. Am. Chem. Soc. 109 (1987) 5526–5528. [80] E.S. Forzani, H. Zhang, L.A. Nagahara, I. Amlani, R. Tsui, N. Tao, Nano Lett. 4 (2004) 1785–1788.

113

[81] D. Nilsson, T. Kugler, P.O. Svensson, M. Berggren, Sens. Actuators B 86 (2002) 193–197. [82] D. Nilsson, M. Chen, T. Kugler, T. Remonen, M. Armgarth, M. Berggren, Adv. Mater. 14 (2002) 51–54. [83] M. Hamedi, R. Forchheimer, O. Inganas, Nat. Mater. 6 (2007) 357–362. [84] D. Nilsson, N. Robinson, M. Berggren, R. Forchheimer, Adv. Mater. 17 (2005) 353–358. [85] P. Svensson, D. Nilsson, R. Forchheimer, M. Berggren, Appl. Phys. Lett. 93 (2008). [86] U. Lange, V.M. Mirsky, Anal. Chim. Acta (2011), doi:10.1016/j.aca.2010. 12.008.