Electrochemical impedance spectroscopy studies of

Trends in Analytical Chemistry, Vol. 24, No. 1, 2005

Trends

Electrochemical impedance spectroscopy studies of polymer degradation: application to biosensor development C esar Fernandez-S anchez, Calum J. McNeil , Keith Rawson Electrochemical impedance spectroscopy (EIS) has emerged as a useful analytical tool for the development of sensor devices in a wide variety of configurations. We focus this review on the particular application of EIS to the study of degradation phenomena taking place at polymer-coated substrates, which have primarily been of great interest in the study of corrosion protection; more recently, it has led to the successful construction of versatile polymer-coated transducers for sensor development. Impedance analysis of breakdown processes of polymer coatings on electrochemical transducers through the direct or indirect action of biomolecules constitutes a feasible detection protocol for the fabrication of generic integrated biosensors. We give a detailed description of such applications, and present a particular view on sensor devices reported so far and ideas that bring significant improvements to this promising technology. ª 2004 Elsevier Ltd. All rights reserved. Keywords: Degradation; Enzyme sensor; Immuno sensor; Impedance spectroscopy; Polymer-coated electrodes

1. Introduction C esar Fern andez-S anchez, Calum J. McNeil* School of Clinical and Laboratory Sciences, The Medical School, University of Newcastle upon Tyne, Framlington Place, Newcastle upon Tyne NE2 4HH, UK Keith Rawson Cambridge Life Sciences, Cambridge Business Park, Angel Drove, Ely, Cambridgeshire CB7 4DT, UK

*Corresponding author. Fax: +44-191-2227991; E-mail:

Electrochemical impedance spectroscopy (EIS) appears to be an excellent technique for the investigation of bulk and interfacial electrical properties of any kind of solid or liquid material connected to or being part of an appropriate electrochemical transducer [1]. Any intrinsic property of a material or a specific process that could affect the conductivity of an electrochemical system can potentially be studied by EIS. It is a non-invasive technique that does not require complex or expensive instrumentation and is easy to operate, thus allowing its current applications not only in research laboratories but also as a tool for the control of processes such as the performance of batteries, semiconductors,

0165-9936/$ - see front matter ª 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.trac.2004.08.010

or thin-film technology, or monitoring corrosion [2]. Using EIS, a vast range of coatings have been widely tested as effective barriers against corrosion of metal surfaces in the last few decades. Paints and other organic and inorganic emulsions deposited on the metal surface break down creating pinholes, craters and other defects when corroded. In other words, water and any other existing free ions are allowed to penetrate into the polymer. This so-called ionic attack alters the insulating structure of the polymer, which modifies the impedance characteristics of the overall metal/polymer element. Microbial colonies and marine water have been reported as common corrosion agents, their effect on different polymers being assessed with EIS at set frequencies. EIS has also been increasingly successful in the design and the development of sensor systems. We need to bear in mind the vast range of materials that may be used in their fabrication, as these may confer specific electric properties to the resulting device. Also, a great number of methodologies can be performed to integrate the appropriate recognition element with the transducer, which, in turn, gives rise to the generation of different interfaces with specific electric properties. In this context, the versatility of EIS allows its application to the control and monitoring of the different stages necessary for the fabrication of the sensor and its eventual characterisation. Also, EIS has been used successfully as an 37

Trends

analytical tool for the measurement of electric property changes of the sensor system in the presence of increasing concentrations of the corresponding analyte. Based on the applications outlined above, the impedance analysis of deterioration processes taking place on polymer-coated transducers under controlled experimental conditions appears to be an excellent transduction protocol in the development of electrochemical sensors. To date, degradation mechanisms going from pH changes to hydrolysis of the side chains of the polymer to temperature gradients have all been monitored by EIS and applied to the development of sensor systems for a few specific applications. Nevertheless, those resulting devices can be envisaged as generic platforms for the measurement of an extensive range of analytes of medical and environmental interest.


such circuits, a resistance ideally describes a conductive path, such as that generated by the bulk conductivity of the system or the charge-transfer step due to an electrode reaction, whereas a capacitance generally describes space-charge-polarisation regions within the system as well as modification of an electrode surface due to adsorption processes or polymer-layer deposition. The Randles circuit is the simplest equivalent circuit that describes an electrochemical cell where a single-step Faradaic process in the presence of diffusion may occur (Fig. 1(a)). It combines three components, namely the electrolyte resistance between working and reference electrodes ðRe Þ, the double-layer capacitance ðCdl Þ and the Faradaic impedance due to the charge-transfer process ðZf Þ at the working electrode-electrolyte interface. Zf is normally subdivided into a charge-transfer

2. Theoretical background

Cdl

(a)

EIS measures the impedance of a system (Z) as a function of frequency of an applied perturbation. When working with electrochemical systems, this perturbation is normally an AC voltage of small amplitude (typically 5–10 mV peak-to-peak) and the response is a current that differs in amplitude and phase (phase difference, /) with the applied voltage. The ratio of applied voltage to measured current is the impedance of the system ðZ ¼ E=IÞ, which is easily calculated over a wide frequency range, thus giving a spectrum where processes with different kinetics that may occur at the system under study are dominant at different frequency regions. In this context, dipolar properties will be reflected at high frequencies, whereas bulk and surface properties will be evident at intermediate and low frequencies, respectively. From a physical point of view, impedance is just a totally complex resistance (measured in Ohms, X) that appears when an AC current flows through a circuit made of resistors, capacitors, inductors or any combination of these. This magnitude shows a complex notation, with a resistive or real part attributable to resistors (in phase with the applied voltage) and a reactive or imaginary part attributable to the contribution of capacitors (out of phase with the applied voltage by þp=2) and/or inductors (out of phase with the applied voltage by p=2): pffiffiffiffiffiffiffi Z ¼ Z 0 þ jZ 00 ¼ R jX ; X ¼ 1=xC; j ¼ 1; where R is the resistance (measured in X), X the reactance, C the capacitance (measured in Farads, F), and x the applied angular frequency (measured in rad/s; x ¼ 2pf , f is the frequency measured in Hz). Experimental impedance data of an electrochemical cell can be easily fitted to the impedance of an equivalent circuit mainly comprising resistors and capacitors. In 38

http://www.elsevier.com/locate/trac

Re Zf

Rct (b)

Zw

- Z’’ Charge transfer-limited process

Diffusion-limited process

ωmax = 1/ Rct Cdl

ω π/4

Z’ Re

Rct Re+Rct-2σ2Cdl

(c)

Zw

log Z

90 Rct

φ Cdl

45 Re 0

log f Figure 1. (a) Randles equivalent circuit, (b) Nyquist Plot, and (c) Bode Plot that describe impedance behaviour of a simple electrochemical cell involving a single Faradaic process. Each dot of the Nyquist Plot represents the impedance at a given frequency.


resistance ðRct Þ and the so-called Warburg Impedance ðZw Þ, which reflects the influence of the mass transport of the electroactive species on the total impedance of the electrochemical cell. Thus, for those diffusion-limited processes, Zw becomes dominant, whereas for those charge-transfer-controlled processes, Zf is just Rct . All these components can be identified by studying the variation in impedance of an electrochemical cell over a wide frequency range (normally 100 kHz–0.1 Hz). A common way of showing the resulting data is the complex plane or Nyquist Plot (Fig. 1(b)), in which the real ðZ 0 Þ versus the imaginary ðZ 00 Þ components of the impedance are plotted. In this plot, two separate processes are very well differentiated, that is, a semicircle relating to a charge-transfer-controlled process, the intercept of which with the X axis gives Re and Rct values, and a straight line with a slope of 1 due to Zw , whose extrapolation to the X axis allows calculation of the Warburg coefficient ðrÞ, from which the diffusion coefficients of the electroactive species can be estimated. From the frequency at the top of the semicircle, where Z 00 is maximum, the time-relaxation constant ðsÞ for the Faradaic process can also be calculated. Another way of presenting impedance data is the Bode Plot, in which the logarithm of the absolute value of Z and the phase ð/Þ are plotted against the logarithm of frequency ðf Þ. Fig. 1(c) shows the Bode Plots for the Randles circuit. Unlike the Nyquist Plots, these data presentations give direct information about f and / that help ascertain the different constituent phases of the system more easily. Thus, in those frequency regions where a resistive behaviour is dominant, a horizontal line is observed for the log Z– log f representation and a / close to 0 is measured. Also, capacitive behaviour within a frequency region is described by a straight line with a slope of )1 in the log Z– log f plots and a / around 90, whereas diffusion-controlled phenomena (Warburg Impedance) would give a straight line with a slope of )1/2 and a / of 45. A detailed mathematical description of all the above parameters can be found elsewhere [3,4]. Impedance behaviour described by Nyquist Plots with more than one semicircle or Bode Plots exhibiting different capacitive/resistive regions may describe electrochemical systems with several phases and/or more complicated Faradaic processes (i.e., coating formation/ deterioration or adsorption phenomena on the transducer as well as coupled chemical reactions). Indeed, EIS has proved to be an excellent technique to verify the properties of polymer coatings as insulating materials and follow their deterioration processes under specific experimental conditions [5]. Fig. 2 shows the impedance spectrum of an insulated polymer-coated electrode in contact with an electrolyte solution. It can be modelled by a resistance, the electrolyte resistance ðRe Þ, in series with a capacitance, the geometric capacitance ðCg Þ. The

Trends

Rb

(a)

Re

Cg

(b) - Z’’

ω

Z’ Re (c)

Rb

Rb

90

φ

log Z Cg

45

Re

0

log f Figure 2. (a) Equivalent circuit that best describes the impedance behaviour of an insulating polymer-coated transducer and, also, of a partially degraded system following a delamination process (dotted-line circuit element describes possible imperfections on the surface of the polymer coating). (b) Nyquist Plot (black dots indicate intact polymer; white dots partially degraded polymer). (c) Bode Plot (solid lines indicate intact polymer; dashed lines partially degraded polymer). No contribution of Zw has been considered.

polymer-coated electrode itself theoretically resembles a pure capacitor, the capacitance of which ðCg Þ can be calculated with the following equation: Cg ¼ ee0 A=d; where e is the dielectric constant of the polymer, e0 is the permittivity of the free space, A is the electrode area and d is the thickness of the polymer layer. In practice, a bulk resistance ðRb Þ in parallel with Cg normally appears and accounts for initial conductive pathways across the polymer ascribed to small defects in the coating. This resistance is very high, so the resulting RC parallel combination is not reflected in the Nyquist Plot as a semicircle but as a bend of the straight line, which is characteristic of pure capacitive behaviour. By contrast, it is easily identified in the Bode Plot by a horizontal line at low frequency values. http://www.elsevier.com/locate/trac

39

Trends

Degradation processes occurring at this polymercoated transducer are easily detected in the impedance spectra, which also allow different processes of deterioration to be differentiated. Delamination of polymer coatings will lead to a decrease of the film thickness ðdÞ and a consequent increase in Cg of the system, together with a decrease in the Rb that is now more easily detected at low frequency values. This is reflected in the Nyquist and Bode Plots, as shown in Fig. 2. The same equivalent circuit shown in Fig. 2 for an intact polymercoated transducer can be used to model the degradation process. By contrast, polymer breakdown due to the formation of pores would give rise to significant differences in the impedance spectra. Nyquist and Bode Plots at different stages of degradation, together with the equivalent circuit that best fits these results, are shown in Fig. 3. Initially, the appearance of pores on the surface of the coating gives rise to a decrease in Rb , which, in this case, is a measurement of the resistance of the electrolyte filling such pores. The in-depth growth of those pores would allow the electrolyte to penetrate the polymer and eventually reach the electrode surface and spread over it. This new electrode–electrolyte interface is easily detected in the impedance spectra by the appearance at low frequencies of a second semicircle (characterised by a second s), which is fitted to a resistor/capacitor parallel combination in the equivalent circuit, namely a chargetransfer resistance ðRct Þ and a double-layer capacitance ðCdl Þ, respectively. For both degradation mechanisms described above, the eventual dissolution of the polymer would lead to impedance spectra similar to those previously described for a bare electrode and fitted by the Randles equivalent circuit. It should be pointed out that the impedance spectra and related equivalent circuits described above do not always fit real situations. In such cases, the so-called distributed circuit elements are used to account for those deviations from ideal behaviour. Those most commonly used are the Warburg Impedance ðZw Þ mentioned above and the constant-phase element (CPE). Zw was first introduced to explain diffusion-limited behaviour of electrode processes at low frequencies by considering just an ideal infinite-length diffusion line of electroactive species from the bulk of the solution to the electrode surface. But an electrochemical cell is always finite in dimensions, so the Warburg Impedance had to be modified to fit more adequate finite-length diffusion processes. The CPE is another generic distributed circuit element, which stems from the fact that the electrode surface is normally rough and/or the bulk properties of the electrode material are inhomogeneous. This also applies to polymer coatings in those real applications described below. CPEs are normally added in equivalent circuits instead of ideal capacitors and are identified in 40



Cg

(a)

Rct Re Rb

Cdl

(b) - Z’’

ω

Z’ Rb

Re

Rct

(c)

90

Rb

φ

log Z Cg

Rct Cdl

45

Rb Cg

Re

0

log f Figure 3. (a) Equivalent circuit that best describes the impedance behaviour of a polymer-coated transducer partially degraded by formation of pores and their corresponding, (b) Nyquist Plot (black dots indicate intact polymer; white dots and black triangles partially degraded polymer), and (c) Bode Plot (solid lines indicate intact polymer; dotted and dashed lines partially degraded polymer). Semicircles 1 and 2 represent the partially degraded coating and exposed transducer interface beneath the coating, respectively. No contribution of Zw has been considered.

impedance spectra due to the appearance of depressed semicircles in Nyquist Plots (with their centre below the real axis) and phase-angle values lower than 90 in Bode Plots. A thorough mathematical and physical description of distributed circuit elements is available [2,4, and references therein].

3. Applications 3.1. Corrosion To date, most of the research carried out on the degradation of polymer coatings using EIS has aimed to

Moisture adsorption and pore formation Rb Microorganisms Stainless steel Polyimide


Rb , Bulk (pore) resistance; Rct , Charge transfer resistance; Cg , Geometric capacitance; Cdl , Double-layer capacitance; fb , Breakpoint frequency. Harrison’s solution: 0.35% (NH4 )2 SO4 + 0.05% NaCl. n.r., Not reported.

[22,23]

[18,19]

Performance of primers, midcoats and topcoats on steel Integrated circuits fb

Poly(3-octyl pyrrole) and poly(3-octadecyl pyrrole) conducting polymers Alkyd, epoxy polyamide, latex, polyurethane

Natural and artificial seawater Steel

Pore formation

[17] Aerospace industry Z

Rb

Rb , Cg

Phosphatised mild steel Detergent solution and high temperatures Aluminium alloy Harrison’s solution, salt fog and UV light Aluminium alloy Harrison’s solution Acrylic electrodeposited coating Sol–gel

Rb , Rct , Cg , Cdl Copper

Al and Mg alloys

Plasma coating from hexamethyldisiloxane Lacquer and epoxy resin

n.r.

[14–16] Aerospace industry

[10] Domestic appliances

[9]

[7,8]

Transport and aerospace industry n.r.

Pitting on defective zones of the coating Water uptake and ionic diffusion within the polymer voids Blister formation in the coating pores and high temperatures n.r. Rct , Cg , Cdl

[6] Architectural coatings Pore formation Rb , Rct , Cg , Cdl Steel Polyester

Artificial UV radiation (270–390 nm) NaCl or Na2 SO4 solution NaCl solution

Application Degradation mechanism Electric parameter Degradation agent Substrate Polymer

Table 1. Main features of several EIS studies on polymer degradation due to corrosion

address the behaviour of different polymers as protective barriers against corrosion. The use of polymers to protect different materials, such as steel and other metals/alloys, against the attack of the environmental conditions to which they are exposed has been very popular, with many different polymers being designed for this purpose. Table 1 summarises several examples of corrosion processes studied by EIS. This reflects the variety of polymers tested against corrosion, the different electric properties calculated from the impedance studies as well as the different range of industrial sectors where the setup of a suitable procedure for monitoring corrosion phenomena has become very important. Polyesters are often used as decorative architectural coatings and, at the same time, as barriers against corrosion. Because of the former role, they have usually been doped with pigments that negatively influence their protective properties. Mild steel panels were electrostatically sprayed with clear and pigmented polyesters and exposed to UV light for several time periods [6]. Impedance measurements were recorded as Nyquist Plots over a wide frequency range. The modelling of the system with an appropriate equivalent circuit allowed calculation of bulk resistance ðRb Þ, charge-transfer resistance ðRct Þ, coating (geometric) capacitance ðCg Þ, and interfacial (double-layer) capacitance ðCdl Þ. From the values obtained, the deterioration of the polymer coatings was assessed and related to an increased level of porosity of the films. Similar studies were carried out on the evaluation of plasma-polymerised coatings using hexamethyldisiloxane monomer on aluminium [7] and magnesium [8] alloys as effective pre-treatments against corrosion. An appropriate equivalent circuit was modelled; it allowed evaluation of Cg and calculation of the water volume within the film at the different stages of the degradation process, thus suggesting initial delamination and blistering phenomena that would subsequently result in a pitting process and failure of the protective coating. Working in a similar way, lacquer- and epoxy-resincoated copper plates were exposed to artificial seawater and their corrosion behaviour tested in terms of a change in Cg , Cdl , the coating (pore) resistance ðRb Þ and Rct [9]. From this data, a three-stage degradation mechanism was suggested, namely water saturation of the free volume of the coating, followed by electrolyte penetration into the metal-coating interface that intensified the corrosion process and caused a final loss of adhesion of the film on the substrate. Cg and Rb changes were also useful in assessing the protective behaviour against detergents of acrylic coatings electrodeposited on phosphatised mild steel used in domestic appliances [10]. Those applications shown above are described by two well-differentiated processes shown in the corresponding Nyquist Plots as two semicircles with different time-

Trends

References


41

Trends

relaxation constants ðs1 =s1 > 20; 0:2 < ðR1 =R2 Þ < 5Þ, which represent the coating and substrate (metal) impedances at high and low frequency ranges, respectively. This fact allowed a suitable description of the polymer deterioration in terms of Cg , Cdl , Rb and Rct . However, it is quite usual to have more complex impedance spectra with superimposed semicircles, which reflect processes with closer s, and/or straight lines (slope ¼ 1) at low values of frequency, which relate to diffusion-limiting phenomena within the system under study (Warburg Impedance). The latter could also be superimposed on other processes appearing in the spectra at low-frequency ranges. All this makes the calculation of the different parameters and the interpretation of those processes cumbersome. Interestingly, this drawback was easily circumvented in the study of the degradation of an alkyd polymer coating under saline and high humidity environments [11]. All Nyquist Plots exhibited an apparent diffusion tail at low frequencies, the slope of which increased as the coating degraded, reaching a final value of 1 when the coating detached from the substrate. Another process appeared to be overlapped at this frequency region and was responsible for this apparent change in the Warburg Impedance, which was easily related to the different stages of the corrosion process. The study of the protective behaviour of an organomineral polymer-based coating used as a topcoat in the automotive industry introduced another useful correlation between the so-called breakpoint frequency (fb , frequency at which / first falls to 45 when scanning the spectrum from high to low frequencies) and the exposure time of the coated substrate to a corrosive environment [12]. This method enabled the quick detection of any deterioration of the coating. Unlike the examples described above, where an evaluation of a specific coating was carried out to some extent, some other reported studies made use of impedance measurements just to compare the feasibility of applying different coating materials and procedures to a specific technology [13]. In this context, sol-gel based coating systems [14–16] and electroactiveconducting polymer coating systems [17] were envisaged as excellent alternatives to hazardous chromate conversion coatings currently used in the aircraft industry as surface treatments (primers) to prevent corrosion on aluminium alloys just by recording and comparing Nyquist and Bode Plots of the different systems over long periods of time. Following similar studies, several polymer-coating systems on steel were tested upon exposure to natural and artificial seawater [18,19]. It is remarkable that these tests were carried out for more than two years at remote test places using an experimental set-up that allowed the collection of impedance spectra once a week via modem. Coatings included primer, midcoat 42



and topcoat layers of different materials (e.g., alkyd solid, epoxy polyamide, latex and polyurethane). A large amount of impedance data was collected and interpretation of corrosion phenomena was carried out in the first place by simply comparing the shape of Bode Plots. A general decrease in impedance was an indication of larger delaminated areas within the substrate due to corrosion. Further evaluation involved estimation of pore resistance, where the phase-angle plot showed a minimum ðRbo Þ, and the breakpoint frequency ðfb Þ, from which the delaminated or corroded area was calculated. Rbo decreased steadily with exposure time up to a certain time, at which it increased because pores were plugged by corrosion products and also bacteria. In an another study from the same authors, impedance spectra in the form of Bode Plots were collected to help assess the growth of bacteria colonies on corrosion products developed on polymer-coated steel when exposed to marine water, independent of the coating formulations [20]. Other related works have provided evidence of the active role that colonies of microorganisms on the surface of coated materials play in degradation [21]. The biosusceptibility of polyimide coatings was assessed by measuring the decrease in Rb with time observed in changes in the shape of Bode Plots at lowfrequency ranges [22,23]. Fungi caused an enhancement of the transport of water and ionic species into the polymer matrix, thus changing the dielectric properties of the material and accelerating corrosion phenomena of the coated substrates. This was very important in the electronic field, where polyimides have been widely used as insulating layers of integrated circuits. Similar investigations were carried out with a wide range of materials with potential uses in the aerospace industry [24]. As described above, EIS has provided useful information about corrosion phenomena occurring evenly on a coated substrate. However, it is well documented that corrosion is a localised process that tends to start at areas of lowest ionic resistance, then spread very quickly along the material surface. Local EIS (LEIS) is a recently developed technique that has the potential to measure localised degradation processes [25]. The design of displaced micro-probes containing reference and counter electrodes in a variety of arrangements has proved very useful for recording impedance measurements locally and creating three-dimensional surface-impedance maps at a single frequency, which allow location of those defects responsible for corrosion on the coated substrate. From all the works referred to above, it is evident that EIS plays an important role in evaluating and monitoring the protective performance of polymer coatings against corrosion. Such is the significance that it has led to the creation of two artificial neural networks for the

Z , Impedance; Rb , Bulk (pore) resistance; Rct , Charge transfer resistance; Cg , Geometric capacitance; Cdl , Double-layer capacitance; IDEs, Interdigitated electrodes; hIgG, Human immunglobulin G; PSA, Prostate specific antigen. n.r., Not reported.

a-Chymotrypsin activity [31,34] Collagenase activity [35,36] Z Z a-Chymotrypsin enzyme Collagenase enzyme

Delamination Delamination

[33] n.r. Z

Oxidase activity

[30] Swelling

Methyl vinyl ether-maleic anhydride copolymer (EMAC)/carbon IDEs Glycidylmethacrylate methylmethacrylate copolymer/Gold disc Poly(ester amide)/Gold Gelatin/Gold IDEs

Rb , Cg

Methyl methacrylate-methacrylic acid copolymer (Eudragit S100)/Gold Eudragit S100/carbon IDEs

Rb , Cg

Cg

Urea, creatinine

[32,41] PSA Pore formation

hIgG, urea

Inorganic salts

Overoxidation and nucleophilic substitution Pore formation Polythiophene/platinum wire

High positive potentials (>1.2 V vs. Ag/AgClsat ) Urease-induced increase of pH (pH > 7) Urease-induced increase of pH (pH > 7) Urease-induced increase of pH (pH > 7.6) H2 O2

Rb , Rct , Cg , Cdl

Analyte/s Degradation mechanism Electric parameter Degradation agent Polymer/transducer

Table 2. Main features of EIS studies on polymer degradation applied to sensor development

3.2. Sensors Impedance measurements of degradation processes occurring at polymer-coated transducers have recently emerged as a means of detecting biological/chemical interactions directly or indirectly responsible for the degradation of those coatings; thereby they have been exploited for analytical purposes. Table 2 gathers several representative examples of these novel applications. A few reports have dealt with a thorough description of the impedance of polymers deposited on electrode surfaces and their mechanism of degradation, the ultimate objective being to use these polymers in sensor development. A good example was the study of different phenomena occurring at polythiophene conducting films, doped with CF3 SO3 , on platinum electrodes by EIS [27]. This polymer was proposed as a good candidate for the development of electrochemical sensors for inorganic ions. The redox exchange behaviour of polythiophene films was explained by studying the variation of Rb , Rct , Cg and Cdl with applied polarisation potential and film thickness in aqueous solutions containing the doping ion. As an example a significant increase of Rb at high positive polarisation potentials accounted for a deactivation of the polymer by overoxidative degradation and a total loss of its redox exchange properties. The thorough characterisation of two different polyacids, which underwent pH-dependent degradation processes, gave insight into their potential use for the development of biosensor devices involving enzymerelated systems, whose corresponding reactions altered the pH of the reaction media. Methyl methacrylate/ methacrylic acid copolymer (Eudragit S100, R€ ohm Pharma) and methyl vinyl ether/maleic anhydride copolymer modified by esterification with 1-octanol (EMAC) degraded at pH above a threshold value because of ionization of the carboxylic side groups within their structure. The former broke down at pH > 7, while the latter remained stable up to pH 7.6. Impedance studies with gold screen-printed electrodes either spray or spin coated with Eudragit S100 [28,29], and with gold or carbon screen-printed electrodes coated with EMAC by either spin or screen-printing processes [29,30], were carried out at different degradation stages by halting the degradation process in buffered saline solutions of pHs below their degradation-pH threshold value. Impedance spectra in a wide frequency range were shown in the form of Bode Plots. Both polymers exhibited capacitive behaviour over a large frequency region, which is an indication of their good insulating properties. However, their degradation behaviour was completely different. Bode Plots of Eudragit S100-coated gold electrodes (Fig. 4) and the calculation of derived electric parameters,

References

automatic classification of polymer-coating quality on steel substrates based on phase angle-log f and log Z- log f data (Bode Plots) [26].

[28,29]

Trends

[27]



43

Trends

Figure 4. Bode Plots of (a) intact and (b)–(e) partially degraded Eudragit S100 polymer-coated transducers, and (f) bare transducer. Spectra from (b) to (e) were obtained as the polymer film was more extensively degraded. Degradation processes were carried out at pH 7.8 and halted by rinsing the transducer thoroughly. Measurements were then recorded at pH 5.2. Pore formation can be inferred from the changes in the impedance spectra. (Reprinted with permission from [28]. 1995 American Chemical Society.)

such as Cg , Cdl , Rb and Rct , accounted for the formation of pores on the surface, which, at a certain stage, reached the electrode surface and allowed the electrolyte to penetrate and spread over it [28,29]. From these


values, the relative porosity and the fraction of wetted electrode area were estimated. The eventual dissolution of the polymer layer was observed. More recently, combined quartz crystal microbalance and electrochemical impedance studies on Eudragit S100 spin-coated gold electrodes were in agreement with this breakdown mechanism [31]. Studies carried out with EMAC-coated electrodes showed that this polymer did not dissolve but swelled, remaining as a film on the electrode surface [29,30]. Swelling is due to the incorporation of water and electrolyte cations that compensate the negative charges created by deprotonation of the carboxylic moieties within the polymer. The percentage of swelling of the polymer could be estimated by comparing the Rb of EMAC films before and after different breakdown stages. The catalytic hydrolysis of urea by the enzyme urease gives carbon dioxide and ammonia, thus causing a pH increase in the reaction medium. This enzymatic reaction was easily coupled to the breakdown process taking place either on Eudragit S100- or EMAC-coated electrodes at alkaline pH for the development of electrochemical single-use biosensors for urea and human immunoglobulin G (hIgG) [28], and urea and creatinine [30], respectively. Disposable Eudragit S100 spray-coated gold screenprinted electrodes were put in close contact with activated cellulose membranes previously loaded with urease and inserted into a Perspex well template [28]. The well was filled with a certain volume of an electrolyte solution at pH 6.5 containing a specific urea concentration. All measurements were carried out in a two-electrode arrangement using a glassy carbon counter electrode (Fig. 5). Impedance measurements at a

Figure 5. Sensor arrangement for the determination of urea and human immunoglobulin G (IgG) based on the capacitance measurement of the degradation of Eudragit S100 polymer-coated electrode.

44



single frequency were recorded. More precisely, the increase of the electrode capacitance (extracted from the imaginary component of the impedance) with time was monitored at a set frequency of 20 kHz, where the electrode behaviour was purely capacitive. Using Eudragit S100 polymer layers of 1 lm thickness, their degradation yielded a four orders of magnitude change in capacitance, which proved the excellent sensitivity of the device. Calibration curves for urea in a simulated serum matrix over a concentration range of 1–100 mM were obtained when plotting the ratio of the capacitance value recorded at a specific time to the initial value of capacitance for each urea concentration. The same set-up was used for the determination of hIgG. Ureaselabelled anti-hIgG was used as a tracer for the development of the analytical signal. Competitive and non-competitive immunoassays were carried out with activated cellulose membranes loaded with hIgG and anti-hIgG, respectively. Following the several steps of the corresponding immunoassay, the membranes were placed over the electrode surface, inserted into the Perspex well and measurements recorded in 100-mM urea solutions in the same fashion as above. HIgG was measured over a concentration range 0.0001–100 lg/ ml using both assay formats. Because of the low pH-breakdown threshold of Eudragit S100 (pH > 7.0), measurements of urea with the urease device in whole blood, serum or plasma at physiological pH cannot be carried out. When working with immunoassays using urease as a label, all the required steps have to be performed on the cellulose membrane before placing it over the coated electrode, which makes each experiment more tedious and the

Trends

resulting device less appealing for commercial purposes. However, the same research group has successfully circumvented the above drawback by coupling a similar sensitive impedimetric transducer with lateral flow immunoassay formats developed on membrane strips (Fig. 6) and demonstrated the approach for the development of a fast, easy-to-use, single-use immunosensor device for prostate specific antigen (PSA), the most reliable tumour marker for the early diagnosis of prostate cancer [32]. The non-specific degradation of the polymer-coated transducer due to the natural pH of biological samples was easily avoided by keeping samples in contact with the transducer for just a few seconds. The anti-PSA antibody capture zone of a nitrocellulose membrane strip was put in close contact with the transducer, as depicted in Fig. 6. A certain volume of a sample containing PSA was mixed with anti-PSA urease conjugate and allowed to flow along the strip, then reacting on the antibody capture zone and eventually reaching the sink pad. The further rapid addition of a urea solution on the opposite edge of the strip, required for the detection of the immunocomplex formed on the antibody capture zone, simultaneously washed away the unbound species and the sample matrix from the strip, thus preventing the non-specific breakdown of the Eudragit S100 polymer layer. The overall process took less than 2 min. Indeed, the sample reached the sink pad in around 45 s and the washing step took just 30 s. Capacitance measurements were started immediately after adding the urea solution under the same experimental conditions described for the capacitive sensor approaches above. Fig. 7 provides a fine example of those measurements recorded for increasing concentrations of

Figure 6. Construction of an impedimetric sensor system for PSA based on a lateral-flow immunostrip and Eudragit S100 polymer-coated carbon interdigitated electrodes (IDEs).


45

Trends


900

[PSA]

ngml-1

0 800

2 10

700

30

C / C0

600 500 400 300 200 100 0 0

200

400

600

800

1000

Time / s

Figure 7. Capacitance responses of the sensor depicted in Fig. 6 for several concentrations of PSA. C =C0 is the ratio of the recorded capacitance (C) to initial capacitance ðC0 Þ. This ratio helps correct the minimal variation of C0 of intact Eudragit S100 polymer-coated transducers.

PSA, including concentration levels required for the early detection of prostate tumours. The higher pH at which EMAC polymer degrades made this polymer a good feasible alternative to Eudragit S100 for the construction of sensors for clinical applications [30]. In order to address its performance, single-use urea and creatinine sensors based on the use of EMAC-coated-carbon screen-printed interdigitated electrodes (IDEs) were developed. Pads of a wicking material were loaded with either a urease solution or a creatinine deiminase solution and stuck over the working area of the coated IDE system

(Fig. 8). Thereafter, an adequate volume of the corresponding unaltered sample containing analyte was added to the pad. Simultaneously, impedance measurements at a single frequency were recorded following the same experimental procedure as in the devices described previously. Urea calibration curves were obtained over a concentration range of 5–30 mM. In addition, this analyte was successfully determined in normal serum. However, this device was 5 times less sensitive than that developed using Eudragit S100-coated transducers. Moreover, the calibration range was narrower. Creatinine was measured at concentration levels as low as 2 mM. Nevertheless, this device was not sensitive enough for clinical analysis. The degradation of acrylate bulk polymers by the action of the hydrogen-peroxide product of an oxidase enzyme has also been proposed as a feasible transduction method for the development of biosensors based on impedance measurements [33]. However, to our knowledge, no true applications have yet been reported. Some biodegradable polymers break down by the direct action of enzymes (i.e., they act as the enzyme substrate) and constitute another excellent approach for the development of impedimetric biosensors. In this context, a poly(ester amide) polymer, which is degraded by the proteolytic enzyme a-chymotrypsin, was spun onto a gold electrode and used as an impedimetric transducer for the detection of the enzyme activity in standard solutions [31,34]. Impedance spectra were recorded over a frequency range 5 Hz–100 kHz every 30 s during the course of degradation, using a network analyser that allowed fast collection of data. Results were presented as log Z- log f Bode Plots (Fig. 9). Resistive behaviour was observed at high frequencies (electrolyte resistance, Re ), while, at lower frequencies, the increase of impedance with decreasing frequency indicated capacitive behaviour (geometric capacitance, Cg ). Degradation of the film by the enzyme led to a

Figure 8. Sensor arrangement for the determination of urea and creatinine based on capacitance measurement of EMAC polymer-coated carbon interdigitated electrodes (IDEs).

46



Trends

Although not very suitable for the design of biosensors to be used in electrolyte solutions, the same approach was proposed for the detection of collagenase in air with more or less success [36].

4. Summary and future trends

Figure 9. Impedance spectra (Bode Plots) of poly(ester amide) films recorded during its degradation by the enzyme a-chymotrypsin (electrolyte solution pH 7.3). A delamination process is easily inferred from the shifts of the overall film impedance with time. (Reprinted with permission from [31]. 2002 American Chemical Society.)

decrease of the impedance in the low-frequency range, corresponding to an increase in the geometric capacitance, which was related to the thinning of the polymer film. When the polymer film was dissolved completely, the observed capacitance in the spectrum equalled the double-layer capacitance of the bare electrode. In this way, a-chymotrypsin concentrations in the nanomolar range were easily detected. Similar experiments were carried out with a dextran hydrogel polymer, which was degraded by the enzyme dextranase [31,34]. However, the biodegradation process that this polymer underwent could not be monitored by EIS. Unlike the good insulating properties of poly(ester amide) films, intact dextran hydrogels were easily filled up with electrolyte molecules that wetted the electrode surface. Thus, the geometric capacitance of the system equalled the double-layer capacitance before the degradation took place, which led to the recording of the same impedance spectra before and after its enzymatic dissolution. The need to use electrically insulating polymers susceptible to undergoing degradation processes narrowed the range of materials used in these approaches. Indeed, it has constituted one of the major drawbacks of application of this methodology to biosensor development. The same problem was encountered in the development of a biosensor for collagenase based on impedance measurements of the enzymatic digestion of gelatin deposited on interdigitated gold electrodes [35]. Collagenase could be detected only in solutions at very low electrolyte concentrations (