Redox competition and generation-collection modes based scanning electrochemical microscopy for the evaluation of immobilised glucose oxidase-catalysed reactions

Abstract

Redox competition (RC-SECM) and generation-collection (GC-SECM) modes of scanning electrochemical microscopy were applied for the evaluation of a glucose oxidase (GOx)-modified non-conducting poly(methyl methacrylate) surface. The current vs. distance curves in RC-SECM mode were registered at −600 mV vs. Ag/AgCl to determinate local O₂ concentration, taking into account that the O₂ is consumed in the GOx-catalysed enzymatic reaction. This measurement was performed in phosphate–acetate buffer at pH 6.6 with 0–30 mmol L⁻¹ of glucose using a platinum ultramicroelectrode (UME) as the moving working electrode in a three-electrode electrochemical cell. The UME current, which is related to oxygen reduction rate, decreased when glucose was added to the solution. Another part of the investigation was performed in GC-SECM mode at +600 mV vs. Ag/AgCl in order to measure local H₂O₂ concentration, which is formed during the GOx-catalysed enzymatic reaction. The same SECM mode was used for imaging the GOx-catalysed reaction without any redox mediator. The imaging distance was chosen based on both the RC-SECM and GC-SECM experimental results. The RC-SECM and GC-SECM modes are described, and the processes that occurred on the UME- and GOx-modified surfaces are revealed.

---

Introduction

Biosensors are systems composed of enzyme(s) or other materials of biological origin that are immobilised on surfaces; the biological part is responsible for the selective reaction with an analyte and the generation of electrical signal. However, certain technical problems are encountered during the development of biosensors that should be solved in order to increase the applicability of these analytical devices. Among these problems, increasing the efficiency, stability, reliability and analytical applicability of biosensors are the most important tasks. Therefore, the localised evaluation of the bio-electrocatalytic activity of redox enzymes immobilised on the surface could be very attractive for biosensor design^1,2 or the development of biofuel cells.³ Scanning electrochemical microscopy (SECM) is an innovative method that could be applied for the surface-activity analysis of enzymatic biosensors.^4–6 Initially, SECM was designed as a method for the investigation of electrochemically active surfaces.^7,8 The most important part of SECM is an ultra-microelectrode (UME) with a radius ranging from a few nm to 25 μm.⁹ The UME is usually moved by positioners in three directions, x, y, and z, in the solution close to the surface of interest. The UME is mostly switched as a working electrode in the electrochemical system consisting of two, three or four electrodes.⁹ One of the most informative SECM modes is based on the vertical movement of the UME vs. the sample because it allows changes in current to be registered vs. distance over the sample. From the curves measured in feedback (FB) mode, the distance of the UME from the sample surface can be determined, the electrochemically active surfaces can be evaluated, and reaction kinetics can be calculated.^10–14 Current flow in FB mode is caused by oxidation/reduction reactions occurring at the UME. The feedback mode can be positive or negative, depending on the changes in current when the UME is approaching the sample surface. In negative feedback mode, the current signal is decreasing due to the blocked diffusion of redox compounds. In positive feedback, the current signal is increasing because redox compounds are formed and/or regenerated at the sample surface.⁹ In the present study, the evaluation of an unmediated glucose oxidase (GOx)-catalysed reaction by SECM shows that the current vs. distance dependence, which is based on the O₂ reduction current, exhibits negative feedback behaviour. When the O₂ is consumed on the surface, feedback can be consumptive.^15–18 For this phenomenon, the redox competition mode (RC-SECM), in which the UME and the sample compete for the same analyte in solution, was suggested by Schuhmann's group.¹⁹ During the experiment, oxygen reduction current mostly remains constant unless the UME is approaching the oxygen consumption area. This effect could be measured at a bi-potentiostatic mode, where both the UME and surface are held at the oxygen reduction potential.^18,20 Schuhmann's group has reported several works on the application of redox competition modes of SECM (RC-SECM), particularly: (i) the RC-SECM mode was used to characterise the performances of a biosensor employing the local electrocatalytic activity of GOx immobilised within a polymer hydrogel matrix on the top of Prussian blue-modified glassy carbon electrodes to which a particular potential was applied;²⁰ and (ii) the local bio-electrocatalytic activity in RC-SECM mode was evaluated when GOx was immobilised on a biofuel cell cathode and a particular potential was again applied to the electrode.³

Another important regime of SECM is generation-collection (GC-SECM) mode. In this mode, the UME only registers currents caused by the reaction products.^17,21,22 Usually, the UME passively detects the redox compounds, which are generated at the surface. The problem is that the reaction on the sample occurs continuously, independently of the operation of the UME. However, after some adaptations, it is possible to measure the concentrations of reaction products in real time.^23–25 In our study, GC-SECM-based measurements were performed by registering the H₂O₂ oxidation current, where the H₂O₂ is the product of glucose oxidase (GOx)-catalysed reactions.

Both RC-SECM and GC-SECM experiments can be carried out in constant height and constant distance modes. In constant height mode, the UME is moved only laterally in the x and y directions, while in constant distance mode, the UME can be moved in x, y and z directions.²⁶ Constant height mode is appropriate for the evaluation of smooth surfaces (roughness smaller than the UME radius).²⁷ In this mode, the UME current depends on the distance between the UME and the surface of interest and on the reactivity of compounds immobilised on the surface. Resolution studies of SECM in constant height mode show a quantitative correlation by which resolution decreases as the distance between the UME and the sample increases.²⁸ To determine the most suitable distance for the appropriate resolution of SECM constant height mode measurements, the current vs. distance dependence can be measured in feedback mode by approaching the UME to the surface of interest; the distance between the UME and the sample can be calculated from SECM theory, where i_T/i_T,∞ (the ratio of the UME current and steady-state current far from an electrochemically active surface) can be related to d/a (the ratio of the distance between the sample and the UME and the UME radius).²⁶

The main aim of this work was to identify the appropriate glucose concentration and UME-surface distance for the SECM-based imaging of GOx-modified surfaces. The RC-SECM mode was used to determine the UME-surface distance for the evaluation of a GOx-catalysed reaction. The O₂ reduction current was registered, and a decrease in current as the UME approached the GOx-modified surface was observed. This current decrease indicates oxygen consumption by GOx. The generation-collection mode was used to image the GOx-catalysed reaction at a chosen distance from the surface of interest. The most suitable UM-surface distance for imaging was chosen after evaluating the measurement results of both RC-SECM and GC-SECM modes. It should be noted that there are some principal differences between this work and previous reports^18–20 published by Schuhmann's group: (i) we have used the redox competition mode for the evaluation of enzymatic reactions while GOx was immobilised on an insulating surface, and no external potential was applied; and (ii) in studies conducted by Schuhmann's group, a redox mediator was used, while we have performed the SECM measurements without any redox mediator.

Experimental

Materials

Glucose oxidase (EC 1.1.3.4, type VII, from Aspergillus niger, 215.3 units per mg per protein) and 25% glutaraldehyde solution were purchased from Fluka Chemie GmbH (Buchs, Switzerland). D-(+)-Glucose was obtained from Carl Roth GmbH & Co. (Karlsruhe, Germany). Before investigations, glucose solutions were allowed to mutarotate overnight. All solutions were prepared using distilled water. Sodium acetate trihydrate, potassium chloride, monopotassium phosphate, and sodium dibasic phosphate were obtained from Reanal (Budapest, Hungary) and Lachema (Neratovice, Czech Republic).

Immobilisation of glucose oxidase

A cylindrical poly(methyl methacrylate) (plastic) cell surface was kept in a closed vessel over a 25% solution of glutaraldehyde for 10 min. Subsequently, 1.6 μL of 10 mg mL⁻¹ GOx solution was dropped onto the surface, covering a surface area of 1.13 mm². The surface was then dried at room temperature to obtain a 14 μg mm⁻² GOx layer. The modified surface was then kept in a closed vessel over a 25% solution of glutaraldehyde for 10 min at room temperature followed by washing with buffer.

Measurements by SECM

SECM and a disk-shaped Pt UME from Sensolytics (Bochum, Germany) were used for experiments. The platinum wire (diameter 10 μm, purity 99.99%) was sealed in borosilicate glass. SECM measurements were performed in both RC-SECM and GC-SECM modes in buffer without mediator. A three-electrode electrochemical cell was applied with a UME-based scanning probe that was switched into a three-electrode circuit as a working electrode. A Pt electrode was used as a counter electrode, and Ag/AgCl in 3 M KCl was applied as a reference electrode. Current vs. distance dependences in RC-SECM mode were registered while applying a potential of −600 mV vs. Ag/AgCl. First, the UME was moved with a speed of 1 μm s⁻¹ in the vertical direction until it touched the unmodified plastic surface. From this measurement, the distance was calculated using eqn (1). In this case, negative FB was observed due to hindered diffusion. Second, the UME was retracted to a distance 200 μm from the surface of interest and positioned in another place that was modified by GOx; the UME then approached the GOx-modified surface. These measurements were performed in phosphate–acetate buffer at pH 6.6 with glucose concentrations ranging from 0 to 30 mmol L⁻¹. Each measurement was repeated three times, and the mean value was used for further calculations. Current vs. distance dependences in GC-SECM mode were registered while applying a potential of +600 mV vs. Ag/AgCl by approaching the GOx-modified surface from 1 mm to the calculated ‘zero’ distance. The imaging of the GOx-modified surface in GC-SECM mode was performed at +600 mV vs. Ag/AgCl at a distance of 40 μm from the surface.

Results and discussion

Registration of ‘approaching’ curves in RC-SECM mode

The SECM-based measurements could be divided into two principally different modes based on positive or negative feedbacks.^9,26 The distance between the UME and the surface of interest can be determined by recording approach curves.⁹ In these kinds of experiments, measurement results can be plotted as the dependence of normalised current i_T/i_T,∞ on normalised distance d/a, where: i_T – measured current while UME is approaching the sample; d – distance; i_T,∞ – the steady-state current when the UME is placed very far from the surface; and a – UME radius. Steady state current is expressed as i_T,∞ = 4nFDCa, where n is the number of electrons transferred per molecule, F is the Faraday constant, and D and C are the diffusion coefficient and the initial concentration of the measured substance (e.g., oxygen), respectively.

In negative FB mode, the current decreases when the UME is approaching the surface of interest. When the same analyte is consumed on both the UME and on the sample surface, the process is called RC-SECM mode.²⁰ This mode can be applied for the evaluation of the local bio-electrocatalytic activity of enzymes; for example, an enzyme immobilised on a biofuel cell cathode has been reported by Schuhmann's group.³ Moreover, the characterisation of the performances of a biosensor employing the local electro-catalytic activity of GOx immobilised within a polymer hydrogel matrix on the top of glassy carbon electrodes has been successfully performed in a similar manner.²⁰ In both of these cases, the enzyme was immobilised on a conducting surface, and SECM investigations was simultaneously conducted at a selected potential. In contrast, GOx was immobilised on an insulating surface in this study, and therefore no potential was applied to the GOx-modified surface. Despite this, the competition of two processes (O₂ consumption on the GOx-modified surface with 2-electron transfer and O₂ consumption on the UME with 4-electron transfer) during the SECM measurements could be described as RC-SECM mode. The processes that occur on the UME and GOx-modified surface in RC-SECM mode when negative potential is applied to the UME are revealed in Fig. 1. In the solution without any redox mediator, the reduction of dissolved O₂ occurs on the UME, and the O₂ is also consumed by the GOx-catalysed reaction. Therefore, the O₂ reduction-based UME current decreases when UME approaches the surface. However, in this case, another factor such as the blocked diffusion of O₂ to UME also has a significant influence on the measurement of current vs. distance.

Fig. 1 Schematic of the processes occurring during SECM measurements on both GOx-modified and UME surfaces in RC-SECM mode without any redox mediator. Gluconolactone and glucose are abbreviated as GLL and GLC, respectively.

In order to determine the distance of UME from the surface, the O₂ reduction current is usually measured while the electrode approaches the insulating surface.²⁶ The current vs. distance dependence was registered in the buffer while a potential of −600 mV vs. Ag/AgCl was applied and the UME approached the unmodified plastic surface (Fig. 2, buffer); during this process, the distance was calculated by eqn (1). Further measurements were performed in the same fixed x–y position while adding glucose to the solution; therefore, the SECM measurement results were mostly affected by two factors: (i) the hindered diffusion when the UME is close to the surface of interest; and (ii) the consumption of O₂ by GOx-catalysed reactions.

Fig. 2 The dependence of normalised current on normalised distance at different glucose concentrations in buffer at a UME potential of −600 mV vs. Ag/AgCl.

Current vs. distance dependences were registered in RC-SECM mode at different glucose concentrations in order to determine which factor has a more significant influence on the current signal. Fig. 2 shows the O₂ reduction current dependence at initial glucose concentration. Approximately 250 μmol L⁻¹ of O₂ is initially present in the solution, which is exposed to air; this dissolved O₂ is responsible for the generation of the UME background current in RC-SECM mode. In order to avoid current shielding effects, we compared currents that were normalised by applying eqn (1). Since the O₂ is consumed in the enzymatic reaction, the addition of glucose to the solution facilitates the enzymatic reaction. The consumption of O₂ is registered when glucose is added to the solution; current decreases faster compared to measurements in the absence of glucose. If the decrease in current was primarily related to hindered diffusion, the current vs. distance dependence should be the same. However, the results show that the layer of consumed O₂ increases with consecutive additions of glucose to the solution. Hence, the most significant influence on the change in current is O₂ concentration rather than blocked diffusion. This notion is also evidenced by the fact that the current remains at the same level (i_T/i_T,∞ = 0.1) when glucose concentration ranges from 10 mmol L⁻¹ to 30 mmol L⁻¹ and UME is close (from 0 to 4 d/a) to the surface of interest. In this case, the layer, which contains a lower concentration of O₂, is thicker due to the much faster O₂ consumption. The dependence of O₂ consumption on glucose concentration at different distances from the GOx-modified surface is linear (Fig. 3). At shorter distances, the current decreases by 25–100%, which is clear evidence of O₂ consumption.

Fig. 3 The dependence of normalised current on glucose concentration in buffer without any redox mediator at a UME potential of −600 mV vs. Ag/AgCl.

GC-SECM mode-based measurements

The H₂O₂ oxidation current on UME was registered in GC mode (Fig. 4). The highest concentration of H₂O₂, which is formed during GOx-catalysed reactions, is observed close to the GOx-modified surface. Therefore, the UME current increases significantly as it approaches the GOx-modified surface, and this increase is related to the rate of the enzymatic reaction.

Fig. 4 Schematic of the SECM processes occurring on GOx-modified and UME surfaces in GC mode without any redox mediator. Gluconolactone and glucose are abbreviated as GLL and GLC, respectively.

The H₂O₂ concentration profile (Fig. 5) was determined by registering the current vs. distance dependence. However, the estimation of the registered current vs. distance dependence in GC mode has some disadvantages: (i) the dependence of current vs. distance changes over time because the enzyme is continuously consuming both substrates (glucose and O₂), and the concentrations of products (H₂O₂ and gluconolactone) in solution increase over the course of the reaction; (ii) the current increases as the UME approaches the GOx-modified surface modified within a certain distance range, in which the hindered diffusion effect still does not occur; and (iii) it is not possible to estimate the exact surface-UME distance from the current vs. distance dependence curves, therefore, the UME could crash, or the sample could be damaged by the UME as it approaches the surface. To avoid these negative effects, the measurement was performed immediately after the addition of glucose, and the measurement distance was chosen from the negative feedback dependence of current vs. distance measured while approaching the plastic surface at −600 mV (Fig. 2, in buffer). The measurement of GC current vs. distance dependence was started at a UME-surface distance of 1 mm. It was determined that the current in GC mode decreases more slowly when the UME approaches the surface of interest compared to that in RC-SECM mode. This phenomenon can be explained as follows: H₂O₂ diffusion from the GOx-modified surface is fast, and the increase in current compared to the measurement without any glucose can therefore be observed, even at a distance of 1 mm. However, the concentration of H₂O₂ is highest at the closest point, which can be related to the continuously proceeding enzymatic reaction that produces the H₂O₂. Here, the effect of hindered diffusion is not observed because the measurement distance was carefully calculated from the negative feedback measurement to avoid sample damage. Thus, both modes are important for the determination of the most suitable distance for imaging.

Fig. 5 Current vs. distance curves registered while approaching the GOx-modified surface in the presence and absence of glucose. The UME potential was +600 mV vs. Ag/AgCl.

A distance of 40 μm was chosen from both approaching curves (Fig. 2, in buffer and Fig. 5) for the following reasons: (i) the current related to O₂ reduction at the UME is 0.85 of the normalised steady-state value (applied potential was −600 mV vs. Ag/AgCl); and (ii) the current related to H₂O₂ concentration is at maximal value (applied potential was +600 mV vs. Ag/AgCl). Similar results were obtained inside a biosensor based on GOx;²⁹ the authors found that the distance of maximal current while measuring H₂O₂ concentration by GC-SECM mode and for the change in maximal current while measuring O₂ concentration by RC-SECM mode is the same.

These conditions allow scanning at a distance at which hindered diffusion does not occur; the current related to O₂ reduction at the UME (Fig. 2, in buffer) at this distance does not differ by more than 20% from the steady-state value. On the other hand, the H₂O₂ concentration is at the maximum level (Fig. 5); therefore, it allows the measurements to be performed at highest resolution.

Horizontal scanning performed at a constant distance of 40 μm (d/a = 8) from the GOx-modified surface illustrates that the UME current is low (0–3 nA) in the absence of glucose and increases to 24 nA in the presence of 10 mmol L⁻¹ of glucose (Fig. 6). The imaging was performed immediately after the addition of glucose; however, the diffusion of H₂O₂ from the GOx-modified surface is very fast, and the current after the addition of glucose increases not only in close proximity to the GOx-modified surface (24 nA, x-coordinate from 100 to 200 μm), but also in the surrounding area (9 nA, x-coordinate from 0 to 100 μm). At the same time, in the current vs. distance curve (Fig. 5), the current changes from 8 nA when UME is far (600–1000 μm) from the GOx-modified surface to 20 nA when it is close (20 μm) to the GOx-modified surface. The trends in the current changes of the approaching curve (Fig. 2) and the 3D images of the UME current registered at the interface between the GOx-modified and unmodified surfaces (Fig. 5) are very similar, indicating that an appropriate distance was chosen for imaging.

Fig. 6 UME current registered at the interface between the GOx-modified and unmodified surfaces in the presence and absence of glucose in buffer; the UME was operating in GC mode at a potential of +600 mV vs. Ag/AgCl at a distance of 40 μm (d/a = 8).

Conclusions

SECM is a powerful tool for the investigation of enzyme-catalysed reactions. By comparing two different SECM operation modes (RC-SECM and GC-SECM), we have determined that both modes are suitable for the investigation of GOx-modified surfaces in the absence of redox mediators. In RC-SECM mode, a negative potential was applied for the determination of O₂ consumption on a GOx-modified surface, and the most robust O₂ consumption was found at the highest glucose concentration. Moreover, the current vs. distance curves show that the thickness of the layer decreased with O₂ concentration approximately 20 μm from the GOx-modified surface. From the same experiment, the dependence of current vs. glucose concentration was calculated, and the highest current change, which indicates the highest O₂ consumption rate, was found close (at 20 μm) to the GOx-modified surface.

Another part of the investigation was performed in GC mode at positive UME potential. The current vs. distance curve illustrates that the H₂O₂ concentration differs significantly at 0 to a distance of 600 μm from the GOx-modified surface, while the O₂ concentration varies significantly at 0 to 100 μm (20 d/a) from the surface. This indicates that H₂O₂ diffuses very far from the surface after the enzymatic reaction, while O₂ is consumed close to the GOx-modified surface. In contrast, the concentration of H₂O₂ is highest close to the surface, where the enzymatic reaction takes place. If a horizontal SECM scan is performed, the choice of an appropriate distance is a very important factor because the measurement results can be distorted due to hindered diffusion or fast O₂ consumption at closer distances. The selection of a suitable distance for the horizontal scan should take into account the concentrations of both compounds (consumed O₂ and formed H₂O₂), which appear close to the GOx-modified surface. A UME-surface distance of 40 μm was found to be optimal for horizontal scanning while taking both phenomena into account. The GC mode-based horizontal scan measurements show the most significant increase in UME current (from 0 to 24 nA) when glucose is added to the solution. This can be explained by the formation of H₂O₂ during the enzymatic reaction.

Acknowledgements

The work was supported by the Research Council of Lithuania, Support to research of scientists and other researchers (Global Grant), Enzymes functionalised by polymers and biorecognition unit for selective treatment of target cells (NanoZim's), Project Nr. VP1-3.1-ŠMM-07-K-02-042.

References

1. I. Lapenaite, A. Ramanaviciene and A. Ramanavicius, Crit. Rev. Anal. Chem., 2006, 36, 13–25.
2. M. Maciejewska, D. Schafer and W. Schuhmann, Electrochem. Commun., 2006, 8, 1119–1124.
3. K. Karnicka, K. Eckhard, D. A. Guschin, L. Stoica, P. J. Kulesza and W. Schuhmann, Electrochem. Commun., 2007, 9, 1998–2002.
4. R. Lei, L. Stratmann, D. Schafer, T. Erichsen, S. Neugebauer, N. Li and W. Schuhmann, Anal. Chem., 2009, 81, 5070–5074.
5. G. Wittstock and W. Schuhmann, Anal. Chem., 1997, 69, 5059–5066.
6. I. Morkvenaite-Vilkonciene, I. Astrauskaite and A. Ramanavicius, 8th International Conference ITELMS'2013, Panevezys, Lithuania, 2013.
7. G. Wittstock, M. Burchardt, S. E. Pust, Y. Shen and C. Zhao, Angew. Chem., Int. Ed., 2007, 46, 1584–1617.
8. M. V. Mirkin, W. Nogala, J. Velmurugan and Y. Wang, Phys. Chem. Chem. Phys., 2011, 13, 21196–21212.
9. A. J. Bard and M. V. Mirkin, Scanning Electrochemical Microscopy, Marcel Dekker, NY, 2001.
10. M. Burchardt, M. Träuble and G. Wittstock, Anal. Chem., 2009, 81, 4857–4863.
11. M. Pellissier, D. Zigah, F. Barriere and P. Hapiot, Langmuir, 2008, 24, 9089–9095.
12. D. Zigah, M. Pellissier, B. Fabre, F. Barriere and P. Hapiot, J. Electroanal. Chem., 2009, 628, 144–147.
13. M. Burchardt and G. Wittstock, Bioelectrochemistry, 2008, 72, 66–76.
14. D. T. Pierce, P. R. Unwin and A. J. Bard, Anal. Chem., 1992, 64, 1795–1804.
15. D. Zhan, X. Li, A. B. Nepomnyashchii, M. A. Alpuche-Aviles, F.-R. F. Fan and A. J. Bard, J. Electroanal. Chem., 2013, 688, 61–68.
16. M. Nebel, S. Grützke, N. Diab, A. Schulte and W. Schuhmann, Angew. Chem., Int. Ed., 2013, 52, 6335–6338.
17. H. Shiku, T. Shiraishi, H. Ohya, T. Matsue, H. Abe, H. Hoshi and M. Kobayashi, Anal. Chem., 2001, 73, 3751–3758.
18. A. O. Okunola, T. C. Nagaiah, X. X. Chen, K. Eckhard, W. Schuhmann and M. Bron, Electrochim. Acta, 2009, 54, 4971–4978.
19. K. Eckhard, X. Chen, F. Turcu and W. Schuhmann, Phys. Chem. Chem. Phys., 2006, 8, 5359–5365.
20. L. Guadagnini, A. Maljusch, X. X. Chen, S. Neugebauer, D. Tonelli and W. Schuhmann, Electrochim. Acta, 2009, 54, 3753–3758.
21. G. Wittstock, Fresenius. J. Anal. Chem., 2001, 370, 303–315.
22. R. E. Gyurcsányi, G. Jágerszki, G. Kiss and K. Tóth, Bioelectrochemistry, 2004, 63, 207–215.
23. X. H. Liu, M. M. Ramsey, X. L. Chen, D. Koley, M. Whiteley and A. J. Bard, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 2668–2673.
24. S. Amemiya, A. J. Bard, F. R. F. Fan, M. V. Mirkin and P. R. Unwin, Annu. Rev. Anal. Chem., 2008, 1, 95–131.
25. A. J. Bard, X. Li and W. Zhan, Biosens. Bioelectron., 2006, 22, 461–472.
26. A. J. Bard and J. Kwak, Anal. Chem., 1989, 61, 1794–1799.
27. J. P. Li and J. G. Yu, Bioelectrochemistry, 2008, 72, 102–106.
28. K. Borgwarth, C. Ricken, D. G. Ebling and J. Heinze, Fresenius' J. Anal. Chem., 1996, 356, 288–294.
29. B. Csoka, B. Kovacs and G. Nagy, Biosens. Bioelectron., 2003, 18, 141–149.

---