UNIVERSIT DEGLI STUDI DI BARI

FACOLT DI SCIENZE MATEMATICHE FISICHE NATURALI

DIPARTIMENTO DI CHIMICA
Tesi di Dottorato di Ricerca in Chimica dei Materiali Innovativi XIV Ciclo- A.A. 1998-2001

Enzyme Modified Electrodes in Amperometric Biosensors

Dr. Walter Vastarella

To my father

Acknowledgements

I would like to thank firstly the man who gave me the possibility to complete this research work, Privat Dozent Dr. Uwe Spohn. He showed always his inclination to encourage me and to follow me step by step. He rendered pleasant my period stay in Halle, especially at the Biotechnology and Biosensors Institute where we shared most of the time. Secondly, I acknowledge Dirk Janasek, my wonderful colleague and generous companion of many days in the laboratories of the institute of Halle. I have to thank also my Italian advisor Prof. Agostiano, for her constant support and suggestion during the PhD period at the university of Bari and the co-workers at the institute of Physical-Chemistry in Bari. Last but not least, I am grateful forever to my parents, who gave me the possibility to study, to reach the degree in chemistry and for their stimulation in following the way of the scientific research. The electroanalytical techniques and FIA methodology were available at the Biochemistry and Biotechnology Institute of the Martin Luther Universitaet in Halle- Wittenberg. All microscopy techniques, including scanning and transmission techniques were made at the Fraunhofer Institut fr Werkstoffmechanick, with the supervision of Dr. Heillmann.

CONTENTS

Introduction

1.- Biosensors, general principles

1.11.1.1-

10 10 10 12 17 21 21 23 25 25 26 27 29 30 30 31 31 33 33 36 36 37

Biosensors, state of the art A short historical survey Performance criteria: characteristics of the biosensors response Enzymes

1.1.2- Classification and recommended definitions

1.1.3-

1.2-

1.2.1- General information 1.2.2- Kinetic parameters of enzymatic reactions

1.3-

Immobilization in enzyme based biosensors

1.3.1- Immobilisation methods 1.3.2- Adsorption 1.3.3- Covalent binding 1.3.4- Entrapment in gel lattices 1.3.5- Encapsulation 1.3.6- Cross-linking 1.3.7- Practical considerations 1.3.8- Inner and outer membranes: the problem of interference 1.4Theoretic aspect of enzyme based biosensors 1.4.1- Mass transport of analytes 1.4.2- Binding and conversion of the bio-element 1.4.3- A kinetic model for the electronic transfer in biological systems 1.4.4- Mechanisms of electronic transfer 2. Flow Injection Analysis 2.1- General scheme and main components of FIA

40 40

2.2- Partial dispersion as foundation of FIA 2.2.1- FIA signals 2.2.2- Physical expressions in FIA 2.2.3- Influence of various factors on the dispersion 3. Supports for immobilization 3.1- Pre-activated immunodyne membranes 3.2- Nanoporous alumina: application as membrane 3.3- Controlled porous glasses for covalent immobilisation 3.4- Nanoclusters of semiconductors CdS 3.4.1- Quantum dot semiconductor properties 3.4.2- Quantum dot semiconductor synthesis 4. Enzymes characterization 4.1- Glucose oxidase: structure and general function 4.2-Sarcosine oxidase: structure, function and the application to creatine and creatinine determination 4.3- Kinetic characterisation of formaldehyde dehydrogenase (FDH) 4.3.1- Chemico-Physical properties of formaldehyde dehydrogenase 4.3.2- Reactivity of NAD dependent FDH 4.3.3- Enzyme activity in aqueous and microemulsive solutions: experimental results 4.3.4- Reactivity of NAD dependent FDH in presence of CdS nanoclusters 5. New electrodes for an amperometric glucose biosensor 5.1- Selectivity in amperometric glucose biosensors 5.2- Experimental section 5.2.1- Chemicals and Reagents 5.2.2- Measuring set-up 5.2.3- Plasma deposition procedures on gold electrode 5.2.4- TEM investigations on the Ru/Rh/gold electrodes 5.2.5- Enzyme immobilisation and measurements of enzyme activity 5.3- Results and discussion 5.3.1- Detection of hydrogen peroxide in FIA 5.3.2- Selectivity and lifetime 5.3.3- Detection of glucose

43 43 45 49

52 52 55 57 59 59 61

64 64

66 70 70 71 73 77

80 80 83 83 83 85 85 88 89 89 93 95

5.4- Conclusions

97

6. Amperometric sarcosine, creatine and creatinine biosensor 6.1- Sarcosine detection: state of the art 6.2- Materials and methods 6.2.1- Chemicals and Reagent 6.2.2- Measuring set-up 6.2.3- Enzyme immobilisation and preparation of the enzymatic reactor 6.3- Result and discussion 6.3.1- Adjustment of the experimental parameters 6.3.2- Sarcosine, creatine and creatinine detection 6.3.3- Long term stability of the biosensor 6.4- Future perspective for enhanced biosensor stability 6.5- Conclusions 7. Enzyme/ nanocluster semiconductors combined systems 7.1- Photo-electrochemical study of CdS nanocluster in presence of NAD dependent formaldehyde dehydrogenase 7.1.1- Photoelectrochemistry at the interface solution/semiconductor 7.1.2- Enzymatic photo-activity with CdS: experimental set-up 7.1.3- Enzymatic photo-activity with CdS: results and discussion 7.2- Concluding remarks References

99 99 101 101 102 103 105 105 107 110 113 114

116

116 116 118 120 126

128

Introduction

The abundant literature related to the keywords Biosensor proves without doubt that this is an attractive fields of research, at the junction of chemistry, biology and physics [1]. As a matter of fact, biosensors couple the capability of the biological compounds to recognize selectively and sensitively a substance or group of substances with the power of the detection systems development. Several attempt to use nanostructurized systems (physic and chemistry of materials), micro-electronic devices (engineering and bioelectronic), or electrochemical knowledge were made for different purposes. Chapter 1 was dedicated to the main general information and features of biosensor, mostly amperometric enzyme based biosensor: definition, components, theoretical aspects in the electronic transfer. The synergic use of both conductive and semiconductive nano-sized material and enzymes is probably the foremost challenging application in developing combined biochemical structures for biosensor or bioelectronic devices [1]; this requires the correct integration between the biocatalyst and the conductive support. For this purpose, in the chapter 3 four different matrices as support materials and some method for immobilizing enzymes onto a suitable electrode support were analyzed. On chapter 4 enzymes of the oxidase and dehydrogenase families were examined, according to their characteristic, properties and activity under working experimental condition. Specifically, glucose and sarcosine oxidase and formaldehyde dehydrogenase were used in this thesis. Major attention was devoted to the selective amperometric investigation of hydrogen peroxide, glucose (chapter 5), sarcosine, creatine and creatinine (chapter 6) at ruthenium/rhodium modified nano and micro-structured materials deposited onto an indicator gold electrode, for the response of the analytical species. These two different applications of amperometric biosensor were performed, by coupling the enzyme catalytic activity with the best working electrode according to the following features: Background signal, S/B ratio; Active surface of the electrode Sensitivity on the basis of voltammetric investigations.

Pre-activated and nanoporous membranes or porous controlled glasses, in which the enzyme was covalently bonded, were used as supporting materials in amperometric three electrode set-ups, to improve the sensor response. Their application in a glucose oxidase (chap. 5) and sarcosine oxidase (chap. 6) modified biosensing devices, under flow conditions, were achieved. Sensitivity, long term stability and specificity were principally evaluated by using Flow Injection Analysis as main investigation method. In chapter 2 the physico-chemical and analytical concepts of FIA methodology were also reviewed, to furnish a basic comprehension for the evaluation of the main sensor parameters. In chap. 7 photo-electrochemical assays were performed upon an enzyme/semiconductor combined system. Using a classic three electrode electrochemical cell the catalytic activity of NAD dependent formaldehyde dehydrogenase in microemulsion micellar solution was tested under illumination. The indicator electrode consisted of CdS semiconductor deposited on gold substrate for the amperometric detection. The prototypic third-generation system was designed with two principal aims: semiconductor nanomaterial should replace the co-factor NAD+/NADH as charge carrier in the redox process, due to its lower costs and better stability at normal experimental conditions in comparison to NAD+/NADH; the enzyme should retain its activity even in micellar solution and in presence of CdS semiconductor, for monitoring formaldehyde in food and agricultural applications. Actually this project is yet at the earlier stage of preparation.

CHAPTER 1 BIOSENSORS, GENERAL PRINCIPLES

1.1- BIOSENSORS, STATE OF THE ART

1.1.1- A short historical surveyThe development of chemical sensor has received a great deal of scientific interest in the last two decades. Not only the chemical industry may benefit from these sensors but also the food industry, bio-industry, medicine, environmental control because of their capability to give continuously and reversibly a selective and fast response to the presence of a specific compound in a complex mixture of components, without perturbing the system. Biosensors combine the power of analytical detection techniques with the specificity of biological recognition system and therefore are the most promising devices today about this selectivity [2,3]. Furthermore, biosensors possess many unique features such as compact size, simplicity of use, one-step reagentless analysis, absence of radioactivity, etc., that make them very attractive alternatives to conventional biological analytic techniques. Sensors can be generally divided into physical, chemical and biochemical sensors. Physical sensors measure traditional parameters such as temperature, pressure, humidity, or flow rate; sometimes they are not considered to be real sensors. According to the common nomenclature proposal (IUPAC) [4-6], chemical sensors are miniaturised compact devices that transforms a chemical information, such as a variation of concentration, into an analytically useful and measurable signal. In other words, the selective and reversible detection is accompanied by a concentration dependent electrical signal. As reported in a further discussion, different electronic signals were employed to translate the information of a variation: current, potential changes, capacitance, impedance, piezoelectric transduction, and field effect transistors. However, chemical and physical sensors are beyond the scope of the present thesis. Biosensors are special chemical sensors in which the recognition system utilises a biochemical mechanism. The biological recognition system, usually a receptor protein,

10

antibody, enzyme, translates the information from the biochemical domain into a chemical or physical output signal, with a defined sensitivity. Due to the intrinsic, highly selective properties of the biomolecular species in comparison to the inorganic catalysts, biosensors are the most selective one. They are based on either the catalytic transformation of a substance or the detection of enzyme inhibitors or mediators. They normally suffer from one severe drawback, namely instability of the biological receptor molecules, that are subject to denaturation. In fact increasing the stability of the biomolecule is one of the major issues in biosensor research. Biological activity can be preserved by storing the biosensor under the right conditions, e.g., buffer pH, low temperature. This may give long off-line lifetimes, improving the so-called shelf-stability. Of greater interest is the lifetime under working conditions (the operational stability), which is still very limited in many cases and more difficult to be controlled. The stability of the biomolecule can be enhanced by immobilization procedure. A large range of immobilization techniques have been described in literature [7-13] and some of these techniques will be used and discussed in this thesis. A classical chemical sensors is the pH-electrode, already invented at the beginning of the XXth century, when the pH-sensitivity of certain glass membrane was discovered [14]. Even though a pH-electrode is nowadays an indispensable instrument, at that time, the importance of this sensor was not recognised. After its invention many other ion selective electrodes followed and actually several cations and anions can be measured selectively by potentiometric measurements with these electrodes [15]. Other chemical sensors, based on semi-conducting materials, have been developed for the selective detection of gases, ions and biomolecules [16,17]. A large number of detection principles have been described in the literature [18 and ref. therein], but their description is outside the scope of this thesis. Biosensors were first mentioned as such in 1977 [19]. Before that time, devices that contained a biological sensing element were simply referred to as electrodes with a description of the kind of sensing component, e.g., enzyme electrode [20]. In the last decades the potential use of biosensors in a great number of disciplines has led to a great deal of research activity and their growing interest has dramatically increased the number of publications and patents about bio-analytical sensing devices for several applications. More recently, a large number of papers were presented at the main meetings and conferences devoted to biosensors, whilst a recent market survey forecasts the total world market of almost $ 900 million in the year 2000, of which medically related consumers account for $ 300 millions in sales.

11

Nowadays some interesting commercial biosensors are already available for detecting different substrates like glucose, lactate, urea and penicillin, but their number is still relatively very small in comparison to the large theoretical efforts in their design and development. The capability to have a disposable and miniaturizable device has been another topical issue for their diffusion. The successful design and development of biosensors requires the involvement and the right combinations of different know-how and disciplines: biochemistry is useful to improve selectivity and sensitivity of the biological components, whereas physics, electrochemistry and engineering technology provide the information with regard to microelectronics and computers [21] in the transduction systems. As already mentioned medicine has the major interest in biosensors, because stable and reliable devices would make possible on line and in-vivo monitoring of important biocomponents, such as metabolites, hormones and drugs. This interest may lead, for instance, to the integration of an in-vivo blood sugar monitor with an automatic insulin delivery system, for patients who are affected by diabetes mellitus. In the field of biotechnology, biosensors have been used for fermentation control by following fluctuations in the nutrients and adjusting their level if necessary; in environmental technology they can be used to monitor pollutants, herbicides and pesticides; in food technology the concentration of various nutrients, the quality and the freshness, contamination by micro-organisms, or the presence of prohibited components could be detected for all kinds of foodstuffs or beverages. In space research biosensors may facilitate the monitoring of certain metabolic processes under influence of reduced gravity. Last but not least the military industry is interested in biosensors for chemical and biological defence.

1.1.2- Classification and recommended definitions Biosensors are made up of three different but strictly connected elements: the selector, the transducer and the detector. a) the selector is the part that selectively binds and recognises the compound to be detected; b) the transducer transfers the signal from the output domain of the recognition system to a physically measurable signal; c) the detector permit to amplify, process and display the chemico-physical signal in a suitable form.

12

A biosensor is precisely defined, according to IUPAC definition, as a self-contained integrated device, capable of providing specific quantitative or semi-quantitative analytical information using a biological recognition element, which is retained in direct spatial contact with a transduction element [22]. It is usually made by attaching a biologically-sensitive material (the selector), which is responsible for the recognition of the test species in the mixture, to a suitable transducing system. When biological material, such as enzyme, organelle, membrane component, a bacterial cell or other whole cell, an antibody or antigen, or whole slice of mammalian or plant tissue, interact specifically and reversibly, there is a change in some physico-chemical parameters associated with the interaction, before the transduction. In table 1.1 some of numerous available transduction methods are reported. One of the most useful is the transformation of a bio-chemical reaction into an electrical potential or a current signal, also because is relatively easy to construct and handle electrochemical devices based on this kind of transduction and relatively feasible their miniaturization.

SELECTOR
Enzyme Antibody DNA, RNA Receptor

TRANSDUCER
Current Potential Surface Plasmon Impedance

DETECTOR
Amperometric electrode Potentiometric electrode Laser light reflectometer Conductometer

Table 1.1- Basic elements of a biosensor: some example

Biosensors may be schematically classified according to the biological specificity conferring mechanism, or to the mode of signal transduction or, alternatively, a combination of two. Therefore these might be distinguished as amperometric, potentiometric, field effect or

13

conductivity sensors [6]. Amperometric biosensors are based on the measurements of the current resulting from the oxidation or reduction of an electroactive species, by keeping a constant potential at the working electrode (Pt, Au, Carbon, etc.) or array of electrodes with respect to a reference electrode, which may also act as the auxiliary electrode at low current. The resulting current is correlated to the bulk concentration of the electroactive substance or its reaction within the adjacent bio-catalytic layer [23]. Potentiometric biosensors involve the determination of the potential difference between an indicator and a reference electrode (or two reference electrodes separated by a permselective membrane) without significant flowing of current. The transducer may be an ion selective electrode as recognition element [24], e.g. pH electrodes, ion or gas electrodes. Due to their typical electrochemical features the response of both amperometric and potentiometric biosensors is never an equilibrium, but either steady state or transient response. Many enzyme reaction (e.g. urease) and many biological receptors may be also monitored by ion conductometric or impedimetric devices, using interdigitated microelectrodes [25] to perform either a conductance or impedance differential measurement. Finally, field effect transistors may be also used as transduction devices. Alternatively, it is possible to distinguish amperometric biosensors of first, second and third generations, according to the different mechanism of the electronic transport. First generation amperometric biosensors work by means of the direct detection of electroactive species that are enzymatically produced or consumed [26]. In the most common first generation biosensors that need of a co-substrate in the reaction environment, the enzyme was physically entrapped near the electrode surface. A good example is the traditional glucose sensor, which can detect the hydrogen peroxide produced in the oxidation process catalyzed by a glucose oxidase enzyme (indicated as GOd):

Glucose + O 2 GOd Gluconolactone + H 2 O 2

(1.1a)

During the catalytic cycle the enzyme is first reduced and then regenerated by oxidation with the molecular oxygen in the sample solution. The operative mode of a glucose sensor based on the above sequence is the electrochemical oxidation of the hydrogen peroxide to regenerate oxygen and to complete the cycle at the working electrode:

14

 H 2 O 2 O 2 + 2e + 2 H
electrode

(1.1b)

The most important drawbacks of this sensor are the denaturation of the GOd by H2O2 and the strong dependence of the electrical signal on the co-substrate concentration. The second generation biosensors make use of an artificial electron carrier, or mediator, in place of oxygen, which shuttles the electron involved in the redox process from the active center of the enzyme to the electrode or viceversa [21, 27-32]. Both mediator and enzymatic substrate must be in the analytical solution. One example can be schematised by the following reactions, where M indicates the mediating species:

Glucose + GOdox gluconolac tone + GOdred

(1.2a)

GOdred + Mox GOdox + Mred

M red n e + M ox

(1.2b)

(1.2c)

Examples of mediators are low molecular weight redox couple as ferrocene [33], ferrocene derivatives [34], bipyridil complexes [35]. Ideally it is very practical and preferable to have a material that communicates directly with a redox enzyme, in order to remove the complication due to the presence of mediators in the analytical solution. In that case no cosubstrates (oxygen or mediator) are required and the electrons involved in the process are measured directly at the electrode. Enzyme electrodes based on this principle are called third generation biosensors [21]. Interest about third generation biosensors grew a few years ago when it was found out that direct electrochemistry with redox proteins is possible without using a co-substrate [36]. The foremost work in this field used mercury electrodes, onto which enzyme adsorption was strong and lead to molecule denaturation. In order to make the metal electrode compatible with redox enzymes and the biological receptor, surface modification of the bare electrode has been required [35, 37]. Nevertheless it has been demonstrated that small redox protein like cytochrome C can shuttle electrons directly to bare glassy carbon electrodes [38]. For large redox enzyme, like glucose oxidase, this is difficult to accomplish as they have thick

15

insulating protein shells and their catalytic centers are buried deep inside and protected from the surroundings [39]. Therefore in these cases they must be immobilised on a compatible modified electrode surface in a way that makes the electron transfer from the catalytic center to the electrode feasible without the occurrence of denaturation [37, 40, 41]. Many research group addressed their work to use highly conducting organometallic complexes as charge transfer salts of tetracyanoquinodimethane and tetrathiafulvalene, for achieving direct electronic contact with redox enzymes. At the state of our knowledge this was successfully accomplished only with GOd [42] and colinesterase [43]. Further improvement towards new third generation biosensor are progressing faster and faster.

16

1.1.3- Performance criteria: characteristics of the biosensors response

For any sensor based on molecular recognition, it is important to characterise a response. This is even more topical when operating parameters may indicate the nature of the rate determining step and facilitate optimisation of enzyme-based biosensor in a given matrix. When performance criteria are not specific to biosensors but common to most types of chemical sensors or analytical methods, as precision, accuracy, inter-laboratory and interpersonal reproducibility, it is recommended that standard IUPAC definitions be followed [44]. Most of the discussion below related to enzyme-based biosensor can be considered valid for other types of biosensing devices. One of the methods for assessing the main parameters consists in measuring the receptor specific activity, i.e. the ratio of the number of active molecules over the total number of immobilised molecules. This figure is dependent on the mode of immobilization such as the molecular orientation, or the number of points of attachment. However both the rapid proliferation and the diversity of biosensors has led to a lack of rigour in defining performance criteria. It is still a great demand to establish standard protocols for some characteristics of biosensor that are common to different types of electrochemical sensor, in accordance with standard analytical IUPAC protocols.

Calibration characteristics.

Sensor calibration is generally performed by adding standard solutions of the analyte and plotting the steady state response (Rss) corrected for a background signal (RB), versus the analyte concentration or its logarithm, log c/c0, where c0 refers to a 1 mol l-1 reference concentration. Transient responses are less significant for continuous monitoring. A convenient way to perform such calibrations is to indicate the maximum rates of variation of the sensor response (dR/dt)max. As we will see in a further discussion, under well defined hydrodynamic conditions the biosensor can be placed in a FIA system for sequential sample analysis. The sensitivity and linear concentration range of calibration curves are determined plotting the ratio (Rss - RB )/c versus log c/c0 and (dR/dt)max/c versus log c/c0 for a transient response. This method is more concise than to plot only the numerator versus log c/c0, since it gives the same weight to low and high concentration results. Sensitivity is to be determined within the linear concentration range of the calibration curve.

17

Electrochemical biosensors have an upper limit of the linear concentration range, directly related to the biocatalytic or biocomplexing properties of the biochemical or biological receptor. The compromise for reaching an extension in the linear range is the decrease of sensor sensitivity. The local substrate concentration within the reaction layer can be at least two orders of magnitude lower than in the bulk solution. In relation to the usual parameters for Michaelis Menten kinetics, Km and Vmax, which will be expressed in the next paragraphs, enzyme based biosensors are particularly characterised by their apparent Km and (Rss - RB)max. The former is the concentration yielding a response equal to half of its maximum value, and the latter represents the infinite analyte concentration. As for any electrochemical sensor, composition and number of replicates and how the sample matrix is duplicated must be always state, especially for disposable biosensors. The sensitivity is the slope of the calibration curve (Rss - RB ) versus c or its logarithm. The limit of detection takes into account the blank and the signal fluctuations and therefore it can not be considered as a specific to biosensors. It is better to determine the working concentration range determined by the lower and the upper limits of quantification., which may be larger than the linear concentration range.

Selectivity and reliability.

Biosensor selectivity depends strongly upon the choice of biological receptor and transducer. It is known that bacteria, yeast or tissue cultures are naturally non-specific. Many enzyme are specific, nevertheless class enzymes, such as alcohol, group sugar, peroxidase, laccase, tyrosinase, glucose NAD-dehydrogenase, ceruloplasmin, etc., have been used for development of non-selective biosensors, such as for example in phenols determination sensors. Metal electrodes are often sensitive to numerous interfering substances and hence they have a low selectivity, which can be improved with pH electrode or ISFETs transducers. When transducer interfering substances are well identified, such as ascorbate and urate in glucose sensors based on H2O2 detection, their influence may be restricted by means of two different solutions. Alternatively a compensating sensor may be introduced in the set-up without biological receptor on the surface. Within various methods for selectivity determination, two are recommended, depending on the aim of its measurement. The first one consists in measuring the biosensor response to interfering substance addition: a calibration curve for each interferent is compared to the analyte calibration curve, under identical experimental operating conditions. Selectivity is

18

expressed as the ratio of the signal output with the analyte alone and with the interferent alone at the same concentration as that of the analyte. In the second procedure interferents are added into the measurement cell, already containing the analyte, at the mid range of its expected concentration value. Selectivity is then expressed as the percentage variation of the signal. The second method is more easily quantifiable than the first one even though it is more restricted in application and significance, because dependent on the concentration range that one would determine. Definition of reproducibility is the same for electrochemical biosensors as for any other analytical device: it is a measure of a scattering in a series of observation and results performed over a period of time and it is generally determined for the analyte concentration in the usable range. The reliability of a biosensor depends on both its selectivity and reproducibility and is necessary for accuracy assessment. It must be determined under actual operating conditions, in the presence of possible interfering substances. In order to be reliable, the biosensor response should be directly related to the analyte concentration and should not vary with the fluctuation of the interferent concentration within the matrix. In this way the reasonable interference and how to quantify its influence should be simultaneously stated.

Response time and sample throughput

Steady state response time is easily determined into the measurement cell where is the analyte solution. It is the time necessary to reach 90% of the response, whereas transient response time corresponds to the time necessary for the derivative of the signal (dR/dt)max to reach the maximum value after the analyte addition. Both response times depend upon the analyte, the co-substrate and product transport rate through different layers or membranes, but also depend directly upon the activity of the receptor (enzyme). Therefore the thickness and permeability of layers and membranes for enzyme immobilization are essential parameters. Finally they also depend on the mixing condition of the sample into the batch measurement cell, a factor which is not negligible. A simple way to define such hydrodynamic conditions in the biosensor vicinity is using a FIA system: as we will show in the next chapter, biosensors in FIA set-up are defined and characterized as other common FIA detector; if the analyte concentration is varied stepwise transient or steady state time are defined as in the batch, whilst if it is introduced into the circulating fluid, only transient response time is available.

19

When biosensor is used for sequential measurements the sample throughput is a measure of the number of individual samples for unit of time. It take into account not only the response time, but also the time needed for the signal to return to its base line, commonly indicated as T. This features depend generally upon the sample composition, the analyte concentration or the sensor history. Theoretical modelling of biosensor operation enables a better understanding of the relative importance of all these factors on the response time and sample throughput, but modelling are somewhat limited by the knowledge of a very large numbers of sensor parameters, such as partition and diffusion coefficient of layers, thickness of membranes distribution of biocatalytic activity within the layer or the membrane, transducer operating properties, etc. [45]

Biosensor stability and lifetime

Stability is divided into two different categories: operational and shelf-stability. The operational stability of a biosensor may vary according to the sensor geometry, method of preparation, as well as on the applied enzyme or transducer. Furthermore is strongly dependent on the response rate limiting factor, i.e. external or inner diffusion of the substrate or biological recognition reaction. It varies considerably depending upon the operational conditions. To evaluate the biosensor stability under using, consideration of the analyte concentration, the continuous or sequential contact with the analyte solution, the temperature, the buffer pH, the eventual presence of organic solvents buffer and the matrix composition must recommended. For storage stability assessment, significant parameters are the state of storage, i.e. dry or wet, the atmosphere composition (air or nitrogen, or other noble gases), pH, buffer composition and presence of additives. While it is relatively easy to determine the laboratory bench stability of biosensor, both during storage and operation, procedures for assessing their behaviour during several days of introduction into industrial reactors is much more complex and difficult to handle. In laboratory or industrial set-ups it should be always necessary to specify both the stability, along with the storage and operational conditions, and the substrate concentration, as compared to the Km. Knowledge of the biosensor rate limiting factor is especially important to understand stability properties. Finally, the mode of lifetime assessment should be specified, by reference to the initial sensitivity. The commonly recommended definition is that lifetime is the storage or

20

operational time necessary for the sensitivity, within the linear concentration range, to decrease by a factor of 10% (indicated as tL10) or sometimes of 50% (indicated as tL50). For the storage lifetime determination it is better to compare different storage times of biosensors from the same production batch.

1.2- ENZYMES

1.2.1- General information

Enzyme is a biological catalyst with extremely high specificity and efficacy. It must be remembered that a catalyst permit to reach easily the equilibrium without modifying its position. However, most of the enzymatic reactions take place in a short time at normal temperature, without using dramatic value of pressure and pH, and work at much higher rates in comparison to the common chemical organic and inorganic catalysts. Enzyme-catalyzed reactions are normally from 103 to 107 faster than the same noncatalyzed reactions. They are extremely specific and selective for the substrate which they interact with. Enzymes generally have a variable specificity degree, catalysing either a group of substrates that have correlated structures or a single molecule. Some kind of them assess a good degree of stereo-specificity, as they catalyze only one of two substrate stereoisomer. The foremost general feature of enzymes is the reaction specificity. In fact they do not generate useless by-products of the reaction and give as high yield in the enzymatic reactions as almost 100%. In this thesis we will concentrate on redox enzymes, such as flavoprotein, rather than redox proteins (for example cytochrome C). The redox proteins are relatively low molecular weight proteins containing a redox active prosthetic group, located towards the periphery of the protein itself. They act as simple carrier of electrons and catalyse no specific biotransformation. Redox enzymes have the main functional capability of the globular proteins to bind specifically to one or more molecules by means of their active sites. These redox centres lie normally between two sub-units of the molecule, buried in a slit or in a crevice of it and are responsible to catalyse the oxidation or reduction of a particular substance in the

21

cell. In general the redox proteins react with larger redox partners than the redox enzymes, with some consequence for the heterogeneous electrochemistry. Enzymes are generally bigger than the substrate they bind, so that only a little portion of substrate is effectively involved in the enzymatic reaction. The molecular recognition of the substrate is achieved by the well known lock and key principle between the respective receptor area and the analyte to be recognized. The I.U.B., International Union of Biochemistry have recently published a schematic classification of the enzymes in six different general categories according to their principal reactions they catalyse, in order to establish a systematic nomenclature on reward (tab 1.2). This classification system attribute a typical code number to every enzyme, to identify the main category and sub-category as well as the reaction it catalyses. Recent classifications have been also made according to the nature of the electronic acceptor or the prostetic group of the enzyme (such as flavodehydrogenase, pyrrolo quinoline quinone, and so on).

Classification
1. OXIDOREDUCTASE 2. TRANSFERASE 3. IDROLASE 4. LIASE 5. ISOMERASE 6. LIGASE Redox reaction

Type of reaction

Functional group migration Hydrolysis Elimination to get double bonds Isomerisation Bond formation along with ATP hydrolysis

Table 1.2- Basic classification of enzymes classes

22

1.2.2- Kinetic parameters of enzymatic reactions

Enzyme are especially characterised by kinetic studies about their activity. This is accomplished monitoring the modification of the reaction velocity after varying reagent concentrations and experimental conditions. In simple enzymatic reactions, the reagent (here indicated as S) conversion is enzyme catalyzed to give one or more products (P). The following reactions:

k2 E + S k1 ES E + Product

k 1

(1.3)

can be expressed with the general equation of Michaelis-Menten which correlates the velocity to the substrate concentration:

v0 =

k 2 [E]0 [S] K M + [S]

(1.4)

where k1 and k-1 are respectively the association constant of S and E and the dissociation of S from ES, k2 is the catalytic turnover constant, [S] [E]0 are respectively the substrate and enzyme concentration, KM the Michaelis-Menten constant and:

v max = k 2 [E]0

(1.5)

KM correspond to the substrate concentration as the reaction velocity is an half of vmax and it gives an indication of the affinity degree between the enzyme and its substrate: the higher is KM the weaker is the ES stability. It varies considerably with the nature of the substrate and the enzyme, but also with pH and temperature conditions. The highest catalytic turnover is obtained at relatively low substrate concentrations for enzyme with a small value of KM. In the construction of enzyme electrodes, it is desirable to obtain the lowest KM and the highest vmax. They can be measured either by varying the substrate concentration and keeping constant the enzymatic concentration, through the analysis of the initial velocities, or by the well known Lineweaver-Burk equation, obtained making the reciprocal form of equation (1.4).

23

Deviation from the linearity of the Lineweaver-Burk equation reveals whether mass transport is important in enzyme-based biosensor systems. Figure (1.1) shows the dependence of the reaction rate upon substrate concentration, together with KM and vmax. The initial rate increases with the substrate, until a non-limiting excess of substrate is reached, after which additional substrate causes no further increase in the rate. A levelling-off of calibration curves is expected at substrate concentrations above the enzymatic KM. Accordingly, low KM values result in a narrower linear range, which reflects the saturation of the enzyme active sites.

Fig. 1.1- Dependence of velocity of an enzymatic reaction on the substrate concentration, at a constant level of enzyme activity (Vm corresponds to vmax)

Experimentally, the linear range in amperometric biosensing devices may exceed the concentration corresponding to KM, because the local substrate concentration in the electrode region is often less than the bulk concentration. The next sections will be also dedicated to amperometric probes coated with diffusion limiting membranes. The level of co-substrate may also influence the linear range, through the stoichiometric limitation of equation (1.3). An improved sensitivity can be achieved by coupling in series different enzymatic reactions in a chain, cycling, catalytic mechanism, for example by enzymatic recycling of the analyte in systems made of two enzymes. Such an amplification scheme generates more than a stoichiometric amount of product and hence large analytical

24

signal for low levels of analyte. Finally, a second enzyme can be used to generate a detectable species, from a non-electroactive product of the first reaction.

1.3- IMMOBILIZATION IN ENZYME BASED BIOSENSORS

Enzyme-based electrodes combine the specificity of an enzyme for its substrate with the analytical power of electrochemical devices. The choice of the sensing electrode depends primarily upon the enzymatic system employed. Amperometric probes are highly suitable when oxidase or dehydrogenase enzymes, generating electro-oxidizable peroxide or NADH species, are employed. The success of an enzyme biosensor ultimately relies on how well the enzyme bonds to the sensor surface and remains there during use. Immobilization between matrix and bioreceptor has been showed as the topical procedure for improving the enzyme stability and thus general biosensor performances. On the next pages several physical and chemical schemes to immobilize the enzyme onto the electrode will be reviewed.

1.3.1- Immobilization methods

Since the development of the enzyme-based sensor for glucose, first described by Clark in 1962 [20], in which glucose oxidase was entrapped between two membranes, an impressive literature on methods of immobilization and related biosensor development has appeared. After that, research work has provided a bewildering array of support material and method for immobilization, several of which have been extensively reviewed elsewhere [21, 22, 46, 47]. Biological receptors, enzymes, antibodies, tissues with a high biological activity, can be immobilised in a thin layer at the transducer surface by using different methods. Taking into account the relative use and application, both the choice of the support material and the main method of immobilization can influence not only the activity and the half-life of a biotransformation but also the operational stability of the biosensor.

25

In particular enzyme molecules in solutions behave as any other solute in that they are dispersed in the solution and have a complete freedom of movement. Enzyme immobilization is generally a technique specifically designed to restrict this freedom of movement. There are five principal methods for immobilising enzyme/cell on support materials: simple adsorption, covalent binding, entrapment, encapsulation, cross-linking (see figure 1.2). For amperometric devices the bulk modification of entire electrode material [48] must be take into consideration as further different technique of biological receptor immobilization onto an electrode support. Excepting the last one, the merits of each procedure will be briefly discussed below.

1.3.2- Adsorption

The immobilization by adsorption is the less complex method, for it involves reversible surface interactions between enzyme and support material. The forces involved are mostly electrostatic, as Van der Waals forces, ionic and hydrogen bonds, although sometimes hydrophobic bonding can become significant. These forces are weak, but sufficiently large in number to enable reasonable binding. Existing surface chemistry between enzyme and support is used, so in this method neither activation or chemical modification are required and little damage is done to enzymes. The simple and cheap procedure consists of mixing together enzyme and support under suitable conditions of pH and ionic strength, for a period of incubation, followed by collection of the immobilised material and extensive washing to remove non-bound biological component. The most significant disadvantage is the leakage from the support with probable desorption and contamination of the solution. Physical factors as flow rate, bubble agitation particle-particle abrasion can affect the desorption of the biocatalyst from the support, therefore this method alone has not been used under flow conditions. Simple adsorption onto a clay, glass or ion exchange resins surface can attach enzymes only for a single nonrepeatable measurement.

26

Fig.1.2- Principal methods of immobilization

1.3.3- Covalent binding

This method involves the formation of a covalent bonds. Functional chemical groups belonging to amino acid residues on the surface of the enzyme, which are not essential for activity, may be attached covalently to chemically activated supports (glasses, cellulose or synthetic polymers). A number of amino acid groupings are suitable for participating in covalent bond formation. Those mainly involved are the amino group of lysine or argirine, the carboxyl group of aspartic and glutamic acid, the hydroxyl group of serine and threonine, the sulfydril group of cysteine [49]. Many support materials are available to form covalent bonds. This large plenty of support reflects the fact that no ideal support exists and respective advantages and disadvantages

27

must be taken into account when considering possible procedures for a given enzyme immobilization. Although several factors influence the selection of a particular support, research work has shown that hydrophilicity is the main factor for maintaining enzyme activity [50]. Polysaccharide hydrophilic gel polymers are popular support materials for immobilization: the sugar residues in these polymers contain hydroxyl groups as ideal functional groups for chemical activation, in order to provide covalent bond formation and to create hydrophilic environment in the support by means of hydrogen bonds. Because the polysaccharide supports are susceptible to microbial/fungal disintegration and organic solvents can cause shrinkage of their gels, they are preferably used in bead form. Other popular support for enzyme immobilization are porous glasses and silica: porous silica consist of small spherical particles of silica fused together in such a way as to form micro-cavities and small channels. The support is normally sold in bead form with a well known micro-architecture, and is very strong and durable. Sintered glasses may be tempered to form a system of uniform channels whit a diameter dependent on the tempering conditions. Porous glasses are also durable and resistant to solvent distortion, even though less hydrophilic than polysaccharides. There are many reaction procedures for coupling an enzyme and a support in a covalent bond. However most reaction fall into four categories: 1. Formation of an isourea linkage 2. Formation of a diazo linkage 3. Formation of a peptide bond 4. An alkylation reaction It is important to choose a method that will not inactivate the redox site of the enzyme. So if an enzyme employs a COOH group at the active site for participation in catalysis, it is wise to choose a reaction that involves an amino group for the covalent bonding with the support. Before the enzyme is added in a coupling reaction, functional groups onto the support should be activated, normally to render them electrophilic, by a specific reagent. Covalent binding of enzymes on surfaces activated by means of functional groups or spacers is performed by glutaraldehyde, carbodiimide, SAMs, silane-compounds, most of them commercially available. As example, cyanogen bromide (CNBr) is often used to activate the hydroxyl groups in polysaccharides, joining enzyme and support via an isourea linkage, whilst carbodi-imide group is used for activating supports containing carboxyl functional group, joining enzyme and support via a peptide linkage.

28

It is also possible to chemically modify the normal functional groups in order to produce a range of derivatives containing different functional groups and to extend the possibility of immobilization methods (e.g. cellulose derivatives).

1.3.4- Entrapment in gel lattices

Immobilization by entrapment differs from adsorption and covalent binding in that enzymes are free in solution, but restricted by the lattice structure of the entrapment system. It is possible to distinguish between three general methods: 1- Entrapment behind a membrane: a solution or suspension of enzymes, cells, a slice of tissue is confined by an analyte permeable membrane as a thin film covering the detector; 2- Entrapment of biological receptors within self assembled monolayers (SAMs) or bi-layer lipid membranes (BLMs); 3- Entrapment of biomolecules within a polymeric matrix membranes (such as polyacrylonitrile, agar gel, polyurethane, or polyvinyl-alcohol), redox gels, sol-gels with redox centres [51]. The most used technique is the entrapment in polymeric film (e.g. polypyrrole, Nafion) via casting or electropolymerisation, and in redox gel lattice. In the latter approach, gel porosity must be controlled to ensure that the structure is tight enough to prevent enzyme leakage, but at the same time to allow free movement of substrates and products. Inevitably, the support will act as a barrier to mass transfer with serious implications for reaction kinetics, but it has useful advantages, since harmful biological compounds are prevented from interaction with the immobilised biocatalyst and, as well described in the following discussion, some dangerous interference are avoided. Additional improvement can be achieved combining several membranes or coatings, in order to extend the linear range (via reduction of the local substrate concentration) and to reject the potential interference of coexisting electroactive species.

29

1.3.5- Encapsulation

Encapsulation of receptors can be achieved by enveloping the biological components within various form of semipermeable membranes [52]. It is similar to entrapment method in that enzymes are free in solution, but restricted in space. Large proteins cannot pass out or into the capsule, but small substrates and products can pass freely across the semipermeable membrane. Many materials have been used to make microcapsules varying from 10-100 m in diameter, for example nylon and cellulose. The problem associated with diffusion is more acute and may result in rupture of the membrane if products from a reaction accumulate rapidly. A further problem is the similar density of the enzyme to that of the bulk solution with consequent disadvantages in reactor configuration designing, in flow dynamics, and so on. It is also possible to use biological cells as capsules, as for example of erythrocytes, red blood cells. The membrane of erythrocytes is normally only permeable to small molecules, even though when placed in hypotonic solutions, they swell rapidly, increasing the permeability of the membrane, and causing the diffusion of the erythrocytes protein out and of the enzyme into the cell. Returning the swollen protein to an isotonic solution enable the cell to return to its normal state, and the trapped enzyme do not to leak out. A distinct advantage of this method is the co-immobilization of enzyme/cell in any desired combination.

1.3.6- Cross-linking

This procedure is support-free and involves joining the receptor to each other to form a large, three-dimensional complex structure (see the bottom of figure 1.2). Crosslinking or cocrosslinking can be achieved by chemical or physical methods [53]. Chemical method normally involves covalent bond formation between the enzyme by means of bi- or multifunctional reagent, such as glutaraldehyde and toluene diisocyanate. Both albumin and gelatine has been proved as good molecular spacers to minimise the close proximity problems that can be caused by crosslinking a single enzyme. Physical crosslinking of cells by flocculation is well known in biotechnology industry and does lead to high cell densities. Flocculating agents, such as polyamines, poly-ethyleneimine, polystyrene sulfonates, phosphates, have been extensively used and well characterised. Crosslinking is rarely used alone ass technique of immobilization, because the absence of mechanical properties and poor stability are severe limitations for biosensor development.

30

This one is often used to enhance other methods of immobilization, normally in order to reduce cell leakage in other systems.

1.3.7- Practical considerations

In biosensing systems with the best response and performance criteria, receptors are commonly immobilised either alone or mixed with other proteins such as bovine serum albumin, as co-crosslinker agent, either directly on the transducer surface, or on a polymer covering it. In the latter case, pre-activated organic membranes can be used directly for the biomolecule immobilization without further chemical modification of the membrane. Covalent attachment, cross-linking procedures or a combination of both are more complicated than the simple entrapment ones, but are useful in cases where sensor is so small that the appropriate membrane must be fabricated directly on the transducer, in order to obtain more stable and reproducible activities. Furthermore, the biosensor long term stability is largely improved with the combination of two different methods, because the enzyme is both strongly bonded to the support and cross-linked (sometimes also cocrosslinked) in a three dimensional lattice.

1.3.8- Inner and outer membranes: the problem of interference

Besides the reacting layer or membrane many biosensors, especially those designed for clinical and biochemical applications, can incorporate one or several inner or outer layers with three different functions: a) Protective layer. The outer membrane prevents large molecules (proteins or biological sample cells) from interfering with the reaction layer. It also reduces leakage of the layer component into the sample solution. Its function is important for example in implanted glucose sensors, since the glucose oxidase is of non-human origin and may cause immunological reaction. Furthermore, a properly chosen membrane can exhibits permselective properties which could be useful to biosensor function. It may decrease the influence of possible interfering species detected by the transducer. For instance, most in
vivo or ex vivo glucose biosensors present a negatively charged inner cellulose acetate

membrane in order to decrease the interfering effect of ascorbate or urate,

31

electrochemically detected with enzymatically generated hydrogen peroxide [54-57]. After the enzyme casting onto the cellulose acetate membrane, an outer polycarbonate film is deposited on the bioreceptor. b) Diffusional outer barrier. Most enzymes follow some form of the Michaelis-Menten kinetic (eq. 1.4), therefore enzymatic reaction rates have nonlinear behaviour at high concentrations. Nevertheless, linear dynamic ranges may be enlarged if the sensor response is controlled by the substrate diffusion through the membrane and not by the enzyme kinetic process. This control is achieved by means of an outer membrane placed over a highly active enzyme layer [58]. Normally the thinner is the membrane the shorter is the biosensor response time. Furthermore, such diffusional barrier also makes the sensor response independent of the amount of active enzyme present, causing a sensitive improvement of the response stability. c) Biocompatibility and biostability of the surface. Biosensors are considered biocompatible if their implantation does not affect the normal functioning of the host medium and, vice-versa, the host medium does not materially affect the normal operation of them. As biosensors are in contact with biological matrices, such as tissues or fluids or cell cultures implanted in vivo, they are subject to two sets of modification: firstly, a modification of the host biological sample caused by biosensor introduction and toxicity, mutagenicity, thrombogenicity, carcinogenicity of its own elements. Secondly, a modification of operational biosensor properties, e.g. outer layer or inner detector fouling, substrate or co-substrate transport rate towards the biorecognition area. Normally an outer layer is essential to improve the biosensor stability, after the implantation (apart from molecular recognition systems which require direct contact sample-bioreceptor). According to the dimensional request of the sensor, pre-cast membranes or polymeric material deposited by dip- or spin-coating may be used. Some micro-sized biosensors diffused for disposable measurements are prepared by electro-polymerisation or general membranes to entrap the enzyme.

32

1.4- THEORETIC ASPECTS OF ENZYME-BASED BIOSENSORS

In developing amperometric enzyme-based biosensors the following processes should be taken into consideration: the transport of the analyte to the sensor surface; analyte binding to the redox enzyme incorporated in the sensor; convection of the analyte and electron transfer across the electrode. A theoretical description of each different aspect is useful for experimental data analysis and, after there, for effectively optimising the response of a biosensor.

1.4.1- Mass transport of analytes

In order to be detected by the receptor molecule the analyte must be transported from the bulk solution to the biosensor electrode. It must be continuously replenished to keep steady state conditions. There are three general modes of transport: migration, diffusion and
convection (schematically depicted in figure 1.3).

Migration is the movement of a charged species due to the gradient in the electrical potential. This potential gradient, and therefore the mechanism of migration, can be fortunately neglect in well-defined electrochemical conditions. Biosensor measurements are generally carried out in solutions, by working with an excess of supporting electrolyte which suppresses any potential gradient that might be formed. Diffusion is the spontaneous movement of particles as a result of a concentration gradient, depending on the random motion of the molecules. The formal description of diffusion is given by first and second law of Fick in one dimension [59]:

J = - D (dc

dx

(1.6)

dc

dt

= D d2 c d x2

(1.7)

where J is the net flow of the matter, D the diffusion coefficient [m2 s-1] and dc/dx the concentration gradient. These equations describe the transport mechanism to a biosensor surface at a given rate, because they show respectively the rate of the transport of the analyte

33

towards a face and the exchange of the substrate at the biosensor surface. To solve the second equation, initial and boundary conditions must be defined. An appropriate initial condition is the bulk substrate concentration, indicated with cB. When the enzymatic reaction is sufficiently fast the surface concentration, c0, reduces effectively to zero and this gives a boundary condition. The difference cB-c0 is the concentration gradient leading to mass transport diffusion. As long as there are no other means of transport the reaction rate is under diffusion control. Equation (1.7) can be solved to give the following equation: c = cB erf (x/2 D1/2 t1/2)
(1.8)

in which the bulk concentration cB is also the concentration at infinite distance from the electrode surface and erf the error function. In the treatment of diffusional problem, the error function or integrated normal error curve is often encountered to describe the advancement to a unit limit as the distance reaches infinity [60]. Equation (1.8) indicates the increase of the analytical concentration at the increasing distance from the surface (x in numerator) and the decreasing distance of the reaction time (t in denominator). From the practical point of view, under diffusion control a decaying response of the amperometric biosensor can be expected. The derivative of (1.8) with regard to the distance at x=0 can be substituted in (1.6) to obtain the Cottrell equation which can be used to follow the chrono-amperometric response under diffusion control [61]: iD = n F A D1/2 cB / 1/2 t1/2
(1.9)

where iD is the diffusion limited current, n the number of the electrons, A the electrode area,
D the diffusion coefficient, F the Faraday constant, cB the bulk concentration of the

electroactive species. Convection is the physical transport of particles by a macroscopic movement. Such fluid flow occurs with stirring or flowing solution and with rotation or vibration of the electrode (forced or induced convection) or due to gradient of density (natural convection). Applying convection in many cases is very favourable for biosensors analysis, because it result in a more stable and reliable response. Convective mass transport is achieved by moving the electrode with respect the solution (e.g. for the rotating disk electrode), or by forcing the solution to move past the electrode (e.g. flow injection systems).

34

Convection keeps the concentration of all species as the bulk concentration value, achieving a quick and stable response rather than a decaying signal. In a very thin layer on the surface, the convection is not possible and mass transfer is controlled by diffusion. The stagnant layer is called the diffusion layer, , where a steady-state concentration profile exists: in this region the surface reaction is in equilibrium with mass transport (see the vertical dotted line on figure 1.3). Biosensors often contain membranes with a well defined thickness. If care is being taken to keep it as small as possible, concentration of polarisation can reach a minimum, simply by applying convection during the measurements.

Fig. 1.3- The three modes of mass transport

35

1.4.2- Binding and conversion of the bioelement

An important step that takes place in biosensor is the recognition of the analyte of substrate (already indicated as S) by the biological receptor molecule (E= enzyme). Formation of the complex ES generally is very fast and limited by diffusion processes. As the receptor is a redox enzyme (oxidase, dehydrogenase, reductase) the complex formation is followed by an oxidation or reduction reaction. In typical amperometric biosensors the enzyme reaction can be simply described by Michaelis-Menten equation, particularly when the concentration of S is sufficiently higher than the concentration of the complex ES. In all the next discussion, experimental conditions have been varied in order to achieve a large excess of substrate concentration.

1.4.3- A kinetic model for the electronic transfer in biological systems

In all living systems electronic transfer or transport plays a vital role. In many biological processes the transport of electrons occurs sequentially in reaction chains that evolve to allow the living organism to capture and utilise energy from the light and food. If the electrons transfer is not specific, the net process will ultimately come to a halt; every individual step in the chain are characterised by a high degree of specificity. From the point of view of the electrochemist, the challenge in studying the various component of these electronic transfer chains lies in the specificity of the interactions between the components. Their redox reactions at clean metal electrode have a commonly relatively slow kinetics, as the electrode surface does not resemble the natural redox partner [62]. For this reason the electrode surfaces modifications have been developed in a variety of ways either with monolayers or multilayers (vide infra). This developing ability to design new electrode surfaces by covering with substrate and to control their chemistry has exciting the possibility to enhance the electrochemistry of the biological redox components and the selectivity of the process. Considering the typical dimension of redox enzymes, the distance between active sites and electrode is an important factor affecting the electronic transfer, provided any additional factor such as its irreversible absorption or denaturation of the electrode surface do not take place. Recently, theoretical studies resulted in a kinetic relationship:

36

' k' = K E (r) exp G RT

(1.10)

where k is the heterogeneous kinetic constant, KE the diffusion constant, the effective frequency along the reaction coordinate, the electron transfer coefficient, dependent on the distance r and G0 the activation energy. In electronic transfer reaction of small molecules

=1 as the electron transfer is adiabatic. In the case of interaction between biological redox
systems and macroscopic electrodes, where the transfer takes place at longer distances, must be take into account this relationship at =1:

= 0 exp (r r 0 )

(1.11)

with r0 is the distance of saturation. A value of around 12 nm-1 in the equation (1.11) was obtained from kinetic of electronic transfer studies on different non-biological reactions, where the distance can be systematically varied. Obviously, it depends upon the material through which the electronic transfer takes place. In biological redox systems additional factor must be considered: firstly, the non-spherical symmetry of the molecules; secondly, preferential pathways of electronic transfer can exist, due to the non-homogeneity of the protein around the active site. Surely, the transfer probability decrease, increasing the distance r. Finally distance affects the dependence of the probability of electronic transfer upon the orientation on the electrode surface. A large amount of papers have been devoted to theoretical studies about the electrochemical modification of electrode surface, the control of the protein orientation on the support by means of interaction between the electrode species and aminoacidic residues of the protein [63].

1.4.4- Mechanisms of electronic transfer

Let us consider the case of an enzyme-based electrode biosensor. Any electrical variation at the support as a result of the biocatalytic process (e.g. O2 depletion, NH3, CO2 formation) provide routes for the electronic transduction of the signal. In the next chapters we will apply the amperometric detection of enzyme-generated products to one of the most used products of biocatalytic reactions, H2O2. The resulting current reflects the turnover rate of the

37

electronic exchange between substrate and biocatalyst and is directly correlated to the substrate concentration. Particularly, the transfer rate depends upon the biosensor features and a measurable signal in amperometric biosensor with a high sensitivity can be obtained, provided electron transfer between electrode and redox enzyme is sufficiently fast. In first approximation the simplest model to describe the electronic transfer is called
direct tunnelling mechanism: no mediators are necessary to transfer the electric charge.

According to the general Marcus theory, the transfer rate between donor-acceptor is completely defined by the following equation [62]:

G 0 + (d d 0) = exp exp 4RT

(1.12)

where G is the free energy variation of the electronic transfer reaction, is the reorganisation energy, d0 is the Van der Waals distance, whereas d is the real distance donoracceptor (i.e. between the enzyme prosthetic group and the electrode), and the electronic coupling coefficient. This relationship shows that the separation distance rules the electronic transfer rate and the value of the coefficient . In fact, the redox centers are often buried inside the cores of the proteins and therefore they are so electronically insulated that it can not be easy to observe a direct electronic transfer across the system electrode-enzyme. That is the case where the separation distance is very large and the redox enzyme lack electrical contact with the support. In order to assess a fast electronic transfer in amperometric biosensors, preferential pathways between the enzyme redox site and the electrode surface must be designed, where the biofunctions of the protein are driven and stimulated by electron transfer or the perturbation originated from the electrical contact is directly used as a transduction signal of biochemical event. In every case, two main parameters are required: Assembly of biomaterial onto the conductive support with a well-defined immobilization method Design of the appropriate electronic communication between support and redox protein.

38

39

CHAPTER 2 FLOW INJECTION ANALYSIS

OVERVIEW OF THE CHAPTER

Flow injection analysis is one of the most powerful analytical tool to monitor automatically a large variety of on-line processes for multi-components systems. In this section the principal advantages of FIA technique have been reviewed. The application of this methodology in biosensing systems has been showed very useful to investigate the main sensor parameters. The chapter aims also to provide a short description of the essential parts of an FIA system. Theoretical considerations about the dependence of the signal upon the experimental parameters, i.e. the diffusion flow rate and the geometric configuration, have been also performed compared to continuous flow techniques.

2.1- General scheme and main components of FIA

Flow injection analysis is a methodological innovation of the common analytical techniques, characterised by its simple basis, relative inexpensive equipment, handy operation and great capability for achieving rapid accurate and precise results. The extreme versatility of this methodology makes it stands out from most new analytical methodologies, that seldom allows their users to apply their ingenuity, and that make use of sophisticated instruments with high cost of purchase and maintenance. The term FIA was coined by Rzicka and Hansen in 1975, studying the first injection system which made use of a hypodermic syringe to inject the sample into a reagent stream [64]. It consists of a continuous flow analytical method, originally designed for the analysis of manually injected samples. Due to their properties it is possible to use small volume of sample with poor reagent consumption. Physical foundation of FIA is related to the concept of dispersion, i.e. the dilution of a volume of sample into a carrier non-segmented stream. The main features of FIA, which

40

permit to distinguish it from the conventional segmented flow analysis (SFA) and high performance chromatography (HPLC) techniques, are: No presence of air bubble in the flow stream (differently to segmented flow system, which is based on the segmentation of the flow just by air bubbles) Direct injection of samples into the stream Partial and controlled dispersion: the uncompleted mixing along the carrier originates a time dependent concentration gradient (the signal) Flexibility of geometric and hydrodynamic conditions Non steady state detection, due to the non-homogeneity of the stream

On figure (2.1) a comparison amongst these three techniques is reported. Inspection reveals that FIA resembles chromatography more than the classical segmented flow methods do. In short, FIA may be considered a hybrid of SFA and HPLC (high performance liquid chromatography) techniques [65].

Sample Liquid reservoir

S FA
Propelling unit

Deareation

Reactor/ Separator

DETECTOR

air

FIA
Sample Propelling unit Injection system

Liquid reservoir

Reactor/ Separator

DETECTOR

HPLC
Liquid reservoir

Propelling unit

Sample Injection system

Column

DETECTOR

Figure 2.1- Schematic diagram of the components and principles of continuous flow analysis, flow injection analysis and liquid chromatography. The brown blocks correspond to the common elements between the three techniques

41

FIA systems do not require the use of certain mechanism for a reproducible aspiration or deareation of the bubbles, do not give rise lag time in recording signals due to the pulsing effect, which segmented technique originates through the presence of air bubbles. Furthermore it is possible to replace the glassware commonly used in other analytical techniques by tubes, pumps and valves, with clear advantages in term of the assay time, minimal contact between operators and toxic reagents, protection of both reagents and samples from contamination of atmospheric gases. In FIA system flow rate are generally lower and the tubing diameter are smaller than in SFA. The sampling time is much longer and the sample volume much larger in SFA, with the consequence that in FIA the sample throughput is higher and the sampling procedure, usually filling of a valve loop, is more reliable and reproducible. Finally, the wash-out cycle is essential in SFA to avoid cross-contamination between samples in different phases, with further complication of the technique, whilst this operation is unnecessary in FIA, since the carrier stream continuously flushes the system. The major difference between HPLC and FIA is their principle, since in chromatography there must be always an interface which affects the separation of a mixture of substances passing through the column. The working pressure is a technical factor responsible for their significant differences: FIA uses normally lower pressures than those used in HPLC, where a dual-piston pump must supply high pressures to overcome the hydrodynamic resistance of packed columns, generally packed with an ion-exchange resin, redox material or an immobilized enzyme, to improve the efficiency of the separation process. Despite the fact some FIA methods have been developed with the aid of HPLC components, FIA systems need not be designed to withstand high pressure, therefore they are much less expensive than HPLC instruments. The scope of application of these techniques are very different, since chromatography remains a separation method, whereas FIA is mainly devoted to determine single species in a large number of samples. The basic scheme of an FIA system consists of four essential components: a) A propelling unit which produces the flow of one or more solutions, either containing reagents or merely acting as the carrier(s).The propelling system must force the carrier stream through the units, in a perfectly reproducible, pulse-free and as constant as possible flow rate. This function can be performed by a peristaltic pump (the most used in FIA applications, up to now), a gas-pressure system or even gravity-based units; b) An injection system which allows the reproducible insertion or introduction of an accurately measured reproducible sample volume into the flow, without stopping it;

42

c) A length of tubing, commonly, and sometimes improperly, called the reactor, along with the transport operation takes place, with or without an additional process. The role of the reactor can be played by a straight, coiled or knotted tube (which may also be packed with inert beads), or by a mixing chamber or a tube packed with a chemically active material, such as redox, ion-exchange resin, immobilised enzyme, etc.; d) A flow cell, accommodated in a detector (which can be a colorimeter, photometer, fluorimeter, potentiometer and mostly amperometer) which transduces some analytic property into a continuous signal to a recorder and/or microcomputer. The flow emerging from the sensing system usually goes to waste, although it sometimes recirculates through the peristaltic pump to achieve constancy in the flow rate, or to allow use of recently developed additional techniques. The automatization of FIA requires the incorporation of a sampling system, a sample withdrawing system (generally involving the use of the peristaltic pump itself) and an electrically controlled injection system working in a co-ordination with the sampler. A micro-computing system with active interface allows the easy programming of this operation.

2.2- Partial dispersion as foundation of FIA

2.2.1- FIA signals

As with the foremost analytical techniques, the response of the FIA detection unit is a transient signal which result from two consecutive processes, one of which time-dependent. In SFA methods, readings are made under stability conditions and the plug passing through the detector is practically homogeneous. Therefore, SFA yields curves with a plateau. Conversely, FIA measurements are made when equilibrium condition has not been reached. A schematic diagram of two recordings in FIA experiments is shown in figure (2.2). These are plots of analytical signal (absorbance, fluorescence, potential, etc.) as a function of time, expressed in seconds or minutes (fast or normal recordings). The record on the right side corresponds the an usual well-established FIA measurement, registered at much lower chart speed than that on the left of the figure. In order to define an FIA recording, at least semiquantitatively, it is necessary to know its essential features:

43

Figure 2.2- Signal recording in FIA measurements

a) peak height, h, which is related to the concentration of the component in the injected sample; the peak area is less used because it requires an integrator analogous to those used in gas chromatography; b) residence time, T, which is defined as the span elapsed from injection till the maximum signal is attained; c) travel time, ta, which is the period between the injection and the start of the signal, i.e. 1 or 2% increase above baseline; d) baseline-to-baseline time, t, defined as the interval between the start of the signal and its return to the baseline. This parameter is an actual measure of the dispersion or dilution of the analyte, for it represents the time taken by the sample to pass through the detector. It should be pointed out that the signals are by no means Gaussian curves, therefore the above mentioned parameters do not completely define the profile of the peak, especially considering its second part of the curve, which distinguishes it from the transient signals originated in other analytical techniques. The dispersion, or dilution, of the sample injected into the stream, which can be expressed by a dilution coefficient D, is given by the position and the shape of the analytical signal. The interest of FIA experiment is focused on determining the influence of every feature of the

44

assembly on the signal to optimise the system. An FIA peak is well characterised, at least quantitatively, by its position, defined by the travel time ta, by its bandwidth (t), and the coordinates of the maximum band (T, Cmax). Peak height h and the return time T give completely indications about two decisive characteristics of the analytical method: sensitivity and sample throughput. It is important to know how both these properties are influenced by the FIA parameters. The derivation of a mathematical expression relating these parameters permits a more accurate optimisation of the system.

2.2.2- Physical expressions in FIA

It is interesting to know the characteristic of the profile as the sample passes through the detection cell. The recorder output from the detector is the representation of the dispersion at a such point and can be used to assess the extent of dispersion. It is widely accepted that the physical basis of FIA has not been accomplished so far, as well as a general expression fully describing the sample concentration (signal) profile as a function of time [66, 67]. Various attempts to derive a general expression C= f(t), relating the profile features (travel, residence time and peak height or area) to the FIA experimental parameters (flow rate, reactor length, etc.), succeeded in empirical relationships. Difficulties involved in theoretical predictions of the behaviour of an injected plug are due to practical factors. First of all, it is not an easy task to define the contribution of different elements such as the injection system, connectors, flow cell, geometry to the dispersion: in fact, the injection is not a mere insertion of a plug into the flow, the reactor is not made simply of a cylinder of a specific size. FIA systems use different channels, requiring the evaluation of the effect of the merging points (geometry, dimensions) and therefore complicates the theoretical model. The mechanisms contributing to the dispersion of the injected sample are: a) Convective transport, under laminar flow conditions. If we consider a tube along which the stream flows, this mechanism yields a parabolic profile with sample molecules having zero velocity at the tube walls and a velocity at the centre of the tube equal to 2/3 of the average velocity.

45

b) Diffusional transport, due to the concentration gradient in the convective transport regime, which gives rise to axial and radial diffusion. The former is due to horizontal gradients of concentration at the leading and trailing edge of the injected zone and it contributes insignificantly to the overall dispersion. The latter, resulting from the concentration differences perpendicularly to the flow direction, gives a strong contribution to the net D. If the flow is considered to be made up of a large number of superimposed fluid cylinder travelling convectively at different speeds, radial diffusion tends to balance the concentration in such a manner that the molecules on the tube wall tend to move to the center and vice-versa. Radial diffusion can be considered to play a similar role to that of air bubbles in SFA, provided that the operational conditions give rise to this effect. c) Density gradient transfer: this mechanism do not contribute significantly to the final dispersion, since density differences between the fluid and the sample zone are generally negligible. The concentration profile depends upon the time, since all the encountered contributions to the various types of transport do depend upon it. The feasibility of achieving the above described signal profiles depends critically both on the reactor length, indicated as L, and on the flow rate, vD. If the residence time is expressed in terms of the theoretical volume of the system and the volumetric flow rate, the following relationship is valid:
2

T mean

reactor volume flowrate

R L vD

(2.1)

The ratio L/vD is usually expressed in cm min ml-1 and can be computed from the mean residence times calculated for each set of conditions. Table (2.1) lists the time parameters for each mode of transport, as well as their L/vD ratio. It can be seen that, under usual FIA conditions range, the transport is governed by a combination of convection and diffusion processes. For very short reactors and/or low flow rates (long T), it is diffusion phenomena which prevail. Several models have been developed to define theoretical principles of FIA and derive a mathematical expression of the form C= f(t) accounting for the physical behaviour of an injected plug. It must be stressed that the signal recorded under usual conditions is not Gaussian and its profile cannot be defined in terms of normal distribution, although several approximations are based on it.

46

Prevailing type of transport

Tmean (sec.) mean residence time

ta (sec.) travel time

L/vD (cm min ml1)

Convective Convective diffusional (usual in conventional FIA) Diffusional

0.25 6.25-37.5

0.12 3.12-18.75

2.1 53-318

50

25

425

Table 2.1- Parameters defining the three general type of dispersion in FIA (travel and residence time refer to R=0.025 cm and D= 1 10-5 cm2 sec-1)

Taylors model

The application of this model under typical FIA experimental conditions has been discussed by several authors [64, 68, 69]. This approach only holds for low flow rates and very long reactors, which help to compensate for radial concentration changes and favour the prevalence of diffusion phenomena. Taylors model is applicable to a Gaussian distribution defined by C =f(t) in a form which depends on the chosen parameters. This model is only applicable if the injected volume is negligible compared to the reactor volume (Vi<<<Vr). The unsuitability of the Taylor model has been irrefutably demonstrated by Painton and Mottola by comparing the concentration-time curves simulated by a computer program based on this model with those obtained under usual FIA working conditions. From their analysis it could be concluded that the lower the ratio between length and flow-rate, the greater the differences between the theoretical and experimental curves. This difference raises for both the maximum and the tailing portion of the signal.

Tanks-in-series model

This model was firstly discussed by Rzicka and coworkers [64], analogously to the description of liquid chromatography in terms of theoretical plates. It relies on the

47

assumption that the fluid flow passes sequentially through a large number of micro-chambers in which stirring is instantaneous. There is no practical difference between the application of the Taylor and the tanks-in-series model for large value of N, but the suitability of this model is questionable for small N values, i.e. for not long reactors. It has been showed to be closer to the experimental facts than that of Taylor in many ordinary FIA systems [71].
The most general model

The expression that takes into account both convective and diffusional transport, best describing the overall physic al dispersion phenomena, is a differential mathematical model. It can be derived applying the balance of mass to a differential element of volume in a fluid and takes into consideration axial and radial concentration gradient, as well as flow-profiles under a laminar flow regime. There have been various attempts to provide numerical solutions to the second-order differential equation defining the dispersion, but this is neither easy nor rapid. Ananthakrishnan et al. [72] have used the alternating direction mathematical method, which consists of finite difference approximation, but applied to a single interface between carrier and injection sample. Vanderslice has adapted this contribution to the double carrier-sample-carrier interface. Painton et al. [70] have applied a mathematics solution to a computational program in order to simulate dispersion curves of the form C= f(t); the likeness to experimental curves is more satisfactory than that found applying Taylors Gaussian model, even for low residence times (short length and relatively low flow-rates). The tailing portion of the simulated curve differs from that of experimental curves, for various equivalent lengths of the system (155 cm <L<400 cm) and keeping as constants radius (0.025 cm), flow rate (0.86 ml min-1) and injected volume. The fall of the experimental curves is more gradual and the tail are, therefore, more pronounced than predicted. This divergence apart, both sets of curves are practically coincident and can be attributed to a departure of experimental conditions from the ideal due to effects from connectors, injection of sample, shape of the detection flow cell, etc., all of which undoubtedly modify the theoretical model and contribute to the dispersion or dilution of the trailing portion of the injected plug, thus delaying its appearance at the detection point. Several correction and accommodation factors according to various approaches have introduced into the equations to allow the adaptation of the simplified theoretical expression to the experimental facts [73].

48

2.2.3- Influence of various factors on the dispersion

The dispersion coefficient, also called dilution factor D, is the earliest parameter used to characterise passage of the sample through the system. It was proposed by Rzicka et al. in 1977 [74], to measure the extent of dilution undergone of a particular part of sample zone between injection and detection. It is defined as the ratio of the concentration before (C0) and after transport through a given FIA system:
D = C0 /C (2.2)

Thus each point on the signal record has a corresponding dispersion coefficient. At the signal maximum,
Dmax = C0 /Cmax (2.3)

Since there is a direct relationship between the property used for detection (potential, pH, fluorescence, absorbance), the magnitude of the transducer signal recorded and the concentration of the sample or its reaction product, D can be determined from the ratio of the height of the signal found in absence of dispersion (continuous flow of undiluted samples) to that of the FIA signal. The experimental procedure for determine D in an elementary FIA system was clearly showed in Valcarcel et al. [75]. From its definition it follows that D always exceeds unity. Quantitatively, the greater the dilution, the flatter the resulting curve: in routine FIA Dmax ranges from 15 and 1 and depends upon the hydrodynamics and geometry of the system. It can be compared with sensitivity and sampling rate in analytical methodology: the larger D, the less intense the signal and the broader the curves recorded, and hence, the smaller the number of the bands obtained per unit time (sample throughput). The overall dispersion is the result of the dispersions originating in the three main part of the FIA system, i.e. the dispersion due to the sample injection, to the transport and to the detector:
D = Dinjec + Dtransport + Ddetector (2.4)

where Dinjec includes the contribution of the sample, Dtransport (the most significant to the overall dilution) that of the reactor geometry and flow rate, Ddetector denotes the contribution of the flow cell geometry (shape and dimension) to the dilution. The last term has been

49

thoroughly studied in HPLC. They all include also contribution from factors, such as dead volumes in connectors. The experimental parameters have been classified to whether they are representative of the sample, geometry of the system or hydrodynamic working conditions. A more detailed study of the influence of sample volume, hydrodynamic and geometric factors on the dilution (D) is reported in different text and reviews [64, 70-73]. Most researcher attempted to make a FIA classification according to the value of dilution factor. -FIA with limited dispersion (D<3) are used when measuring a parameter of the analyte directly in absence of processes other than transport. Since the residence time is rather short, convection will prevail over the radial diffusion. Therefore, to minimize D these system must adopt either high flow-rates (>2 mil min-1) or large sample volumes or small reactor length and diameter. Application of these condition affords high sampling rates and better analytical sensitivity. Typically an atomic absorption, spectrometer or an electrochemical flow cell for amperometric o potentiometric sensing are used as detector in these systems: -FIA system with medium dispersion (3<D<10-15) are designed for studying processes such as chemical reaction in addition to the transport. It should be necessary to mix partially the injected sample zone with the stream containing the reagent, by using long tubes, low flowrates, several channel with various mixing point, etc. Detection is usually photometric, fluorimetric or by flow cell. Sample throughput and sensitivity are lower than the previous class of systems. -FIA system with large dispersion (D>10-15) are characterised by the high degree of mixing between the carrier reagent and the sample, resulting in a well-defined concentration gradient. The residence time is so long that chemical equilibrium can also be attained. These system incorporates a mixing minichamber with stirrers or very long reactor tubes. It is not common in FIA conditions, excepting FIA titration or special gradient modes of operations. FIA set-up developed for this thesis falls into the first category. The tendency to decrease the dilution factor is strongly useful to have a relatively good sensitivity in the analytical detection for biosensing devices.

50

51

CHAPTER 3 - SUPPORTS FOR IMMOBILIZATION

OVERVIEW OF THE CHAPTER

Amongst the large variety of support matrices available in literature for the immobilization procedures, in this thesis four different systems have been principally used. 1- Porous nylon membranes for enzyme entrapment; 2- Controlled porosity glasses (covalent interaction); 3- Electro-synthetized nanoporous alumina membranes; 4- Q-dots cadmium sulphide nanoclusters. Semiconductor nanoparticles layers/bioreceptor combined structures. Special attention have been devoted to the first two supports, which were showed as good matrices for immobilising flavin oxidase enzymes. Due to the high stability and the strong interaction of the enzymes immobilized into these systems, amperometric biosensor devices were successfully designed for glucose, sarcosine and creatine, creatinine detection.

3.1- Pre-activated immunodyne membranes

Highly porous nylon membranes are special organic support with specific group for immobilization, by means of covalent interactions. They normally consist of a polymeric structure with outer functional groups that assess a covalent bonding with the protein in addition to the simple physical entrapment into the pores. Pre-activation by organic functional group renders hydrophilic their surface for the electrostatic interaction with the polar group of enzymes that do not belong to redox active sites. Immunodyne (ID) membranes are often used as versatile solid phase supports for covalent immobilization of polyclonal and monoclonal antibodies, enzymes, purified peptide antigens and amino-terminated oligonucleotides. They are well suited also for in vitro immuno-diagnostic, immersion or flow-through dot immunoassays, wet and dry chemistry enzymatic tests, protein assays and common biosensing measurements. Immobilization procedures for proteins and biologically active macromolecules simply require contacting

52

the pre-activated membrane with a solution of the desired protein. Spot wetting or line coating of solution onto membrane, or immersing simply the membrane into the solution best accomplishes this. The membrane may also be pre-assembled into suitable devices to facilitate either spot wetting or flow-through of immobilization solution. Pre-activation result in a covalent linkage with nucleophilic groups on proteins or enzymes. These methods can be readily scaled-up for commercial processing and manufacture. Immunodyne ABC membranes are commercially available in a range of pore sized from 0.45 m to 5.0 m. This allows optimal selection for specific applications, depending on the binding area, flow rate, diffusion time and assay sensitivity. The microporous structure of immunodyne membranes provides an available immobilization area of 300 cm2 for each cm2 of planar membrane. They possess the following main chemico-physical characteristics: a) High binding capacity: the strong attraction of proteins to immunodyne surface enables the immobilization of more protein ligands in comparison with the traditional tubes, multi-plate wells, latex or polystyrene beads previously used for biochemical immobilization; b) Mechanical durability: the nylon 6,6 membrane structure is formed around a nonreactive, non-woven, polyester fabric matrix which confers high tensile strength, toughness, flexibility and resistance to tearing, cracking and puncture; c) Chemical reactivity: an improved stability of enzyme is provided by the covalent linkage activity of the ID membrane surface; primary reactivity is showed with amine groups at pH around 7; d) Controlled porous structure: their narrow size distribution ensures uniform binding of enzymes across the membrane surface. The 75-85% void volume provides for minimal flow resistance and high diffusion rates. Uniform porosity yields consistent, well-defined spot geometry for dot immunoassays. Membrane flow features are also consistent from spot-to-spot and roll-to-roll to ensure reproducibility of kinetic data; e) Hydrophilicity: they exhibit instantaneous wetting on contact with low-ionic strength aqueous solutions. High capillarity and rapid absorbent wicking are obtainable without surface modifications by wetting agents. Pre-wetting is not required prior to contact with enzyme solution; f) Visible surface properties: the white surface, with a nominal narrow thickness of 150 m enables an excellent colour discrimination and sensitive visual detection of dot

53

immobilization and transfer; in fact, the uniform texture is ideal for optical detection systems; g) Dimensional and chemical stability: expansion of ID membrane on initial exposure to water is less than 0.3% with no further changes on prolonged fluid exposure. Less than 0.3% contraction occurs on drying. They are unaffected by a wide range of reagents, solvent, under different experimental conditions. In most cases the conditions to keep the biochemical enzyme activity are far more restrictive than those to keep the integrity of the membrane itself. Once formed, covalent linkage between enzyme and ID membrane is highly resistant to dissociation and hydrolysis. Due to their chemical, mechanical stability and versatility, immunodyne membranes will be principally used in the following chapters as solid porous support for immobilising flavine dependent oxidases, specifically glucose oxidase and sarcosine oxidase. Thereafter the enzyme-based membrane will be deposited onto an indicator electrode for electro-analytical, amperometric detection of glucose and sarcosine.

54

3.2- Nanoporous alumina: application as membrane

Alumina films have attracted much interest as a key material in the fabrications of several functional devices at nanometer-scale composition, due to their typical micro and nanoporous structure [76]. It consists of an ordered and packed array of columnar hexagonal cells at the centre of which a straight hole is located. Its nano-pores diameter normally range from 200 to 1 nm with an ideally cylindrical shape (high aspect ratio) and a narrow size distribution. They aligned parallel to each other and normal to the surface [77], even though its real structure may present many irregularities and imperfection and its degree of periodicity is normally very low [78]. The dimension of the layers depends on the experimental conditions of the electrosynthesis (T, Vbias, treatment time, electrolytic solution), adjusting that, tailored porous structures can be obtained. For example the interval of pores increase in proportion to the applied voltage of anodization, whilst the pores depth in proportion to the anodization time [79]. Porous alumina films with ordered cells have been used in many fields of science and technology (as filters, catalysts, recording media, 2D photonic band-gap structures, functional electrodes), but some application has been restricted to specific uses because of their low stability in aqueous solution and insufficiency of mechanical strength [80]. In fact, the anodic film often comprises amorphous alumina which in water solvent can swell by hydration, resulting in pores plugging. In order to overcome this disadvantage, chemically stable porous films with other material had also obtained [81 and ref. therein] by using a two-step replicating process. This synthetic route consists of the first formation of a duplicated negative-type of metal and the sequential formation of positive microporous metal, after the alumina template dissolution. Masuda and co-workers [82] reported the process of microporous Au films, using a negative type prepared by electrochemical deposition into micropores and eliminating metallic dendritic residues by Ar etching. However, the non-uniformity of ion etching influenced the bad uniformity over a large area of the negative metal. They improved this uniformity of the Au film, adopting a procedure in which the negative-type structure comprising an array of metal cylinder was first fabricated and then the positive-type was reproduced onto it.

55

The first step was the anode oxidation of a well-polished Al plate and a current recovery treatment with a stepwise voltage decrease, in order to deposit uniformly the metal into the pores. This stepwise change of voltage resulted in a dendritic pore formation on the bottom of the structure not useful for the fabrication of positive type. Thus, this part must be removed to obtain a straight cylinder. Rather than using the non-uniform sputtering methods, the dendrites could be eliminated by selectively dissolving two kinds of metals in the pores, the first one on the bottom (e.g. electrodeposition of Fe or controlled deposition of Pd catalyst) and the second one on the top (electroless iron catalysed deposition of Ni or polymer filling). Thereafter the alumina and the metal which filled the dendritic part of the pores were removed, to leave the negative type. The last step was the electrochemical deposition of positive type (e.g. gold) into the cylindrical structure of negative type metal. In other works [83] anodic porous alumina was used as evaporation mask to the fabrication of nano-dots array: for instance gold can be deposited onto a Si substrate by a two-step anodization of alumina, to improve regularity of the cavities configuration and by subsequent etching process to remove alumina layers. Finally, gold was evaporated on a suitable substrate, using alumina as a mask adhered to Si in an electron-beam evaporation apparatus, under high vacuum conditions (5 10-5 torr). SEM measurements showed also in this case a typical hexagonal texture well correspondent to that in the holes of the mask [8284]. By this synthetic two-step replication model all the membranes with an identical structure of porous alumina could be fabricated (Au, Ni, Pt, also semiconductors of TiO2 by means of electrodeposition of TiCl3 in acid solution with NaHCO3). Hence these systems can be applied as electrode in many electrochemical applications in which a textured morphology of the surface is required. Important results were obtained in the field of the enzyme immobilization into the micropores of anodic alumina: M. Maida and co-workers have investigated the behaviour of GOd directly entrapped into microporous alumina [85], without improving the sensor stability. In order to have a much high performance of electrochemical sensors, conductive porous electrode with a large surface area are required. To this purpose, immobilization of GOd into Pt electrode micropores with a texture surface periodicity was examined, by using an H2O2 amperometric detector [86]. It was also possible to immobilise enzymes in the inner wall of microtubule or micropores of membranes, after the removal of alumina, used as masking system [87], for creating biosensing devices with a relatively long term stability of the bioreceptor. Up to now, using template systems as support for bioreceptor immobilization is only a challenging way to be

56

explored. In this work an electrochemical synthesis route to prepare nanoporous alumina membranes by anodization, and the following enzyme immobilization into their pores have been tempted.

3.3- Controlled porous glasses for covalent immobilization

By interfacing biosensor system with flow injection analysis, the rapid and selective enzymatic detection of low molecular analytes can be automated and performed with high precision. From the analytical and economical point of view the immobilization plays the key role, in improving significantly even the operational stability of biosensors. The use of controlled pore glasses (CPGs) has been showed as a powerful method to immobilise covalently one o more enzymes. Hydrolases, dehydrogenase and oxidase have been already experimented as enzymes for CPGs-dependent biosensing devices [88]. Trisoperl CPGs purchased from Schuller GmbH (Steinach, Germany) as carrier agents has proven very useful in our experiments of immobilization to construct a multi-enzymatic reactor. They consist of porous amino-silylated glasses with small spherical particles fused together in such a way as to form small channels, as depicted in figure 3.1. The support is normally sold in bead form with a well known micro-architecture and different available particle size (in the M range), and it is durable and resistant to solvent distortion. The enzyme is normally immobilized by means of totally covalent bond formation or hydrogen bonds between its non-prosthetic groups and chemically activated hydroxyl groups on the surface of CPGs.

57

Figure 3.1- Schematic picture of the micro-structured Trisoperl controlled pore glasses (CPGs)

Trisoperl controlled porous glasses (CPGs) have been used in biosensing purposes for two principal reasons: a) The much higher specific surface, compared to the other carrier types, allows for a very high loading of immobilised enzymes; thus a high long term operational stability of the biosensor should be achieved; b) Their spherical shape reduces the back-pressure which can take place in packed enzyme cartridges, especially designed in analysis under flow conditions. This allows for a longer using of the cartridge, and a carrier material made of smaller particle size [88, 89].

58

3.4- Nanoclusters of semiconductors CdS

3.4.1- Quantum dot semiconductor properties

Nanostructured materials have proven as one of the most powerful tool in new trend of technology and research, due to their absolutely peculiar properties at nanometer size scale. Many studies have shown that there exists a gradual transition from semiconductor or metal bulk properties to molecular properties, as the size of a crystal is successively decreased in the nanometer range. In this transition region, optical, mechanical, photo-catalytic and transport properties of semiconductors drastically changes [90]. Nanomaterials properties are due to the following effects: 1. Geometric effect: the surface features have a strong influence as the mean diameter of the particles is in the exciton size regime (i.e. < 10 nm). This means that the number of the surface atoms is of the same order of magnitude of that of the atoms inside the crystal. 2. Electrical/quantum size effect: the energetic band structures of bulk solids vary as the mean dimension decrease. Particularly by decreasing their size, a gradually variation from the typical non-structured band-edge absorption to the molecular (excitonic) bands of nanocluster, as well as a blue-shift of the absorption onset towards lower wavelengths is observed. Many properties of the nanocluster semiconductors find applications in luminescence, nonlinear optic, catalysis, optoelectronics and light energy conversion fields. Recently, much interest have been devoted to redox photo-catalysts for interfacial reactions. Peculiar photoelectronic properties of nanocluster-modified biomaterials could successfully develop new optoelectronic devices, as new photocells to convert the light energy in electric energy or as optoelectronic sensor. Even though in the early stage of preparation, especially from the point of view of the scientific results, semiconductors can be exploited also in biosensor development. They may improve the sensor performances, especially electrochemical sensor, or replace the function of other fundamental components of the system. New combined structures made of nanoclusters along with enzymes have been already designed: relevant examples exist for the photo-formation of hydrogen, in NAD(P)+ photoreduction, in enzyme catalysed photosynthesis activated by semiconductors, for photo-

59

catalytic reduction of CO2 by means of surface modified CdS, lactic acid semiconductorphoto-catalysed conversion, or ZnS photo-catalysed reduction of CO2 to methanol in presence of methanol dehydrogenase. In such systems electron and holes generated respectively in the conduction and valence bands can be used by the enzyme, for the substrate reduction (oxidation) through electron acceptor (donor). The electronic transfer mechanism of hole and electron towards the interface enzyme/nanocluster depend strongly upon the redox potential of semiconductor, the enzymatic reaction and upon the presence of a substance which can be reversibly oxidised and reduced. In low dimensional regime semiconductors, the dependence of their properties on the diameter size is visible. In the case of CdS, band gap energy shift from 2.5 to 4.5 eV, fusion temperature change from 1600 to almost 400 C, and pressure change of phase transition can be observed as the particles size are in the excitonic dimensional regime. Furthermore, a large energy variation is required to add or remove a charge onto nanocrystals, so that even electric charges transport strongly depends upon the nanocrystal size [91-95]. Tuning fundamental properties can be made without altering their intrinsic chemical composition, thus in principle it should be also possible to obtain nanomaterials with the same chemical composition and different electrochemical potential, simply by modulating their size. Nanoparticles exhibit also lower ionisation potential as particle size is decreased beyond the threshold for quantum effect. Hence, photo-catalytic redox potentials of electron and hole can be tuned controlling the particle size, to achieve increased redox power, for selective chemical reaction. Moreover semiconductor nanoclusters that provide a high surface area to volume ratios improve the efficiency of most photocatalytic processes. Nanoparticles architectures on electrode support attracted research efforts towards the development of electronic devices (for example semiconductor arrays assembled on electrodes as active interfaces for photocurrent generation). The synergic use of both semiconductive nano-sized material and biomolecular units in a correct orientation, opened the way to interesting perspectives in the fast and continuous growing field of science, namely bioelectronic. Fundamental features of a bioelectronic device are the immobilization of a biomaterial (such as proteins, DNA fragments, antigenes or antibodies, cofactors, or low molecular weight species that have affinity with other biomaterials) on to a semi-conductive matrix and the transduction of the function associated with the biological element. One of the most recent project in biosensors and bioelectronic is using multi-enzymatic arrays as bio-active interfaces, in order to improve the sensor stability and sensitivity. This requires the development and the modeling of efficient integrated

60

systems made of biomaterials and conductor matrices, where electronic transduction of enzyme-substrate interaction is the common analytical mean of detection. So far, chemical methods of assembling support with biomaterial make use of bridgemolecules such as polymers, membranes or sol-gels materials. Extensive research has been devoted to the surface functionalization, by monolayers of organic species or chemicophysical modification or biomaterials addition through grafting procedure, to improve the immobilization and the device stability. The concept of self assembling monolayers has been extended to insert metal and/or nanosized semiconductors onto the surfaces: Sampath et al. studied the affinity in the interaction between biomaterial and metallic nanoparticles, in order to design three dimensional super cross-linked structures on transducer elements as news sensing matrices [96]. Another relevant example of combined structures has been the application of fluorescent semiconductor and bioreceptor as biological labeling [91]. We attempted to use CdS nanosized (II-VI) semiconductors as new charge transfer agents in cofactor dependent enzyme combined systems, that consist of nanosized semiconductors and enzymes, as a base for a new third generation biosensor. Semiconductor would shuttle the electron towards or from the electrode surface, after suitable photoactivation. Enzymes must be used to catalyse with high selectivity and specificity a large amount of redox processes photosensitised by an inorganic semiconductor. Particularly formaldehyde detection by formaldehyde dehydrogenase enzyme will be attempted, for its application as a component in food and agricultural applications. The basic parameters which must be kept into account to successfully construct enzymebased nanoclusters electrode in amperometric biosensors are: a) The electrochemical potential modulation of nanoclusters, hence their size modulation; b) the capability to not denaturate the enzyme after its immobilization in a biosensing device.

3.4.2- Quantum dot semiconductor synthesis

A large variety of synthetic methods have been carried out in nanomaterial preparation, including typical chemical techniques and physical preparation methods, such as metallorganic chemical vapour deposition, molecular beam epitaxy, nanolithography. The

61

problem can be solved designing nanometric structures with the so-called top-down approach, starting from larger template objects to tailor the final structure. The opposite procedure, often preferred by chemists, is called bottom-up method, consisting of assembling nanosized systems atom by atom, from the smallest particle. Enormous improvements has been made in preparing and characterising semiconductors of II-VI, III-V and IV type, with or without many metals, by means of host-guest inclusion chemistry or colloidal synthesis. Recently new synthetic methods have been developed, based on thermal decomposition of precursors in co-ordinating solvent at relatively high temperature [93]. Nanostructured CdS preparation in quaternary water in oil microemulsive systems, characterisation, and immobilization on conductive supports, has been showed as one of the most simple and successful synthetic method, as extensively reported in several papers [97-99]. Without discussing extensively about the synthesis and characterization methodologies, it should be sufficient to stress the high capability to control and regulate at will the particle size according to the reverse micelles approach, tuning even the semiconductor nanoparticles properties.

62

63

CHAPTER 4 - ENZYMES CHARACTERIZATION

OVERVIEW OF THE CHAPTER

The principal features of two different flavin-dependent enzymes (glucose oxidase, sarcosine oxidase) have been reviewed, in order to make a complete understanding of the electron transfer mechanisms between the redox enzymatic groups and conductive supports used in our applications. The case of NAD-dependent formaldehyde dehydrogenase (FDH) has been also considered: enzymatic assays were performed about the FDH enzyme both in aqueous and micro-emulsive solutions. Enzymatic activity has been tested in absence and in presence of Q-dots CdS semiconductor, aiming to combine directly the peculiar nanostructured semiconductors properties with the catalytic power of the enzymes, for future combined third generation biosensing application.

4.1- Glucose oxidase: structure and general function

Glucose oxidase from Aspergillus niger (-D-glucose: O2 oxidoreductase, E.C. 1.1.3.4) is one the redox enzyme that is used as the selector in this thesis. Before oxygen played a crucial role in life, bio-electrochemical processes took place in the narrow potential range of 700 mV. When oxygen became available as an energy source, this range extended to 1300 mV and enzyme containing flavin could evolve. Glucose oxidase (GOd) belong to the family of flavin-containing enzymes. It is an aerobic dehydrogenase, which catalyses the oxidation of -D-glucose to D-gluconolactone. Under natural conditions, this oxidation is accompanied by the reduction of molecular oxygen to hydrogen peroxide (see reaction 1.1). The flavin group is strongly associated or bound to the protein. The cofactor, flavin adenine dinucleotide (FAD), also called prosthetic group, is responsible for the enzyme properties without leaking from it at any time during the catalytic cycle. Glucose oxidase contains two

64

FAD-units per molecule [100], that are firmly, but not covalently, bound to the peptide chain. The iso-alloxazine system of FAD can exist in three redox state: the oxidised one, which is showed in figure (4.1), and in two reduced states which are attained respectively by the uptake of one or two electrons.

Figure 4.1- FAD with its iso-alloxazine ring system in the oxidized state

Despite the fact that this enzyme is abundantly used, there is very little structural information available on it. As with most flavin enzymes [101] no X-ray structure has yet been solved, for the large carbohydrate content (16%) of the native molecule, which prevents crystallisation of the enzyme [102]. It is know that the flavin prosthetic group is anchored in a small slit, formed by the protein structure and therefore is only partially accessible to water [103]. Attempts to accomplish electronic interaction with this redox protein and location of the prosthetic group inside the protein molecule should be taken into consideration. Glucose oxidase has a high affinity for molecular oxygen [104], which will always be a competing electron acceptor when present during experiments in first generation biosensors. Careful design of experimental analysis, where a direct communication is tested, should be, therefore, necessary. In table (4.1) some physical and chemical parameters of glucose oxidase are presented [103-105]

65

Parameter

Value

Molecular weight Number of subunits Co-factor Carbohydrate content Isoelectric point Electrophoretic mobility Diffusion coefficient Turnover number Michaelis-Menten constant (glucose) Michaelis-Menten constant (O2)

150,000 dalton 2 2 FAD 16 % 4.5 2.2 10-5 cm2 s-1 v-1 4.12 10-7 cm2 s-1 500 s-1 140 mM (the highest reported in literature) 0.33 mM

Table 4.1- Some physico-chemical properties of glucose oxidase from aspergillus niger

4.2- Sarcosine oxidase: structure, function and application to the creatine and creatinine determination

Phosphocreatine plays a unique role in the energetic of muscle and other excitable tissues, such as brain and nerve. This compound serves as a reservoir of high energy phosphate groups and can transfer them to ATP in a reaction catalyzed by creatine kinase. Creatine can be synthetized in liver and kidney from amino acids such as argirine, glycine, and methionine that transfer amidine group through a complex reaction sequence [106]. Most of creatine is transported to muscle through vein and then phosphorylated to form phosphocreatine. Both the compounds are converted to their anhydride, whereas creatinine diffuse non-enzymatically from the muscles. The serum creatinine composition varies inversely with the glomerular filtration rate, and therefore its concentration is widely used as an indicator of renal function. On the other hand, much of creatine is reabsorbed renally and it should not secreted in urine under normal condition. Determination of the creatinine levels

66

in human blood serum and creatine levels in skeletal muscle have been, therefore, widely used for laboratory diagnosis of renal clearance and muscular function [107, 108]. Sarcosine oxidase plays a critical role in the most reliable enzymatic methods for creatine and creatinine determination. The combined use of a multiple enzymatic system became possible when bacterial sarcosine oxidase was found to be commercially available [109]. These methods are based on the following principle to obtain a precise and specific determination of creatine in serum and urine: the degradation of creatinine and the resulting by-products is sequentially catalyzed by creatininase (CA, also called creatinine amidohydrolase), creatinase (CI also called creatinine amidino-hydrolase), and sarcosine oxidase, according to the cycle of reactions (4.14.3):
CA creatinine + H 2 O creatine

(4.1)

creatine + H 2 O CI sarcosine + urea

sarc. oxidase

(4.2)

sarcosine + H2 O + O2 glycine + H 2 O2 + HCHO

(4.3)

Hydrogen peroxide produced can be detected amperometrically by its reduction at a working electrode [110], or spectrophotometrically by the peroxidase-coupled color development; also formaldehyde produced can be detected by NAD-dependent enzymatic oxidation. In a following chapter an up-to date overview of the main methods to determine creatine content will be explored. Sarcosine oxidase catalyses the oxidative demethylation of sarcosine (reaction 4.3) using molecular oxygen as an electron acceptor, similarly to the behaviour of sarcosine dehydrogenase. Involvement of sarcosine oxidase in the metabolism of creatinine was clearly shown in specialised literature [111]. Sarcosine oxidising enzymes so far studied contain covalently bound flavin, which has been showed to be linked via the 8 -position of the isoalloxazine ring to the imidazole N(3) of the histidine residues (see the first structure of the following figure) [112].

67

Figure 4.2- Chemical structure of flavin-amino acid bonds found in flavoproteins, used in oxidase enzyme [113]

As table (4.2) shows, the sarcosine oxidases can be classified into three different groups, referring to their sub-unit composition, that is, the monomer, heterodimer, and heterotetramer. They all contain a polypeptide of approximate molecular weight of 50,000 Dalton; however, their primary sequence was observed among different groups, since no homologous sequence was observed between the Corynebacterium type sub-unit B and the complete primary structure of Bacillus enzyme.

68

Sarcosine oxidase

Molecular weight (Mr)

Prosthetic group

KM (mM) Michaelis-Menten

Reference

Cylindrocarpon didum M-1 Bacillus sp. B-0618 Bacillus sp. NS-129

45,000 (monomer) 42,000 (monomer) 42,955 (387 amino acid residues)

Covalent FAD Covalent flavin Covalent FAD

1.8 12.2 N.E.

[114] [115] [116]

Streptomyces KB210-8SY Alcaligenes denitrificans

spec.

44,000 (monomer) 190,000 (hetero-dimer) 174,000 (hetero-tetramer) 185,000 (hetero-tetramer)

-------Flavin Covalent FAD Non covalent FAD Covalent flavin Noncovalent flavin

0.91 4.2 3.4 6.4

[117] [118] [119] [120]

Corynebacterium Arthrobacter denitrificans

Table 4.2- Different sarcosine oxidising systems; biochemical and physical features

Sarcosine oxidase gene from Bacillus sp. NS-129 was cloned for the first time by Koyama et al. and the nucleotide sequence of the gene was determined along with the amino acid sequence [121]. The monomeric enzyme is composed of 387 amino acid residues and contains the covalently bound flavin adenine di-nucleotide (FAD), although the amino acid residue at the binding site is not yet determined. This is the only enzyme that a complete sequence of sarcosine oxidase has been published in the journal so far. Amongst the different sarcosine oxidases, Corynebacterium enzyme has been the most extensive studied [122]: it contains both non covalently and covalently bound FADs with a twofold functionality, namely dehydrogenase flavin as non-covalent function and oxidase as covalent one. Conversely, sarcosine oxidase from Bacillus species has been preferred in the following experiments for its simplicity to be solubilized in phosphate buffer and its availability on the biochemistry market.

69

4.3- Kinetic characterisation of formaldehyde dehydrogenase (FDH)

4.3.1- Chemico-Physical properties of formaldehyde dehydrogenase

Formaldehyde dehydrogenase (FDH) from Pseudomonas putida was selected as cofactor dependent enzyme in this part of thesis, for the basic construction of hybrid systems made of an inorganic semiconductor and an enzyme. It is also called NAD (nicotine amide adenine dinucleotide) dependent formaldehyde dehydrogenase, belonging to the oxido-reductase enzymatic family. It catalyses the oxidation of formaldehyde to formic acid in presence of the stoichiometric reduction of NAD+, according to a sort of ping pong mechanism:
FDH HCOOH + NADH + H + (4.4) H 2 CO + NA D + + H 2 O

In this ping pong reaction the product formation takes place before that every substrate can be bound to the enzyme. Nowadays the correct reaction sequence is under intensive studies. It may possible that hydrated form of formaldehyde would be bound firstly to the enzyme and then formic acid would be produced; in the following step NAD+ bound to the reduced form of enzyme would be reduced. Formaldehyde dehydrogenase from Pseudomonas putida, as well as that derived from Methylophilus methylotrophus or Arthrobacter bacteria, is not highly specific to formaldehyde, due to its sensitive activity with respect to other aldehydes with longer aliphatic chain. It is know that FDH activity decreases by increasing the length of the alkyl chain of the substrate, up to be deactivated with butyric aldehyde. Enzyme activity can be influenced by inhibitors that either bond or form complexing systems to the redox site and reduce the number of enzymatic turnover, i.e. the substrate amount which is consumed in one second by an enzyme. FDH is strongly inhibited by several substances, such as: Ni2+, Pd2+, Hg2+, pchlorine-mercury benzoate, 2,2-bipyridine, phenyl methane sulphonil fluoride, H2O2, whilst it is partially inhibited by Zn2+, Mn2+, EDTA. Structural characterisation showed that FDH is a dimer of two sub-units of 75,000 Dalton in molecular weight. Every sub-units contains two atoms of zinc, one of which takes part to the catalytic activity, the other one keeps the native conformation of enzyme. NAD+ as coenzyme is an organic molecule which act as principal electron acceptor in biochemical oxidation [123a, b]. The reactive part is the nicotine-amide ring, which capture a proton and

70

two electrons to give NADH. Coenzymes are chemically modified in the reaction where they take part, in order to complete the catalytic cycle. In the case of transitorily bound cofactors, such as the case of NAD+, the regeneration process can be enhanced by a different enzyme.

4.3.2- Reactivity of NAD dependent FDH

The principal aim to use FDH with Q-dot nanostructured CdS was to successfully replace the NAD+/NADH nucleotidic cofactor, by exploiting the redox properties of the semiconductor as charge carrier. In fact, the redox active cofactor plays a central role as electron relay component in the electron transfer chain between the biochemical active site and conductive support. NAD cofactor is biologically active by diffusional route and its reduction to NADH occurs by temporary association in the vicinity of the active site. The NAD+/NADH couple exhibits poor electrochemical oxidation of NADH and the electro-reduction of NAD+ are kinetically unfavoured. The most redox enzymes NAD+/NADH dependent reveal fundamental difficulties due to their poor stability, high costs and electrical contacting with electrode that prevent their broad use in biosensing devices. Hence, semiconductor nanocrystals represent a possible route to achieving an increased electron/hole transfer efficiency, overcoming the problems due to the NAD+/NADH dependence. Firstly, FDH enzyme have been tested in its enzymatic activity at the right pH and temperature condition. Using the classic Michaelis-Menten equation (see eq. 1.4 in the first chapter) the velocity was followed as a function of [S], obtaining directly KM and vmax either by using exponential fitting programs, or the linear form of Lineweaver-Burk equation. In the case of FDH, the coenzyme reduction was measured following the spectrophotometric variation at 340 nm. For the enzymatic assay nine points were collected, corresponding to different formaldehyde concentration from 10 to 1000M, keeping constant the -NAD composition. The initial velocity v0 can be derived from the slope of the experimental absorbance as a function of the time, considering the well-known LambertBeer law: A=lc
(4.5)

71

Fig.4.3- Initial v0 determination by Lineweaver Burk equation

where l= 1 cm is the cuvette thickness, c the substrate concentration, NAD= 6.23 103 cm2/mol is the molar extinction coefficient of NAD. In figure (4.3) kinetic curve of 200M formaldehyde concentration with the corresponding fitting shows a linear behaviour up to the first 80/100 seconds. Over this value a deviation from the linearity takes place. FDH have been characterized in different systems to verify its real efficiency and activity. Measurements have been carried out in: 1. aqueous solution varying the substrate concentration and keeping as constant the NAD+ concentration; 2. in ternary microemulsion system made of AOT-isooctane-water; 3. in quaternary micro-emulsion: CTAB/n-hexane/pentanol/water (primary system for biosensing applications); 4. in quaternary micro-emulsion: CTAB/n-hexane/pentanol/water with the addition of synthetized nanoclusters CdS, according to the methodology in reverse micelles [118120].

72

4.3.3- Enzyme activity in aqueous and microemulsive solutions: experimental results

Two series of enzymatic kinetic at the same experimental condition were followed in buffered aqueous solution at pH 7.5, using MOPS (monohydrate sodium salt of morpholin propane sulphonic acid) as buffer. The concentration of the substrate was varied from 10 up to 1000 M. The study of FDH in aqueous solution, at a constant temperature of 25C, gave no-significant different values of KM and vmax between the two set of measurements. From the low level of KM constant it seems that the complex ES in aqueous solution is stronger than in microemulsive system.

KM (M)

vmax (M/sec)

Test 1 Test 2

1.699 1.890

1.753 10-7 1.635 10-7

Enzymatic kinetic in a simple ternary microemulsive system was performed. Isooctane is the organic continuos phase, water the dispersed phase and AOT, which stand for bis-2 ethylhexyl-sulphoccinate, the surfactant. The following KM and vmax were obtained: KM = 7.383 M vmax= 7.483 10-8 M/s The increase of KM compared to that obtained in aqueous solution, can be ascribed to the less stability of the ES complex, proportionally with the increase of vmax. On below, the graphic kinetic behaviour in ternary microemulsive systems is reported.

73

8 10-8 7 10-8 (M/s) 6 10-8

V0
5 10-8
-8

4 10

3 10

-8

200

400

600

800

1000

1200

[Formaldehyde] ()

Fig.4.4- General kinetic exponential fitting of FDH enzyme in AOT/water/isooctane ternary microemulsion

The enzymatic activity in quaternary microemulsion was also studied for its importance in the synthesis of nanosized semiconductors. A quaternary w/o microemulsion of cathionic surfactant (CTAB, hexadecil-trimethyl ammonium bromide), pentanol (co-surfactant), nhexane and water has been used for the first time in the synthesis of quantum dots of CdS [118, 119], due to the synthesis simplicity and the well defined properties of the reverse micelles. The composition of the four component microemulsion was completely defined by three basic parameters: the molar ratio between the water concentration and the surfactant concentration (W0= H2O/CTAB), the ratio between alcohol and surfactant (P0 = H2O/POH) and the molar concentration of CTAB. In fixed experimental conditions (CTAB at 0.1M, P0 as pentanol content in the range between 8 and 14, the water content in the range 10-80) it was demonstrated that the system consists of spherical reverse micelles, where it is possible to modulate the water droplets dimension and surface dynamics either varying the water or the alcohol content [119]. Thus, kinetic measurements, alternatively varying P0 (8.65/10/14) and W0 (10/20/30), were performed. From the elaboration of the experimental data by using again the MichaelisMenten equation, the following results were obtained:

74

P0= 8.65 KM (M) vmax (M/s) P0= 10 KM (M) vmax (M/s) P0= 14 KM (M) vmax (M/s) 4.343 3.643 4.014

W0=10 2.444 3.751 10-8 W0=10 3.095

-8

W0=20 4.817 5.616 10-8 W0=20 3.659

-8

W0=30 4.818 10-8 W0=30 4.565 10-8 W0=30 ----------------

4.269 10

3.927 10 3.091

W0=10 3.979 10-8

W0=20 3.477 10-8

Table 4.3- Experimental data of FDH enzymatic reaction in quaternary microemulsion at different P0 and W0

The further increase in KM confirms the less stability of ES complex in micellar solutions, followed by an increase in vmax. Comparing the values at different W0, with P0 fixed, and vice-versa from table (4.3), it can be confirmed that FDH was not significantly affected by parameters variation. In particular, vmax do not change after variation of water pool (W0), keeping P0 as a constant. This allow for working in the optimal condition of the CdS synthesis, without the enzyme activity can be influenced. In order to have a complete scientific comprehension in the data processing, in figure (4.5) is reported an example of enzymatic study by using the double reciprocal LineweaverBurk relationship for a quaternary microemulsion at P0 =8.65 and W0 =30. The values inferred through this method, compared to that of the direct Michaelis-Menten method, did not show significant differences:

Method
Lineweaver-Burk Michaelis-Menten 4.512 4.817

KM (M) 4.773 10-8 4.818 10-8

vmax (M/s)

75

4 107 3.8 107 3.6 107

1/V (M s)

-1

3.4 10

3.2 107 3 107 2.8 107 2.6 107 2.4 10

7

Metodo dei doppi reciproci 0 0.01 0.02 0.03 0.04

-1

0.05

0.06

0.07

1/[Form.] (M )
Fig.4.5- Example of enzymatic study according to the double reciprocal Lineweaver-Burk relationship for FDH in a quaternary microemulsion at P0 =8.65 and W0 =30

8e-8

-1 -1 vmax (mol l sec )

6e-8

4e-8

2e-8

10

15

20

25

30

35

W0

Fig.4.6- Vmax dependence upon W0, keeping as a constant P0 =8.65

76

4.3.4- Reactivity of NAD dependent FDH in presence of CdS nanoclusters

It should be necessary to study the kinetic behaviour of FDH in presence of nanosized semiconductor microemulsive solutions, in order to gain information with respect to a combined system (enzyme-nanocluster semiconductor) in the same solution. As well explained in the following pages, NAD-dependent formaldehyde dehydrogenase activity were tested in presence of both nanocluster CdS and cofactor, to avoid the possibility that the enzymatic process would be inhibited by the presence of semiconductor. This should be the basis to construct and develop a third generation biosensing device for formaldehyde detection. As in the simple ternary and quaternary microemulsion, kinetic measurements at different W0 and P0 were performed again. One example of kinetic behaviour in presence of CdS is reported on below (figure 4.7 refers to CdS synthetized in quaternary microemulsion at P0=8.65, W0=20). From the table (4.4), the analysis of the obtained values of KM and vmax shows that both a high unmodified enzyme-catalytic activity and the lack of the dependence upon the parameters of microemulsive composition were confirmed.

5 10 4.5 10

-8

-8

4 10

-8

V (M/s)

3.5 10

-8

3 10 2.5 10

-8

-8

Cinetica enzimatica in presenza di CdS P =8 W =20

0 0

2 10

-8

200

400

600

800

1000

1200

[Formaldeide] (M)
Fig. 4.7- Kinetic curves in presence of a quaternary microemulsion CdS at P0=8.65, W0=20

77

P0= 8.65 KM (M) Vmax. (M/s) P0= 10 KM (M) Vmax. (M/s) P0= 14 KM (M) Vmax. (M/s)

W0=20 4.910 4.300 10-8 W0=20 2.909 5.030 10 W0=20 4.021 3.785 10-8
-8

W0=30 7.610 4.200 10-8 W0=30 4.447 4.509 10-8 W0=30 ----------------

Table 4.4- Experimental data of FDH enzymatic reaction in quaternary microemulsion with CdS nanoclusters at different P0 and W0

At a first sight, kinetic measurements carried out in different microemulsive quaternary system, varying the water (W0) and cosurfactant (P0) content, have confirmed the enzymatic activity is fully retained. Hence, formaldehyde dehydrogenase can be used in different ambient, especially in quaternary microemulsive systems, where a sufficient variety of (IIVI) semiconductors can be easily synthetized, without resulting in a denaturation of the enzyme properties. This represents the first step in developing bioreceptor/semiconductor combined system for biosensing applications.

78

79

CHAPTER 5 NEW ELECTRODES FOR AN AMPEROMETRIC GLUCOSE BIOSENSOR

OVERVIEW OF THE CHAPTER

Hydrogen peroxide has been detected amperometrically with gold electrodes modified with thin layers of ruthenium and rhodium. The ruthenium layer was radio frequency (r.f.) magnetron sputtered and rhodium layers were made by vacuum evaporation. Using the cathodic reduction at a potential of -100 mV vs. Ag/AgCl/ 0.1M KCl electrode, between 2 M and 10 mM H2O2 has been successfully detected under FIA (Flow Injection Analysis) conditions. By a combination of these electrodes with a glucose oxidase modified membrane, glucose concentrations ranging from 2 M to 10 mM can be measured amperometrically under flow conditions. Although the response is dependent on the oxygen concentration, the electrodes has been showed high operational stability and high selectivity against many electroactive non-enzymatic substances as well.

5.1- Selectivity in amperometric glucose biosensors

For the development and application of amperometric biosensors the selective and sensitive detection of hydrogen peroxide plays a key role. The majority of first generation amperometric enzyme electrodes utilise molecular oxygen as the natural electron acceptor and produce hydrogen peroxide as by-product of reaction, according the reaction (1.1a-b). As a result two electrochemical methods can be used to follow the catalytic reaction. Indeed a large number of electrode devices can either monitor the consumption of oxygen or the production of hydrogen peroxide, but this mode of operation has different disadvantages: firstly, the response can become limited by the ambient oxygen concentration;

80

secondly, high overpotential are required to oxidise hydrogen peroxide at simple metallic electrodes so there can be interference problems; thirdly, the hydrogen peroxide can degrade the enzyme with a consequent strong decrease of the operational devices stability.

These problems become topical for use in real biological samples where enzyme degradation is high and interfering substances are in a large number. Several approaches were employed to reduce or minimise these difficulties, principally by decreasing the detection potential in a range where the number of the oxidizable species is lower. The main works devoted to this aim are: a) Using interferential membranes b) Novel electrodic configuration; electrode modification In a) polymeric and cellulose membranes can be remembered as interferential sieves that permit only the passage of low molecular weight proteins or positively/negatively charged molecules; specially designed membranes can achieve a diffusional transport mechanism which allow for a potential detection decrease. For example ferricyanide or ferrocene derivative with or without soluble redox mediators [124], peroxidases of the protoporphyrine IX type were immobilised on different electrode materials with [125-130] and without redox mediators [131-134], just in order to achieve H2O2 detection potentials in the electrochemically optimum range between -100 and +50 mV [135]. Excepting the electrodes proposed by Heller et al. [127, 128], so far all these electrodes showed relatively low operational stability, especially under flow conditions. In this thesis the approach (b) has been used to decrease the detection potential of H2O2 by a suitable electrode modification, as we will see on the next discussion. Gorton and Svensson [136] investigated for the first time carbon electrodes modified with sputtered thin layers of palladium/gold and achieved a considerable decrease of the potentials both for the anodic and the cathodic detection of hydrogen peroxide. Especially encouraging results were obtained at potentials more negative than 0 mV vs. Ag/AgCl reference electrodes. Thereafter Wang et al. [137-140] proposed the use of carbon electrodes modified with particles of rhodium [139], ruthenium [138, 140] and platinum [137] and achieved very low potentials for the detection of hydrogen peroxide by electrochemical reduction. Recently, O'Connell [141] used successfully the cathode deposition of ruthenium and rhodium layers on glassy carbon electrodes in order to achieve similarly low detection potentials. They gave little or no information about the ratio of the signal to background current.

81

To this purpose at first we have been investigated Rh/Ru modified gold electrodes prepared by plasma vacuum deposition. Using this new working electrode to achieve low potentials for the detection of H2O2, combined with high operational stability and minimised background currents, is due to the fact that there is still a great demand for amperometric H2O2 and glucose (but also sarcosine) sensors with an improved selectivity. Thereafter, as showed in this chapter, a glucose oxidase modified membrane was placed on a selected Rh/Ru modified gold electrode to construct a glucose electrode with relatively high sensor performance criteria.

82

5.2- Experimental section

5.2.1- Chemicals and Reagents

Ascorbic acid, uric acid, paracetamol, -NADH, glucose, glycine, sodium sulfite, EDTA, lithium-potassium acetyl phosphate, cysteine, glucose oxidase (GlucOd) E.C. 1.1.3.4, type II from Aspergillus niger (cat. no. G-9010, approx. 1600 U ml-1) were purchased from Sigma (Deisenhofen, Germany) and used without further purification. All other chemicals were of analytical grade and from Merck (Darmstadt, Germany). 0.1M potassium phosphate buffer solutions containing 1mM EDTA were prepared from bidistilled water and used as the carrier solution. EDTA was necessary to suppress the ionic contribution.

5.2.2- Measuring set-up

Fig. (5.1) shows the FIA (Flow Injection Analysis) set-up, which was used to investigate the H2O2 and the glucose detection.

Figure 5.1. FIA set-up: P1 peristaltic pump, P2/P3 piston pumps, IV injection valve, MC mixing coil, D flow detector cell, S sample solution, B and C buffer and carrier solutions, W waste

83

The sample solutions were injected into the carrier solution C by a pneumatically actuated injection valve, indicated as IV (Rheodyne 9010, Cotati, USA), with a sample loop of 20 l. The carrier was mixed with the buffer solution B in a tightly knotted PTFE tube with a inner diameter of 0.5 mm and a length of 20 cm. The amperometric flow detector D consisted of two Plexiglas plates into one of which a groove with a depth of 0.1 mm, a width of 2 mm and a length of 10 mm was milled resulting in an effective geometric area of 20 mm2. The Rh/Ru modified foil electrode (indicator electrode) was lodged inside the groove and inclined by an angle of 45, to prevent the fixation of gas bubbles. On below is reported a photograph of the homemade detection flow cell mounted in the flow injection system.

The enzyme modified membranes were layered on the indicator electrode without additional chemical treatment. The Ag/AgCl/0.1 M KCl electrode served as the reference electrode, whereas the counter electrode was a stainless steel capillary (inner diameter of 0.5 mm, length of 7.4 mm) mounted at the outlet of the detector cell. The cyclic voltammograms were recorded with a potentiostat/galvanostat model 263A (Princeton Applied Research). Hydrodynamic voltammograms were recorded with the potentiostat CPE-1 (Institut fr Technische Biochemie e.V., Halle, Germany) under FIA conditions.

84

5.2.3- Plasma deposition procedures on gold electrode

To prepare the indicator electrodes, thin gold foils (purity of 99.995% and thickness of 0.1 mm) were covered with two different metallic layer. At first, the ruthenium layer was viaplasma deposited on gold foil by radio frequency magnetron (13.56 MHz) sputtering procedure in a ULVAC two electrode apparatus (SH33D model). The ruthenium deposition was carried out by using argon as feed gas, with a power of 150 W and at a residual pressure of 1 Pa. The thickness of the Ru layer was adjusted by varying the sputtering time (generally from 0.3 to 30 minutes) and measured by the quartz oscillation method. Thin layers with a thickness of 12 nm were used for our following amperometric measurements. Additional rhodium layers with various thickness were deposited on the ruthenium by thermal evaporation at high vacuum conditions (<10-3 Pa), by means of a vacuum evaporator Shimadzu Ltd. EA-400S with two different deposition angles of 20 and 90.

5.2.4- TEM investigations on the Ru/Rh/gold electrodes

TEM (transmission electron microscopy) investigations were performed by using a Philips CM 20 instrument, available at the Fraunhofer Institute in Halle-Wittenberg. From the experimental observation some information could be gained. It can be assumed without doubt that the ruthenium forms a continuous layer on the gold surface, as showed in figure (5.2), where TEM image of plasma polymer coated Ru/Rh modified gold electrode is reported. Only for this cross-sectional TEM investigations a thicker layer was deposited.

85

Figure 5.2- Cross-sectional TEM micrograph demonstrating the vertical set-up of the Ru/Rh modified gold electrode. The plasma polymer layer (PP) and the platinum layer were only deposited for preparative reason using the FIB technology

Because of the surface roughness of the ruthenium layers and the small amounts of deposited rhodium, a non-continuous additional layer of Rh should be formed by thermal evaporation on the Ru. Figure (5.3) shows three TEM micrographs of three rhodium layers of different layer thickness. The TEM bright-field micrographs were obtained at 200 keV. For these investigations it was necessary to deposit the rhodium thin layers on carbon-coated copper grids. The effective thickness in figure (5.3) are 1 nm, 4 nm, and 10 nm; the discontinuous structure is depicted clearly on the left micrograph, weakly on the middle one. A thin discontinuous metal layer is supposed to be formed in the early stage of nucleation and growth of the rhodium layers. Due to the discontinuous structure, the thickness information can only be given as apparent values. It could be expected that both ruthenium and rhodium are able to catalyze the amperometric detection of H2O2. An almost closed metallic structure is formed for the rhodium layer at a thickness of 10 nm (right TEM picture on figure).

86

Figure 5.3- TEM pictures of three rhodium layers having effective thickness of 1 nm (left), 4 nm (middle) and 10 nm (right)

In order to investigate the vertical structure of these ruthenium/rhodium modified gold electrodes, cross-sectional TEM investigations were carried out. The focused ion beam (FIB) technology was applied for the high-precision cross-sectional preparation. For this purpose, the ruthenium/rhodium layer was deposited on a Silicon wafer, which was previously covered with a relatively thick r.f. sputtered gold layer. An outer thin platinum layer has been deposited, to protect the multilayer system during the preparation procedure. In order to render visible the Rh film, a plasma polymer layer was deposited before the Pt deposition, using as feed monomer precursor hexamethyl-disilazane. More details about the FIB preparation technique and the plasma polymerisation are given in ref. [142] and [143] respectively. The TEM micrograph shown in Fig. (5.2) images the vertical set-up of the resulting multilayer system. The ruthenium layer has a thickness of 80 nm and is substantially thicker as used for our sensor experiments. The rhodium layer (thickness 10 nm) is visible on top of the ruthenium layer and covers it completely. At lower effective thickness of the rhodium layer, as demonstrated in fig. (5.3), the island growth of the rhodium atoms must be considered and it becomes obvious that the ruthenium layer is not completely covered [144]. The Ru/Rh modified gold electrode did not respond to glucose if the membranes had not previously been treated with enzyme catalyst. Therefore it can be concluded that nonspecific oxidation of glucose at electrode surface does not take place and this can be used as working electrode in our measurements.

87

5.2.5- Enzyme immobilisation and measurements of enzyme activity

GlucOD was immobilised covalently onto a pre-activated nylon membrane (Immunodyne, Pall, Dreieich, Germany, mean pore size 0.45 m) by the following procedure: 400 l of enzyme solution were purified by ultrafiltration, dissolved in 400 l of 0.1 M potassium phosphate buffer, pH 8.0, and concentrated up to 100 l. Thereafter, 25 l were dropped onto the membrane and dried. The same procedure was also tested onto a nanoporous alumina membrane the peculiar features of which have been described. The highly porous membranes treated with GOd-solution were already used in an electrochemical assay based on the rotating disk electrode (RDE) technique, in order to have an indication of the amount of biochemically active redox enzyme present in the biosensor [145]. The actual current response of an amperometric biosensor may be obscured by interfering electroactive species in the sample solution, and/or the solution impedance, or the capacitive charging of the electrode, so an independent assay of this enzyme activity would be helpful in the biosensor analysis. In that assay the oxygen was replaces by an artificial electron acceptor such as benzoquinone, which has been showed to give the best results. The following reactions took place:

gluconic acid + GOd glucose + GOd (ox) (red)

(5.1)

benzoquino ne + GOd hydroquino ne + GOd (red) (ox)

(5.2)

The formation of hydroquinone can be amperometrically detected at RDE, where the enzymatically produced form is re-oxidized to benzoquinone:

hydroquinone benzoquinone + 2 H + + 2 e

(5.3)

The angular velocity of the rotating electrode was chosen such that no diffusion limitation took place. The measurements were conducted at pH 7.5 and the glucose concentration was chosen as high as possible without inhibition of the substrate. The benzoquinone concentration was 5 mM, which was a good compromise to avoid the reaction rate variation

88

and to minimise, at the same time, the current increase due to the spontaneous oxidation of the glucose. The measurements were performed in an anaerobic three electrode cell to exclude the mediation of oxygen. The measured current has been showed directly proportional to the enzyme activity, irrespective if whether the enzyme is in solution or is immobilised.

5.3- Results and discussion

5.3.1- Detection of hydrogen peroxide in FIA

In continuous flow analysis the detection unit is situated in a flow cell, and a continuous flow of carrier solution and a analytical solution are conducted along the detector (in our case the analyte is the substrate for the applied enzyme onto the membrane). In amperometric analysis with continuous flow technique, the current of a sensor is measured under steadystate conditions, in the presence of various substrate concentrations (the analytical concentration at the biosensor surface is not changing at the moment of the current registration). The analytical signal is compared with the current that flows when no substrate is present (background current). The difference between the analytical and background current is a relative indication of the biosensor activity. In flow injection analysis technique, where a small volume of a sample solution is introduced into the carrier stream, a well defined reproducible concentration transient is produced at the detector. The current response in FIA is the sum of various parameters (as showed in chap. 2), and the best results are obtained when the current response have a peak height of half the value of the steady state response. This means for enzyme based biosensors that not all the molecules are occupied with substrate and there no current limitation by reaction kinetics. The main consequence is that under FIA condition the biosensor dynamic range is normally larger. Figure (5.4) shows a hydrodynamic voltammogram (A) and the signal to background current ratio (B) recorded for the H2O2 detection at a flow rate of 1 ml min-1 in the detector cell by using FIA set-up depicted in fig. (4.1). The signal current reaches a slightly inclined plateau

89

in the potential range from 0 to +100 mV, whereas the highest signal to background ratio can be measured between -100 and -50 mV. Between +200 and +300 mV the direction of the signal current is reversed. To achieve maximum selectivity against easily reducible substances, such as ascorbate, the following experiments were performed at potential E=-100 mV (the very high signal to baseline ratio value at 80 mV can be due to some unpredictable behaviour and cannot be considered as a good working potential).

Figure 5.4- (A) Signal current and (B) ratio of signal current to background current as a function of the potential; 1 mM H2O2 in 0.1 M potassium phosphate buffer and 1 mM EDTA, pH 7.0, vD = 1 ml.min-1

Figure (5.5) shows the dependencies of the signal and background current, and the ratio of the signal to background current on the buffer pH, respectively. The background current increases by decreasing pH values, because of the thermodynamic promotion of the cathodic reduction of dissolved oxygen (see chapter 1). The signal to background ratio is increasing with increasing pH. Taking into consideration the combination of H2O2 detection with the glucose oxidase catalysed analyte conversion, the further investigations were performed at buffer pH 8.0.

90

Figure 5.5- Dependence of (A) the signal, background current and (B) ratio of signal current to background current on buffer pH; [H2O2]= 1 mM, vD = 0.2 ml.min-1 , E = -100 mV

Furthermore cathodic current dependence upon the solution flow rate has been examined. According to figure (5.6) the signal current is increasing only slightly with increasing flow rate, under continuous flow conditions. The peak current h, recorded under FIA conditions, is decreasing with the flow rate, indicated as vD, according to the following expression:
h = const vD 0.5 (5.4)

91

Figure 5.6- Dependence of the H2O2 signal current for the Ru/Rh modified gold electrode on the flow rate under continuous flow () and FIA () conditions; [H2O2] =1 mM, pH = 8.0, E = -100 mV.

It can be assumed that the decrease of h is predominately caused by increasing dispersion factor D, which was previously defined by means of the Rzicka and Hansen relationship in chapter 2 [146]. At D=1 the electrode response self seems to be predominantly controlled by diffusion to the electrode surface. Referring to figure (5.7), a double logarithmic calibration graph was measured between 2M and 10mM H2O2, under flow conditions. In the range between 5M and 2mM, the graph can be described by the following equation: lg (h/A) = (0.891 0.010) lg (c/M) + (2.53 0.04)
(5.5)

92

with r2=0.999, =0.05, the number of standard concentrations m=10, the number of replicates n=4, the peak height h and the analyte concentration c. A detection limit of 2M was achieved at a signal to noise ratio of 3 under FIA conditions.

Figure 5.7- Calibration curve of the amperometric hydrogen peroxide detection under flow injection conditions; vD = 0.2 ml.min-1; pH = 8.0, E = -100 mV.

5.3.2- Selectivity and lifetime

The long term operational stability was tested by sequential injection of 1mM H2O2 at a flow rate of vD= 0.2 ml.min-1 and a frequency of 20 injections per hour. After a period of 2.5 hours, during which the electrode surface has been stabilised, a relatively stable peak signal was achieved for the next 25 hours, resulting in an operational stability better than 90% related to the initial peak height. After more than two or three days of continuous use under these conditions the rhodium/ruthenium layer was sometimes stripped off by corrosion.

93

The selectivity of many amperometric biosensors is restricted by the anode oxidation and sometimes also by the cathode reduction of interferents. The resulting side reactions can cause electrode fouling and poisoning. To verify the selectivity of the proposed indicator electrode, the influence of electrochemically active compounds was investigated in the presence and the absence of H2O2. To this purpose a 1mM H2O2 solution was mixed continuously with 0.1 M phosphate buffer, at pH 8.5, either containing 2mM of the corresponding or no interferent, at a flow rate ratio of 1 just before the injection valve IV. Table 5.1 summarises the results. The measured peak heights were related to the peak height measured for 1mM H2O2 in the absence of any interferent.

relative peak height

relative peak height in the absence of H2O2 %

relative decrease of the H2O2 signal after injection of interferent %

Interferent (1 mM) Without Ascorbic acid NADH Glucose Glycine Formaldehyde Sarcosine Paracetamol Acetyl phosphate Ethanol Methanol Cysteine Sulphite

in the presence of H2O2 %

100.0 1.1; n=5 85.0 1.7; n=6 96.8 0.6; n=6 99.7 0.3; n=10 95.6 0.6; n=5 96.6 0.7; n=4 97.3 0.8; n=5 100.4 0.4; n=5 99.4 0.4; n=4 99.8 0.7; n=5 99.4 1.6; n=4 57.1 1.0; n=6 91.8 0.9; n=6

0.0 0.0; n=3 -10.8 0.4; n=11 0.0 0.0; n=6 0.0 0.0; n=6 0.0 0.0; n=5 0.0 0.0; n=4 0.0 0.0; n=5 0.0 0.0; n=5 0.0 0.0; n=4 0.0 0.0; n=5 0.0 0.0; n=4 -8.5 0.7; n=4 -5.1 0.5; n=6 0.0 2.0; n=7 3.0 0.5; n=6 1.9 0.4; n=6 0.0 0.8; n=4 0.7 1.4; n=6 0.0 0.8; n=5 1.2 0.4; n=5 0.0 1.1; n=3 1.0 3.0; n=5 0.0 2.0; n=5 36.0 1.0; n=6 0.0 0.9; n=7

Table 5.1- Selectivity of the amperometric hydrogen peroxide detection; n-number of repeated measurement; [H2O2] =0.5 mM; vD = 0.2 ml.min-1

94

With the exception of ascorbic acid, cysteine and sulphite no significant response on the H2O2 detection was observed to other components, when they were present in the millimolar concentration regime. Cysteine and, after a longer contact, also sulphite caused a significant loss of sensitivity for the following H2O2 detection indicating a poisoning of the electrode surface. It should be noted that in comparison to earlier described H2O2 indicator electrodes, ascorbic acid had a relatively small effect on the response. However in real sample the ascorbate concentration is usually lower (e.g. milk 0.1mM) than those used in this selectivity assay [147]. Finally, paracetamol did not show any significant effect both in presence and absence of H2O2.

5.3.3- Detection of glucose

The determination of glucose was performed under the same conditions as for the H2O2 detection, but at flow-rate vD = 0.3 ml.min-1, using the same FIA set-up previously depicted at the applied potential of 100 mV vs. Ag/AgCl/KCl reference electrode. The working electrode (12nm Ru and 7.5nm Rh deposited on gold foil) was covered by a glucose oxidase modified immunodyne membrane, according to the procedure in 4.2.5. As showed in figure (5.8), a detection limit of 2M was achieved under flow conditions, similarly to that in the hydrogen peroxide determination.

10 0

Cathodic current [A]

10 -1

10 -2

10 -3

10 -4 10 -6 10 -5 10 -4 10 -3
-1

10 -2

10 -1

G luc ose co n c [m o l l ]

Figure 5.8- Glucose Calibration curves under FIA conditions (vD= 0.2 ml min-1, E= -100 mV vs. Ag/AgCl; Working electrode: GOD on ID-membrane/12 nm Ru/7.5 nm Rh/Au)

95

The double logarithmic regression function can be described by equation (5.6) in the range from 0.01 to 1mM glucose (r2=0.998, =0.05, m=7, n=4): lg (h/ A ) = (0.962 0.019) lg (c/ M) + (2.16 0.06)
(5.6)

An operational stability better than 90 % was obtained during a period of more than 20 hours under FIA conditions at a frequency of 18 injections of 0.1mM glucose per hour. The selectivity of the glucose was investigated under the same conditions as those of the H2O2 detection, by mixing continuously 1mM (the clinically relevant value) glucose solution with interferent solutions at different concentrations. Hence, the sensitivity of the glucose biosensor to ascorbic acid, NADH, paracetamol, uric acid and cysteine was separately tested. Table 5.2 summarises the results: only cysteine caused a significant negative bias and a poisoning effect below the range of mM concentration.

Relative peak height in the presence of glucose %

Relative peak height in the absence of glucose %

Relative decrease of the glucose signal after injection of interferent %

Interferent

Without Ascorbic acid (10 M) NADH (1 mM) Paracetamol (1 mM) Uric acid (1 mM) Cysteine (10 M)

100.0 1.8; n=3 100.1 0.1; n=3 100.6 0.2; n=4 97.9 0.2; n=6 99.1 0.2; n=4 71.9 0.1; n=4

0.0 0.0; n=3 1.7 0.1; n=3 0.0 0.0; n=4 0.0 0.0; n=4 0.0 0.0; n=4 -52.5 0.1; n=4 0.0 0.1; n=3 1.9 0.2; n=4 4.9 0.2; n=3 5.1 0.1; n=4 4.9 0.1; n=4

Table 5.2- Selectivity of the amperometric glucose detection under FIA conditions; n= number of repeated measurements; [glucose]= 0.5 mM; vD= 0.3 ml min-1

96

5.4- Conclusions

Gold electrodes with a thin layer of ruthenium and thereupon a porous layer of rhodium was successfully used to detect hydrogen peroxide amperometrically at potentials in the range between -100 and 0 mV vs. Ag/ AgCl/ 0.1 M KCl, by monitoring the cathode reduction of H2O2. Both a relatively high operational stability and a high selectivity against many electroactive substances are achievable. Therefore, multilayer electrode has been used as an indicator electrode behind oxidase modified membrane layers for the selective detection of glucose. A relatively good linear detection of glucose has been achieved along with a high operational stability and selectivity, especially under flow conditions, but at the present stage of research no significant improvements were made in comparison with the main disposable glucose sensors of successful commercialization. Similar lifetime curve were also obtained before in literature for other biosensors. The biosensor described here can be easily constructed and used as a simple device; our amperometric measurements were carried out at low potential, for prolonging the sensor selectivity in the case of every enzyme which catalyze the hydrogen peroxide production. Operational stability could increase by replacing the common immunodyne membranes with a well-protected nanoporous membrane for the enzyme immobilisation, for prolonging the enzymatic activity even beyond one week.

97

98

CHAPTER 6 AMPEROMETRIC SARCOSINE, CREATINE AND CREATININE BIOSENSOR

OVERVIEW OF THE CHAPTER

Selective detection of sarcosine, creatine and creatinine is still a topical issue in clinical applications. The possibility to detect amperometrically and selectively the hydrogen peroxide cathodic reduction at ruthenium and rhodium modified gold electrode at potential of 100 mV vs. Ag/AgCl/0.1 M KCl was early showed. In this chapter sarcosine has been selectively detected in the concentration range between 5 and 1000 M, by using a sarcosine oxidase modified membrane onto the multilayer electrode, after an accurate selection of optimal flow rate and pH conditions. Combining the sarcosine sensor with a packed coimmobilised creatinase and creatininase enzyme reactor, the detection of creatine and creatinine has been performed respectively between the range 201000 M and 101000 M, with a relatively good operational stability under flow conditions. This represents a good basis to enhance novel biosensing systems in order to detect selectively and sensitively creatine and creatinine.

6.1- Sarcosine detection: state of the art

The selective and sensitive detection of sarcosine, as proof of the presence of creatine and creatinine, have attracted much attention during the last decade in the development and application of amperometric biosensors, due to its importance for clinical and biomedical purposes [148]. Nowadays, most clinical laboratories and commercial analyzers continue to use for creatine detection a spectrophotometric procedures based on Jaff reaction, where the absorbance of the addition complex, resulting from the reaction between creatinine and

99

picrate, is monitored in alkaline solution [149]. The reaction is not specific for creatinine, because many chromogenic compounds interfere in this assay, thus the selective creatinine evaluation remains a problem [150]. Consequently, alternative methods using high performance liquid chromatography [151] or enzyme-based methods were invented with a view to overcoming these problems [152, 153]. Sarcosine oxidase has been widely used in the enzymatic methods. Recently new systems with integrated miniaturised sensing elements were developed, even using microelectronics techniques for electrodes fabrication, but so far all these electrodes showed relatively low operational stability, rather high costs and long response time. Tsuchida and co-workers [154] designed new multi-enzymatic amperometric sensors based on the catalytic sequence of reactions already described in chap.3:
CA creatinine + H 2 O creatine

(6.1)

creatine + H 2 O CI sarcosine + urea

(6.2)

sarcosine + H 2 O + O2 Sarc. oxidase glycine + HCHO + H 2 O2 (6.3)

where reactions were catalysed respectively by creatininase (reaction 6.1), creatinase (6.2) and sarcosine oxidase (6.3). When three enzymes are coupled, the current responses resulting from the catalytic reaction are diminished, thus affecting the sensitivity and detection limit of these methods. On the basis of the work of Tsuchida, new commercial electrode-based creatinine sensors were designed as clinical analysers (Nova 16 ). After much efforts, a disposable biochip sensors for the creatinine determination was developed by Maradas and co-workers [155], in which electro-polymerized perm-selective films were used. In their publication the fabricated sensors were tested in a flow injection set-up and detection limits; long term operational stability and selectivity were also reported. One of the drawbacks of these miniaturised sensors is the very low signal of creatinine for samples with high creatine content, because of the high sensitivity for creatine. The topic demand to have a creatine and creatinine sensor with both good operational stability and high specificity remains. We have seen that recently a large amounts of metal-dispersed modified electrodes was prepared in order to reduce the interferences of several electroactive substances at the detection potential of hydrogen peroxide [137-139, 156]. Guilbault and co-workers used for

100

the first time the cathodic deposition of ruthenium and rhodium layers on glassy carbon electrodes in order to achieve low working detection potentials [141], without giving information about the background current and the ratio of the signal to background current. We have already shown electrochemical investigations of ruthenium and rhodium layer deposited after each other on gold electrodes, onto which a glucose oxidase modified membrane was placed. The same FIA methodology for the enzymatic detection of sarcosine, based on amperometric investigation by means of a sarcosine oxidase modified electrode, has been used, with a suitable modification of the FIA set-up and addition of a multienzymatic cartridge-reactor. Sarcosine oxidase (SOD) catalyzes the reaction (6.3) to give glycine, formaldehyde and hydrogen peroxide as products. The cathodic reduction of H2O2 has been showed to be successfully detected by Ruthenium/Rhodium on gold electrode (chapter 5), at a potential of 100mV vs. Ag/AgCl/0.1M KCl, where the influence of several electroactive non-enzymatic substances is minimized.

6.2- Materials and methods

6.2.1- Chemicals and Reagents

L-ascorbic acid, uric acid, paracetamol, glycine, glutaraldehyde (GA) 25% aqueous solution, sarcosine, sarcosine oxidase from Bacillus Species (EC 1.5.3.1, cat. Nr. S-7897; with 47 and 50 units/mg solid), creatine, creatinine and bovine serum albumin (BSA, cat. nr. A-9647) were purchased from Sigma (Deisenhofen, Germany) and used without further purification. Creatinase (EC 3.5.3.3) and creatininase (EC 3.5.2.10) were from Sigma Aldrich. Controlled porous glasses (CPGs) sphere of the sieve fraction from 140 to 160mm were obtained from Schuller (GmbH, Steinach, Germany). All other chemicals were of analytical grade and from Merck (Darmstadt-Germany). 0.2M sodium/potassium phosphate buffer solutions containing 0.2M KCl were prepared from doubly-distilled water and used as the carrier solution.

101

6.2.2- Measuring set-up

Figure (6.1) shows the FIA set-up which was used to investigate the sarcosine, creatine and creatinine detection.

Sample

MC
Carrier

ER
IV
W

D
Y

Figure 6.1- FIA set-up: Sample is injected by peristaltic pump, IV: injection valve, ER: enzyme reactor, MC: mixing coil, D: flow detector cell, W: waste

Sample solutions were injected into the buffer carrier solution by a pneumatically actuated injection valve IV (Rheodyne Cotati, USA) with a sample loop of 30 l and reacted in the enzyme reactor ER. At the mixing point Y and in the mixing coil MC the carrier was mixed with the sample and pumped in the detection cell D. The electrochemical signal was monitored and scanned using a software controlling the pump and the injection valve simultaneously, as better showed in a generic schematic process control on figure (6.2). The amperometric detector D was a thin layer flow cell consisting of two Plexiglas plates, into one of which a groove was milled for lodging the indicator electrode. The effective geometric area for the indicator electrode was 20 mm2.The counter electrode was a stainless steel capillary mounted at the outlet of the detection cell described in the earlier chapter. Every potential indicated in the following experiments was measured with respect to a Ag/ AgCl/0.1M KCl reference electrode. Hydrodynamic voltammograms and cathodic current were recorded again with the home-made potentiostat CPE-1, under flow conditions. To prepare the indicator electrodes thin and clean gold foils of thickness of 0.025mm were modified by Ru/Rh layers, as previously described. The plasma chemical deposition process as well as the microscopic measurements about layers thickness and roughness were performed at the Fraunhofer Institute of Material Science Freiburg/Halle. All the next

102

experiments were performed onto the multilayers electrode, sputtering 100 nm of Ru after deposition of 7.5 nm thick Rh on gold.

Figure 6.2- FIA equipment used in general multi-enzymatic detection: P1-P3: pumps; IV: injection valve; Y: mixing point; MC: mixing coil; D: detector; S: sample solution; TL A/ TL B: carrier and reagent solutions; W: waste; ER2: enzyme reactor [from courtesy of ref. 159]

6.2.3- Enzyme immobilisation and preparation of the enzymatic reactor

Sarcosine oxidase from Bacillus species was immobilized covalently onto a pre-activated Nylon ID-membrane (Immunodyne, Pall, Dreieich, Germany) by using the following procedures: the content of 50 unit per mg of SarcOD was dissolved in 400 l of 0.9 mol l-1 trehalose in 0.1 M sodium/potassium phosphate buffer at pH 8.0. Thereafter a concentration procedure was performed to obtain 120 l and a solution of 20 l were dropped onto the IDmembrane strip. The co-immobilised creatine and creatinine enzymatic reactor was designed compatibly with the flow conditions and was packed in a cartridge according to the method proposed by Spohn et al. [157]. 200 mg. of aminosilylated Trisoperl CPGs were suspended in 5 ml of

103

2.5% m/m glutaric aldehyde in 0.1 M phosphate buffer at pH 6.5 and were activated by water jet in vacuum for 30 min and under normal pressure for further 30 min. The suspension was filtered over a fritted glass with doubly distilled water under vacuum conditions. A total amount of 18.4 mg of lyophilized cretininase/creatinase enzyme mixture, necessary to have respectively 189.6 unit of creatininase and 192 units of creatinase, was dissolved in phosphate buffer at pH 8.0. Pre-activated CPGs was added to the enzyme solution. The obtained suspension was kept under reduced pressure for 30 min at ambient pressure and then at a temperature of 4C for further 15 min. Everything was packed into the enzyme cartridge to give an enzyme reactor of 2cm length, which is schematically showed in the following picture. Finally the enzyme reactor was connected by tubing, upward of the detection cell.

Polypropylene grid

Immobilized Enzyme

Catrigde

Fitting

Figure 6.3- Detailed scheme of a complete packed enzyme cartridge

104

6.3- Result and discussion

6.3.1- Adjustment of the experimental parameters

Figure (6.4) shows the dependency of both the cathodic current on the detector flow rate, as the sarcosine concentration and the applied potential were kept constant respectively to 0.5mM and 100mV. The signal current is increasing with increasing the flow rate in the range between 0.05 and 0.15 ml min-1, whereas decrease in the range from 0.15 and 0.50 ml min-1. Taking into account that the maximum is reached at 0.15 ml min-1 the following experiments were performed at this flow rate value.

0,15 0,14

Cathodic current [A]

0,13 0,12 0,11 0,10 0,09 0,08 0,0 0,1 0,2 0,3
-1

0,4

0,5

0,6

Flow rate [ml min ]

Figure 6.4- Cathodic current dependence on the flow-rate. Sarcosine oxidase modified Ru (100 nm)-Rh (7.5 nm) onto gold membrane under FIA conditions.-[Sarcos.]= 0.5 mM in 0.1 M sodium/potassium phosphate buffer (pH 8.0); Eelectr =-100 mV; injection volume:50 l

105

In figure (6.5) is reported the buffer pH dependence of sarcosine detection on the current signal; although the sensitivity is relatively high in a wide pH range, the highest signal was reached at pH 8.0, by using a 0.2 M sodium/potassium phosphate buffer containing 0.2M KCl, to confirm the right buffer conditions we have always used.

0.16 0.14 0.12 0.10

Current [A]

0.08 0.06 0.04 0.02 0.00

pH

Figure 6.5- Sarcosine signal dependence on the buffer pH. Sarcosine oxidase modified Ru (100 nm)Rh (7.5 nm) onto gold membrane under FIA conditions; flow rate D = 0.15 ml min-1; injection volume: 50 l, Er = -100 mV, [Sarc]= 0.5 mM; ()= 0.2 M sodium-potassium buffer + 0.2 M KCl, ()= 0.2 M TRIS buffer + 0.2 M KCl

The selectivity of the proposed indicator electrode in the amperometric biosensor is restricted by the anodic oxidation and sometimes also by the cathodic reduction of interferents, which can cause electrode fouling and poisoning. Hence the influence of electrochemically active compounds was investigated both in presence and absence of sarcosine. In order to evaluate the selectivity of the electrode, 0.25 mM of sarcosine solution was mixed continuously with 0.5 mM of interferent solutions and/or phosphate carrier buffer, under flow conditions, at a flow rate ratio of 1:1 just before the injection valve. Table 6.1 summarizes the results. The relative peak heights measured both in absence and in presence of sarcosine were related to the peak height measured for 0.25 mM sarcosine, in absence of

106

any interferent. Only ascorbic acid showed anodic interference, because of its probable reoxidation process. Glycine showed neither electrode poisoning effect neither significant anodic interference.

Relative peak height in the presence of Interferent sarcosine %

Relative peak height in the absence of sarcosine % 0.0

Relative decrease of the glucose signal after injection of interferent %

Without

100.0

Ascorbic acid (10 M)

82.1 0.2

-41.65 0.13

0.0 0.2

n=6

n=5 0.0 0.0 0.00 0.13

Glycine (1 mM)

99.89 0.10

n=6

n=4 0.0 0.0 1.5 0.2

Formaldehyde (1 mM)

90.51 0.11

n=9

n=8 0.0 0.0 0.58 0.07

Paracetamol (1 mM)

100.41 0.06

n=5

n=5

Table 6.1- Selectivity of the amperometric sarcosine detection under FIA conditions; flow rate vD = 0.15 ml min-1, level of confidence = 0.05, n= number of repeated measurements

6.3.2- Sarcosine, creatine and creatinine detection

Figure (6.6) shows the double logarithmic calibration graphs of sarcosine, creatine and creatinine, supporting by table (6.2) where the main calibration parameters are summarized. The detection of sarcosine was performed after placing onto the indicator electrode the Sarcoxidase (SarcOD) modified membrane, which was prepared according to the method

107

described in the previous section. The calibration curve showed a linear dynamic range between 10 M and 500 M.

100

Cathodic current [A]

10-1

10-2

10-3 10-6 10-5 10-4 10-3

-1

10-2

10-1

Concentration [mol l ]

Figure 6.6- Calibration graph of the amperometric sarcosine, creatine and creatinine detection; flow rate D= 0.15 ml min-1, E= -100 mV; ()= sarcosine, ()=creatine, ()=creatinine

108

Detection limit [mol lt ]

-1

Detection range [mol lt ]

-1

R2

(=0.05) [%] 99.97

Sarcosine

2 10-6

1 10-55 10-4

0.93

2.34

m=7; n=4 99.74

Creatine

1 10-5

10-5 10-3

0.94

2.07

m=8; n=3 99.34

Creatinine

1 10-5

10-5 10-3

0.95

2.10

m=8; n=3

Table 6.2- Sensor parameters for sarcosine, creatine and creatinine calibrations. Regression equation: lg (h /mV)= a lg (c /mol l-1) + b

109

The creatine and creatinine detection was measured after the enzyme cartridge preparation and recorded in the range between 10M and 1mM: this represents a good linear range in comparison to the other sarcosine sensors [154, 155]. All the measurements were performed under flow conditions with an injection volume of 50 l. The double logarithmic regression functions can be described by equation (6.4, 6.5 and 6.6):

lg (h/mV) = (0.933 0.016) lg (c/M) + (2.34 0.04) lg (h/mV) = (0.938 0.011) lg (c/M) + (2.07 0.04) lg (h/mV) = (0.95 0.03) lg (c/M) + (2.10 0.04)

(6.4) (6.5) (6.6)

where h is the peak height and c the analyte concentration. The detection limits were evaluated at the dispersion factor of Dmax= 3.3, which was determined according to Rzicka and Hansen relationship [146]. The high overlapping between creatine and creatinine calibration curves showed that the hydrolase enzyme should work quantitatively in reaction (6.1). Conversely, the difference between creatine/creatinine curves compared to the sarcosine calibration demonstrated that reaction (6.2) do not work quantitatively. In every case, enzyme activity should not to decrease under flow conditions in the enzymatic reactor.

6.3.3- Long term stability of the biosensor

The SarcOd dependent membrane onto the indicator electrode showed a high long term operational stability, which was determined by the sequential injection of 0.5mM sarcosine. After a period of around 5 hours (100 injections), during which the electrode surface was forming and the signal was not stabilized, a relatively stable peak signal height was achieved for the next 24 hours (figure 6.7). A long term test resulted in a slightly stronger signal decrease than in the case of the protected glucose oxidase enzyme membrane, thus confirming the good capability of the immunodyne membranes to keep the quality of the enzyme properties.

110

105 100 95

time [hours]

90 85 80 75 70 65 60 0 2 4 6 8 10 12 14 16 18 20 22 24

% signal Time [hours]

Figure 6.7- Operational stability of the immunodyne protected sarcosine sensor; 0.5 mM sarcosine at a frequency of 20 injections/hs

The following chart shows only a part of the total signal sequence for the protected sarcosine oxidase sensor, starting from ten minutes after the stabilization of the signal. The peak width was only around 40 s, depending on the injected sarcosine concentration, allowing a frequency of 20 injections per hour. When using smaller flow cell and sampling volume or doubling the flow rate, injection frequency would be enhanced, enabling fast calibration during on-line analysis, but at the same time, the signal sensitivity would undergo a strong decrease. The lifetime of the sensor seemed to be also relatively good, but further assays must be performed in the future, especially onto real biological samples. Considering that the novel system should be subjected to further improvements, at the present state it could be used for more than 400 determinations over a period of two-three week.

111

Figure 6.8- Long term stability of the immunodyne protected sarcosine-sensor. 100 injections of 0.5 mM sarcosine have been showed only after the first 5 hours of stabilization of the signal

Hence, the stabilization of the enzyme into the pores of the immunodyne membrane result in a relatively considerable operational stability. Lifetime shelf stability can be improved simultaneously improving the immobilization methodology of both creatininase and creatinase in the enzyme cartridge and sarcosine oxidase in nanoporous systems.

112

6.4- Future perspective to enhance the biosensor stability

In the forth chapter we showed that glucose oxidase and sarcosine oxidase can be immobilized efficiently on a new type of pre-activated nylon membrane. In order to improve the operational long term stability of the described glucose and sarcosine biosensors, the possibility to replace porous ID-membranes commonly used on our working electrode with nanoporous alumina membranes was taken into account. To this purpose it was necessary a preliminary comparison of nanoporous alumina membrane in the same amperometric condition of ID-membrane. The results of this study can be used to propose a future model for the interaction between oxidase enzyme and nano-porous membranes. Nanoporous alumina was firstly prepared from the Fraunhfer Institute of Mechanics of Material in Halle, using procedures indicated in many references, by anodic oxidation of Al in acidic electrolytes solution (oxalic, sulfuric, phosphoric). Then it was used as alternative membrane on which glucose oxidase and sarcosine oxidase can be entrapped, before to place the same membrane onto the Ruthenium and Rhodium modified electrode. AFM (atomic force microscopy) observations were performed about a sample of alumina electro-synthetized in 0.3 M of phosphoric acid at T=273K, Vbias of 190 V for a treatment time of 2 minutes. It was not possible to detect the regular microstructural array, made of relatively straight holes located at the center of hexagonal cells. Pore diameters in the range from 150 to 220 nm have been estimated, with a relatively low size distribution. We failed to obtain atomic resolution of the AFM images with our samples, due to the fact that the current was always noisy. Trials to scan with higher current (the tip closer to the sample surface) led to the so-called dragging effects and to many oscillations. It means that sample touched the sample material and successively moved part of the molecules over the substrate surface during the scanning process. As a consequence, better AFM images could not be obtained.

The electrochemical detection cell depicted in chapter 4 can be also used in combination with nanoporous alumina layered onto Ru/Rh modified working electrode. Cyclic voltammogram measurements, by using a potentiostat/galvanostat model 263A, were recently performed under this condition, after the injection of H2O2 in 0.1M phosphate buffer with 1 mM EDTA, in order to compare the sensitivity of the amperometric hydrogen peroxide detection from nanoporous alumina with those obtained from the ID-membranes.

113

At the moment in which this thesis is written, no significant increase of the signal to baseline ratio of the H2O2 detection was yet obtained, by replacing alumina membrane to porous nylon membranes. It may be possible that improvements will be reached after protection of nanoporous membranes with plasma-polymer layers, similarly to what Janasek et al. made for peroxidase modified highly porous nylon membranes used in chemiluminescence sensors [158]. Plasma polymer depositions on different conductive substrate are actually in progress at the Fraunhfer Institute of Mechanics of Material in Halle-Wittenberg and in Freiburg (Germany).

6.5- Conclusions

After deposition of 100 nm Ru followed by approximately 7.5 nm Rh layers, modified gold electrodes can be used to detect sarcosine amperometrically at potential of 100 mV vs. Ag/ AgCl/0.1 M KCl. Even in this case, both a relatively high operational stability and a high selectivity against many electroactive substances were achieved. The proposed electrode was used in combination with a multi-enzymatic creatininase/creatinase flow-cartridge, to detect not only sarcosine, but also creatine and creatinine, in biosensing systems where indicator electrode is covered by sarcosine oxidase modified membrane layers. The sensor activity is relatively high in a broad range of pH and flow rate. The optimum pH of 8.0 is favourable for a number of possible applications. All the experiments described in this chapter have shown that a stable first generation biosensor based on sarcosine oxidase immobilized on highly porous membranes and creatinase-creatininase multienzymatic cartridge can be constructed. The present biosensor can compete very well with other first generation biosensor for sarcosine detection described in literature.

114

115

CHAPTER 7 ENZYME/ SEMICONDUCTOR NANOCLUSTER COMBINED SYSTEMS

7.1- Photo-electrochemical study of CdS nanocluster in presence of NAD dependent formaldehyde dehydrogenase

7.1.1- Photoelectro-chemistry at the interface solution/semiconductor

In order to have a good understanding of the semiconductors behaviour in solution, a short description of the mechanism of electronic transfer between semiconductor electrode surface and solution should be taken into account. Keeping in contact an n-type semiconductor, where electrons are the major charge carriers, with a couple Ox/Red in solution, the electron potential (as indicated in figure 7.1) tends to reach the same value, that is also the Fermi levels reach the same value. This process results in a charge transfer for both the phases. As the Fermi level of semiconductor is higher than in solution, electrons flow from the semiconductor towards the solution. The charging effect is concentrated in a small region of charge spatial, analogously to what happens in the double layer diffusion of a solution. For n-type semiconductors, the electric field in the charging spatial region results in a shift of the Fermi levels to higher value than the bulk semiconductor level. Electrons will move to the electric field of bulk and holes toward the interface electrode/solution, up to a zero potential, when the charge effect is depleted (figure 7.1). As the interface is irradiated by means of energetic radiation larger than the energy gap, commonly indicated as Eg, the photons generate an electron towards the bulk semiconductor and a hole towards the electrode surface, resulting in a redox species oxidation.

116

Fig. 7.1- Energy levels and electrochemical potential at the junction semiconductor surface/solution

The electron and the hole above the charging region tend recombine themselves by producing heat, whereas at the charging region will migrate towards the opposite side. A net current in the external circuit is generated by the photo-anodic current. In our case, semiconductor modified metallic electrode was used, therefore the system is more complex, due to the fact that nanoclusters are surrounded by electrolytic solution (see figure 7.2). After irradiation, the hole migrates to the surface in order to interact with the

117

solution, whereas electron reaches the substrate across the nanoparticle layers. The probability of recombination between electron and hole is proportional to the number of the semiconductor layers onto the electrode [98]. Experimentally, a strong increase in the number of semiconductor layers could generate a drop of photoanodic peaks.

Fig. 7.2- Scheme of the charge movements in substrate covered by semiconductor nanoparticles in solution

7.1.2- Enzymatic photo-activity with CdS: experimental set-up

NAD dependent formaldehyde dehydrogenase enzymatic activity was extensively tested in quaternary microemulsion and in presence of CdS nanoparticles (chapter 4), in order to check some possible inhibition process of the enzyme activity, due to the presence of both formaldehyde substrate and NAD+. Preliminary measurements of CdS nanoclusters photoactivity were already performed in MOPS (mono-hydrate sodium salt of morpholin-propane sulphonic acid) buffer, using both electrodes where CdS was immobilised on dithiol substrates by self assembly techniques, and electrodes where CdS nanoparticles were deposited on gold substrate by spin-coating. Amongst various immobilisation techniques,
self assembling seems to be the most effective to provide highly ordered semiconductor

118

monolayers with strong adherence of nanoparticles onto the electrode, so far [99]. These features are crucial in studying mechanism of electronic transport within the monolayer and from the nanocrystals to the substrates. For electrochemical and photo-electrochemical studies on Q-dots CdS and enzymecombined systems, chrono-amperometric techniques have been used, measuring the current which flows through an electrode in electroactive solution as a function of the time. A controlled tuneable potential is applied on the working electrode. The electrode must be highly polarised, therefore it should possess as small surface as it may assume the correct external potential. Chrono-amperometric curves have a typical exponential decay. If we suppose that a generic redox system:

Ox + n e- Red

(7.1)

starting (at t=0) instantaneously from the initial potential E1, where no process takes place at the electrode, up to potential E2 where a reduction takes place, an instantaneous increase of current which immediately decreases in time can be observed. This is due firstly, to the continuous demand of reductant species which comes from more and more remote layers of solution and, secondly, to the decrease of analyte solution. In order to obtain the current dependence upon the net concentration of analyte and to avoid the dependence on the instantaneous concentration of the electrode, E2 should be in the range of the limit current. Applying the law of diffusion and suitable boundary conditions, the Cottrell equation can be derived (reported as 1.9 in the first chapter). Chronoamperometric technique has generally some limitation in versatility, due to the increase of capacitive current, to the inevitable formation of density gradient and the presence of vibration in long term period which can result in the layer dissolution. Photo-elettrochemical measurements on CdS nanoclusters immobilised on gold supports have been carried out by a PC-controlled potentiostat Autolab PGSTAT10, using a home-made three electrode cell at ambient temperature (20 2C), the scheme of which is reported in the following figure:

119

solution Reference electrode (Ag/AgCl) Counter electrode (Pt)

Plexiglas Working electrode steel

Fig. 7.3- Three electrode cell for photo-electrochemical measurements

The amperometric cell consists of an inox steel support onto which was layered the working electrode (CdS deposited on gold) and two Plexiglas plates on the upper side, where a circular groove was made to contain the solution. Both a platinum counter electrode and a Ag/AgCl/KCl reference electrodes were electrically connected to the plates. An optical fiber tungsten lamp (250 W) was directly interfaced to the electrochemical cell. CdS deposition on gold was carried out principally by means of a spin-coating apparatus (Chemat Tecnology Spin coater KW-4B).

7.1.3- Enzymatic photo-activity with CdS: results and discussion

Preliminary measurements of the photoactivity of CdS immobilized on gold substrate were carried out to test the potentiality of the support as photo-catalytic devices. In figure (7.4) the current signal as a function of the time, after photo-activation (ON), is reported. The initial current reached a readily constant value after photoactivation. As the light has switch off, the current slightly returned to the initial value. Photo-anodic currents showed an exponential behaviour of a typical charging/discharging process due to the photo-activated electron injection in the conduction band and the following recombination with the holes. This trend

120

is slower than a simple capacitive electrode process decaying according to t-1/3, because of the semiconductor carriers diffusion.

Fig. 7.4- Photocurrent response for CdS deposited on gold electrode :(on) illumination ,(off) dark

The same photocurrents have been measured onto an electrode in MOPS buffered formaldehyde solution. As the nanocluster modified gold electrode was illuminated (on), an exponential photocurrent decay has been observed. Further experiments with FDH enzyme in formaldehyde solution have been carried out (figure 7.5). The blue-dark curve indicates the photocurrent behaviour in presence of formaldehyde alone, whilst the red curve is in presence of formaldehyde and enzyme. Under illumination of enzyme solution a strong increase of photocurrent has been observed, in accordance with

121

the possibility that enzymatic reaction in electrochemical cell took place. The decay behaviour in the two curves is different: this could be due to a different process in the cell as the enzyme has been involved. In order to have a confirmation of this hypothesis further measurements onto an electrode in an enzymatic MOPS buffered solution have been made, modifying the conditions of the sample addition.

Fig.7.5- Photocurrent in presence of formaldehyde (dark curve) and in presence of enzyme with formaldehyde (clear curve), after illumination of the system

On figure (7.6) the typical current increase can be observed after illumination. In a) little amounts of formaldehyde have been injected through the cell; after addition of substrate a strong current variation took place, due to the activation of enzymatic reaction. The current, following the usual exponential decay, returns to the initial value of photo-current. In the point b) and c) two further additions of formaldehyde have been made under light and in both cases a small current variation took place. In off the light was switched off and current return to its initial value. To confirm that the cathodic current increase is due to the catalytic reaction after each addition of substrate, the same assay was performed without photo-excitation.

122

on

off
a b c

Fig.7.6- Photocurrent in presence of enzyme after three different additions of formaldehyde

O n2 O ff 1

a on b

c O ff 2

Fig. 7.7- Photocurrent in presence of enzyme after formaldehyde addition; (on1)light, (a) first addition, (off1) dark, (b) formaldehyde addition in dark, (on2) light, (c) further add., (off2) dark

123

On figure (7.7) further information were obtained after conducting similar experiment of that previously made, but without photo activation by light energy, in order to confirm that the cathodic current increase, indicating the substrate reaction, is due to the photocatalytic activation of the semiconductor. Under illumination (a) the addition of substrate results in the expected current variation. In off1 the current returned in initial position. After the switching off in b), every formaldehyde addition did not cause any significant current variation. The reason could be that the process is photo-catalysed by the semiconductor CdS. Finally, we tried to add an aliquot of enzyme to the formaldehyde solution onto a CdS modified electrode, obtaining the following photocurrent behaviour.

124

Fig.7.8- Photo-current for CdS electrode in presence of enzyme after illumination and further enzyme addition (in correspondence of the strong peak increase)

The amperometric response of the system when enzyme is added is substantially increased as compared to that without enzyme, probably due to the enzymatic reaction. All such results on the photo-electrochemical measurements seems to show an apparent enzymatic response and to indicate that nanosized semiconductor could function in photocatalytic combined oxidation of formaldehyde. The enzymatic assays confirm that the activity is nearly completely retained as the reaction is carried out in presence of CdS nanocrystals, under rapid illumination. Further experiments on the long term stability of the system have been carrying out during the next time. Therefore they are beyond the goal of this thesis.

125

7.2- Concluding remarks

Photocurrent time course on gold electrode modified with CdS nanoclusters immobilised by
self assembling techniques, both in presence of formaldehyde and of enzyme solution, were

respectively recorded, under illumination. The amperometric response of the system in presence of FDH enzyme solution, by using a home-made three electrodes cell, was increased as compared to that without the enzyme, under illumination. Photocurrents appears after formaldehyde addition into the cell, already containing FDH solution onto CdS/Au electrode. Current increases whenever the system is photo-activated, confirming that a photo-catalytic effect exists. It should be ascribed to the powerful properties of the nanosized semiconductors in the enzymatic reaction. Consequently, this first photo-electrochemical study carried out on the CdS nanoparticles, linked on gold substrate, in presence of FDH enzyme and formaldehyde pointed out that nanoclusters, once photo-activated, could act as charge carrier. The study tested the effectiveness of CdS nanoparticles to replace the cofactor couple NAD+/NADH as charge transfer in the enzymatic oxidation of formaldehyde to formic acid by FDH. The results obtained on quantum sized CdS immobilised onto gold electrodes in close contact with enzyme and substrate showed that semiconductors could effectively perform the enzymatic oxidation of formaldehyde to formic acid. Further conclusive proves must be gained, especially evaluating the condition of the enzyme after several cycles of photoelectrochemical processes (protein denaturation, inhibition of the redox site, non-correct orientation are the main hurdles to be overcome). The optimisation of enzyme immobilisation procedure onto modified working electrode should improve the sensitivity of the current signal and the enzyme activity in the formaldehyde recognition biosensing. The next steps of the research are going on. Thus, long term stability assays, correct engineering and eventual scale-up of this combined system are necessary for successfully exploiting the amperometric detection to monitor and quantify formaldehyde in real systems as product of interest in agriculture and food industry.

126

127

References:

1. 2. 3. 4. 5. 6. 7. 8. 9.

Willner I.; Angew. Chem. Int. Ed.; (2000) 39, 1180-1218 Brook, S.L., Higgins, I.J., Newmann, J.D. and Turner, A.P.F., Enzyme Microb. Technology, 1991, 13; 946-655 Turner A.P.F., Analytical Proceedings, (1991) 28; 376-377 Cammann, K., Lemke, U., Rohen, A., Sander, J., Wilken, H. and Winter, B., Angew. Chem., (1991) 103; 519-541 Cho, Y.K., and Bailey, J.E., Biotehnology and Bioeng., (1978) 20; 1651-1665; Thevenot, D.R.; et al.; Technical report on Biosensors & Bioelectronics, (2001) 16; 121-131 Kozhukharova, A.; Kirova, N.; Popova, Y.; Batsalova K.; Kunchev, K.; Biotehnology and Bioeng., (1984) 26; 245-248 Bartlett, P.N.; Whitaker, R.G.; J. Electroanal. Chemistry, (1987) 224; 37-48 Sonawat, H.M.; Phadke, R.S.; Govil, G., Biotehnology and Bioeng., (1984) 26; 1066-1070

10. Kaetsu, I., Kumakura, M.; Yoshida, M.; Biotechnology, (1979) 11; 847-861 11. Hsiue, G.; Wang, C., Biotehnology and Bioeng., (1990) 36; 811-815 12. Beh, S.H.; Moody, G.J.; Thomas J.D.R.; Analyst., (1989) 114; 1421 13. Galiatsatos, C.; Ikariyama, Y.; Mark, J.E.; Heineman, W.R.; Biosensors & Bioelectronics, (1990) 5; 47-61 14. Cremer, M.; Z. Biol.; (1906) 47; 562 15. Cammann, K.; Das arbeiten mit Ionselktiven Elektroden; II ed.; Springen, Berlin (1977) 16. Taguchi, T.; Naoyoshi, K.; Patent (1971) 11520; US-A 17. Schasfoort, R.B.M.; Kooyman R.P.; Bergveld, P.; Greve J.; Biosensors & Bioelectronics, (1990) 5; 103-124 18. Cass.A.E.G.; Biosensors: A practical Approach, Oxford Univ. Press, NY (1990) 19. Cannmann, K.; Fresenius Z. Anal. Chem. (1977) 287; 1-9 20. Clark Jr., L.C.; Lyons, C.; Ann. N. Y. Acad. Science, (1962) 102; pp. 29-45 21. Turner, A.P.F.; Karube, I.; Wilson, G.S.; Biosensors: Fundamental and Applications, Oxford Univ. Press, NY (1989) 22. Blum , J.L.; Coulet P.R.; Biosensor: Principle and Application, Oxford Univ. Press, NY (1989) 23. Clark Jr., L.C.: Lyons, C.; Ann. N. Y. Acad. Science, (1962) 102; pp. 29-45 24. Bidan, G., Deronzier, A.; Moutet, J.C.; J. Chem. Soc. Chemical Commun., (1984); p. 1185 25. Degani, Y.; Heller, A.; J. Amer. Chem. Soc., (1989) 111; pp. 2357-2358

128

26. Dicks, J.M.; Hattori, S.; Karube, I.; Turner, A.P.F.; Yokozawa, T.; Ann. Biol. Clin., (1989) 47; pp. 607-619 27. Pishko, M.V., Michael, A.C.; Heller, A.; Anal. Chem., (1991) 63; pp. 2268-2272 28. Bartlett, P.N.; Bradford, V.Q.; Whitaker, R.G.; Talanta, (1991) 38; pp. 57-63 29. Schuhmann, W.; Ohara, T.J.; Schmidt, H.; Heller, A.; J. Amer. Chem. Soc., (1991) 113; pp. 13941397 30. Bourdillon, C.; Majda, M.; J. Amer. Chem. Soc., (1990) 112; pp. 1795-1799 31. OHare, D.; Parker, K.H.; Winlove, C.P.; Bioelectrochem. Bioenerg., (1990) 23; pp. 203-209 32. Beh, S.K.; Moody G.J.; Thomas, J.D.R.; Analyst, (1991) 116, pp. 459-462 33. Yokohama, K.; Tamiya, E.; Karube, I.; Anal. Let., (1989) 22; pp. 2949-2959 34. Pishko, M.V.; Katakis, I.; Angew. Chem. Int. Ed. Engl., (1990) 29; pp. 82-84 35. Betso, S.R.; Klapper M.H.; Anderson, L.B.; J. Amer. Chem. Soc., 94; pp. 8197-8204 36. Armstrong, F.A.; Cox, P.A.; Journal of Electroanalytical Chem., (1987) 217; pp. 331-366 37. Hagen, W.R.; Eur. J. Biochemistry, (1989) 182; pp. 523-530 38. Heller, A.; Acc., Chem. Res., (1990) 23; pp. 128-134 39. Detaxis Du Poet, P.; Miyamoto, S.; Anal. Chimica Acta, (1990) 235; pp. 255-263 40. Armstrong F.A.; Acc. Chem. Res., (1988) 21; pp. 407-413 41. Boutelle, M.G.; Neurosci. Lett., (1986) 72; pp. 283-288 42. Kulys, J.; DaCosta, E.J.; Biosensors & Bioelectronics, (1991) 6; pp. 109-115 43. Albery, W.J.; Electrode kinetics; Oxford univ. Press, London; (1975) 44. Inczedy H.; Ure A. M. Eds., 1998, IUPAC compendium of analytical nomenclature, 3rd ed. Blackwell Science, Oxford 45. Eddowes, M.J.; Cass, A.E.G. ED. Biosensors: a practical approach, IRL Press, (1990) Oxford, pp.211 46 Guibault, G.G.; (1984) Handbook of iummobilized enzymes, M. Dekker, New York 47 Goepel, W.; Chemical and Biochemical Sensor in Sensor- A Comprehensive survey, (1991) vol. 2-3, VCH, NY 48 Rajagopalan, R.; J. Phys. Chem., (1996) 100, pp. 3719 49 Srere, P.A.; Uyeda, K.; (1976), Method in Enzymology, vol. XLIV (mosbach, k. Eds.), Academic, NY, pp. 19-45 50 Gemeiner P.; Enzyme engineering, Ellis Horwood- Gemeiner eds.; NY (1992); pp. 13-119 51 Gorton, L.; (1995) Electroanal., 7; p. 23 52 Kierstan M.P.; Coughlan M.P.; (1991) Protein Immobilization (Taylor, R.F. eds.), Marcel Dekker, NY, pp.13-71; Nilsson, K. (1987), Trends Biotechnology 5, 73-78 53 Broum , G.B.; (1976) Methods in Enzymology, vol. XLIV (Mosbach, K. Eds.), Academic press, NY, pp.263-280 54 Schumann, W.; Biosensors & Bioelectronics; (1993) 8; pp. 191-196 55 Heller, A.; Maidan, R.; Wang, D.; Sens. Actuators B; (1993) 13-14; pp. 180-183 56 Jos. Wang; Analytical Electrochemistry, (2000) cap.6, II ed. J. Wiley and S. inc., NY

129

57 Willner, I.; Katz, E.; Willner, B.; Electroanalysis, (1997) 9, pp.965 58 Scheller, F.W.; Pfeiffer D.; Z. Chem., (1978) 18; p. 50 59 Bard, A.J.; Faulkner, L.R.; Electrochemical Methods, Fundamental and Applications, J. Wiley & Sons, NY (1980) 60 Cass, A.E.G.; Biosensors, A Practical Approach, Oxford Univ. Press, NY (1990) 61 Wang J.; Analytical Electrochemistry, J. Wiley & Son, inc., NY (2000) 62 Bartlett, P.N.; Tebutt P.; Whitaker R. G.; Progr. Reaction Kinetics, vol.16 (1991); pp. 55 63 Marcus R.A.; Sutin, N.; Biochim. Biophys. Acta; 811 (1985), pp. 265 64 Rzicka, J.; Hansen, E.H.; Flow Injection Analysis; (1981) Wiley eds., NY 65 Ranger, C.B.; Anal. Chem.; (1981) 53; p. 20A 66 Painton, C.C; Mottola, H.A.; Anal. Chimica Acta; (1984) 158; p.67 67 Stewart, K.K.; Anal. Chemistry; (1983) 55; p. 931A 68 Rzicka J.; Hansen, E.H.; Anal. Chimica Acta; (1978) 99; p. 37 69 Harris, J.M.; Anal. Chemistry; (1982) 54; p. 2337 70 Painton, C.C.; Mottola, H.A.; Anal. Chimica Acta; (1981) 53; p. 1713 71 Rejin, J.M.; Vander Linden E.; Poppe, H.; .; Anal. Chimica Acta; (1981) 123; p. 229 72 Ananthakrishnan, V.; Gill, W.N.; Barduhn, A.J.; A.I. Chem. E.J.; (1965) 11; p.1063 73 Vanderslice, J.T.; Stewart, K.K.; Talanta; (1981) 28; p.11 74 Rzicka J.; Hansen, E.H.; Zagatto, E.A.; Anal. Chimica Acta; (1987) 88; pp. 1 75 Varcarcel A.; Flow Injection Analysis, chap. 3, pp.74-98 76 Keller, F.; Hunter, Robinson; Journal of Electrochem. Soc., 100 (1953), p.411 77 Masuda H.; Thin solid film, vol. 223 (1993); pp.1-3 78 Broughton, J.; J.Membr.Sci., 106 (1995); p. 89 79 Masuda, H., Science 268 (1995), p.1466 80 Itaya K., J. Chem. Eng. Jap., 17 (1984), p.514 81 Masuda, H.; Bull, Chem, Soc.Jap.; vol. 66 (1993), pp. 305-311 82 Masuda, H.; Chem lett, (1990) vol. 621 83 Masuda, H.; Satoh, M.; Jpn. J. Appl. Phys.; vol. 35 (1996), pp. L126-L129, part2 84 Masuda, H.; Fukuda, K.; Science; 268, (1995) p.1466 85 Bourdillon, C.; Maida, H., J. American Chem.Soc., vol. 112 (1990), p. 1795 86 Masuda, H.; J. of Electroanalytical Chemistry, vol.368 (1994); pp.333-336 87 Parthasarathy R.; Martin C.R.; Nature, vol. 369 (1994) 88 Janasek, D.; Spohn, U.; BIOforum International (1998); 2, pp.108-109 89 Spohn, U.; Preuschoff, F.; Blankestein, G.; Janasek, D.; Kula, M.; Hacker, A.; Anal. Chimica Acta (1995), 303; pp. 109-120 90 Henglein, A.; Chem. Rev. (1989), vol.89, pp.1861-1873 91 Colvin, V.L.; Schlamp, M.C.; Alivisatos, A.P.; Nature, (1994); vol.370 92 Bruchez M.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A.P.; Science; 281 (1998); p. 2013

130

93 Peng, X.; Manna, L.; Yang, W.; Wickham J.; Scher, E.; Kadavanich, A.; Alivisatos, A.P; Nature, vol. 404 (2000) 94 Eastoe, J.; Warne, J.; Curr. Opinions Colloid. Interface Science (1996), 1,800 95 Curri, M.L.; Agostiano, A., Manna, Della Monica, Synthesis and characterisation of CdS Nanoclusters in a Quaternary Microemulsion: the Role of the Cosurfactant,; J. Phys. Chemistry B; 30 june 2000 96 Sampath S.; Ovadia, L.; Adv. Material (1997), vol. 9; 410 97 Agostiano A., Curri M.L.; New Compounds and Materials. Syntheses and Methodologies in Inorganic Chemistry, vol.3 (1998), 261, ed. Vigato 98 Hagfeldt, K.; Gratzel, E.; Chemical Reviews (1995), 95; p.50 99 Colvin V. L., Goldstein A. N. and Alivisatos A. P., J. Am. Chem. Soc. 114 (1992), p. 5221 100 Swoboda, B.E.P.; Massey, V.; J. Biol. Chem. (1965) 240; pp. 2209-2215 101 Wingard, L.B.; Narasimhan, K.; Meth. Enzymology; (1988) 137; pp.103-111 102 Kalisz, H.; Hecht, H.; Journ. Molec. Biol.; (1990) 213; pp.207-209 103 Schopfer, L.M.; Massey, V.; Cliborne, A.; Journ. Molec. Biol.; (1981) 256; pp.7329-7337 104 Frederick, K.R. et al.; Journ. Molec. Biol.; (1990) 265; pp.3793-3802 105 Barker, S.A.; Shirley, J.A.; Economic Microb.; vol. 5; edited by Rose, A.H. Academic press, NY (1980), pp.455-465 106 Suzuki, H.; Aminoacids, (1994) 7; pp. 27-43 107 Maradas, M. B.; Buck R.P.; Anal. Chemistry, (1996) 68, pp.3832-3839 108 Shi, Y.; Biosensors & Bioelectronics, (1997) vol. 12 n.7, pp. 655-659 109 Suzuki, M.; Purification and some properties of Sarc. oxidase from Corynebacterium sp. U-96; J. Biochem.; (1980) 89; pp.559-607 110 Kinoshita, T.; Hiraga, Y.; Chem. Pharm. Bull.; (1980) 28; pp.3501-3506 111 Kim, J.M.; Shimizu, S.; Yamada, H.; Crystallisation and characterisation of sarcosine oxidase from Alcaligenes denitrificans; Agric. Biol. Chem. (1987) 51; pp. 1167-1168 112 Cook, R.J.; Identification of the covalently bound flavin of dimethylglycine dehydrogenase and sarc. Dehydrogenase from rat liver mitochondria; J. Biol. Chem. (1980) 259, pp.12475-12480 113 Decker, K.; Flavins and Flavoproteins; Elsevier N. Holland; Amsterdam, Massey and Williams CH eds. (1982); pp.465-472 114 Mori, N.; Sano, M.; Tani, Y.; Yamada, H.; Agric. Biol. Chem.44; pp.1391-1397 115 Matsuda, H.; Chem. Pharm. Bull.; (1982) 35; pp.711-717 116 Koyama, K.; Agric. Biol. Chem., 55; pp. 1259-1263 117 Inoue, A.; Chem. Pharm. Bulletin; 35; pp.4194-4202 118 Kim, J.M.; Agric. Biol. Chem.; 50; pp 2811-2816 119 Suzuki, K.; J. Biochem.; 89; pp.559-607 120 Ogushi, S.; Chem. Pharm. Bulletin; 36; pp.1445-1450

131

121 Koyama, Y.; Yamamoto-Otake, H.; Suzuki, M.; Nakano, E.; Cloning and expression of the sarcosine oxidase gene from Bac. Sp. NS-129 in Escherichia coli, Agri. Biol. Chem. (1991) 55; pp.1259-1263 122 Hayashi, S.; Steady state kinetics and spectral properties of Corynebacterium sarcosine oxidase, Bioch. Biophysica Acta, (1983); 742; pp.630-636 123 (a) Ogushi, S.; Ando, M.; Tsuru, D.; J. Biochemistry; 96 (1984); pp. 1587-1591 (b) Ando, M.; Yoshimoto, T.; Ogushi, S.; Rikitake, K.; Shibata, S.; Tsuru, D.; J. Biochemistry; 85 (1979); pp. 1165-1172 124 Schumann, W.; Biosensor & Bioelectronics; 8 (1993); pp.191-196 125 Dominguez Sanchez, P.; Miranda Ordieres, A.; Costa Garcia, A.; Blanco, P.; Peroxidaseferrocene modified carbon paste electrode as an amperometric sensor for the hydrogen peroxide assay; Electroanalysis; 3 (1991); pp. 281-285 126 O'Hara, T.; Vreeke, M.; Battaglini F.; Heller, A.; Bienzyme sensors based on "electrically wired" peroxidase, Electroanalysis; 5 (1993); pp. 825-831 127 De Lumley-Woodyear, T.; Rocca, P.; Lindsay, J.; Dror, Y.; Freeman, A.; Heller, A.; Polyacrylamide-based redox polymer for connecting redox centers of enzymes to electrodes, Analytical Chemistry 67 (1995); pp. 1332-1338 128 Kenausis, G.; Chen, Q.; Heller, A.; Electrochemical glucose and lactate sensors based on "wired" thermostable soybean peroxidase operating continuously and stably at 370 C, Analytical Chemistry; 69 (1997); pp. 1054-1060 129 Sadeghi, S.; Gilardi, G.; Cass, A.; Mediated electrochemistry of peroxidase effects of variations in protein and mediator structures, Biosensors & Bioelectronics; 12 (1997); pp. 1191-1198 130 Qian, J.; Liu, Y.; Liu, H.; Yu, T.; Deng, J.; Immobilization sensor, Biosensors & Bioelectronics 12 (1997); pp. 1213-1218 131 Yaropolov, A.; Malovik, V.; Varfolomeev, S.; Berezin, I.; Electroreduction of hydrogen peroxide and an electrode with immobilized peroxidase, Dokl. Akad. Nauk. SSSR 249 (1989); pp. 13991401 132 Jnsson, G.; Gorton, L.; An electrochemical sensor for hydrogen peroxide based on peroxidase adsorbed on a spectrographic graphite electrode, Electroanalysis 1 (1989); pp. 465-46810 133 Popescu, I.; Zetterberg, G.; Gorton, L.; Influence of graphite powder, additives and enyzme immobilization procedures on a mediatorless HRP-modified carbon paste electrode for amperometric flow injection detection of H2O2; Biosensors & Bioelectronics 10 (1995); pp. 443461 134 Ruzgas, T.; Csregi, E.; Emneus, J.; Gorton, L.; Marko-Varga, G.; Peroxidase-modified electrodes: fundamentals and application, Analytica Chimica Acta 330 (1996); pp. 123-138 135 Gorton, L.; Carbon paste electrodes modified with enzymes, tissues, and cells, Electroanalysis 7 (1), 1995; pp. 23-45 of horseradish peroxidase withregenerated silk fibroin membrane and its application to a tetrathiafulvalene-mediating H2O2

132

136 Gorton, L.; Svensson, T.; An investigation of the influence of the background material and layer thickness of sputtered palladium/gold on carbon electrodes for the amperometric determination of hydrogen peroxide, Journal Molecular Catalysis 38 (1986); pp. 49-60 137 Wang, J.; Naser, N.; Angnes, L.; Wu, H.; Chen, L.; Metal-dispersed carbon paste electrodes, Analytical Chemistry 64 (1992); pp. 1285-1288 138 Wang, J., Fang, L.; Lopez, D.; Tobias, H.; Highly selective and sensitive amperometric biosensor of glucose at ruthenium-dispersed carbon paste enzyme electrodes, Analytical Letters 26 (9) (1993); pp. 1819-1830 139 Wang, J.; Chen, Q.; Pedreno, M.; Highly selective biosensing of lactate at lactate oxidase containing rhodium dispersed carbon paste electrodes, Analytica Chimica Acta; 304 (1995); pp. 41-46 140 Sakslund, H.; Wang, J.; Lu, F.; Hammerich, O.; Development and evaluation of glucose microsensors based on electrochemical codeposition of ruthenium and glucose oxidase onto carbon fiber microelectrodes; Journal Electroanalytical Chemistry 397 (1995); pp. 149-155 141 O'Connell, P.; O'Sullivan, C.; Guilbault, G.; Electrochemical metallization of carbon electrodes, Analytica Chimica Acta; 373 (2) (1998); pp. 261-270 142 Heilmann, A.; Altmann, F.; Katzer, D.; Mller, F.; Sawitowski, T.; Schmid, G.; Determination of the pore size and the vertical structure of nanoporous aluminum oxide membranes, Applied Surface Science 144-145 (1999); pp. 682-685 143 Homilius, F.; Heilmann, A.; Werner, J.; Grnwald, W.; The influence of the polymer matrix on the properties and the structure of plasma polymer silver composite films, Physica status solidi (a) 137 (1993); pp. 145-154 144 Vastarella, W.; Janasek, D.; Spohn, U.; submitted to Analytical & Bioanalytical Chemistry on november 2001 145 Koopal, K; Third Generation Amperometric Biosensors; PhD thesis (1999) 146 Rzicka J.; Hansen, E.H.; Flow Injection Analysis; 2nd Edition, J. Wiley & Sons, New York, (1988) 147 Clark Jr., L.C.; Methods in Enzymology, vol. LVI, Ac. Press, NY (1979); pp. 448-479 148 Jeppesen, M.T.; Hansen E. H.; Anal. Chimica Acta (1988), 214, pp.147-155 149 Jaff, M. Z.; Physiol. Chem.; (1886) 10, pp. 391-400 150 Soldin, S.; Henderson, L.; Hill, J.G.; Clin. Biochemistry, 11 (1978), p.82 151 Brown, ND.; Clin. Chem.; (1977) 23, pp.1281-1283 152 Suzuki, H.; Amino Acids; (1994) 7; pp.27-43 153 Moss, G.A.; Clin. Chem.; 46 (1974); p.246 154 Tsuchida, T.; Yoda, K.; Clin. Chemistry (1983) 29; pp. 51-55 155 Maradas, M. B.; Popescu, I.C.; Ufer, S.; Buck R.P.; Analytica Chimica Acta (1996) 319, pp.335345 156 Sakslund, H.; Wang, J.; Lu, F.; Hammerich, O.; Journal of Electroanalytical Chemistry; (1995) 397, pp.149-155

133

157 Spohn, U.; Van der Pol, J.; Eberhardt, R.; Joksch, B.; Wandrey, C.; Anal. Chimica Acta; (1994) 292, pp.281-295 158 Janasek, D.; Spohn, U.; Kiesow, A.; Heilmann, A.; Sensor and Actuators B; 78 (2001), pp 228231 159 Janasek, D.; Spohn, U.; Porous Glass as Carrier in Enzyme immobilization for Flow Injection Analysis; proceeding BIOforum International 2 (1998); pp. 38-39

134