ANALYST, AUGUST 1992, VOL. 117 1215 Chemically Modified, Carbon-based Electrodes and Their Application as Electrochemical Sensors for the Analysis of Biologically Important Compounds A Review Stephen A. Wring Glaxo Group Research ltd., Division of Drug Metabolism, Park Road, Ware, Hertfordshire SG12 ODP, UK John P. Hart Faculty of Applied Sciences, Bristol Polytechnic, Coldharbour lane, Frencha y, Bristol BS 16 IQY, UK Summary of Contents Introduction Overpotential and Electrocatalysis Organometallic Mediators Ferrocenes Phthalocyanines Hexacya nofer rate( 111) Ruthenium Oxide Complexes Meta I loporphyri ns Phenoxazines, Phenathiazines and Phenazines Organic Mediators Phenoxazine mediators Meldola Blue Nile Blue Phenathiazine mediators Phenazine mediators Qu i none-H yd roqu i nones Tetrat h iaf u Iva lene and Tetracyanoq u i nodi methane Tetrat h iafu Iva lene (TTF) Tet racya noq u i nod i metha ne (TCN Q) Phenylenediamine (PD) and Tetramethylphenylenediamine (TMPD) Conclusions References Keywords: Sensor; mediator; carbon-based electrode; biological and clinical compounds; electrochemistry; review Introduction There is a perceived and increasing demand for simple, inexpensive and rapid analytical tests for the determination of trace concentrations of biologically and clinically important compounds.Electrochemical techniques, employing sensitive amperometric sensors, are particularly suited for these appli- cations. Amperometric sensors exploit the use of an applied potential, between a reference and working electrode, to cause the oxidation or reduction of an electroactive species.This process gives rise to either an anodic (oxidation) or cathodic (reduction) current which may be related to the concentration of a compound in solution, Working electrodes may be constructed from different materials; for example, noble metals such as platinum'-3 and gold4.5 have been used successfully for several applications. Recently, much interest has centred on the use of carbon as an inexpensive substrate for electrochemical sensor design; unfortunately, it rarely lends itself to direct biomedical analyses owing to the often impractically high activation overpotential required for the oxidation or reduction of biomolecules at its surface. This is an important disadvantage because specificity is inversely related to the magnitude of the applied potential; therefore, it is generally desirable to avoid the use of extreme potentials whenever possible.Fortunately, promising advances towards improved selec- tivity in carbon-based electrochemical sensors have been achieved through judicious surface modification of the work- ing electrode with redox mediators, which facilitate the charge-transfer between the electrode and an organic species or enzyme in solution at much lower potentials than would otherwise be possible. This review considers the concept, design, development, application and analytical performance of these carbon-based devices. The first part of the review briefly describes overpotential and the methods by which it may be significantly reduced using electrocatalytic processes.This section is followed by the main part of the review which describes the use of electron mediators in the different electrode designs, and critically discusses the factors affecting their selection, prior to the construction and application of individual sensor devices. In some instances, a comparison is made between the per- formance of different mediators for the detection of a particular biomolecule; it is hoped that this approach will provide useful practical information for researchers intending to work in this important area. The electrocatalysts described in this review have shown particular promise, or success, for important practical applica- tions in biomedical, environmental and biotechnological analyses; these include, organometallic mediators, such as the ferrocenes, phthalocyanines, ruthenium oxides and metallo- porphyrins, and several organic mediators, including quinone, tetracyanoquinodimethane, tetrathiafulvalene, phenazine,1216 ANALYST, AUGUST 1992, VOL.117 phenoxazine, phenathiazine and phenylenediamine-based compounds. The preparation and manufacture of carbon as an electrode substrate material is deemed to be beyond the scope of this review and readers are referred to the excellent reviews and monographs already available .6-8 Overpotential and Electrocatalysis Overpotential can be investigated using cyclic voltammetry, an electroanalytical technique that may be used for mechanis- tic studies during the development and optimization of an electrochemical sensor.It can rapidly provide information on the magnitude of any overpotential and the reversibility, or otherwise, of a particular electrochemical (EC) process. In cyclic voltammetry, a triangular potential waveform is applied to the working electrode as indicated in Fig. l(a). When the electrochemical process is reversible the cyclic voltammogram exhibits current peaks on both the forward and reverse scans, [Fig. l(6)l and their separation will be 0.059In V ( i e . , AE, = E,, - E,, = 0.059In; where, n is the number of electrons transferred). For a quasi-reversible process the AE, value is greater than 0.059/n V. If the reaction is irreversible only a single peak will be observed on one of the potential scans. For a perfectly reversible system E4 = Eo; {where: Eo is the standard potential [E" = (E,, + E,,)/2] and EJ is the half-wave potential (in a reversible system, this is the potential when both the oxidized and reduced forms of the compound are present at the electrode surface in equal concentrations)}.However, in the event of any deviation from ideal behaviour there will be a certain overpotential (q) where q = E+ - Eo. Therefore, the magnitude of a voltammetric or amperometric current response will be governed by a certain overpotential (qeffective) that, in protic media, will be composed of the important components presented below and summarized in Fig. 2.9-11 Cycle 1 Cycle 2 ----- I (a' 81 Ti me/s PotentialN versus SCE Fig. 1 ( a ) Potential-time profile used for cyclic voltammetry. ( b ) The current-potential response for a reversible species using cyclic voltammetry E I ectrode ne Electrode surface reclion Bulk solution Chemical reactions '- Chemical reactions 'A - I I Mass transfer - O s u x " Obulk I I I I I I I 1 I R,U,?--- I Rbulk Fig.2 ref. 9 Pathway of a general electrode reaction. Reproduced from llmass transfer This arises because in quiescent solutions a concentration gradient develops between the electrode surface and the bulk solution. To overcome this gradient an additional potential is required, and in an extreme case, with a fast reaction and high current densities, a limit is reached determined by the maximum rate of transfer of ions to the electrode. This type of overpotential may be removed by stirring the test solution, and may lead to the differences seen in applied potentials required for quiescent and hydrodynamic techniques.b a c t i o n This term describes the overpotential associated with either a preceding o r follow-on chemical reaction, or adsorption, at the electrode surface. qactivation When the activation energy barrier for electrolysis is increased at a given electrode material the additional potential required to sustain the reaction at a given rate is termed the activation overpotential. In electrochemical studies using carbon elec- trodes it is generally activation overpotential that predomi- nates and it is this that must be addressed to improve method sensitivity and selectivity. During the development of any electrochemical method, or sensor, it is imperative to consider the effects of overpotential, particularly when electrochemical detection is considered for the sensor-based applications.Indeed, the development of sensors has especially benefitted from the modification of electrodes with electrocatalytic species because the minimiza- tion of activation overpotential has allowed increased ana- lytical selectivity through the use of lower operating poten- tials. 1 ~ 1 3 Fig. 3(a) shows a schematic representation of the electro- catalytic charge-transfer process for a soluble substrate molecule (or enzyme) undergoing chemical oxidation by a mediator which is subsequently re-oxidized at the electrode surface. This mechanism can be described as a CE (chemical then electrochemical) process; in other applications the mediator itself may be electrochemically oxidized prior to its reaction with the substrate species [e.g., with ferrocene,14 see Fig.3(6)] this is known as an EC process. In both instances the electron transfer sequence incorporates a catalytic regenera- tion mechanism where the mediator is replenished and hence becomes available for further reaction with any remaining substrate molecules. I act0 n e Fig. 3 ( a ) Schematic diagram of a mediated charge transfer process. ( 6 ) Charge transfer process for the determination of glucose at a ferrocene-modified electrode. Reproduced with permission from ref. 15ANALYST, AUGUST 1992, VOL. 117 1217 t c.l C ?? 3 0 I I I I I I 0 0.1 0.2 0.3 0.4 0.5 PotentialN versus SCE Fig. 4 A, d.c. cyclic voltammogram of ferroccne monocarboxylic acid (0.50 mmol dm-3) at pH 7 and 25 "C, in the presence of D-glucose (50 mrnol dm-3) at a scan rate of 1 mV s-1.B, as for A, but with the addition of glucose oxidase. Rcproduced with permission from rcf. 15 Organometallic Mediators Ferrocenes The ferrocene group of mediators have been successfully applied to the quantitation of several compounds and have been particularly important in the development of sensors for the determination of glucose in a variety of sample matrices. Cass et uE.15 reportcd one of the earliest applications using an amperometric enzyme electrode for the determination of circulating levels of glucose in blood samples collected from patients suffering from diabetes mellitus. For their study the authors used glucose oxidase (GOD) to convert glucose to gluconolactone; the reduced form of the cnzyme was subse- quently re-oxidized using the electrochemically generated dimethylferricinium ion [Fig.3(6)]. This latter stage gave rise to an anodic current response which was proportional to glucose in the concentration range 1-31) mmol dm-3 of glucose. For these electrodes,lS individual ferrocenes were dc- posited from solution (15 mm3 of 0.1 mol dm-3 in toluene) onto 4 mm discs of graphite foil which had been sealed into glass tubes; whereupon, the GOD was covalently bound to the graphite surface using the carbodiimide reaction. In these devices the ferrocenes were employed as low molecular mass mediators, replacing oxygen, for the re-oxidation of the flavin adenine dinucleotidc (reduced) (FADH) prosthetic group within GODreduced and to facilitate the electron transfer between the enzyme and the working electrode'" (Fig.4). Dimethylferrocene was selected as the mediator of choice foliowing a study employing cyclic voltammetry to determine the redox potentials of several ferrocene derivatives (Table 1). The data in Table 1 clearly indicate that functionalities present on the ferrocene molecule will cause variations in the electrochemical bchaviour which may be exploited to improve the selectivity of an electrode's response. The electrode design of Cass et al. has also been adapted for the determination of glucose in other sample matrices, p . g . , molasses,17 where excellent correlation was achieved between the sensor and a standard method using gas-liquid chromato- graphy (correlation coefficient = 0.98).Indeed. thc clectro- chemical method was superior because it did not require any sample preparation. Another, and undoubtedly the most succcssful , application of the ferricinium-glucose oxidase dctection scheme, has been the dcvelopment and commercial- ization of the ExacTech disposable glucose sensorls-1" for the Table 1: Formal redox potentials for ferrocene and its derivatives. (Data from ref. 15) E"'/mV vmus SCE at pH 7 Ferroccne derivative 1, I '-Dimethyl 100 Ferrocene { bis (n-cyclopentadieny1)iron; [(CsHs)2-]Fe2+) 165 Vinyl 250 Carbox y 275 1 ,l'-Dicarboxy 285 400 (Dirnct h yiami no)met h yl Conductive Working PVC substrate silver track electrode \ ,Contacts 1 I f I Conductive Dielectric carbon track layer I Ag-AgCI reference electrode Fig.5 with permission from ref. 18 ExacTech disposable glucosc clectrode strip. Reproduced personal monitoring of whole blood glucose concentrations in diabetics20 These bionsensors are fabricated using screen- printing technology to deposit electrode substrates and enzyme reagents onto an inert poly(viny1 chIoride) (PVC) support material (Fig. 5 ) . Individual electrode strips are inserted into a small 'pen-shaped' measuring device that provides the applied potential and converts current response information to a digital glucose concentration value. In other studies, substituted ferrocenes have been devel- oped to improve the long-term stability of electrodes which are based on the adsorption of mediators onto graphite surfaces. Jonsson et 01.21 reacted hydroxymethylferrocene with anthracene-9-carboxylic acid to produce a novel anthracene substituted ferroccne. The authors suggest that the enhanced lifetime of the electrode is achieved due to the anthracene moiety providing an 'anchor' for the adsorption of ferrocene onto the graphite.Howevcr, the E"' of the new species was +295 mV versus a saturated calomel (reference) electrode (SCE) (pH 7) which was significantly morc positive than the value for ferrocene, and thus the concomitant reduction in selectivity may render the mediator unsuitable for applications using biological samples. In addition to covalent attachment, enzyme electrodes have been prepared by simply mixing GOD and a ferrocene into carbon paste or carbon-epoxy resin substrates. For example, Wang et al.22 produced modified carbon-paste electrodes (CPEs) with very fast response times (t9slyo = 18 s) which were achieved owing to the intimate contact between enzyme, mediator and sensing sites. However, in some instances the electrodes displayed a loss of activity ( ~ 2 0 4 0 % ) over the first 6 h of use, although they were stable after this period. Gorton et al. 23 redressed this problem using a ferrocene-containing siloxanc polymer which was mixed into the carbon paste together with the enzyme. Selectivity was also enhanced by coating the surface of the electrode with a poly(ester-sulfonic acid) cation-exchanging polymer (Eastman AQ-29D) which excluded potential interferents such as ascorbic and uric acids, and protected thc clectrode from fouling caused by bovine serum albumin (BSA) and endogeneous species present in urine.These coated electrodes retained their initial current responses to glucose (+1%) over the three-week period studied. More recently, Hale et ul.24 have shown that the response of this type of glucose electrode can be enhanced by systematic modification of the ferrocene-siloxane polymer backbone. This study indicated that the optimum polymeric1218 ANALYST, AUGUST 1992, VOL. 117 Table 2 Further applications of ferrocene modified electrodes Analyte Electrode Glucose Amperometric enzyme electrode Aromatic amines Differential-pulse (diphenylamine) using a modified polymer electrode peroxide graphite electrode Hydrogen Amperometric Cholesterol Amperometric enzyme, carbon- paste electrodes Glycolate Amperometric enzyme, carbon- paste electrodes Galactose, Amperome tric glycolate and enzyme, carbon- L-amino acids paste electrodes Electrode construction and application Discs (of 6 mm diameter) were sealed into glass tubes, and electrical contact made with silver-loaded epoxy resin.A 5 mm3 solution containing 20 mg cm-3 of dimethyferrocene in toluene was deposited onto the graphite and allowed to dry. Nicotinamide adenine dinucleotide phosphate (oxidized) [NAD(P)+] independent glucose dehydrogenase (GDH) was immobilized onto the graphite using a condensation reagent followed by chemical cross- linkage. The calibration range was up to 15 mmol dm-3 glucose (dependent on the amount of enzyme, immobilized). Rate constants, response times (tgsy0, 10-20 s) and currents densities were superior to similar GOD based devices.The enzymic reaction of GDH is also independent of oxygen A polymer was prepared (from divinylbenzene) containing ferrocene and styrene sulfonate groups; the latter species served to preconcentrate the amines via an ion exchange mechanism. The limit of detection was =lo-8 mol dm-3 diphenylamine Ferrocenemonocarboxylic acid and horseradish peroxidase were used in the aqueous phase. The reaction involved a 2e- oxidation of the peroxidase by Hz02, whereupon the reduced enzyme was regenerated by ferrocene used in its reduced form. The ferricinium ion produced was subsequently reduced amperometrically at +89 mV versus SCE ( E O ’ for ferrocenemonocarboxylic acid in this system was +275 mV). The linear calibration range was 5 X 10-8-6 X mol dm-3.Redox conversion of peroxidase is not observed at graphite electrodes in the absence of mediator Kohlrabi (Turnip cabbage) has also been used as a source of peroxidase activity for ferrocene modified CPEs Three approaches have been employed for the determination of total cholesterol; however, each method used ferrocene derivatives to facilitate the mediated electron transfer between an enzyme and the electrode: (1) cholesterol was oxidized by cholesterol oxidase and current response was recorded in a manner similar to the determination of glucose using GOD, [Fig. 3(b)]; (2) cholesterol was oxidized by cholesterol dehydrogenase, the nicotinamide adenine dinucleotide (reduced) (NADH) produced was determined using diaphorase linked to the ferricinium ion; and (3) cholesterol was oxidized by cholesterol oxidase and the hydrogen peroxide produced was monitored using the method above.In each instance total cholesterol was monitored after its dissociation from lipoprotein complexes and esters using surfactants and cholesterol esterase, respectively Carbon-paste electrodes were prepared containing a dimethylferrocene modified siloxane polymer and glycolate oxidase; the overall reaction mechanism is similar to that described for GOD [Fig. 3(b)]. The magnitude of the electrode response was determined at different positions along the hydrodynamic wave, and the maximum currents were recorded at +300 mV which corresponded to a point shortly before the plateau on the hydrodynamic wave Carbon-paste electrodes were prepared containing dimethylferrocene and the individual oxidase enzymes for each analyte; a 20 mm diameter, 30 nm pore size, Nucleopore membrane was placed over the end of the electrode to exclude air.The authors suggest that this general method of biosensor construction should be widely applicable to oxidase enzymes Ref. 30 31 32 33 34, 35 36 37 electron relay system is a balance between intimate ‘enzyme- polymer’ contact obtained with increased polymer flexibility, and improved current response obtained by decreased spacing between the individual ferrocene electron relay sites. Hale et al.25 have also evaluated CPEs containing glucose oxidase and a ferrocene-modified poly(ethy1ene oxide) polymer which was similarly dependent on the density of bound ferrocene moieties along the polymer chain.Further examples of novel CPEs have been described by Beh et ~ 1 . 2 6 and Okuma et al.27 who have developed modified enzyme electrodes using cellulose triacetate as a support and binding matrix. Other important design factors which are likely to affect the long-term stability of redox modified carbon-paste enzyme electrodes have been investigated by Amine et aZ.28 More recently, Wang and Varughese29 have described polishable enzyme electrodes prepared in a similar manner to their CPEs using a graphite-epoxy resin matrix. The electrodes could be used without any deterioration in response for several weeks, although a fresh surface was produced by polishing prior to each experiment; these electrodes could also be used in some organic solutions (50 + 50 v/v methanol + phosphate buffer).Several other recent applications of carbon-based elec- trodes which have been chemically modified with ferrocenes are described in Table 2. Phthalocy anines Each of the transition metal complexes of the phthalocyanine macrocycles is electroactive ,3340 and several have been investigated for numerous types of application,41,42 e.g., electrocatalysis, photovoltaics, photoconductivity, etc. However, cobalt phthalocyanine (CoPC) has shown the most promise for the electrocatalytic determination of biologically important compounds. One of the earliest reports regarding the application of CoPC for electrocatalytic applications was described by Zagal et aZ. ,43 who employed chemically modified pyrolytic graphite electrodes, mounted in Kel F, for the mediated oxidation of cysteine (which requires applied potentials greater than +SO0 mV at unmodified graphite electrodes). The modified elec- trodes were prepared by adsorption of cobalt phthalocyanine tetrasulfonate (CoTSPC) onto the polished graphite surface from an aqueous solution containing 10-5 mol dm-3 concen- trations of the mediator.After 20 min of immersion the electrodes were removed from the solution and rinsed with purified water prior to use. The behaviour of the modified electrodes for the mediated oxidation of cysteine was subse- quently studied by means of cyclic voltammetry in both alkaline (0.1 mol dm-3 Na2C03) and acidic (0.2 mol dm-3 NaH2P04) solutions. In each of the plain electrolyte solutions, a reversible couple was recorded at Eo’ -500 mV versus SCE which was attributed to Co+TSPC e Co2+TSPC + e- (1) On the addition of l mmol dm-3 cysteine, an extra irreversible wave was recorded in both alkaline and acidic solutions (Epa, 0ANALYST, AUGUST 1992.VOL. 117 1219 and = + 450 mV versus SCE, respectively). The mechanism proposed for the oxidation of cysteine in alkaline media is R-SH C RS- + H+ RS- -+ RS' + e- RS' + RS' -+ R-SS-R where R-SH is cysteine and R-SS-R is cystine. The irrevers- ible anodic wave was thought to involve an EC mechanism involving Co". In later reports, Baldwin and co-~orkers44,45 mixed insoluble CoPC directly with graphite and Nujol to produce modified carbon-paste electrodes which they employed for the mediated electro-oxidation of hydrazine, several cysteine compounds and reduced glutathione (GSH) .Their results, using cyclic voltammetry, indicated that the thiol containing compounds underwent a mediated oxidation at = +800 mV versus Ag-AgCI (0.05 mol dm-3 H2S04) which corresponded to the potential of the Co'I oxidation wave recorded in plain electrolyte solutions. The modified carbon- paste electrodes were then employed for the amperometric determination of cysteine and glutathione in human plasma and whole bl00d,46 following their separation by reversed- phase high-performance liquid chromatography (HPLC), ( Eapplied, + 750 mV versus Ag-AgC1) . The present authors have also been interested in the possibility of measuring GSH with CoPC modified CPEs. In one study, we performed systematic investigations using cyclic voltammetry to elucidate the electrochemical behaviour of GSH at modified and unmodified electrodes and to optimize the solution conditions for quantitative analysis .47 These investigations revealed the presence of an additional anodic wave for the oxidation of GSH occurring at potentials less anodic (=O V versus SCE; Britton-Robinson buffer, pH 12) than those described by earlier investigators.This wave is thought to arise from the chemical reduction of Co" to Col by GSH, followed by an electrochemical re-oxidation of Co' to Co", [Fig. 3(a)]; at unmodified electrodes applied potentials of greater than +1 V are required to oxidize the thiol. Subsequently, a supporting electrolyte solution containing 0.1 mol dm-3 phosphate buffer (pH 5.0) was found to be optimal for quantitative analysis of GSH by differential-pulse voltam- metry (DPV). An attempt to use the optimized modified CPEs and solution conditions for the determination of GSH by flow injection (FI) proved unsuccessful owing to leaching of the mediator from the electrode matrix.To redress this problem a re-usable CoPC modified graphite-epoxy resin electrode was developed. This electrode material is easy to prepare and was investigated as an amperometric sensor for GSH determination using stirred solutions in a conventional voltammetric cell and in a thin-layer flow through cell. In the former mode it was found that a potential of +0.5 V versus SCE enabled a linear calibration graph for GSH to be constructed over the concentration range of from 3.9 yg cm-3 to 1.69 mg cm-3 and the limit of detection was 10 ng cm-3.The thin-layer flow through cells containing the CoPC modified graphite-epoxy resin electrodes were initially used as stable detectors for FI; we were able to perform repeat injections of 15 yg of GSH over 270 min without any significant loss in sensitivity with a relative standard deviation (RSD) of 3.1% ( n = 22). These results indicated that the CoPC sensor material could prove to be a useful detector for HPLC-EC determinations of GSH. Indeed, using this technique we were able to determine endogeneous GSH levels present in normal samples of human plasma48 and whole bl00d.49 In a separate study50 the same electrode material was employed to deter- mine ascorbic acid in single- and multivitamin preparations.The CoPC mediator lowered the applied potential necessary for the electro-oxidation of the vitamin by -150 mV to +0.25 V versus SCE. This allowed determinations to be performed without the need for any chromatographic stage, by spiking solutions prepared from the tablets into 0.05 (2) (3) (4) rnol dm-3 phosphate buffer (pH 5 ) within the voltammetric cell. Amperometry was then performed in stirred solutions and quantitation was carried out by the method of standard additions. Whilst we were preparing our initial reports, Wang et aL51 briefly reported a similar electrode material which they had used for the amperometric detection of cysteine, oxalic acid, penicillamine and hydrazine in individual standard solutions.Cobalt phthalocyanine is also an ideal mediator for incor- poration into screen-printed carbon electrodes (SPCEs) as it can be ground into a very fine powder and has minimal solubility in aqueous media. Using these properties the present authors have described a simple method for the rapid and reproducible production of disposable sensors; these have been evaluated by performing cyclic voltammetry and amper- ometry in stirred solutions on 0.05 mol dm-3 phosphate buffer (pH 5 ) solutions containing either ascorbic acid, GSH or coenzyme A.52 Using the former technique it was found that the overpotential for the electro-oxidation of ascorbic acid could be reduced by 350 mV, and by at least 600 mV for GSH and coenzyme A. These results were confirmed by construct- ing hydrodynamic voltammograms using the amperometric technique.The performance of the CoPC modified SPCEs for the quantitative analysis of ascorbic acid and GSH was investigated using amperometry in stirred solutions and DPV. Using the former technique the limit of detection (LOD) was 5 x 10-8 and 1.48 x 10-7 rnol dm-3 for ascorbic acid and GSH, respectively, based on a signal-to-noise ratio of 3 : 1. The calibration graphs were linear from the LODs to 2 mmol dm-3 concentrations. The differential-pulse voltammo- :a) 25 pl of 0.01 no1 dm-3 GSH added 2 min [b) Potential applied to the SPCE 1 Enzyme and tea-butyl hydro peroxide added 1 min I Time - Fig. 6 Amperometric current response rofiles for: (a) a 0.05 rnol dm--? phosphate buffer solution (pH 77 spiked at the indicated point to give 50 pmol dm-3 GSH; and (b) a haemolysate sample prepared from human whole blood after pre-treatment by ultrafiltration with a molecular mass cut-off filter of 30 OOO.Reproduced with permission from ref. 551220 ANALYST, AUGUST 1992, VOL. 117 Table 3 Further applications of CoPC modified electrodes Analyte Thiopurines Oxalic and pyruvic acids Oxalic and ascorbic acids and cysteine Dopamine Carbohydrates Nucleosides Electrode carbon-paste electrodes Modified Modified carbon- paste electrodes in a thin-layer flow through cell Modified, cellulose acetate coated, glassy carbon electrodes in a thin-layer flow through cell Modified, Nafion coated, glassy carbon electrodes Modified, carbon- paste electrodes Modified carbon- paste electrodes Electrode construction and application Ref.Cyclic voltammetric studies were performed using several thiopurine compounds; voltammo- 57 grams shown for 6-mercaptopurine (6-MP) indicated that the overpotential for its oxidation at unmodified glassy carbon and carbon-paste elctrodes could be reduced from = +480 mV and -+540 mV versus Ag-AgCI (pH 7) to -320 mV using the CoPC modified carbon-paste electrodes. The 6-MP and 6-mercaptopurine riboside in plasma were determined by HPLC-EC (Eapplied, +600 mV versus Ag-AgC1) Oxalic acid and pyruvic acid were determined in human urine samples collected from normal 58 subjects and patients having a tendency to form renal stones. The use of the CoPC modified electrodes reduced the overpotential for the oxidation of these species by 350 mV, to +750 mV versus Ag-AgC1; this decrease in applied potential effectively removed interfering peaks present using unmodified electrodes operated at + 1.1 V Glassy carbon electrodes were coated with a hydrolysed cellulose acetate film, containing the CoPC, which acted as a permselective membrane allowing the passage of each analyte whilst excluding albumin which poisoned the unprotected electrodes 59 Glassy carbon electrodes were coated with a Nafion film containing the CoPC; the 60 cation-exchange film improved selectivity by excluding ascorbic and oxalic acid and improved sensitivity by preconcentrating dopamine into its matrix The cyclic voltammetric behaviour of several mono- and disaccharides were investigated at CoPC modified carbon-paste electrodes using 0.15 mol dm-3 NaOH as the supporting electrolyte.The E,, values for their oxidation were typically in the range from 400 to 500 mV versus Ag-AgCI; no anodic current response was recorded at unmodified electrodes scanned up to +700 mV. A method, using HPLC-EC, was developed for the determination of glucose and fructose concentrations in soft drinks Alditol and acidic carbohydrate derivatives were studied in a separate investigation; the electrocatalysis was similar to the simple carbohydrates with E,, values in the range from +400 to +600 mV versus Ag-AgCI (0.15 mol dm-3 NaOH) ribonucleosides; cytidine, uridine, adenosine and guanosine, cyclic voltammograms were recorded, at the modified electrodes, for the sugar ribose and the base cytosine; only the former showed any electroactive behaviour over the potential range studied (from - 100 to +700 mV; Epa, 450 mV versus Ag-AgC1) neither compound exhibited electroactive behaviour at unmodified electrodes.These results suggest that it is the ribose moiety of the nucleosides that gave the anodic response at the CoPC modified electrodes. A method employing HPLC, with amperometric detection at the CoPC modified electrodes, was developed to determine the four nucleosides in a standard mixture and a sample generated by the partial hydrolysis of a commercial ribonucleic acid mixture 61 62 Using cyclic voltammetry an anodic response was recorded for the ribose-containing 63 grams for ascorbic acid and GSH revealed well-defined peak shapes and good resolution, which indicated that excellent electron transfer kinetics were attainable with the modified SPCEs.The calibration graphs of i,, versus concentration were linear for both biomolecules over the range 0-2.22 mmol dm-3 (n = 6). In a subsequent study,53 SPCEs were prepared modified with several iron containing mediators including several substituted ferrocenes. The electrochemical behaviour of these modified electrodes was studied in 0.05 mol dm-3 phosphate buffer solutions (pH 3, 5 and 7) and their suitability as mediators for the determination of GSH was compared. The most promising compounds were ferrocene- carboxaldehyde and iron phthalocyanine; however, neither was superior to the sensitivity and selectivity of CoPC. Recently, we have described the development of an amper- ometric assay which uses the CoPC modified SPCEs and the enzyme glutathione peroxidase to selectively determine GSH in biological fluids.54355 This method uses the enzyme in a subtractive mode to oxidize GSH selectively in test solutions; the rate and extent of peptide removal was monitored amperometrically using the modified electrodes (Fig.6). Preliminary results were presented which indicate that the method might be suitable for the selective determination of GSH in human whole blood. Sklada156 has recently described an interesting method for the determination of organophosphate and carbamate pesti- cides which uses CoPC modified CPEs to detect the produc- tion of thiocholine from butyrylthiocholine by the enzyme cholinesterase. The activity of cholinesterase is non-competi- tively inhibited in the presence of the pesticides which causes a decrease in the rate of thiocholine production and thus a reduction in steady-state current measured following the addition of a pesticide to the standard enzyme-substrate solution.A linear relationship was observed between the relative decrease in current response and pesticide concentra- tion; the detection limits were 0.3 and 80 mg dm-3 for two different commercial pesticide preparations. Several other recent applications using CoPC modified carbon electrodes are summarized in Table 3. Hexacyanoferrate(II1) Potassium hexacyanoferrate(ii1) has been successfully em- ployed as a mediator following adsorbtion onto graphite foil64 and, more commonly, in aqueous solutions65.66 or when electrostatically immobilized with PVP [poly(4-vinylpyri- dine)].In one study67 solution phase, and PVP immobilized, hexacyanoferrate were investigated for the determination of ethanol at carbon-paste electrodes containing viable yeast cells. The yeast cells contained high levels of the enzyme alcohol dehydrogenase, which catalyses the reaction Alcohol + NAD+ -+ aldehyde + NADH ( 5 ) Amperometry in stirred solutions [containing 0.4 mmol dm-3 ethanol, 1 mmol dm-3 NAD+ and 1 mmol dm-3 hexacyanoferrate (111)] was employed to construct hydro- dynamic voltammograms to elucidate the anodic current response arising from the mediated oxidation of NADH at the yeast-modified carbon-paste electrodes. Using this techniqueANALYST, AUGUST 1992, VOL. 117 1221 the maximum current response was recorded at +600 mV versus Ag-AgCl (0.05 rnol dm-3 phosphate buffer, pH 7.4); negligible current flowed in the absence of hexacyano- ferrate(ii1).The steady-state current response was also depen- dent on the amount of yeast in the paste; i . e . , the currents recorded for pastes containing 2, 5 and 10% m/m yeast were 298,405 and 590 nA, respectively (ethanol concentration, 5 x 10-4 rnol dm-3). Calibration graphs were constructed for several primary alcohols, and a linear response was obtained for ethanol concentrations up to 0.3 mmol dm-3. The trend in sensitivity was: ethanol > propan-1-01 > butan-1-01 > pentan-1-01. Further studies indicated that the electrode did not respond to secondary or tertiary alcohols and was suitable for FI.When PVP was incorporated into the carbon paste in order to co-immobilize the hexacyanoferrate(n1) ions, the resulting electrode displayed a rapid current response to ethanol but suffered from poor stability which was associated with the gradual loss of the mediator into solution. In a further application of Fe(CN)& modified PVP-CPEs, Bonakdar et a1.68 developed an amperometric assay using FI in conjunction with a thin-layer flow through cell containing tyrosinase mixed into the modified carbon paste. This electrode was employed to reduce the o-benzoquinone [&pp,ied, -200 mV versus Ag-AgCI, carrier stream, 0.1 rnol dm-3 KCI (pH 7.5)] produced by the enzymic oxidation of phenol. Using commer- cial tyrosinase the calibration graphs were linear for phenol concentrations from the limit of detection at 14 pg dm-3 to 2.5 mg dm-3 (slope, 7.8 nA mg-l cm3); a decrease in electrode response was observed with time, although individual elec- trodes could be used for an entire working day.In contrast, at plain carbon-paste electrodes an applied potential of +900 mV was required for the oxidation of phenol, and initial current response was found t o be halved following 37 injections of a solution containing 0.17 mg dm-3 of phenol owing to the build-up of adsorbed oxidation products. An explanation for the loss of mediator in both of these instances may be derived from the observations of Geno etuZ.69 who systematically investigated the preparation and charac- terization of PVP containing carbon-paste electrodes. These authors employed cyclic voltammetry to elucidate the electro- chemical behaviour of PVP-CPEs used to bind anionic metal complexes such as hexacyanoferrate.In their experiments they showed that the magnitude of the quasi-reversible hexacyanoferrate redox waves (Epa, == +240 mV; E,, == +60 mV versus SCE), recorded in a solution containing 0.5 rnol dm-3 glycine buffer (pH 3.2) and 1 mmol dm-3 F ~ ( C N ) G ~ - , increased with the number of scans. When the electrode was removed from the mediator solution, rinsed, and placed in plain 0.5 rnol dm-3 glycine buffer (pH 3.2) the redox waves were still observed. However, the acidic condi- tions were essential, because washing the electrode with an alkali caused the mediator to leach back into solution; attempts to bind Fe(CN)& under solution conditions more alkaline than pH 3.2 were unsuccessful.The authors subse- quently used the Fe(CN)64- modified PVP-CPEs to study the electrocatalytic oxidation of ascorbic acid and were able to decrease the overpotential by 225 mV which enabled them to determine the vitamin at == +275 mV. In view of its aqueous solubility other applications of hexacyanoferrate(ii1) generally use the mediator in the solu- tion phase, for example the determination of the alkaloid theophyline using theophyline oxidase70 or for monitoring the production of NADH using the enzyme diaphorase.71 Ruthenium Oxide Complexes In a recent study, Leech et uZ.72 prepared ruthenium dioxide modified carbon-epoxy resin electrodes for the determination of the antibiotics; streptomycin, novobiocin and neomycin.Electrodes were constructed by mixing Ru02 into a mixture t u C g 3 0 1 15 min - E I J 80 2 2 40 u 0 0.5 1 .o 1.5 Concentration/mmol dm- 3 -Time Fig. 7 Flow injection response at the Ru02 modified carbon paste electrode to injections of: A, 0.25; B, 0.5; C, 0.75; D, 1.0; and E, 1.25 mmol dm-3 novobiocin. Constant potential operation at 0 V. Flow rate, 1 cm3 min-1; electrolyte, 0.5 mol dm-3 NaOH. Inset shows the current-concentration calibration graph obtained for novobiocin up to a concentration of 1.5 mmol dm-3. Reproduced with permission from ref. 72 typically consisting of an 80 : 20% m/m ratio of graphite-epoxy resin + Ru02. The paste was packed into a glass tube and allowed to cure at room temperature; the hardened surface was then polished using emery paper followed by a fine alumina slurry.The electrodes were washed and ultrasoni- cated before use. Cyclic voltammetry, in 0.5 rnol dm-3 NaOH, revealed two reversible waves at 0 and +450 mV versus Ag-AgC1 which corresponded to the Ru02-Ru203 and Ru042--Ru02 transitions, respectively. In the presence of streptomycin and neomycin the magnitudes of the anodic currents recorded for the more positive wave were increased; the oxidation of novobiocin gave rise to an increased current response at both waves. No anodic current response was observed for the oxidation of the antibiotics at unmodified electrodes over the potential range studied. The modified electrodes were used to construct hydrody- namic voltammograms and calibration graphs using amper- ometry in stirred solutions and FI-EC (Fig.7). Using the former technique, calibration graphs were linear for: strep- tomycin between 1.5 pmol dm-3 and 0.25 mmol dm-3 (slope, 4.43 nA pmo1-l dm3; Eapplied, +350 mv); neomycin between 10 pmol dm-3 and 2 mmol dm-3 (slope, 0.08 nA pmol-1 dm3; &pplied, +350 mV) and novobiocin between 6 pmol dm-3 and 0.4 mmol dm-3 (slope, 1.31 nA pmol-l dm3; Eapplied, +200 mV). In the same, and a subsequent,73 report the authors have indicated that Ru02 modified carbon-paste electrodes can be used for the determination of alcohols and carbohydrates. Cox and Gray74 have reported the determination of insulin, cysteine and glutathione75 at glassy-carbon electrodes modi- fied with a film containing mixed valency ruthenium oxides; the film was prepared by electrolysis in a plating solution containing 2 mmol dm-3 RuCI3 and 2 mmol dm-3 K,Ru(CN)6. Insulin is an organic disulfide and is reported to undergo oxidation at the modified electrodes poised at potentials of greater than +880 mV [Epa, +940 mV versus Ag-AgCI in 0.2 rnol dm-3 K2S04, (pH 2)].Flow injection was1222 ANALYST, AUGUST 1992, VOL. 117 used to construct calibration graphs; these were linear for insulin concentrations between 8.2 and 204 ng injected (sample volume, 7.5 mm3, Eapplied, +960 mV) and the limit of detection was 5 ng injected (for a signal-to-noise ratio of 3 : 1). The thiols, cysteine and glutathione, were oxidized at similar potentials to insulin, and the calibration graphs, obtained using FI, were linear for concentrations between 0.41 pmol dm-3 and 0.2 mmol dm-3 (slopes: cysteine, 50 nA pmol dm-3; glutathione, 36 nA pmol dm-3). Insig- nificant oxidation currents were recorded at unmodified electrodes over the potential range studied.Metalloporphyrins Wang and Golden76 employed cyclic voltammetry, DPV and FI-EC for the determination of ascorbic acid, penicillamine, acetaminophen, NADH, hydrazine, adrenalin, cysteine and oxalic acid at a glassy-carbon electrode modified with an adsorbed layer of manganese(ii1) meso-tetraphenylporphine. In each instance, the overpotential for the oxidation of the individual species was significantly reduced when compared with the unmodified electrode, e.g., using cyclic voltammetry the overpotential for the first five compounds was reduced by ci H / O \ H 0 360, 198, 180, 144 and 127 mV, respectively (supporting electrolyte solution, 0.05 mol dm-3 phosphate buffer, pH 7).Organic Mediators Phenoxazines, Phenathiazines and Phenazines Phenoxazines and their related compounds have been widely used for the determination of the coenzyme NADH (1) generated during analytical methods employing certain de- hydrogenase enzymes, e . g . , lactate dehydrogenase. Direct electrochemical oxidation of NADH at bare metal or carbon electrodes has proved difficult owing to high overpotentials (Epa = + 0.7 V versus SCE, pH 7.1) and adsorption phenomena.77-79 In part, these may be overcome using physical electrode modification techniques such as mechanical polishing80.81 or electrochemical pre-treatment prior to analy- sis.**,83 Indeed, using the former technique Palleschi et al.8 0 ~ prepared an enzyme electrode with immobilized 3-hydroxy- butyrate dehydrogenase and were able to determine hydroxy- butyrate from 5 to 100 pmol dm-3 using an applied potential of +300 mV versus Ag-AgC1 to monitor NADH production. However, electrodes activated in this manner are prone to drift in their current response with time and their use for the 0 II - P-0- P - 0 - R I I OH OH N?Q b HO OH v C O N H 2 1 Nicotinamide adenine dinucleotide 2 Meldola Blue 4 Methylene Blue aND N I CH3 6 Phenazine methosulfate (N-methylphenazinium ion) 3 Nile Blue A 5 Toluidine Blue 7 Tetrathiafulvalene 8 TetracyanoquinodimethaneANALYST, AUGUST 1992, VOL. 117 1223 detection of NADH must be questioned because chemical mediators have proved exceedingly successful.In particular several research groups have reported studies using analogues of phenoxazine, phenathiazine and phenazine compounds. Phenoxazine mediators Meldola Blue. The application of Meldola Blue (MB+, 2; 7-dime thylamino- 1,2-benzophenoxazinium salt) for the deter- mination of NADH has been investigated extensively by Gorton and colleagues."~s5 In their early studies they pre- pared modified graphite electrodes by rotating (OJ, 99.2 rad s-1) bare planar graphite electrodes for approximately 60 s, depending on the coverage required, in phosphate buffer solutions containing 10-4 rnol dm-3 MB+; the electodes were then washed with de-ionized water before use. Elucidation of the electrochemical behaviour of the absorbed Meldola Blue revealed pH dependent reversible redox behaviour (Eo' , +110 mV versus SCE in 0.1 rnol dm-3 HCI).A plot of Eo' versus pH gave a straight line from pH 2.0 to 5.0 with a slope of 60 mV per pH unit; a break occurred at the higher value to give a second straight line between pH 5.0 and 10.0 with a slope of 30 mV per pH unit. This pH behaviour was explained by the reaction schemes given in the following equations: pH >5 MB+ + H+ + 2e- e MBH (6) pH <5 MB+ + 2H+ + 2e- MBH+ (7) These results indicated the presence of a pK, for absorbed Meldola Blue in aqueous buffer solutions at approximately pH 5. Cyclic voltammetry performed using the modified elec- trodes in phosphate buffer solutions (pH 7) containing either NADH or NADPH revealed an increase in the magnitude of the redox waves previously recorded for the mediator in plain solutions (Eo', -175 mV versus SCE).The proposed scheme for the mediated reaction between Meldola Blue and NADH is NADH + MB+ G NADH-MB+ NADH-MB+ + NAD+ + MBH MBH e MB+ + 2e- + H+ The scheme assumes the formation of a coenzyme-mediator complex where the back-reaction of eqn. (9) is negligible owing to the very rapid oxidation of the MBH at the working electrode. In a more recent study,85 using FI, the Meldola Blue modified electrodes were mounted in a wall-jet cell and employed for the determination of glucose. This was facili- tated using an enzyme reactor, containing glucose dehydro- genase [eqn. ( l l ) ] immobilized on porous glass, positioned immediately upstream of the amperometric cell.(8) (9) (10) P-D-glucose + NAD+ + H20 For this investigation, the modified electrodes were prepared by dipping the graphite tip into a cold solution containing 0.1 mmol dm-3 MB+ in 40% ethanol + 1% triethylamine + 59% 0.1 rnol dm-3 phosphate buffer pH 3.5 (by volume). Hydrody- namic voltammograms were constructed for the oxidation of NADH at different potentials using 0.1 rnol dm-3 phosphate buffers (pH 6 and 7) as the carrier stream; the maximum anodic current responses were recorded using the more acidic solution and an applied potential of -50 mV versus Ag-AgC1. The acidic medium was thought to be superior owing to the enhanced overall rate constant of the net reaction at pH 6 as follows: NADH -+ NAD+ + 2e- + H+ The optimized carrier stream for the determination of NADH consisted of a solution containing 0.1 rnol dm-3 phosphate buffer (pH 6) and 1 mmol dm-3 NAD+; the final choice of buffer pH was a compromise between maximum enzyme D-gluconic acid + NADH (11) (12) 103 102 P k 10 I I I I I 1 10 102 103 104 Csubstrate/pmOl dm-3 Fig.8 Calibration graphs for dehydrogenase substrates with a 3-P-naphthoyl-Nile Blue modified graphite electrode, using enzyme suspensions behind a dialysis membrane, a 0.2 mol dm-3 Tris buffer (pH 7-44), 1 mmol dm-3 NAD+, and 0.02% sodium azide. A, Ethanol with 18 U (1 U = 16.67 nkat) of alcohol dehydrogenase from yeast; B, alanine with 7 U of alanine dehydrogenase from bacillus subtilis; C, lactate with 5.5 U of lactate dehydrogenase from hog muscle; and D, glutamate with 4.8 U of glutamate dehydrogenase from beef liver.Reproduced with permission from ref. 90 H n H fi Fig. 9 from ref. 91 Redox species for Nile Blue (3). Reproduced with permission activity at pH 7, and the improved sensitivity and NAD+ stability at pH 6. Calibration graphs constructed for NADH and NADPH were linear between 1 pmol dm-3 and 10 mmol dm-3 (sample volume, 50 mm3) with the latter compound giving slightly poorer slopes owing to its lower diffusion coefficient. For the determination of glucose, the limit of detection was 0.25 pmol dm-3 and calibration graphs were linear to =lo0 pmol dm-3; the experimental conditions were: a carrier stream containing 2.5 mmol dm-3 NAD+; flow rate, 1 cm3 min-1 and applied potential, 0 mV versus Ag-AgC1. During selectivity studies the method was found to be free from interferences arising from uric acid, catechol and1224 ANALYST, AUGUST 1992, VOL.117 ascorbic acid, although to satisfy these conditions for the latter species the applied potential had to be reduced to about -100 mV versus Ag-AgCI. As expected the method was independent of dissolved oxygen. In a separate study Marko-Varga86 employed a similar detection system, using GDH immobilized in a post-column enzyme reactor for the determination of glucose following a chromatographic separation of broth samples collected during the fermentation of penicillin. Recently, Yabuki et al.87 have reported a method for the incorporation of an enzyme (alcohol dehydrogenase), NAD+ and Meldola Blue into a conducting polypyrrole membrane using electropolymerization onto platinum wire.This elec- trode allowed the recycling of the NAD+ cofactor and the method should be worthy of further studies for sensor applications using carbon electrode substrates. Nile Blue. Nile Blue (NBH, 3; 3-amino-7-(diethylamino)- 1 ,2-benzophenoxazine) is another phenoxazine dye which has been used for the determination of NAD(P)H. Early studies using this mediator were performed by Huck88 who, later with colleag~es,89~9~ produced modified electrodes by coating 3-P-naphthoyl-Nile Blue (from 5 mm3 of a 1 mmol dm-3 ethanolic solution) onto the surface of graphite electrodes. The naphthoyl derivative was used for these studies because its reduced water solubility enhanced the working life of the modified electrodes.Using flow injection (Fig. 8), these electrodes were applied to the determination of NADH produced by lactate, glutamate and alcohol dehydrogenases immobilized either behind a dialysis membrane or onto epoxyacryl resin beads within separate enzyme reactors; [carrier stream, 0.2 mol dm-3 phosphate buffer (pH 8); flow rate, 0.5 cm3 min-1; 50 mm3 aliquots of a 10 mmol dm-3 solution of NAD+ were injected every 4-6 min and Eapplied, +lo0 mV versus SCE]. These reactor systems were sub- sequently used to determine lactate in butter, glutamate in beef cube extracts and ethanol in low-alcohol beer. The authors suggest that the electrocatalytic oxidation of NADH using Nile Blue could form the basis of a versatile detection system for general practical applications using suitable de- hydrogenase enzymes.Ni et aL9* have elucidated the electrochemical behaviour of Nile Blue adsorbed onto glassy carbon electrodes, and have proposed that the electrocatalytic oxidation of NADH by Nile Blue A (3) follows an EC mechanism [eqns. (13) and (14)] NBH G= NB+ + H+ + 2e- (13) NADH + NB+ NAD+ + NBH (14) The net reaction has been described previously by eqn. (12) for the Meldola Blue mediated reaction. The pH dependence of the formal potential for the oxidation of Nile Blue is similar to Meldola Blue except the break (pK,) in the Eo’ versus pH plot occurs at pH 6 (Fig. 9). In a separate investigation, Buch-Rasmussen92 employed FI (with a graphite electrode modified with a Nile Blue-tereph- thaloic derivative) for the determination of glucose, lactate and creatinine using the appropriate dehydrogenases for the former two compounds, and an indirect two-enzyme step reaction for the latter.In the first step of this reaction, creatinine was converted to N-methylhydantoin and ammo- nium ion using creatinine iminohydrolase; the ammonium ion product and NADH cofactor were subsequently consumed in the second stage by glutamate dehydrogenase with the loss of NADH being monitored at the modified electrode. A similar Nile Blue derivative has also been used by Polasek et aL93 and Skoog and Johansson94 who constructed sensors for the determination of glucose based on the determination of NADH produced by the enzyme glucose dehydrogenase. These electrodes were used in the amperometric mode and poised at -0.1 to 0 V versus Ag-AgC1; calibration graphs were linear from the limits of detection (0.3 and 1.0 pmol dm-3, respectively) to 2 mmol dm-3 concentrations of glucose and the former workers report that a sample through- put of 200 samples h-1 was attained.Phenathia z ine mediators Ye and Baldwin95 studied the electrocatalytic response of graphite electrodes modified with adsorbed Methylene Blue (4) towards myoglobin and haemoglobin. For their investiga- tions modified glassy carbon electrodes were prepared by a dip-coating procedure where the electrodes were immersed for 60 s in a stirred solution containing 0.01% m/v of the mediator in 0.1 mol dm-3 phosphate buffer (pH 5.3); the electrodes were subsequently rinsed with purified water before use. Investigations to elucidate the electrochemical behaviour of solution-phase Methylene Blue revealed a reversible pair of redox waves (Fig.10; Epc, -110 mV; E,, -70 mV versus Ag-AgCI, in pH 5.45 acetate buffer) and a pH dependence similar to the phenoxazine compounds with a pK, occurring at pH 4.4. No electrochemical response was observed for myoglobin at unmodified electrodes over the potential range studied [from -600 to +500 mV]; however, in the presence of myoglobin an increase in the magnitude of the Methylene Blue waves was recorded. Hydrodynamic voltammograms were constructed (by means of FI) for the reduction of the two proteins at the Methylene Blue modified electrodes and the cathodic current plateau were reached at = -130 mV versus Ag-AgCI. Preliminary results were also given from a method employing size-exclusion HPLC-EC to separate the proteins.Persson96 investigated the application 3-P-naphthoyl-Tol- uidine Blue 0 (3-NTB) and Toluidine Blue (TBO, 5 ) for the determination of NADH using graphite electrodes modified by the dip-coating process. Using cyclic voltammetry, both mediators showed reversible behaviour with Eo’ for TBO at -285 mV and 3-NTB at -135 mV versus SCE (pH 7). The electrocatalytic determination of NADH followed an EC mechanism similar to Nile Blue [eqns. (13) and (14)] and using FI linear calibration graphs were obtained from 1 pmol dm-3 to 2 mmol dm-3. Phenaz ine mediators Jonsson and Gorton97 have investigated the use of the N-methylphenazinium ion (PMS+, 6) as a mediator for the determination of glucose. For this study, glucose oxidase was covalently immobilized onto a graphite electrode using cyanuric chloride and the washed enzyme electrode was modified with PMS+ by dip-coating. The modified electrode showed quasi-reversible behaviour (Epc, --260 mV; Epa, = +60 mV versus SCE; pH 7 ) and a pH dependence where the values of Eo’ changed by 30 mV per pH unit.Subsequently, a similarly modified rotating disc electrode (Eapplied, +50 mV versus SCE) was employed for the mediated determination of glucose, and the calibration graphs were linear from the detection limit at 0.5 to 150.0 pmol dm-3 glucose. However, the method was not independent of dissolved oxygen and solutions had to be de-aerated before analysis. The PMS+ is also reasonably unstable and decomposes to phenazine (Eo’ , -435 mV versus SCE, pH 7 ) ; however, the enzyme electrodes could be used for several months provided they were ‘recharged’ with mediator.H Fig. 10 5.45). Reproduced with permission from ref. 95 Redox species for Methylene Blue in acetate buffer (pHANALYST. AUGUST 1992, VOL. 117 1225 Hydroquinone pBenzoqui none Fig. 11 mediators Redox mechanism for the hydroquinone-p-benzoquinone t 20 E 0 2 4 6 8 1 0 Concentration/lO 7 rnol dm 3 - t Fig. 12 Effect of H202 concentration on the FI signal. Conditions: A, 1.0 X 10-8; B 2.5 x 10-8; C, 5.0 x 10-8; D, 3.0 x 10-7; E, 6.0 X and F, 1.0 x 10-6 rnol dm-3. Carrier stream, 0.1 rnol dm-3 phosphate buffer (pH 7.0) containing 0.1 mmol dm-3 hydroquinone; flow rate, 10 cm3 min-'; sample volume, 20 mm3; applied potential, -0.30 V versus Ag-AgC1.Carrier stream was deoxygenated in an ultrasonic bath for 30 min before use. Reproduced with permission from ref. 103 Phenazine mediators have also been used for the determina- tion of NADH at carbon98 and platinum99 electrodes, and glucose at a goldl(m electrode. Persson and Gorton'o' have recently reported a study performed to compare the suitability of several phenazines, phenoxazines and phenathiazines for the electrocatalytic oxidation of NADH. Their results suggest that the phenazine structure is the least suited to the oxidation of NADH because of its low E"' and pK values. The remaining mediators studied were ranked according to the measured sensitivity (mA mol-1 dm3) towards NADH in buffer solutions greater than pH 6, where: 3-NTB >3-(3-naphthoyl-Brilliant Cresyl Blue (3-NBCB) > 3-(3-naphthoyl-Nile Blue A > 3-anilino- Meldola Blue.From the data presented, graphite electrodes modified with 3-NBCB appear to be the most stable. Quinone-H y droquinones This redox couple has been used as a mediator for both electro-reductive and electro-oxidation processes. Smyth and co-~orkers10*-103 employed the former mode for the sensitive determination of hydrogen peroxide using horseradish peroxi- dase. In their study they demonstrated that the enzyme can reduce hydrogen peroxide, in the presence of hydroquinone, to water. The quinone generated in this reaction can subse- quently be determined by its reduction at a glassy carbon electrode (poised at -300 mV versus Ag-AgC1) in a process involving a two-electron and two-proton quasi-reversible reaction (Fig.11): Epc, -150 mV; Epa, +320 mV versus Ag-AgCI in 0.1 rnol dm-3 phosphate buffer (pH 7). The enzyme electrodes used for this study were prepared by immersing glassy carbon (activated by polishing with alumina), for 30 min, into a solution containing 0.01 mg cm-3 of the enzyme in 0.1 rnol dm-3 phosphate buffer (pH 7). For these studies the mediator was dissolved in the supporting electrolyte solution [O. 1 rnol dm-3 phosphate buffer (pH 7)]; however, the authors indicated that they intend to immobilize the mediator during the next stage of the Fig. 13 Schematic diagram presenting the 4-methyl-o-quinone (4MQ) mediated oxidation of NADH employed for the amperometric determination of 3-hydroxybutyrate (3-OHB).Reproduced with permission from ref. 109 sensor development. The enzyme method was applied to the determination of hydrogen peroxide in standard solutions using DPV and FI (Fig. 12); the limit of detection using the former technique was 1 x 10-8 rnol dm-3 with a linear calibration response from 1.0 x 10-7 to 1.0 x 10-6 rnol dm-3 (slope, 6.89 X l o 7 FA mol-1 dm3). Hydrodynamic voltammograms, constructed using FI, indi- cated that the maximum current response would be recorded at potentials more cathodic than -300 mV versus Ag-AgCI (pH 7 ) ; calibration graphs obtained were linear between 2.5 X 10-8 and 1.0 x 10-6 rnol dm-3 (Fig. 12; slope = 3.35 X lo7 nA mol-1 dm-3) and the limit of detection was 1 x 10-8 mol dm-3. Interference studies were performed using a selection of compounds that may be present in industrial waste water; of the substances investigated only acetic and ascorbic acids caused the electrode response to hydrogen peroxide to deviate by greater than 5%.Another research group104-106 has employed the electro-oxidation of the hydroquinone, produced during reactions of p-benzoquinone (BQ) with reduced oligosaccharide dehydrogenase (ODH), for the determination of a-amylase activity in human serum or plasma. The dehydrogenase enzyme catalyses the oxidation of several sugars whereupon it can be re-oxidized by BQ which acts as a mediator between the enzyme and electrode. Using cyclic voltammetry, this group showed that the quasi-reversible current response of the ODH-BQ electrode [Epc, -30 mV; E,,, +230 mV versus Ag-AgCI (pH 7 ) ] was enhanced upon the addition of maltose into the supporting electrolyte solution.For the determination of a-amylase activity the ODH-BQ electrode was employed in the amperometric mode [Eapplied, +500 mV versus Ag-AgCI (pH 7 ) ] to monitor the formation of maltotriose and maltose produced by the enzymic hydrolysis of maltopentose. Amylase activities were determined in six serum samples and the results obtained using the electrochem- ical method agreed well with a commercial chromogenic method; indeed the imprecision figures for repeated measure- ments using the electrochemical method were superior (RSDs, 1.2% and 4.2%, respectively; n = 5). Quinone mediators have also been applied to the determi- nation of NADH. In early studies Ravichandran and Bald- win107 used cyclic voltammetry to elucidate the electrocata- lytic response of 1,2-naphthoquinone modified carbon-paste electrodes towards NADH [Epa, -100 mV versus SCE (pH 7)].Whilst Jaegfeldt et af. 108 examined naphthalene and ethaneanthracene based catechols adsorbed onto graphite electrodes; the former species showed promise with the mediated oxidation of NADH occurring at +185 mV versus SCE (pH 7 ) . More recently, Batchelor et af. 109 described an amperometric assay for the ketone body 3-hydroxybutyrate (3-OHB) using the enzyme 3-hydroxybutyrate dehydrogenase (HBDH) and 4-methyl-o-quinone as the mediator (Fig. 13). This method involved the use of a carbon-based electrode,1226 ANALYST, AUGUST 1992, VOL. 117 I 15 I 1 3 I I I I 1 0 1 2 3 0 1 2 3 4 5 Time/min - Fig.14 Typical (a) steady-state and (b) flow injection profiles for the TT'F modified carbon paste enzyme electrode. Conditions: GOD loading, 0.25 mg; applied potential, +0.2 V versus Ag-AgC1; steady-state mode flow rate, 0.5 cm3 min-1; FI flow rate, 1 cm3 min-1; sample volume, 25 mm3; and carrier stream, 0.1 mol dm-3 phosphate buffer (pH 7.4). (Numbers given on the curves are concentrations in mmol dm-3.) Reproduced with permission from ref. 122 containing the mediator, NAD+ and HBDH, screen-printed onto an inert support. These disposable amperometric elec- trodes (Eapplied, +350 mV versus Ag-AgC1) were used to determine 3-OHB spiked into blood and the authors suggest that their devices could be usefully employed for monitoring ketotic patients in an emergency room or doctor's office.Miki et al. 11" have also employed quinone mediators for the determination of NADH using diaphorase enzyme electrodes. A further interesting application is the use of mediator cocktails for the detection of viable bacteria.111.112 This method, involves the detection of micro-organisms, in a variety of sample matrices, by monitoring the oxidation of reduced mediators produced by diverting electrons from the respiratory chains of the bacteria. The maximum current responses were obtained using a mixture of hexacyanofer- rate(m) and p-benzoquinone which permitted a detection limit of -105 cfu cm-3 (where cfu = colony forming units, i.e., viable micro-organisms) . Other applications include the determination of glucose, xanthine and lactate using either: benzoquinone adsorbed onto graphite foil discs,113 solution phase quinones,114 or quinone containing polymer modified electrodes115.116 used in conjunction with glucose oxidase, xanthine oxidase and lactase enzymes, respectively.Tetrathiafulvalene and Tetracyanoquinodimethane The charge transfer salt tetrathiafulvalinium-tetracyano- quinodimethane (TTF-TCNQ) is an excellent organic con- ductor,117 and has been used successfully as an electrode substrate for several electrochemical applications.118-12" In addition, the component compounds have been used individu- ally as electrocatalysts incorporated into carbon-based elec- trodes. Tetrathiafulvalene (TTF, 7 ) The electrochemical behaviour, and application, of glassy carbon enzyme electrodes chemically modified with TTF and glucose oxidase have been described by Gunasingham and Tan.121 For their investigations, electrodes were prepared by drop-coating initially TTF (15 mm3 of a 0.25% m/v solution in acetone) and secondly glucose oxidase [ l o mm3 of a 4% m/v solution in 0.1 mol dm-3 phosphate buffer (pH 7.4)] onto a glassy carbon surface.The enzyme layer was subsequently immobilized by the addition of 5 mm3 of a 1 + 1 mixture containing BSA (10% m/v) and glutaraldehyde (2.5% m/v) in 0.1 mol dm-3 phosphate buffer (pH 7.4). Finally, the electrode was covered with a polycarbonate membrane (0.30 pm pore size). Tetrathiafulvalene was reported to undergo two one-elec- tron oxidation processes [eqn. (15)], the first being reversible [E,,! +150 mV versus Ag-AgC1, (pH 7.4)] and suitable for the mediated enzymic determination of glucose [eqns.(16)-(18)]. TTF e TTF+ + e- + TTF2+ + e- (15) Glucose + GODoxidized + gluconolactone + GODreduced (16) GODreduced + 2TTF+ + GODoxidized + 2TTF + 2H+ (17) 2TTF- 2TTF+ + 2e- In a later report,122 these workers used their TTF modified carbon-paste enzyme electodes, in conjunction with methods employing flow injection or stopped-flow analysis, for the determination of circulating blood glucose levels in samples collected from several diabetic patients. The applied potential was +200 mV versus Ag-AgC1 (pH 7.4); using the FI method calibration graphs for glucose were linear from 3 to 80 mmol dm-3 (Fig. 14), sample throughput was approximately 120 samples h-1 and the precision for 100 replicate injections of a standard solution containing 10 mmol dm-3 glucose was 0.6% (RSD).The results obtained using the electrochemical methods compared favourably with an established Reflolux method. More recently, Palleschi and Turner123 have described a sensor for the determination of L-lactate which was produced by dip-coating lactate oxidase and TTF onto carbon foil discs which had been previously bonded onto the end of glass rods. The reactions involved during the enzymic determination of lactate at the modified electrode are similar to those given for glucose [eqns. (16)-( 18)]. The applied potential and temperature were +200 mV versus Ag-AgCI (pH 7.35) and 30 k 0.5 "C, respectively; calibration graphs were obtained between 1 x 10-4 and 9 X 10-3 mol dm-3 lactate. Tetracyanoquinodimethane (TCNQ, 8) Cenas and Kulys124 and Hendry and Turner125 have investi- gated the application of TCNQ for the mediated determina- tion of glucose.Cyclic voltammetry, in acetonitrile, revealed two reversible redox couples (Epal =+220 mV and Ep,2 --0.320 mV versus Ag-AgC1) and the more anodic response was studied by the latter workers for the development of a mediated enzyme electrode containing glucose oxidase. As TCNQ is insoluble in aqueous media, dip-coating was employed to adsorb the mediator (from a solution containing 10 mg cm-3 TCNQ in toluene) onto the surface of graphite foil discs; the enzyme was then immobilized using a carbodiimide reaction scheme. Preliminary results revealed non-linear calibration graphs for glucose concentrations between 0 and 70 mmol dm-3 and were accompanied by very short electrode lifetimes ( l k , 1-1.5 h); the authors suggest the latter problem may be caused by a loss of enzyme activity or leaching of the enzyme or mediator away from the electrode surface.In a more recent application tyrosinase-TCNQ based enzyme electrodes were investigated by Kulys and Schmid126 for the determination of phenol in water samples. Two methods of enzyme immobilization were studied; in one design the enzyme was trapped behind a dialysis membrane whilst the second type used carbodiimide to attach the enzyme to the TCNQ modified graphite electrode covalently. The reaction mechanism for the tyrosinase-TCNQ enzyme elec- trodes is Phenol + O2 + 2TCNQ- + 2H+ - catechol + The measured response arises from the generation of a cathodic current for the electro-reduction of TCNQ TCNQ + e- * TCNQ- H20 + 2TCNQ (19) (20)ANALYST, AUGUST 1992, VOL.117 NH2 NH Fig. 15 Reproduced with permission from ref. 129 Reversible two electron oxidation of phenylenediamine. Calibration graphs were constructed for cathodic current response versus phenol concentration for both designs of electrode; these were linear to 65 pmol dm-3 (slope, 0.36 A mol-' dm3) and 25 pmol dm-3 (slope, 2.2 A mol-1 dm3) for the membrane and covalently immobilized enzyme elec- trodes, respectively [Eapplied, + 130 mV versus Ag-AgC1, (pH 7) data quoted for fresh electrodes]. In each instance, the electrodes were found to be relatively unstable; the membrane electrode response increased for the first 3 d and then decreased, whilst the covalently immobilized tyrosinase elec- trode response decreased by ~ 4 0 % d-1.Loss of activity was attributed to enzyme stripping from the electrode surface. The electrochemical behaviour of phenol at unmodified carbon- based electrodes has been described in the section dealing with hexacyanoferrate . Recently , Kulys et al. 127 have described the development of a novel enzyme electrode containing poly(ethy1ene glycol) enlarged NAD+ (NAD-PEG). 128 These electrodes were prepared by adsorbing NMP+TCNQ- (NMP = N-methyl- phenazium), dissolved in hot acetonitrile onto the end of a graphite rod electrode. After the solvent had evaporated an aliquot of a solution containing alcohol dehydrogenase and the NAD-PEG was dropped onto the electrode surface and entrapped by a dialysis membrane which was fixed by an O-ring.The molecular mass of the NAD-PEG was about 20 000 and was therefore retained behind the membrane when the electrode was placed into test solutions. Following optimization of the solution conditions the enzyme electrode was employed to construct calibration graphs for ethanol concentrations between 0.5 and 7 mmol dm-3. This electrode design may prove useful to other research groups working in biosensor development as it allows enzyme-cofactor systems to be evaluated within a sensor format without having to use chemical immobilization methods. Phenylenediamine (PD) and Tetramethylphenylenediamine (TMPD) Ravichandran and Baldwin129 employed cyclic voltammetry to elucidate the electrochemical behaviour of carbon-paste electrodes modified with each mediator (PD and TMPD), and their application to the determination of NADH and ascorbic acid.The PD underwent a reversible two-electron oxidation to the quinone diimine [Fig. 15; Epa, +180 mV versus SCE (pH 7)]. In the presence of NADH the magnitude of the anodic wave increased. The TMPD underwent two single- electron oxidations; the first (Epa, +lo0 mV) was reversible and able to catalyse the oxidation of ascorbic acid. Whilst the second anodic wave was irreversible (Epa, +400 mV) owing to hydrolysis of the electrochemical product; NADH only showed an electrocatalytic response at this second wave. An EC mechanism was proposed for the catalytic response at both of the modified electrodes.Conclusions In conclusion, it is now universally recognized that sensors and biosensors have an increasingly vital role to play in the pursuit of the more rapid and simple analytical methodologies that are required not only by skilled analysts but also by untrained personnel, who may want to use these devices in hospitals, doctors surgeries, and other workplaces, and also for personal 1227 monitoring. This review has described, and compared many designs of voltammetric and amperometric sensors that have been developed and used for the determination of a wide range of analytes. Undoubtedly, the future development of increasingly selective, sensitive and stable sensors will bring many challenges. It is hoped that the information given in this paper will provide useful practical guidance for researchers working, or intending to work, in this important and exciting area.The authors thank their co-workers and all fellow researchers, whose papers have been cited in this review; they are also grateful to Glaxo Group Research and Bristol Polytechnic for their interest and support. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 References Yao, T., Kobayashi, N., and Wasa, T., Anal. Chim. Acta, 1990, 231, 121. Aizawa, M., Shinahara, H., Chiba, T., and Matsuzawa, M., VTTSymp., 1988,90 (Biosensor symp. 1987), 64; Chem. Abstr., 111,53489. Kitagawa, Y., Kitabatake, K., Kubo, I., Tamiya, E., and Karube, I., Anal. Chim. Acta, 1989, 218, 61. Yokoyama, K., Tamiya, E., and Karube, I., J.Electroanal. Chem., 1989, 273, 107. Birch, B. J . , and Burns, I. W., Methods and Apparatus for Electrochemical Measurements, European Patent Application, 0 255 291 A l , 1988. Bond, A. M., and Scholz, F., Z. Chem., 1990, 30, 117. Kinoshita, K., Carbon-Electrochemical and Physicochemical Properties, Wiley Interscience, New York, 1988. Taraevich, M. R., and Khruscheva, E. I., in Modern Aspects of Electrochemistry, eds. Conway, B. E., Bockris, J. O., and White, R. E., Plenum, New York, 1989, pp. 295-358. Bard, A. J., and Faulkner, L. R., Electroanalytical Methods, Fundamentals and Applications, Wiley, New York, 1980, pp. 1 4 3 . Hall, E. A. H., Biosensors, Open University Press, Milton Keynes, 1990, pp. 97-140. Noninski, C. I., Doki. Bolg. Akad.Nauk., 1988, 41, 59. Wang, J., Anal. Chim. Acta, 1990, 234, 41. Murray, R. W., Ewing, A. G . , and Durst, R. A , , Anal. Chem., 1987, 59, 379A. Frew, J . E., and Hill, H. A. O., Anal. Chem., l987,59,933A. Cass, A. E . G., Davis, G., Francis, G . D., Hill, H. A. O., Aston, W. J., Higgins, J . , Plotkin, E. V., Scott, L. D. L., and Turner, A. P. F., Anal. Chem., 1984, 56, 667. Bartlett, P. N., and Whitaker, R. G., Biosensors, 1987/88, 3, 359. Bradley, J., Kidd, A. J., Anderson, P. A., Dear, A. M., Ashby, R. E., and Turner, A. P. F., Analyst, 1989, 114, 375. Green, M. J., and Hilditch, P. I . , Anal. Proc., 1991, 28, 374. Hilditch, P. I., and Green, M. J., Analyst, 1991, 116, 1217. Mathews, D. R.. Holman, R. R., Bown. E., Steemson, J., Watson, A., Hughes, S., and Scott, D., Lancet, 1987, 1.778. Jonsson, G., Gorton, L., and Pettersson, L., Electroanalysis. 1989, 1, 49. Wang, J.. Wu, L.-H., Lu, Z., Li, R., and Sanchez. J . , Anal. Chim. Acta, 1990. 228, 251. Gorton, L., Karan, H. I., Hale, P. D., Inagaki, T., Okamoto, Y., and Skotheim, T. A., Anal. Chim. Acta, 1990, 228, 23. Hale, P. D., Boguslavsky, L. I., Inagaki, T., Karan, H. I., Lee, H. S., Skotheim, T. A., and Okamoto, Y., Anal. Chem., 1991, 63, 677. Hale, P. D., Lan, H. L., Boguslavsky, L. I.. Karan, H. I., Okamoto, Y., and Skotheim. T. A., Anal. Chim. Acta, 1991, 251, 121. Beh, S. K., Moody, G. J., andThomas, J. D. R., Analyst, 1991, 116, 459. Okuma, H . , Takahashi, H., and Sekimukai, S., Anal. Chim. Acta, 1991, 244, 161. Amine, A . , Kauffmann, J.-M., and Patriarche, G.J., Anal. Lett., 1991, 24, 1293. Wang, J., and Varughese, K., Anal. Chem., 1990,62, 318. D'Costa, E. J., Higgins, I. J., and Turner, A. P. F., Biosensors, 1986, 2, 71.1228 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 Guadalupe, A. R., and Abruiia, H. D., Anal. Letts., 1986, 19, 1613. Frew, J. E., Harmer, M. A., Hill, H. A. O., and Libor, S. I., J. Electroanal. Chem., 1986, 201, 1. Chen, L., Lin, M. S., Hara, M., and Rechnitz, G. A., Anal. Lett., 1991, 24, 1. Ball, M. R., Frew, J. E., Green, M. J., and Hill, H. A. O., Proc. Electrochem. SOC., 1986, 8614,7. Frew, J. E., and Green, M. J., Anal. Proc., 1988,25, 276. Hale, P. D., Inagaki, T., Lee, H. S., Karan, H.I., Okamoto, Y., and Skotheim, T. A., Anal. Chim. Acta, 1990,228,31. Dicks, J. M., Aston, W. J., Davis, G., and Turner, A. P. F., Anal. Chim. Acta, 1986, 182, 103. Owlia, A., and Rusling, J. F., J. Electroanal. Chem., 1987,237, 297. Irvine, J. T. S., Eggins, B. R., and Grimshaw, J., J. Electroanal. Chem., 1989,271, 161. Zagal, J. H., Contrib. Cient. Tecnol., 1988, 18, 107. Lever, A. B. P., CHEMTECH., 1987, 17,506. Green, J. M., and Faulkner, L. R., J. Am. Chem. SOC., 1983, 105, 2950. Zagal, J., Fierro, C., and Rozas, R., J. Electroanal. Chem., 1981, 119,403. Korfhage, K. M., Ravichandran, K., and Baldwin, R. P., Anal. Chem., 1984,56, 1517. Halbert, M. K., and Baldwin, R. P., Anal. Chem., 1985, 57, 591. Halbert, M. K., and Baldwin, R. P., J. Chromatogr., (Biomed.Appls.), 1985, 345,43. Wring, S. A., Hart, J. P., and Birch, B. J., Analyst, 1989, 114, 1563. Wring, S. A., Hart, J. P., and Birch, B. J., Analyst, 1989, 114, 1571. Wring, S. A., Hart, J. P., and Birch, B. J., Talanta, 1991, 38, 1257. Wring, S. A., Hart, J. P., and Birch, B. J., Anal. Chim. Acta, 1990, 229, 63. Wang, J., Golden, T., Varughese, K., and El-Rayes, I., Anal. Chem., 1989,61, 508. Wring, S. A., Hart, J. P., Bracey, L., and Birch, B. J., Anal. Chim. Acta, 1990, 231,203. Wring, S. A., Hart, J. P., and Birch, B. J., Analyst, 1991, 116, 123. Wring, S. A., Hart, J. P., and Birch. B. J., Electroanalysis, 1992,4, 1. Wring, S. A., and Hart, J. P., Analyst, 117, 1992, 1283. Skladal, P., Anal. Chim. Acta, 1991, 252, 11. Halbert, M. K., and Baldwin, R. P., Anal.Chim. Acta, 1986, 187, 89. Santos, L. M., and Baldwin, R. P., J. Chromatogr., 1987,414, 161. Wang, J., Golden, T., and Li, R., Anal. Chem., 1988,60, 1642. Wang, J., and Li, R., Talanta, 1989, 36, 279. Santos, L. M., and Baldwin, R. P., Anal. Chim. Acta, 1988,206, 85. Tolbert, A. M., and Baldwin, R. P., Electroanalysis, 1989, 1, 389. Tolbert, A. M., and Baldwin, R. P., Anal. Lett., 1989,22,683. Schubert, F., Saini, S., and Turner, A. P. F., Anal. Chim. Acta, 1991,245, 133. Ramsey, G., and Turner, A. P. F.. Anal. Chim. Acta, 1988,215, 61. Durliat, H., Causserand, C., and Comtat, M., Anal. Chim. Acta, 1990, 231, 309. Kubiak, W. W., and Wang, J., Anal. Chim. Acta, 1989,221.43. Bonakdar, M., Vilchez, J. L., and Mottola, H. A., J. Electroanal. Chem., 1989,266, 47. Geno, P.W., Ravichandran, K., and Baldwin, R. P., J. Electroanal. Chem., 1985, 183, 155. Wang, J., Dempsey, E., and Ozsoz., M., and Smyth, M. R., Analyst, 1991, 116, 997. Shi-Hua Chen, S., Anal. Chim. Acta, 1990, 190, 129. Leech, D., Wang, J., and Smyth, M. R., Analyst, 1990, 115, 1447. Leech, D., Wang, J., and Smyth, M. R., Anal. Proc., 1992, 29, 25. Cox, J. A,, and Gray, T. J., Anal. Chem., 1989, 61, 2462. Cox, J. A., and Gray, T. J., Electroanalysis, 1990, 2, 107. Wang, J., and Golden, T., Anal. Chim. Acta, 1989, 217, 343. 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 1 I5 116 117 118 119 ANALYST, AUGUST 1992, VOL. 117 Elving, P. J.. Schmakel, C. O., and Santhanam, K. S. V., CRC Crit.Rev. Anal. Chem,., 1976, 6, 1. Persson, B., and Gorton, L., J. Electroanal. Chem.. 1990,292, 115. Gorton, L., J. Chem. SOC. Faraday Trans., 1986, 82, 1245. Palleschi, G., Rathore, H. S., and Mascini, M., Anal. Chim. Acta, 1988, 209, 223. Palleschi, G., Rathore, H. S., and Mascini M., Electroanalysis, 1989, 1, 199. Samec, Z., and Elving, P. J., J. Electroanal. Chem.. 1983. 144, 217. Blaedel, W. J., and Jenkins, R. A., Anal. Chem., 1975, 47, 1337. Gorton, L., Torstensson, A., Jaegfeldt, H., and Johansson, G., J. Electroanal. Chem., 1984, 161, 103. Appelqvist, R., Marko-Varge, G., Gorton. L., Torstensson, A., and Johansson, G., Anal. Chim. Acta, 1985, 169,237. Marko-Vargo, G., J. Chromatogr., 1987, 408, 157. Yabuki, S., Shinohara, H., Ikariyama, Y., and Aizawa, M., J. Electroanal. Chem., 1990, 277, 179. Huck, H., Fresenius’ 2. Anal. Chem., 1982, 313, 548. Schelter-Graf, A., Schmidt, H.-L., and Huck, H., Anal, Chim. Acta, 1984, 163, 299. Huck, H., Schelter-Graf, A., Danzer, J., Kirch, P., and Schmidt, H.-L., Analyst, 1984, 109, 147. Ni, F., Feng, H., Gorton, L., and Cotton, T. M., Langmuir, 1990,6,66. Buch-Rasmussen, T., Anal. Chem., 1990, 62, 932. Polasek, M., Gorton, L., Appelqvist, R., Marko-Varga, G., and Johansson, G., Anal. Chim. Acta, 1991,246, 283. Skoog, M., and Johansson, G., Biosens. Bioelectron., 1991. 6, 407. Ye, J.. and Baldwin, R. P., Anal. Chem., 1988, 60, 2263. Persson, B., J. Electroanal. Chem., 1990, 287,61. Jonsson, G., and Gorton, L., Biosensors, 1985, 1, 355. Tortensson, A., and Gorton, L., J. Electroanal. Chem., 1981, 130, 199. Yao, T., Matsumoto, Y., and Wasu, T., Anal. Chim. Acta, 1989, 218, 129. Yokoyama, K., Tamiya, E., and Karube, I., J. Electroanal. Chem., 1989,273, 107. Persson, B., and Gorton, L., J. Electroanal. Chem., 1990, 292, 115. Dominguez, P., Tuiion, P., Fernandez, J. M., and Smyth, M. R., Anal. Proc., 1989, 26, 387. Sanchez, P. D., Blanco, P. T., Alvarez, J. M. F.. Smyth, M. R., and O’Kennedy, R., Electroanalysis, 1990, 2, 303. Ikeda, T., Shibata, T., and Senda, M., J. Electroanal. Chem.. 1989, 261, 351. Kinoshita, H., Ikeda, R., Senda, M., and Usui, T., Bioelectro- chem. Bioenerg., 1990, 24, 365. Ikeda, T., Shibata, T., Todoriki, S., Senda, M., and Kinoshita. H., Anal. Chim. Acta, 1990.230.75. Ravichandran, K., and Baldwin, R. P., J. Electroanal. Chem., 1981, 126, 293. Jaegfeldt, H., Torstensson, A. B. C., Gorton, L. G. O., and Johansson, G., Anal. Chem., 1981, 53, 1979. Batchelor, M., Green, M. J., and Sketch, C. L., Anal. Chim. Acta, 1989, 221, 289. Miki, K., Ikeda, T.. Todoriki, S., and Senda, M., Anal. Sci., 1989, 5,269. Swain, A., Int. Ind. Biotechnol., 1988, 8, 11. Patchett, R. A., Kelly, A. F., and Knoll, R. G., Food Microbiol., 1989, 6, 159. Hu, J., and Turner, A. P. F., Anal. Lett., 1991, 24, 15. Scheller, F. W., Hintsche, R., Bogdanovskaja, V., and Ohls- son, P., Proceedings of the 3rd International Symposium on Redox Mechanisms. Interfacial Properties of Molecular and Biological Importance 1987 (Published 1988) p. 181. Cenas, N. K., Pocius, A. K., and Kulys, J . J., Bioelectrochem. Bioenerg., 1983, 11, 61. Cenas, N. K., Pocius, A. K., and Kulys, J. J., Bioelectrochem. Bioenerg., 1984, 12, 583. Gemmer, R. V., Cowan, D. 0.. Poehler, T. O., Bloch, A. N., Pyle, R. E., and Banks, R. H., J. Org. Chem., 1975,40,3544. Albery, W. J., Bartlett, P. N., and Cass, A. E. G., Phil. Trans R. Soc. Lond., 1987, B316, 107. McKenna, K., and Brajter-Toth, A., Anal. Chem., 1987, 59, 954.ANALYST, AUGUST 1992. VOL. 117 1229 120 121 Kulys, J . J . , Biosensors, 1986, 2, 3. 128 Lammert, R . , Ogbomo, I., Baumeister, T., Danzer, J . , Gunasingham. H., and Tan, C.-H., Electroanalysis, 1989, 1, Kittsteiner-Eberle, R . , and Schmidt, H.-L., GBF Monogr., 423. 1989, 13 (Biosens: Appl, Med. Environ. Prof. Process Control) 93. Palleschi, G . . and Turner, A . P. F., Anal. Chim. Acta, 1990, Ravichandran, K . , and Baldwin, R. P., Anal. Chem., 1983,55, 234,459. 1586. Cenas, N. K . , and Kulys. J . J . , J . Electroanaf. Chem., 1981,128, 103. Hendry, S. P.. and Turner, A . P. F., Horm. Metab. Res., 1988. 20.37. Kulys. J . , and Schmid, R. D., Anal. Lett., 1990, 23, 589. Kulys, J . J . , Bilitewski, U . , and Schmid, R. D.. Anal. Lett., 1991, 24, 181. 122 123 124 125 126 127 Gunasingham, H . , and Tan, C.-H., Analyst, 1990, 115, 35. 129 Paper 1/01 531 A Received February 19, 1992 Accepted May 28, 1992