ISSN:0003-2654    

DOI:10.1039/AN99217FX031

出版商:RSC    

年代:1992

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN99217BX033

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

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

ISSN:0003-2654    

DOI:10.1039/AN9921701215

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

1230 ANALYST, AUGUST 1992, VOL. 117 ~ ~ ~ ~~ ~ Introduction to Papers Presented at the Meeting ‘Analytical Applications of Chemically Modified Electrodes‘ held in Bristol, UK, January 7-8, 1992 The idea of deliberately modifying electrode surfaces in a controlled fashion was first put into practice some 15 years ago. Since that time, a vast literature has been generated on the subject, illustrated by a review of modified carbon electrodes in this issue (Wring). One of the attractions of the field is that there exists the opportunity for scientists with widely differing expertise-from biological and synthetic chemistry, through physical and analytical chemistry, to chemical physics-to contribute to the subject. One of the perennial problems is that the contributions of these different groups are not always integrated.In a meeting held at Bristol University on January 7-8, 1992, we tried to bring some of these interests together under the umbrella of Analytical Applications of Chemically Modi- fied Electrodes. The meeting was co-sponsored by the Electrochemistry and Electroanalytical Groups of the Royal Society of Chemistry. In accordance with the mutual interests of these groups, our aim was to try and bridge the gap (hence the cover design) between fundamental studies and applica- tion to real problems, with rational design of the electrode/ analyte interface completing the feedback loop. Papers were presented covering areas from the design and synthesis (Higgins, Mortimer) of analytically useful modified electrodes through to their utility for inorganic (Smyth) and biological (Bartlett, Gibson, Gorton) solution phase and gas phase (Slater) sensing. This is backed up by extensive characteriza- tion studies (Hillman, Lyons, Walton, Wang) using a wide range of physical techniques. Over the next few years, as the field of modified electrodes matures, an important question will be asked: ‘Do they have real applications?’ The paper devoted to the production and application of disposable modified electrode sensors (Wring) strongly suggests that the answer is ‘Yes’. Robert Hillman John Hart

ISSN:0003-2654    

DOI:10.1039/AN9921701230

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, AUGUST 1992, VOL. 117 1231 Scanning Probe Microscopies for High-resolution Characterization of Electrochemical Sensors* Plenary Lecture Joseph Wang Department of Chemistry, New Mexico State University, Las Cruces, NM 88003, USA A better understanding of tailored electrodes and electrochemical sensors requires a more detailed picture of their surfaces. New scanning probe techniques, such as scanning tunnelling or scanning bioelectrochemical microscopies, offer unique opportunities for high-resolution in situ characterization of tailored electrode- based sensors. Scanning tunnelling microscopy provides valuable information on the topography of pre-treated surfaces, the heterogeneity of composite electrodes, the morphology of electropolymerized films, the packing arrangement of adsorbed monolayers and the microdistribution of immobilized biological components.Scanning bioelectrochemical microscopy is shown to be extremely useful for the mapping of localized biological activity and the monitoring of dynamic biological events. Valuable insights are achieved by correlating the structural features with the preparation/modification conditions and the subsequent sensing performance. Such correlations can facilitate the predictive design of increasingly better sensors. Keywords: Scanning tunnelling microscopy; scanning electrochemical microscopy; modified electrode; biosensor The ability to control and manipulate the surface properties deliberately can greatly facilitate the development of electro- chemical sensors. 1 Tailored electrodes are very attractive for chemical sensing, as they couple the high sensitivity of amperometry with new dimensions of selectivity and stability provided by the surface modifier.Before many of these sensing applications are realized, numerous unanswered questions concerning the microstructures and function of modified electrodes must be addressed. Surface charcterization can play a very important role in understanding the fundamentals and performance of tailored sensor surfaces. Recently developed high-resolution scanning probe microscopies2 offer unique opportunities and chal- lenges for the characterization and optimization of electro- chemical sensors based on chemically modified electrodes. The term ‘scanning probe microscopies’ is a generic one for a family of techniques based on different types of interactions between the tip and the surface of interest.The opportunities accruing from use of these techniques for characterizing modified electrodes are described and discussed in the following sections. Scanning Tunnelling Microscopic Characterization of Tailored Sensor Surfaces In scanning tunnelling miFroscopy (STM), a sharp tip is brought to within several Angstroms of a sample surface so that a tunnelling current flows when a small bias (2 mV-2 V) is applied between them. The tip is scanned over the surface, while the current is being monitored. The tunnelling current is exponentially related to the tip-to-sample distance, and can be used as a sensitive probe for the structural and electronic properties of interfacial systems. The theoretical and practical aspects of STM are discussed in several reviews.Z4 Scanning tunnelling microscopy has rapidly become a powerful tool in electrochemistry.4 The ability to obtain high-resolution images in real time and space offers unique opportunities for the study of electrode/solution interfaces. Since the first application of STM for in situ observation of electrode surfaces,S several studies have appeared dealing * Presented at the Meeting on Analytical Applications of Chemi- cally Modified Electrodes, Bristol, UK, January 7-8, 1992.primarily with the characterization of electrodeposition or dissolution (corrosion) processes. Our activity has focused on the high-resolution probing of amperometric sensors. Such activity has been facilitated by the incorporation of computer- controlled potentiostats and electrochemical cells in commer- cial STM instruments6 (Fig.1). In situ electrochemical STM can, therefore, provide valuable insights into the nature of tailored electrodes (unattainable by other techniques). Par- ticularly attractive, for optimum sensor design and perfor- mance, is the ability to relate the new surface microstructures with the preparation (modification) conditions and the subse- quent amperometric response. Structure-preparation-perfor- mance relationships obtained in this laboratory for model tailored surfaces are reported below. Activation Processes Procedures for activating and cleaning of solid electrodes have been developed for obtaining reproducible electroanalytical results.’ It has been demonstrated that STM can be a very powerful tool for studying changes in the surface topography associated with electrochemical pre-treatment of glassy car- bon electrodes, and hence for obtaining a better understand- ing of such activation processes.8 Such ability relies on the exponential tunnelling current-gap relationship characteris- tics of STM systems.(The tunnelling current can change by a factor of 2 vr more with a change in the tip-surface separation of only 1 A.) Clear changes in the surface roughness have, therefore, been observed under different treatment paramet- ers (potential, frequency, duration, media, etc.). The surface images corresponded well with the cyclic voltammetric data for background and analyte solutions.Additional information Tunnelling current amplifier Reference electrode Tip c 1 /- Solution 4 v, = T -L Fig. 1 In situ electrochemical scanning tunnelling microscopy1232 ANALYST. AUGUST 1992, VOL. 117 on activation processes can be obtained by atomic force microscopy (AFM), which affords knowledge of the forma- tion of insulating layers (e.g., oxides) during anodic pre- treatments.9 Analogous STM and AFM studies should be useful for probing sensor passivation processes, involving the formation of inhibitory layers. 1 0 Electropolymerization Processes The fabrication of many electrochemical sensors relies on electropolymerization processes.11.12 The electrochemical preparation approach permits precise control of the surface microstructures and hence of the sensor performance.Scan- ning tunnelling microscopy can be used to investigate the morphology of electropolymerized films (on the nanometre scale), and it provides valuable insights into the correlation between the anodization conditions and the growth patterns. For example, it has been shown that the morphology of phenolic films is strongly dependent on the anodization conditions. 10 While a nucleation-and-growth process (leading to discrete, disordered aggregates) appears for potential- scanning experiments, a layer-by-layer growth process charac- terizes the fixed-potential polymerization (e.g., Fig. 2). Additional structural differences were observed via the use of different anodization potentials, scan rates or phenolic monomers. Scanning tunnelling microscopy has also been used to obtain structural-preparation correlations for poly- (aniline) coatings, where the high-resolution images indicate differences in the electropolymerization kinetics.13 Composite Sensor Surfaces Studies on the heterogeneity of solid electrode surfaces represent another example of utilizing STM for characterizing amperometric sensors. In particular, composite (array-like) electrodes possess signal-to-noise advantages when compared with traditional electrodes consisting of a single conducting phase.13 Such improved detectability depends strongly on the distribution of the conductor within the material. Because STM can be applied only to electrical conductors, scans taken over ‘large’ (micrometre) surfaces can provide useful informa- tion on the spatial variation of the conducting and insulating regions. Such tunnelling-on/tunnelling-off effects have been used in this laboratory to characterize the microdistribution of the conducting region in complex composite materials, such as modified carbon paste,” carbon foam composite16 and graphite-epoxyl7 electrodes.Visualization of Biosensors Scanning tunnelling microscopy images of biocomponents on electrochemical transducers can offer valuable insights into the operation of amperometric biosensors. High-resolution STM visualization of biosensors can provide useful insights into the immobilization of biocomponents on electrode surfaces. Careful correlation of the resulting surface micro- structures with the amperometric response can facilitate the fabrication of biosensors under optimum conditions.One promising immobilization avenue is the entrapment of enzymes in electropolymerized films. Scanning tunnelling microscopy (under potentiostatic control) has been shown in this laboratory to offer distinct views of glucose oxidase within a poly(pyrro1e) coating.18 Such imaging capability is attributed to changes in the local work function associated with the presence of the insulating enzyme clusters within the conduct- ing polymeric matrix. The resulting ‘black hole’ (e.g., Fig. 3) thus reflects the microdistribution of the enzyme on the surface. Such distribution, and the subsequent amperometric response, are strongly dependent on parameters of the immobilization/electropolymerizativn procedure. Similar structural/response correlations should facilitate the predic- tive design of biosensors based on other bicomponen t/immo- bilization/transducer systems.Adsorbed Monolayers Self-organized monolayers represent a versatile and powerful approach to tailored interfaces, as they provide new levels of selectivity and can serve as model systems for biomembranes. Scanning tunnelling microscopy can provide valuable insights into the behaviour of monolayers on surfaces. Specifically, Widrig et aL.19 have used STM to reveal the packing arrangements of n-alkanethiolate monolayers spontaneously adsorbed on gold surfaces. In addition, nano-scale defects within organic monolayers can be imaged by a combination of underpotential deposition and STM.20 Similar activity is currently being devoted to STM investigations of Langmuir- Blodgett monolayers and bilayers of phospholipids and fatty acids.Scanning Bioelectrochemical Microscopy Scanning bioelectrochemical microscopy (SBECM) was de- veloped in this laboratory21 to allow in situ mapping of biologically active surfaces. This scanning probe microscopic technique is a variant of scanning electrochemical micro- scopy,2’which involves use of a similar experimental setting as STM, but relies on a different sensing mechanism (the Fig. 2 Three-dimensional STM view of a poly(pheno1) film on a glassy carbon surface. following electropolymerization at 0.9 V for 1 min; reproduced, with permission, from ref. 10 Fig. 3 Three-dimensional STM view of glucose oxidase entrapped within a poly(pyrro1e) coating on a glassy carbon surface; reproduced.with permission, from ref. 18ANALYST, AUGUST 1992, VQL. 117 1233 I Positioner I Fig. 4 Scanning bioelectrochemical microscopy of biological sur- faces. Reactant (R) diffuses into the biological surface where the reaction occurs, producing a product (P) which is measured at the microelectrode (after reaching it by diffusion), resulting in another product (P’) monitoring of Faradaic, rather than tunnelling, currents). It is based on placing a microelectrode tip in close proximity to the biological specimen (in contact with a markedreactant solu- tion) in order to probe amperometrically/voltammetrically the biological consumption of the electroactive species or biogeneration of a detectable product (Fig. 4). For example, mapping of the localized glucose oxidase activity in a carbon paste - biosensor has been accomplished by incrementally moving the carbon fibre tip over the surface and monitoring the oxidation of the enzymically produced hydrogen peroxide.In addition to two-dimensional X-Y images of the surface biocatalytic activity, SBECM can be used for probing the kinetics of biological surface reactions. This is accomplished by repetitive and rapid measurements of biologically con- sumed (or generated) marker, performed at a fixed location above the biosurface, but at short time intervals (10 s ) . Square-wave voltammetry or chronoamperometry are par- ticularly suitable for this task. The dynamics of hydrophobic partitioning of drugs into lipid layers, of metal uptake by an alga-containing surface or the enzymic activity of tissue (mushroom) surfaces has thus been explored.23 Such ability to reveal the dynamics of biological processes is, perhaps, the greatest power of SBECM.The non-destruc- tivehon-damaging SBECM operation is very attractive for studying delicate biological structures (in their active state). Future SBECM work will be aimed at improving its resolution (through the use of nanometre tips to yield molecule-resolved images), and expanding its utility for investigating the structure/dynamics of biomembranes, interactions with deoxyribonucleic acid (DNA) and immunological reactions of affinity-based biosensing surfaces. There is no doubt that, on improving its resolution, SBECM will play an increasing role in studies of biological processes, in general, and for under- standing the performance of biosensors, in particular.Microlithography In addition to in situ characterization, it is possibIe to use scanning probe techniques for the modification of surfaces with high resolution. We are currently exploring the utility of STM and SBECM for the microfabrication of chemical sensors. In particular, we are interested in coupling the controlled (programmed) movement of the tip with modula- tion of the bias voltage for electrodepositing reactive struc- tures of nanometre dimensions. These could include extremely narrow and pre-defined enzyme ‘lines’, pre-defined membrane barriers (molecular sieves) or microelectrode arrays. Alternatively, it could be possible to use the tip to remove, locally, material from the surface. We hope to couple the lithographic and imaging capabilities of these techniques with amperometric testing, and to use the resulting prepara- tion-conditions/structuraVperformance correlations for pro- viding the foundation for nanometre-scale surface modifica- tion.Chemical modification of this sort would play an increasing role in the future fabrication of microsensors. Outlook The techniques of STM and SBECM are excellent tools for probing the structure of chemically modified electrodes. However, extreme caution is often required for the proper interpretation of complex surface objects and for obtaining reproducible images. Many important opportunities remain for further advances in the use of STM, SBECM, AFM and related scanning probe microscopic techniques for the charac- terization of electrochemical sensors and tailored electrodes.The author believes that the most valuable contributions of these imaging techniques will be the in situ investigation of molecular interactions and recognition. These and other studies will undoubtedly result in a rational design of new interfaces with predictable properties. Such developments are both stimulating and encouraging for those working in the field of modified electrodes. The author thanks D. Yaniv, L. McCormick, T. Martinez, N. Naser, R. Li and L. H . Wu for their valuable contributions, and the National Institutes of Health and the American Chemical Society for financial support. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Wang, J., Electroanalysis, 1990, 3, 255.Pool, R., Science, 1990, 247, 634. Hansma, P., and Tersoff, J., J. Appl. Phys., 1987, 61, Rl. Cataldi, T., Blackham, I., Briggs, A., Pethica, J., and Hill, H., J. Electroanal. Chem. Interfacial Electrochem., 1990, 290, 1. Sonnenfeld, R., and Schardt, B. C., Appl. Phys. Lett., 1986,49, 1172. Yaniv, D., and McCormick, L., Electroanalysis, 1991, 3, 103. Mattusch, J., Hallmeier, K., Stulik, K., and Pacakova, V., Electroanalysis, 1989, 1, 405. Wang, J., and Lin, M. S., Anal. Chem., 1988, 60,499. Freund, M. S., Brajter-Toth, A., Cotton, T. M., and Hender- son, E., Anal. Chem., 1991, 63, 1047. Wang, J., Martinez, T. Yaniv, D., and McCormick, L., J. Electroanal. Chem. Interfacial Electrochem., 1991, 313, 129. Wang, J., Chen, S. P., and Lin, M. S., J. Electroanal. Chem. Interfacial Electrochem., 1989, 273, 231. lmisides, M., John, R., Riley, P., and Wallace, G., Electro- analysis, 1991, 3, 879. Kim, Y., Yang, H., and Bard, A. J., J. Electrochem. SOC., 1991, 138, L71. Tallman, D., and Petersen, S., Electroanalysis, 1990, 2 , 499. Wang, J., Martinez, T., Yaniv, D., and McCormick, L., J. Electroanal. Chem. Interfacial Electrochem., 1990, 286, 265. Wang, J., Brennsteiner, A., Angnes, L., Sylwester, A., La Gasse, R., and Bitsch, N., Anal. Chem., 1992, 64, 151. Wang, J., Romero, E., and Ozsoz, M., Electroanalysis, 1992,4, 539. Yaniv, D., McCormick, L., Wang, J., and Naser, N., J. Electroanal. Chem. Interfacial Electrochem., 1991, 314, 353. Widrig, C., Alves, C., and Porter, M., J. Am. Chem. SOC., 1991, 113,2805. Sun, L., and Crook, R., J. Electrochem. SOC., 1991, 138, L23. Wang, J., Wu, L. H., and Li, R., J. Electroanal. Chem. Interfacial Electrochem., 1989, 272, 285. Engstrom, R., and Pharr, C., Anal. Chern., 1989, 61, 1099A. Wu, L. H., Ph.D. Dissertation, New Mexico State University, Las Cruces, 1992. Paper 2100576J Received February 3, 1992 Accepted March 2, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701231

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, AUGUST 1992, VOL. 117 1235 Am perometric Biosensors Based on an Apparent Direct Electron Transfer Between Electrodes and Immobilized Peroxidases* Plenary Lecture Lo Gorton, Gunilla Jonsson-Pettersson, Elisabeth Csoregi, Kristina Johansson, Elena Dominguezt and Gyorgy Marko-Varga Department of Analytical Chemistry, University of Lund, P.O. Box 124, S-22 7 00 Lund, Sweden An apparent direct electron transfer between various electrode materials and peroxidases immobilized on the surface of the electrode has been reported in the last few years. An electrocatalytic reduction of hydrogen peroxide stars at about +600 mV versus a saturated calomel (reference) electrode (SCE) at neutral pH. The efficiency of the electrocatalytic current increases as the applied potential is made more negative and starts t o level off at about -200 mV versus SCE. Amperometric biosensors for hydrogen peroxide can be constructed with these types of peroxidase modified electrodes.By co-immobilizing a hydrogen peroxide-producing oxidase with the peroxidase, amperometric biosensors can be made that respond t o the substrate of the oxidase within a potential range essentially free of interfering electrochemical reactions. Examples of glucose, alcohol and amino acid sensors are shown. Keywords: Biosensor; direct electron transfer; amperometry; electrode; immobilized enzyme Amperometric biosensors have been at the focus of electro- analytical research since the first 'enzyme electrode' for the detection of glucose was reported by Updike and Hicks in 1967.' More than 1000 papers have been published since then on amperometric biosensors for glucose and for a series of other analytes.The field of enzyme based amperometric biosensors was recently reviewed by Bartlett et a1.2 One of the major obstacles to be solved in the construction of enzyme based amperometric biosensors is how to optimize the electron transfer reaction between the cofactor of the redox enzyme used and the electrode. All redox enzymes rely on a cofactor as the redox active compound for activity. In all classes of redox enzymes, except the nicotinamide dependent dehydrogenases, the cofactor is strongly bound within the enzyme structure causing steric hindrances for a direct electron transfer between the active site of the enzyme and the electrode.3 Only in a few cases has a direct electron transfer reaction been claimed to occur and in still fewer has the electron transfer reaction been efficient enough to allow the construction of a sensor.2.4 The electron transfer has mostly been brought about using a soluble species that can diffuse from the active site to the electrode and vice ~ersa.2~4 Sometimes redox mediator modi- fied electrodes have been used to facilitate the electron transfer.2.4 A special case is the use of conducting salt electrodes in combination with redox enzymes .2,5 Peroxidases have been intensively studied for the construc- tion of amperometric biosensors not only for sensing hydrogen peroxide and small organic peroxides but also in the combina- tion with hydrogen peroxide producing oxidases for serving the substrate of the oxidase, e.g.glucose, alcohol, amino acids and xanthine.2 In most cases soluble mediators610 or mediator modified electrodes have been used in conjunction with these sensors.2.I 1 However, in 1977, Yaropolov et al.12 reported on what appeared to be a direct and a very efficient electron transfer from carbon black to adsorbed horse radish peroxidase (HRP, E.C. 1.1.11.7) in the presence of hydrogen peroxide. An electrocatalytic reduction of hydrogen peroxide started to occur at about +600 mV versus a saturated calomel (refer- ence) electrode (SCE) at pH 7. Making the applied potential more negative resulted in a much increased current. Not until about -200 mV did the current level off to a more constant value. Similar effects were reported by Paddock and Bow- den13 on cytochrome c peroxidase adsorbed on edge oriented pyrolytic graphite, by Jonsson and Gorton on HRP adsorbed on spectrographic graphite,14 by Kulys and Schmid on fungal peroxidase on graphite,'5 by Wollenberger et al. on HRP immobilized on graphite and Pt,16 by Gorton et al.on HRP, lactoperoxidase and microperoxidase adsorbed on various graphites, glassy carbon and coal electrodes,17-19 by Wata- nabe et al. on immobilized microperoxidase on Sn02 elec- trodes,20 by Gorton et al. on HRP immobilized in carbon paste,4 and by Wollenberger et al. also on HRP and fungal peroxidase immobilized in carbon paste and in graphite epoxy resins.21 The peroxidase modified electrodes can thus work as amperometric biosensors for hydrogen peroxide detection within the optimum potential range, -200 and 0 mV versus SCE, where the risk for interfering reactions is minimized and also where the background current and noise levels take their lowest values.Sensors for oxidase substrates have also been studied using this effect by co-immobilizing a hydrogen producing oxidase with the peroxidase. Kulys and Schmidls studied sensors for alcohol, choline, and glucose by co-immobilizing alcohol, choline and glucose oxidase together with fungal peroxidase on solid graphite electrodes. Gorton et a1.17318 and Jonsson- PetterssonZ2 reported on glucose sensors based on co-immobi- lizing glucose oxidase with HRP on heat-treated spectro- graphic graphite. Here we report on the latest achievement in this direction from our laboratory.We have been involved in the investiga- tion on the background mechanism, construction of co- immobilized oxidase-peroxidase based carbon-paste sensors, and construction of microsensors based on carbon fibres. * Presented at the meeting on Analytical Applications of Chemi- cally Modified Electrodes, Bristol, UK, January 7-8, 1992. 1- Permanent address; Department of Nutrition and Food Analysis, University of Alcala de Henares, 28871 Alcala de Hcnares, Madrid, Spain. Background Mechanism The reaction between hydrogen peroxide and peroxidases occurs according to the following. In a first 2e- transfer step hydrogen peroxide is reduced to water and the bound cofactor1236 2 H H ~ ~ H202 ANALYST, AUGUST 1992, VOL. 117 hemine IX,, ( HRP )r e- E-proto- Carbon paste hemine IXred electrode (in most cases ferroprotoporphyrin IX) is oxidized.This oxidized form of peroxidase is usually denoted compound-1 ; H202 + HRP + H20 + compound-1 (1) Compound-1 is then reduced in a first 1 e- electron step to form compound-11; Compound-I + e- + compound-I1 Compound-I1 + e- + HRP (2) (3) followed by a second le- step back to HRP; Virtually any reducing agent is capable of donating electrons to compound-I and compound-11. The necessary protons are donated either by the reducing agent or taken from the surrounding media. Examples of commonly used reducing agents in this context are; phenols, o- and p-hydroquinones, pyrogallol, resorcinol, N, N- alkylanilins, o- and p-phenylene- diamines, iodide, hexacyanoferrate(u), ferrocenes, o-toli- dine, o-toluidine , o-dianisidine, 4-aminoantipyrine, etc, used both in spectrophotometric and electrochemical detection systems to follow peroxidase catalysed reactions.It is important to elucidate the mechanism behind the electron transfer between electrode and attached enzyme, because if known it can hopefully be further optimized. One problem encountered with these peroxidase sensors is the restricted linearity of the calibration curves for hydrogen peroxide. Many peroxidases are rather small redox enzymes with relative molecular masses ranging between 40000 and 100 OOO u.23 The amino acid sequence of one of the isoenzymes of HRP is known.24 The enzyme consists of a hemin prosthetic group, 2Ca2+, a single polypeptide chain, and also 8 neutral carbohydrate chains.From a possible structure of this isoenzyme given by Welinder,24 it is reasonable to believe that the bound cofactor is not situated in the centre but rather at a distance more close to the outer surface of the enzyme molecule. The distance between the bound cofactor in the active site of an immobilized HRP molecule and the electrode could thus be small enough to allow a reaction between the cofactor and the electrode surface. Strong evidence for a direct electron transfer from the electrode to adsorbed peroxidases have been claimed by Yaropolov et al. ,l2 by Paddock and Bowden,13 by Kulys and Schmid15 and by Wollenberger et al. 16 One indication in this direction is the appearance of an electrocatalytic reduction current of hydrogen peroxide close to the I?" value of the reactions in eqns.(2) and (3).25 However, cyclic voltammo- grams of peroxidase modified electrodes do not show any significant voltammetric waves in the absence of hydrogen peroxide that can be correlated to oxidation-reduction reactions with the bound ferriprotoporphyrin IX cofactor. Kulys and Schmid15 found a strong correlation between the open circuit steady-state potential of an electrode modified with fungal peroxidase in a solution containing hydrogen peroxide and the E"' values of eqns. (2) and (3). However, the steady-state potentials obtained in similar experiments repor- ted from this laboratory with adsorbed HRP on various graphites and on glassy carbon could not be directly correlated with the I?" values.19 Our findings were that with thermally pre-treated electrodes (700 "C, 1.5 min) higher catalytic efficiencies could be obtained especially for glassy carbon and some high density ultra-pure graphite.Oxidative electrochem- ical pre-treatments of carbon electrodes at high potentials (== +1.5-2.0 V versus SCE) on the other hand strongly decrease the efficiency of the catalytic effect and the amount of HRP that could be adsorbed on the electrode surface. Electrocat- alytic activity of a deactivated electrode could be regained by the deliberate immobilization of strongly adsorbing mediators containing o-quinone or other quinoid functionalities. 19 A possible explanation might be that the thermal pre- treatment procedure, most efficient for obtaining the catalytic effect, introduces oxygen containing functionalities on the For covalent coupling of enzyme(s) I Fig.1 Possible structure of graphite with functionalities responsible for the electrocatalytic effect in combination with immobilized peroxidases and of functionalities that can be used in conjunction with carbodiimide coupling of the enzyme 1 E-oroto- Fig. 2 Reaction scheme for hydrogen peroxide reduction at a carbon paste electrode chemically modified with HRP electrode surface capable of mediating the electron transfer from the electrode to compounds I and I1 [eqns. (2) and (3)]. By analysing the hydrodynamic voltammograms of electrodes modified with HRP, lactoperoxidase and microperoxidase obtained in the presence of hydrogen peroxide it was revealed that at pH 7 the catalytic current increases abruptly at about +200 and -150 mV.19This could be in accordance with having different o-quinone functionalities created on the electrode surface,26 e.g.as depicted in Fig. 1, expected to have values close to +200 and -150 mV, respectively. The electrodes that revealed the most efficient electrocatalytic current for hydrogen peroxide reduction when modified with HRP were also shown to be the most efficient for electrochem- ical oxidation of ascorbate and NADH.19 (The experiments with ascorbate and NADH were performed in the absence of HRP.) It is well known that o-quinones and electrodes modified with o-quinones catalyse the electrochemical oxida- tion of these compounds.26~27 A similar explanation was given by Staskeviciene et a1.28 for the electron transfer reaction between immobilized lactate dehydrogenase (cytochrome b2) and carbon black.However, no such explanation can be given for the peroxidase-modified electrodes based on Pt or Sn02. 16,20 All initial work on reagentless amperometric sensors with peroxidases were reported for solid electrodes. Peroxidases can also be immobilized in carbon paste electrodes while retaining their electrocatalytic effect for hydrogen peroxide reduction.4Jl Fig. 2 shows the reaction scheme for catalytic reduction of hydrogen peroxide at a carbon paste electrode chemically modified with HRP. We found that a short (15 s) thermal pre-treatment of the graphite powder resulted in a slightly higher catalytic current than when taking the powder as it is delivered by the manufacturer (Fluka, Cat. No.50870). The importance of a close coupling between the graphite and the added peroxidase was shown when the enzyme was added to a pre-made carbon paste material. No catalytic current could be traced for this preparation. If, however, peroxidaseANALYST, AUGUST 1992, VOL. 117 1237 0.5 1 1.5 2 2.5 a 0 0.02 0.04 0.06 [HzOzl/mmol dm-3 Fig. 3 Calibration graphs for hydrogen peroxide of two carbon paste electrodes chemically modified with covalently bound HRP. A, using only carbodiimide; and B, using both carbodiimide and glutaral- dehyde. (a) Shows the entire and (b) the lower part of the concentration range investigated 0.25 - 0.15 - f a 0.05 - -0.05 - 400 -200 0 200 400 600 UlmV ! 800 Fig. 4 Hydrodynamic voltammograms of two equivalent HRP modified carbon paste electrodes obtained for 0.1 mmol dm-3 hydrogen peroxide.A, Indicates a starting potential of +600 mV; and B of -200 mV. The electrode surface is also covered with six layers of an ion exchange membrane (Eastman AQ 29D30) was first allowed to adsorb onto the graphite powder [1400 U (1 U = 16.67 nkat) of HRP taken per 100 mg of graphite powder] prior to the addition of the pasting liquid the electrocatalytic properties were retained.4 It was shown too, that the choice of the pasting liquid affected the response to hydrogen peroxide. Comparing the results obtained for paraffin oil, two silicon oils (Dexsil 400 GC and silicon oil GE SF-96), and silicone DC 710 (50% phenylmethylsilicon oil), the last three being common stationary phases in gas chromatography, the best one was found to be the phenyl- methylsilicon oil (40 mm3 per 100 mg of graphite powder) judged on the basis of a high response to hydrogen peroxide and a low background current.All electrodes simply based on adsorbing HRP lacked long-term stability. By first activating the graphite powder with l-cyclohexyl-3-(2-morpholinoethyl)cardodiimide metho- p-toluenesulfonate [4.2 mg of carbodiimide dissolved in 300 mm3 of 0.05 mol dm-3 acetate buffer (at pH 4.8) per 100 mg of graphite, left to react for 2 h, then rinsed with pure water seven times and dried in a vacuum before the addition of the enzyme dissolved in 200 mm3 of 0.1 mol dm-3 phosphate buffer at pH 7.0, left to react for 16 h, and dried before addition of the pasting liquid] both the electrocatalytic response (>2 times) and the long-term stability were much improved.The functionalities on the surface of the graphite 0.24 0.20 4 A a 0.16 0.12 3 4 5 6 7 8 9 PH Fig. 5 Variation of the response with pH to 0.1 mmol dm-3 hydrogen peroxide at -50 mV versus Ag-AgCI for four equivalent carbon paste electrodes chemically modified with covalently bound HRP. The electrode surface is also covered with six layers of an ion exchange membrane (Eastman AQ 29DW) 9 0.08 3 4 5 6 7 8 PH Fig. 6 Variation of the response to 0.7 mmol dm-3 glucose with the pH of the 0.1 mol dm-3 phosphate buffer of four equivalent electrodes modified with both HRP and GOD. The electrode surface is also covered with six layers of an ion exchange membrane (Eastman AQ 29DW) powder used for coupling to carbodiimide are depicted in Fig.1. Further addition of covalent coupling agents, e.g. glutaral- dehyde, to the reaction mixture (0.2% glutaraldehyde) was also beneficial for the long-term stability of the paste electrodes and also had some additional influence on the magnitude of the response current. Fig. 3 shows the calibration characteristics for hydrogen peroxide of HRP-modified carbon paste electrodes with an applied potential of -50 mV versus Ag-AgCI (0.1 mol dm-3 KCl) when using carbodiimide activated graphite and also when using glutaraldehyde. The electrodes (surface area 0.25 cm2) were mounted in a flow-through amperometric cell of the wall-jet type29 connected to a single line flow injection (FI) system with a flow rate of 0.6 ml min-1 and with an injection volume of 50 mm3.The calibration graphs reveal the restricted linear response characteristics. An apparent Michaelis- Menten constant, KMapp, could be evaluated from electro- chemical Eadie-Hofstee plots and gave values of about 0.8-1.0 mmol dm3, which is in accordance with the linear part of the calibration curve seen in Fig. 3(b). When concentrations of hydrogen peroxide of more than about 2 mmol dm3 were used an irreversible decrease in the sensor response was noticed. The electrodes obtained using both carbodiimide and glutaraldehyde lost about 40% in response to hydrogen peroxide after 24 h. The stability could be much increased by covering the electrode surface with an anion-exchange mem- brane of the Eastman AQ type.4.30 Even though the initial response decreased by about 75% on covering with six layers30 of the membrane, the response of the electrodes was virtually constant after 1 d up to about 3 weeks with analyses carried out every 5 d and with storage at 4 "C between experiments.Fig. 4 shows the hydrodynamic voltammograms obtained between -200 and +600 mV for electrodes covered with the anionic membrane. The hydrodynamic voltammograms show clear increases in the catalytic current at about +200 and -150 mV as was referred to above, irrespective of the startingANALYST, AUGUST 1992, VOL. 117 1238 0.6 0.4 0.2 B I 1 I I I 1 2 3 4 5 6 I I I I I [Substratel/mmol dm-3 0 0.02 0.04 0.06 0.08 0.1 Fig. 7 Calibration curves for A, hydrogen peroxide and B, glucose at pH 5 and at -50 mV versus Ag-AgC1 of a carbon paste electrode chemically modified with covalently bound HRP and GOD.The electrode surface is also covered with six layers of an ion exchange membrane (Eastman AQ 29D30). (a) Shows the entire and (b) the lower part of the concentration range investigated R-CHZOH R-CHO Carbon paste electrode I Fig. 8 lized AOD and HRP in a carbon paste electrode Reaction scheme for an alcohol sensor based on co-immobi- potential (-200 or +600 mV) of the experiment. The current versus pH profile is depicted in Fig. 5 showing the highest response to hydrogen peroxide between pH 4 and 5. Four different sensor types have so far been studied in our laboratory based on co-immobilizing HRP in the paste along with hydrogen peroxide producing oxidases, viz. glucose oxidase (GOD), alcohol oxidase (AOD), D-amino acid oxidase (D-AAOD), and L-amino acid oxidase ( L-AAOD).The ratio between the taken amounts of HRP and GOD was studied for the glucose sensor to find the amounts giving the maximum glucose response. A similar immobilization proce- dure was used as when HRP was bound alone. Carbodiimide activation was followed by addition of either 4.5 mg of HRP (1400 U) and 4.5 mg of GOD (680 U) per 100 mg of graphite or 4.5 mg of HRP (1400 U) and 2.3 mg of GOD (347 U) per 100 mg of graphite. Glutaraldehyde was also added to the reaction mixture just at the moment of mixing with the graphite powder. The maximum response for glucose was obtained with an HRP:GOD ratio of 2 : 1. All glucose electrodes were also covered with the anionic membrane, as above.Fig. 6 shows the response versus pH of four equi- valently prepared electrodes. The pH profile reveals clearly the pH effect of GOD, cf. Fig. 5, showing the maximum activity for glucose at pH 5. Fig. 6 also shows the variation in the reproducibility between electrodes prepared from the same batch of chemically modified carbon paste and after membrane formation. Fig. 7 shows the calibration characteris- tics for glucose and hydrogen peroxide obtained at pH 5. As is 4 5 6 7 8 9 1 0 PH Fig. 9 Variation of the response to 1.0 mmol dm-3 ethanol with pH at -50 mV versus Ag-AgC1 obtained for three equivalent carbon paste electrodes chemically modified with covalently bound HRP and AOD. The electrode surface is also covered with six layers of an ion exchange membrane (Eastman AQ 29D30) clear, the response for glucose is much lower than that for hydrogen peroxide and that the response to both substrates is almost completely non-linear over the whole investigated concentration range. Analysing the KMaw value for HRP when co-immobilized with GOD resulted in lower values than when HRP was immobilized alone.A similar effect was noticed in a previous investigation when HRP and GOD were co-immobilized on a solid graphite electrode22 even though an increase in the value might have been expected owing to a higher diffusion hindrance. No explanation to this behaviour can be given at this time. The non-linear calibration charac- teristics for glucose is expected to be partly due to the low KMapp of HRP. Usually linear response curves for glucose are obtained over at least two orders of magnitude when GOD is immobilized directly onto the surface of carbon electrodes or in carbon paste electrodes.30-32 Alcohol oxidase was also co-immobilized with HRP in a carbon paste to produce an alcohol sensor.Although it is a rather unselective enzyme showing activity not only for methanol and ethanol but also for propanol, isopropanol and butanol with a decreasing turnover rate with an increased length of the carbon chain. It belongs to the group of oxidaseb that have low reported reaction rates with alternative electron acceptors, e.g. ferrocinium derivatives, to molecular oxy- gen,ls which are commonly used as mediators to facilitate the electron transfer between various redox enzymes and elec- trodes.2 It is thus of special importance if the problems encountered with sensor construction, using oxidases that have low reaction rates with alternative electron acceptors, can be solved using co-immobilization of these oxidases with a peroxide and hence use can be made of the apparent direct electron transfer mechanism between the electrode and the immobilized peroxidase.The ratio of AOD and HRP taken was guided by the experiments with co-immobilizing GOD and HRP and was 3 : 1 [4.5 mg (137 U) of AOD and 1.5 mg (394 U) of HRP per 100 mg of graphite]. The same immobilization procedure was followed using carbodiimide activation and addition of gluta- raldehyde to the immobilization reaction mixture. Fig. 8 shows the reaction scheme for the alcohol sensor.Fig. 9 shows the variation of the response to ethanol with the pH of the contacting buffer of three equivalently prepared electrodes. Compared with the responses to hydrogen peroxide or glucose a clear shift into the alkaline region is seen, which reflects the preference of AOD to work optimally in an alkaline environ- ment and a pH of 7.5 was chosen for further experiments. Contrary to the GOD electrode the alcohol sensor responded linearly to ethanol between 0.1 and 1 mmol dm-3 (not shown). However, the co-immobilization had a similar effect on the & ~ P P of HRP as when co-immobilizing it with GOD. Fig. lO(a) line A shows FI peaks for ethanol and hydrogen peroxide and shows much smaller tailing peaks for ethanol than for hydrogen peroxide, cf.below.ANALYST, AUGUST 1992, VOL. 117 1239 "i t ' 10min ' (bl t 10min' Time - Fi 10 FI responses to (H) 50 pmol dm-3 hydrogen peroxide and to yi70.5 mmol dm- e thanol at H 7.5 and at -50 mV versus Ag-AgC1; a ) with carbon paste e1ectrod)es chemical1 modified with covalently bound HRP and AOD (b) and in (a) but wiere PEI was also added to the immobilization reaction (0.12%). A denotes electrodes with no electropolymerized layer on the surface whereas B, C and D denote electrodes subjected to 2,5, and 20 tential scans in a 5 mmol dm-3 o-phenylenediamine containing b u g r , res ctively, between 0 and +650 mV versus SCE (scan rate 50 mV s-lp" a s 0 0.04 0.08 0.12 [Substratel/mmol dm-3 Fig. 11 Calibration curves for: A, hydrogen peroxide; B, methanol; and C, ethanol, at pH 7.5 and at -50 mV versus Ag-AgC1 of a carbon paste electrodes chemically modified with covalently bound HRP and AOD, PEI, also covered with a layer of electropolymerized mixture of 1.5 mmol dm-3 rn-phenylenediamine and 1.5 mmol dm-3 resorcinol obtained after 3 scans (50 mV s-l) between 0 and +650 mV versus SCE and with six layers of an ion exchange membrane (Eastman AQ 29Dm).(a) shows the entire and (b) the lower part of the concentration range investigated In the last few years several papers have reported improved biosensor characteristics when covering the electrode surface with electropolymerized layers.33-38 The surface can be protected against interfering and fouling agents, an increased diffusional barrier will be obtained, which increases the value of an unfavourable KMapp of an immobilized enzyme thereby increasing the linear response range of the substrate, and an increased long-term stability can be obtained for the enzyme immobilized either within the layer or directly onto the electrode surface.The electropolymerization reaction with these reagents starts at relatively low potentials (- +650-700 mV versus SCE) and was reported not to destroy the activity of the enzymes investigated. Different monomers were therefore tried in order to form the electropolymerized layers to protect the electrode surface, increase the long-term stability and improve the sensor characteristics. The mono- mers tried were: aniline, 0.1 mol dm-3 in 0.1 mol dm-3 phosphate buffer at pH 7.0;33 pyrrole, 0.1 mol dm-3 in 1.0 mol dm-3 KC1;N phenol, 0.05 mol dm-3 in 0.1 mol dm-3 phosphate buffer at pH 7.0;35 o-phenylenediamine, 5 mmol dm-3 in 0.1 rnol dm-3 acetate buffer at pH 5.2;36*3' and a mixture of rn-phenylenediamine and resorcinol, 1.5 mmol dm-3 of each in 0.1 rnol dm-3 phosphate buffer at pH 6.5.38 Electrodes were dipped into these solutions and a series of cycled potential scans was allowed to proceed between 0 mV and a higher potential registered to allow the monomer to be oxidized and the polymer to be formed.The membrane formation for all of these reagents resulted in an additional increase in the background current and a decreased response (as expected) but also a much more pronounced tailing of the ethanol peaks was noticed when the membrane covered electrodes were operated at -50 mV, i.e.as sensors in a 50 n A I B I I 10 5 tlmin Fig. 12 FI recordin s at H 7.0 and at -50 mV versus Ag-AgC1 of: A, 0.1 mmol dm-f hy&ogen peroxide; and B, 1 mmol dm-3 L-phenylalanine, with a carbon paste electrode chemically modified by deposition of HRP, L-AAOD and PEI 25 r I 0 500 1000 1500 2000 2500 3000 HzOz/pmol dm-3 Fig. 13 Calibration curve for hydrogen peroxide at pH 6.0 of a bundle of carbon fibres chemically modified by adsorbed HRP1240 ANALYST, AUGUST 1992, VOL. 117 the FI system. Positive results regarding increased stability and linear calibration characteristics were only obtained with o-phenylenediamine and with the mixture of rn-phenylene- diamine and resorcinol. Figs. lO(a) B-D show the effect on the response to 50 pmol dm-3 hydrogen peroxide and 0.5 mmol dm-3 ethanol after covering the alcohol sensor with an electropolymerized layer obtained after 2, 5 and 20 scans between 0 and +650 mV vesus SCE in a solution containing 5 mmol dm-3 o-phenylenediamine. A further stabilization of the electrode could be found when the surface was also covered with layers of an anionic membrane, Eastman AQ 29 D4730 after the electropolymerization reaction.When aniline, phenol or pyrrole was used for making the electropolymerized membrane even lower response currents for ethanol and hydrogen peroxide were registered and also much increased background currents. These reagents might either be detri- mental to the enzymes or destroy the electrode material. In a previous study of immobilizing alcohol dehydrogenase and its necessary cofactor, NAD+, in a mediator modified carbon paste,39 we found that the addition of poly- ethyleneimine (PEI)40 had a very beneficial effect not only on the response to ethanol, but also on the stability and gave a much reduced peak tailing.Fig. lO(6) line A shows the effect on the response when adding PEI (Sigma Cat. No. P3143) to the carbon paste electrode compared with Fig. 10(a) line A. A final concentration of 0.12% PEI was taken for the enzyme immobilization reaction mixture and added to the carbodi- imide activated graphite powder. A much higher response current to ethanol was obtained, along with less tailing of the FI peaks, and a significantly less noticeable decrease in current after forming the electropolymerized layer of poly-o-phenyl- enediamine, Figs.10(b) lines b-d. As in the previous case with immobilized alcohol dehydrogenase, no certain explanation to this behaviour can be given. However, at pH 7.5 many enzymes are expected to have a net negative charge. The PEI is positively charged and the electrostatic interaction between PEI and the enzyme is obviously beneficial for the charge transfer reaction and also for the stability. Fig. 11 shows the calibration curves for hydrogen peroxide, methanol and ethanol for the electrodes with a membrane formed from electropolymerization of rn-phenylenediamine with resorcinol and also covered with six layers of the Eastman AQ membrane. As is clear from this figure, much improved sensor characteristics can be obtained by the introduction of proper reagents into the paste.When D- or L-amino acid oxidases were first co-immobilized with HRP using carbodiimide activated graphite powder and with the addition of glutaraldehyde to the reaction mixture no responses to any amino acid could be traced. The response to hydrogen peroxide was, however, retained, indicating that the reagents covalently binding to the amino acid oxidases caused an inactivation of these enzymes. Experiments with only one covalent binding reagent was also negative. Excluding both reagents and just adding PEI to the enzyme solution before addition to the graphite powder resulted in retained activities for these enzymes in the past electrodes. Previous experiments when immobilizing L- and D-AAOD on controlled-pore glass show that the choice of immobilization reagent can be critical.41 Fig.12 shows the FI response to 0.1 mmol dm-3 hydrogen peroxide and to 1.0 mmol dm-3 L-phenylalanine for a carbon paste electrode containing L-AAOD and HRP co-immobilized with the addition of PEI to the paste. L-Phenylalanine was previously shown to be one of the most active substrates for immobilized L-AAOD.~~ Much of the current work in the analytical field is devoted to the miniaturization of the analytical equipment. One par- ticular area is for microelectrodes with their possible use as sensors in microflow systems.42 A series of different carbon fibres was therefore tested to see whether the catalytic effect for hydrogen peroxide reduction could be obtained with carbon fibre electrodes modified with immobilized peroxi- dases. A small effect was traceable on all different types studied except for graphitized carbon fibres where high catalytic currents were obtained.Fig. 13 shows a calibration curve obtained in a beaker for hydrogen peroxide registered using a bundle of about 2&30 fibres (Polycarbon LGR 10-ply Z-twist) 4 mm in length and modified with HRP (6200 U cm-3), by allowing the enzyme to adsorb for 10 min. These electrodes were as expeeted not stable long-term. Some preliminary experiments show that higher response currents to hydrogen peroxide can be obtained with thermally and/or electrochemically pre-treated electrodes and that the stability can be substantially increased with covalent attachment of the enzyme to the electrode surface. This work was financially supported by the Swedish Natural Research Council (NFR), the Swedish Board for Technical Development (STU) and the Swedish National Energy Administration (STEV).The authors thank Ms. Valentina Kacaniklic for the work with the amino acid oxidases and Mr 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 Alexander Rappensberger for drawing the figures. References Updike, S. J., and Hicks, G. P., Nature (London), 1967, 214, 986. Bartlett, P. N., Tebbutt, P., and Whitaker, R. G., Progr. React. Kinet., 1991, 16, 55. Armstrong, F. A,, Hill, H. A. O., and Walton, N. J., Acc. Chem. Res., 1988, 21, 407. Gorton, L., Csoregi, E., Dominguez, E., EmnCus, J., Jonsson- Pettersson, G., Marko-Varga, G., and Persson, B., Anal. Chim.Acta, 1991, 250,203. Kulys, J. J., Biosensors, 1986, 3, 3. Kulys, J. J., Pesliakiene, M. V., and Samalius, A. S., Bioelectrochem. Bioenerg., 1981, 8, 81. Kulys, J. J., Laurinavicius, V.-S. A., Pesliakiene, M. V., and Gureviciene, V. V., Anal. Chim. Acta, 1983, 148, 13. Yao, T., Sato, M., Kobayashi, Y., and Wasa, T., Anal. Chim. Acta, 1984, 165,291. SBnchez, P. D., Blanco, P. T., Alvarez, J. M. F., Smyth, M. R., and O’Kennedy, R., Electroanalysis, 1990, 2,303. Wang, J., Wu, L.-H., and Angnes, L., Anal. Chem., 1991,63, 2993. Frew, J. E., Harmer, M. A., Hill, H. A. O., and Libor, S. I., J. Electroanal. Chem., 1986, 201, 1. Yaropolov, A. I., Malovik, V., Varfolomeev, S. D., and Berezin, I. V., Dokl. Akad. Nauk SSSR, 1979,249, 1399. Paddock, R. M., and Bowden, E. F., J.Electroanal. Chem., 1989,260, 487. Jonsson, G., and Gorton, L., Electroanalysis, 1989, 1,465. Kulys, J., and Schmid, R. D., Bioelectrochem. Bioenerg.? 1990, 24, 305. Wollenberger, U., Bogdanovskaya, V., Bobrin, S., Scheller, F., and Tarasevich, M., Anal. Lett., 1990, 23, 1795. Gorton, L., Bardheim, M., Bremle, G., Csoregi, E., Persson, B., and Pettersson, G., in Flow Injection Analysis (HA)-Based on Enzymes or Antibodies, ed. Schmid, R. D., Gesellschaft fur Biotechnologische Forschung (GBF), Monographs, VCH, Weinheim, vol. 14, 1991, p. 305. Gorton, L., Bremle, G., Csoregi, E., Persson, B., and Jonsson-Pettersson, G., Anal. Chim. Acta, 1991, 249, 43. Csoregi, E., Jonsson-Pettersson, G., and L. Gorton, J. Biotech- nol., submitted for publication. Tatsuma, T., and Watanabe, T., Anal. Chem., 1991,63, 1580. Wollenberger, U., Wang, J., Ozsoz, M., Gonzalez-Romero, E., and Scheller, F., Bioelectrochem. Bioenerg., 1991,26,287. Jonsson-Pettersson, G., Electroanalysis, 1991, 3, 741. Dundford, H. B., and Stillman, J. S., Coord. Chem. Rev., 1976, 19, 187. Welinder, K. G., Eur. J. Biochem., 1979, 96, 483. Hayashi, Y., and Yamazaki, I., J. Biol. Chem., 1979,254,9101. Scheurs, J., van den Berg, J., Wonders, A., and Barendrecht, E., Recl. Trav. Chim. Pays-Bas, 1984, 103,251. Tse, D. C.-S., and Kuwana, T., Anal. Chem., 1978, 50, 1315. Staskeviciene, S. L., Cenas, N. K., and Kulys, J. J., Anal. Chim. Acta, 1991,243, 167.ANALYST, AUGUST 1992, VOL. 117 1241 29 30 31 32 33 34 35 36 37 Appelqvist, R., Marko-Varga, G., Gorton, L., Torstensson, A., and Johansson, G., Anal. Chim. Acta, 1985, 169,237. Gorton, L., Karan, H. I., Hale, P. D., Inagaki, T., Okamoto, Y., and Skotheim, T. A., Anal. Chim. Acta, 1990,228,23. Gorton, L., Scheller, F., and Johansson, G., Studia Biophys., 1985, 109, 199. 40 Jonsson, G., and Gorton, L., Anal. Lett., 1987, 20,839. Shinohara, H., Chiba, T., and Aizawa, M., Sens. Actuators, 41 1988, 13, 79. Aizawa, M., Anal. Chim. Acta, 1991, 250, 249. 42 Bartlett, P. N., and Whitaker, R. G., Biosensors, 1987/88, 3, 359. Malitesta, C., Palmisano, F., Torsi, L., and Zambonin, P. G., Anal. Chem., 1990,62,2735. Sasso, S . V., Pierce, R. J., Walla, R., and Yacynych, A. M., Anal. Chem., 1990, 62, 1111. 38 39 Geise, R. J., Adams, J. M., Barone, N. J., and Yacynych, A. M., Biosens. Bioelectron., 1991, 6, 151. Dominguez, E., Lan, H. L., Okamoto, Y., Hale, P. D., Skotheim, T. A., Hahn-Hagerdal, B., and Gorton, L., Bio- probe, in the press. Bahulekar, R., Ayyangar, N. R., and Ponrathnam, S., Enzyme Microb. Technol., 1991, 13, 858. Dominguez, E., Marko-Varga, G., Carlsson, M., and Gorton, L., J. Pharm. Biomed. Anal., 1990,8,825. Electroanalysis, 1990, 2, Special Issue (Microelectrodes). Paper 21013270 Received March 12, 1992 Accepted April 24, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701235

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, AUGUST 1992, VOL. 117 1243 Novel Conducting Polymers Incorporating Covalently Bound Metal-Tetraazamacrocycle Complexes* Geoff King, Simon J. Higginst and Nikki Price Department of Chemistry, Donnan and Robert Robinson Laboratories, University of Liverpool, P. 0. Box 147, Liverpool L69 3BX, UK The syntheses of a tetraazamacrocycle functionalized at nitrogen with a pendant 3-thiophene, 1 -[2-(3- thienyl)ethyl]-l,4,8,1 I-tetraazacyclotetradecane (cyclamN-CH2CH2-thiophene), and its complexes with Nil1 and Cull, are described. Electro-oxidation of the pure Nil1 complex in acetonitrile electrolytes afforded only soluble oligomers. Novel conducting polymer-modified electrodes were fabricated, however, by electro- copolymerization of the Nil' complex with 3-methylthiophene. These are unusual in that both the metal centre [Ni1l-NiI1l at +I .I 5 V (saturated calomel electrode)] and the conducting polymer backbone show stable, reversible redox behaviour, with no apparent oxidative degradation of the poly(thiophene) backbone. The Ni1I-Ni1l1 redox behaviour appears unaffected by incorporation into the conducting polymer matrix. However, the NilI-Nil wave, observed at -1.40 V in solution, is entirely suppressed in the polymer, which at that potential is an electronic insulator. Keywords: Poly( thiophene); metal complex modified electrode; cyclam; conducting polymer The implications of electrode modification for analytical applications are now well recognized.' We are interested in the modification of electrodes with various types of metal complexes, both electrocatalytic and catalytic.A particularly useful method of electrode modification is the electropoly- merization of a suitable monomer.2J In principle, precise control of the amount of metal complex deposited per unit area is possible, via control of the charge passed in the electropolymerization. The technique allows individual coat- ing of single electrodes in microelectrode arrays.4 It might also allow the sequential deposition of different polymer layers on the same electrode.5 These considerations could be important in device fabrication.4 Broadly, two approaches to electrochemically generated polymer-modified electrodes, incorporating covalently bound metal complexes, have been used. In the first, reductive polymerization of a metal complex, incorporating a vinyl- substituted bipyridine or pyridine ligand, results in a coating of a redox-conducting polymer on the electrode.In order to obtain a reproducible result, the metal complex concerned must be stable at the cathodic potentials used; polymerization occurs as a result of the ligands being reduced to reactive radical anions.3 In the second approach, oxidative poly- merization of a metal complex, incorporating a suitable ligand functionalized with a pendant thiophene or pyrrole moiety, is used [Fig. l(a)]."10 A poly(heterocyc1e) coating on the electrode is obtained. In this instance, the metal complex must be stable at the anodic potential required to oxidize the heterocycle to the radical cation, which is the key intermediate in the subsequent polymerization. In principle, the latter approach has the advantage that in their oxidized form, simple poly(heterocyc1e) films are elec- tronically conducting.If the functionalized polymers were also conducting, this might be valuable in electrocatalytic applica- tions (eg., in electroanalysis) .I1 Additionally, the recent development of a wide range of other functionalized hetero- cycles for electropolymerization should permit copolymeriza- tion to yield polymers with optimized conductivity, hydrophi- licity and electrocatalytic activity. Some attempts at this, using pyrrole-functionalized metal complexes and simple hetero- cycles (thiophene, pyrrole, bithiophene and 3-methylthio- phene), have already been described.11-'3 * Presented at the meeting on Analytical Applications of Chemi- f- To whom correspondence should be addressed.cally Modified Electrodes, Bristol, UK, January 7-8, 1992. To date, most of the work in this field has been performed with ligands of the type shown in Fig. l(a), as the syntheses of these are fairly straightforward. However, homopolymers prepared from metal complexes of these ligands display poor electronic conductivity.10 Also, the electroactivity of the conducting polymer backbone is often adversely affected on repeated electrochemical cycling.8 For some systems, the redox wave(s) due to the metal complex are virtually unchanged in spite of this; presumably, in these instances, the modified electrode material has sufficient redox conductivity to allow the metal centres to communicate with the electrode even after oxidative degradation of the poly(heterocyc1e) backbone.14 More recently, attempts have been made to overcome the poor conductivity found with N-functionalized poly(pyrro1e)- based materials by using 3-functionalized pyrroles or thio- l a R = (CH2InNC4H4 R = CH3 Ib Fig. 1 (a) Pyridine and 2,2'-bipyridine ligands functionalized with pyrrole moieties; ( b ) l-[2-(3-thienyl)ethyl]-1,4,8,11-tetraazacyclo- tetradecane (cyclamN-CH2CH2-thiophene)1244 ANALYST, AUGUST 1992, VOL. 117 phenes. It is known that, provided that a 'spacer' of suitable length separates a bulky functional group from the hetero- cycle, poly(3-functionalized heterocycle) materials, which display high electronic conductivity and good polymer proper- ties, can be synthesized.15 Few materials of this type incorpor- ating metal complexes have as yet been prepared, but those that have'6.17 show interesting properties.Metal complexes of tetraazamacrocyclic ligands are known to act as electrocatalysts for many significant redox reactions, including carbon dioxide reduction,l8 alkene epoxidation'g and reduction of alkyl halides.20 We wished to prepare conducting polymer-modified electrodes containing covalently bound tetraazamacrocyclic ligands and their com- plexes, but at the same time to preserve the electronic conductivity of the conducting polymer. Therefore, it was decided to use 3-functionalized thiophene chemistry17 and the ligand shown in Fig. l(b). During the course of this work, a preliminary report of similar work appeared 2 1 Our results differ from those described in that report in several respects, and we report these results here. Experimental Apparatus Infrared (IR) spectra were measured as Nujol mulls, with use of a Perkin-Elmer 1720X Fourier transform (FT) spec- trometer (Norwalk, CT, USA), from 400 to 4500 cm-1.Electronic absorption spectra were recorded on solutions, by using a Perkin-Elmer 330 spectrophotometer, from 12 500 to 40000 cm-1. Proton nuclear magnetic resonance (NMR) spectra were recorded with use of a Bruker WM 250 MHz FT spectrometer (Billeria, MA, USA). Samples were dissolved in CDCI3. Tetramethylsilane was used as reference. Mass spectra were recorded with a VG 7070E instrument (Stam- ford, CT, USA) operated in the electron-impact (70 eV) mode (intermediates and free ligands) or positive-ion fast-atom bombardment (FAB) mode (complexes); 3-nitrobenzyl al- cohol was used as solvent, and xenon as bombardment gas in the latter instance.Cyclic voltammetry experiments were performed with laboratory-built potentiostat and a conven- tional three-electrode cell configuration. Working electrodes were polished platinum discs, counter electrodes were plati- num gauzes and the reference electrode was a commercial saturated calomel electrode (SCE) separated from the work- ing compartment by a glass frit and a Luggin capillary. All potentials quoted are referred to the SCE. The ferrocene- ferricinium couple was examined routinely to check the stability of the reference electrode and junction potentials. Chemicals Acetonitrile (BDH, Poole, Dorset, UK; HPLC grade) was dried by reflux over, and distillation from, CaH2.Tetraethyl- ammonium tetrafluoroborate (TEAT) was prepared from the bromide (Lancaster Synthesis, Lancaster, UK) and 50% HBF4 (Fluka, Buchs, Switzerland) in water, and recrystallized three times from hot ethanol. It was dried for 48 h at 0.1 mmHg. The 1,4,8,11-tetraaq1cyclotetradecane (cyclam) was prepared by a literature method.22 p-Toluenesulfonyl chloride (Aldrich, Milwaukee, WI, USA) was purified by a literature method.23 All other chemicals were used as received. Syntheses 2-(3- Thieny1)ethyl-p-toluenesulfonate p-Toluenesulfonyl chloride (1.86 g, 9.76 mmol) was dissolved in anhydrous pyridine (3.0 cm3) and left to stand for 15 min at 0 "C. 2-(3-Thienyl)ethanol (1.00 g, 7.80 mmol) was added dropwise, with stirring, at 0 "C. Stirring was continued for 2 h.Ice (10 g) was added and the product was extracted into dichloromethane (3 X 10 cm3). The extracts were washed with HCI (2 mol dm-3; 3 x 10 cm3), then water (3 x 10 cm3) and dried over Na2S04. The solvent was removed (rotary evapora- tor) to leave a white, crystalline solid. Yield: 1.835 g, 83.4%. Proton NMR spectroscopic data [selected 1H NMR data: 6 4.20 (t, 2 H, S020CH2CH2C4H3S)] showed this to be sufficiently pure for synthetic use. A sample was recrystallized from diethyl ether-light petroleum (boiling-range 60-80 "C) for microanalysis and electrochemical experiments. (Found: C, 55.00; H, 4.95. Calc. for Cl3HI4O3S2: C, 55.32; H, 4.96% .) 2.43 (s, 3 H, CH3C6H4-), 3.05 (t, 2 H, -S020CH2CH2C4H3S), 1 - [ 2-(3- Thienyl)ethyl]- 1,4 ,8,11-tetraazacyclotetradecane (lb) ( cyclamN-CH2CH2-thiophene) To a solution of cyclam (7.4 g, 37.0 mmol) in acetonitrile (350 cm3) and pyridine (150 cm3) was added dropwise 2-(3-thienyl)- ethyl-p-toluenesulfonate (4.76 g, 16.86 mmol), with stirring, under reflux.The mixture was heated under reflux for a further 4 h, the solvent was removed under reduced pressure, and the residue was dissolved in 0.2 mol dm-3 NaOH (100 cm3). The solution was extracted with diethyl ether (5 x 100 cm3). (The excess of cyclam could be recovered from the aqueous phase by extraction into dichloromethane.) The organic phase was dried over Na2S04, and the solvent was removed under reduced pressure, leaving a waxy oil. This was dried in vacuo (0.1 mmHg) for 48 h.Proton NMR spec- troscopy: 6 1.70 (complex m, 4 H, -NHCH2CH2CH2NH-), 2.60 (complex m, 23 H), 6.88 (m) and 7.00 (d of d, total 3 H, -C4RH3S-3). Mass spectrometry [mlz 200 (9), 310 (loo), 420 (S)] indicated the presence of small amounts of unreacted cyclam and difunctionalized cyclam. Attempts to purify the crude material by chromatography proved unsatisfactory, but the material was sufficiently pure to allow complexes to be prepared. { 1-[2-(3- Thienyl)ethyl]-l,4,8,ll-tetraazacyclotetradecane} - nickel(i1) perchlorate { [ Ni(cyclamN-CH2CH2-thiophene)] The ligand (0.70 g, 2.26 mmol) dissolved in methanol (20 cm3) was added dropwise to a stirred, hot solution of [Ni(H20)6]- [C1O4I2 (0.91 g, 2.48 mmol) in methanol (60 cm3).A small amount of material precipitated from the orange solution. After 15 min, the solution was filtered hot, and allowed to cool. An orange precipitate appeared, which was filtered off and dried in vacuo. Yield 0.80 g. (Found: C, 33.60; H, 5.33; N, 9.89. Calc. for C16H30C12N4NiOsS: C, 33.83; H, 5.32; N, 9.86% .) FAB-MS: Nil1 complexes of unfunctionalized cyclam {mlz 256, [Ni(L-H)]+}, monofunctionalized cyclam (467, { [Ni(L)]C104}+; 367, [Ni(L-H)]+) and difunctionalized cyclam (577, {[Ni(L)C104} +; 477, [Ni(L-H)]+). The complex was recrystallized three times from hot methanol-ethanol, whereupon FAB-MS data indicated that it was pure. Obser- ved and calculated isotope patterns for the peak at mlz 467 were in good agreement. Yield 0.62 g; 48%. lO-3E,,,l~rn-~ (~,,,/dm-3 mol-1 cm-1) (CH3CN), 22.03 (21).[C10412} [ l-{2-(3-Thienyl)ethyl}-l,4,8,1l-tetraazacyclotetradecane]- copper(i1) perchlorate [ Cu(cyc1amN-CH2CH2-thiophene)]- [CEO412 This was prepared in the same way as the Ni*' complex, as a purple-red solid. Yield 63%. (Found: C, 33.60; H, 5.33; N, 9.89. Calc. for C16H30C12C~N408S: C, 33.83; H, 5.32; N, 9.86% .) 10-3 E,,,lcm-1 (~,,,/dm-3 mol-1 cm-1) (CH3CN), 14.66 (240); 18.68 sh (96). Electro-copolymerization Electrochemical polymerization of 2-(3-thieny1)ethyl-p- toluenesulfonate A 0.015 mol dm-3 solution of the recrystallized tosyl ester in 0.2 mol dm-3 Et4NBF4-CH3CN was used. The workingANALYST, AUGUST 1992, VOL. 117 1245 electrode had an area of 0.85 cm2. After assembly of the cell, freshly activated molecular sieve (4A) was added to the electrolyte, and the solution was de-oxygenated by bubbling 02-free N2.Cycling from 0.25 to +1.95 V (SCE) at 0.05 V s-1 produced a dark-blue adherent film. Electrochemical polymerization of { 1-[ 2-(3-thienyl)- ethyl]-1,4,8,11-tetraazacyclotetradecane}nickel( I I ) perchlorate The conditions used were as for the above experiment, except that the electrolyte contained 0.03 rnol dm-3 3-methylthio- phene and 0.03 rnol dm-3 NilL complex. Cycling from 0.00 to +2.00 V at 0.10 V s-1 produced an adherent electrochromic film, although formation of some soluble material was also evident. Results and Discussion Commercially available 2-(3-thienyl)ethanol is a valuable starting material for 3-functionalized thiophenes.15 Interest- ingly, it has been found that, under rigorously anhydrous conditions (vacuum oven-dried 4A molecular sieves present in the cell), the tosyl ester can be electropolymerized to a conducting polymer with electrochemical characteristics typical of poly(3-al kyl thiophene)~.15 Electropolymerization was carried out by cyclic voltammetry between +0.25 and +1.91 V (SCE) in 0.2 rnol dm-3 TEAT in CH3CN. Conducting polymers of this type could be useful for subsequent functionalization; a patent report mentions the electrosynthesis of a similar polymer and its subsequent reaction with sodium iodide in acetone.24 We have synthesized 1-[2-(3-thienyl)ethy1]-1,4,8,11-tetra- azacyclotetradecane (lb) (cyclamN-CH2CH2-thiophene) from cyclam and the tosyl ester of 2-(3-thienyl)ethanol. Square planar Nil1 and Cu" complexes of the ligand were prepared by using the metal perchlorates.Although these complexes had the correct microanalyses (C, H and N; see under Experimental), FAB-MS data suggested that this was fortuitous; small and variable amounts of complexes of unfunctionalized cyclam and difunctionalized cyclam were invariably present. It was found to be more convenient to purify the complexes, by recrystallization from hot methanol- ethanol, than to purify the free ligand. Electronic spectro- scopy for the pure complexes showed a slight diminution in the ligand field on functionalizing the ligand at nitrogen. Hence, for the Nil1 complex, the broad band due to the overlapping transitions to dx2-y2 is at 21 370 cm-1. In [Ni(cyclam)]2+, this occurs at 22030 cm-1.Tertiary amine donors are weaker field ligands than secondary amine donors.25 This affects the redox potentials for the NiI-Ni" and Ni"-Ni"l couples. For [Ni(cy- clam)]*+ in CH3CN, these occur at -1.46 and +1.08 V, respectively. For [Ni( cyclamN-CH2CH2-thiophene)]2+ , they occur at - 1.40 and + 1.15 V, respectively. This behaviour has been observed previously for N-functionalized macrocycles .26 Attempts were made to prepare modified electrodes by electro-oxidation of the thiophene moieties in cyclic voltam- metry experiments by using the pure (thrice recrystallized) Nil1 complex. When using 0.2 rnol dm-3 TEAT-CH3CN, a quasi-reversible wave at + 1.15 V was observed, correspond- ing to the Nill-Nilll process, and at more positive potentials, an irreversible oxidation (onset approximately + 1.85 V).Anodic of the latter potential, intensely blue material was observed streaming away from the working electrode. Changes in the conditions (electrolyte concentration 0.01-0.35 rnol dm-3; monomer concentration 0.001-0.045 rnol dm-3; anodic poten- tial limit +1.85 to +2.10 V) did not alter the observed behaviour; no polymer was deposited on the working elec- trode. Attempts were therefore made to copolymerize the [ Ni(cyc1amN-CH2CH2- thiophene)]'+ with 3-me thy1 thiophene in the hope that the solubility of the oligomers evidently being generated at > + 1.85 V would be reduced. This strategy was successful: adherent, stable films were Potential - Fig. 2 Cyclic voltammograms of poly(3-methylthio hene)- copoly { [ (cyclamN-CH2CH2-thiophene)nickel( 11'3 perchPorate)- modified platinum disc electrode in 0.2 mol dm- TEAT-CH3CN at: A, 0.01; B, 0.02; and C, 0.05 V s-l, respectively.Scan range, 0.00 to +1.40 V (SCE) produced. The modified electrodes were removed from the growth solution in their reduced state, washed in pure CH3CN and stored in dry air. Subsequent cyclic voltammetry in 0.2 rnol dm-3 TEAT-CH3CN (Fig. 2) showed several features of interest. First, the conducting polymer backbone is electro- active, although its redox wave is of an unusual type for a 3-functionalized thiophene .15 In particular, two distinct processes are evident. Second, by extrapolating the wave for poly(3-R-thiophene) and calculating the charge due to this process and that due to Nill-NitlL, and assuming that the oxidation of the thiophene moieties corresponds to one electron removed per four rings,15 the composition of the copolymer is approximately 3-methylthiophene-[Ni(cy- clamN-CH2CH2-thiophene)]z+ (10 : 1). Third, repeated cycling from 0.00 to +1.40 V has a negligible effect on either the poly(thiophene) or the NiI1-NiI1l process.Fourth, the modified electrode is highly electrochromic, being orange-red when neutral, green after the poly(thiophene) oxidation wave, but before that of NilL-Ni[ll, and intensely blue after the NiII-Ni"' wave. Finally, the Ni1I-Ni1I1 wave Ed is identical, within experimental error, for the process in solution and that in the polymer. This is particularly significant; on oxidation to N P , the process: [Ni(L)]2+ - e- + 2 CH3CN + [Ni(L)(CH,CN)#+ is occurring.Clearly, for the potential to be unchanged within the polymer, there must be sufficient solvent within the film to permit this process to occur. In complexes of this type, it is known that E; for Nill-Nill' is a sensitive function of the species available as axial co-ligands.26 Ellipsometric data for poly- (thiophene)27 suggested that little solvent was present in the as-grown film. Presumably, the copolymer films prepared here contain more solvent as a result of the presence of highly charged [Ni(L)(CH3CN)#+ moieties within the growing film. The scan range was extended to negative potentials. Interestingly, the Ni"-Nil couple, observed in solution at -1.40 V, was not seen for the polymer-trapped complex. Instead, an irreversible reduction (onset potential approxi- mately -1.7 V) occurred.On the reverse scan, an additional oxidation then occurred (Epa approximately +0.83 V), and the Ni"-Ni"' wave was greatly diminished. If the potential was then kept at 0.00 V for 10 min, and another cycle from 0.00 to rt 1.42 V was then recorded, the poly(thiophene) electro- chemistry was almost unchanged, but the charge under the Ni"-Ni"' wave was diminished by 75%.1246 ANALYST, AUGUST 1992, VOL. 117 The irreversible reduction wave is tentatively assigned to partial reduction of the poly(thiophene) to its anionic conduct- ing form. Clearly, in our film there is insufficient redox conductivity to allow communication of the Nil1 complex with the electrode when the poly(thiophene) is neutral (electronic- ally insulating).This contrasts with observations made for poly(pyrro1e) films N-functionalized by bipyridine com- plexes.7-10 Once the poly(thiophene) film becomes partially conducting again at < - 1.7 V, the Ni" complex is presumably reduced to Nil and cannot be re-oxidized again until > +0.85 V. This would account for the decrease in the Ni"-NilI1 charge; Nil-tetraazamacrocycle complexes are labile, and on the voltammetric time scale, significant decomposition of the Nil complex could occur. Currently, in situ electronic spectros- copy and FTIR studies are in progress to investigate this further. While this work was in progress, a preliminary report appeared describing a homopolymer of [Ni(cyclamN- CH2CH2-thiophene)]2+ and its electrochemistry.21 The poly- mer was prepared with use of 0.05 mol dm-3 solutions in 0.1 mol dm-3 Bu4NBF4-CH3CN by cycling from -0.5 to + 2.0 V (versus ferrocene-ferricinium).In contrast to our copolymer, the poly(thiophene) showed no electroactivity, although the NilI-Nilll process was well defined. Attempts to duplicate this result with the recrystallized Nil1 complex failed: only soluble oligomers were generated. However, on using material that was recrystallized only once,21 variable results were obtained; in one instance, an orange polymer film had grown. This had the same electro- chemical characteristics as those described elsewhere.21 Interestingly, this polymer was not noticeably electrochromic. The FAB-MS data on the monomer used in this experiment showed the presence of the complex of a difunctionalized ligand.It is commonly observed that complexes containing more than one monomer unit are more readily electropoly- merized.11-13,28 We tentatively suggest that the presence of this impurity might assist the formation of a polymer-modified electrode; attempts to characterize this process further are continuing. Attempts at mediated electron transfer with the copolymer- modified electrode are in progress, as are attempts to obtain polymers incorporating the Cull complex. The methylation of the remaining amine donors in 1-[2-(3-thienyl)ethyl]-l,4,8,11- tetraazacyclotetradecane is being examined with a view to the preparation of Ru" complexes, and of modified electrodes for alkene epoxidation. 19 We thank Alan Mills for the FAB-MS measurements.S. J. H. is grateful to the Nuffield Foundation for a grant under the Awards for Newly-Appointed Science Lecturers scheme. References 1 Chemically-Modified Surfaces in Catalysis and Electrocatalysis, ed. Miller, J. S., American Chemical Society, Washington, DC, 1982. 2 3 4 5 6 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Handbook of Conducting Polymers, ed. Skotheim, T. A., Marcel Dekker, New York, 1986. Elliott, C. M., Baldy, C. J., Nuwaysir. L. M., and Wilkins, C. L., Inorg. Chem., 1990,29,389. Thackeray, J. W., White, H. S., and Wrighton, M. S.. J. Phys. Chem., 89, 5133. Iyoda, T., Toyoda, H., Fujitsuka, M., Nakahara, R., Tsuchiya, H., Honda, K., and Shimidzu, T., J. Phys. Chem., 1991, 95, 5215. Collin, J. P., Jouaiti, A., and Sauvage, J.P., J. Electroanal. Chem. Interfacial Electrochem., 1990, 286, 75. De Oliveira, 1. M. F., Moutet, J. C., and Vlachopoulos, N., J. Electroanal. Chem. Interfacial Electrochem., 1990, 291, 243. Daire, F., Bedioui, F., Devynck, J., and Bied-Charreton, C., J. Electroanal. Chem. Interfacial Electrochem., 1987, 224, 95. Cosnier, S., Deronzier, A., and Moutet, J. C., J. Electroanal. Chem. Interfacial Electrochem., 1986,207, 315. Cosnier, S., Deronzier, A., and Roland, J. F., J. Electroanal. Chem. Interfacial Electrochem., 1991, 310, 71. Ochmanska, J., and Pickup, P. G., J. Electroanal. Chem. Interfacial Electrochem., 1991,297, 211. Ochmanska, J., and Pickup, P. G., J. Electroanal. Chem. Interfacial Electrochem., 1991, 297, 197. Ochmanska, J., and Pickup, P. G., Can. J. Chem.. 1990, 69, 653. Eaves, J. G., Munro, H. S., and Parker, D., Inorg. Chem., 1987,26, 644. Roncali, J . , Youssoufi, H. K., Garreau, R., Garnier, F., and Lemaire, M., J. Chem. SOC., Chem. Commun., 1990,414. Inagaki, T., Hunter, M., Yang, X. Q., Skotheim, T. A., and Okamoto, Y., J. Chem. SOC., Chem. Commun., 1988, 126. Mirrazaei, R . , Parker, D., and Munro, H. S., Synth. Met., 1989, 30, 265. Fujihara, M., Hirata, Y., and Suga, K., J. Electroanal. Chem. Interfacial Electrochem., 1990, 292, 199. Che, C. M., Tang, W. T., Wong, W. T., and Lai, T. F., J. Am. Chem. SOC., 1989, 111,9048. Bakac, A., and Espenson, J. H., J. Am. Chem. SOC., 1986,108, 712. Sable, E., Handel, H . , and L'Her, M., Electrochim. Acra, 1991, 36, 15. Barefield, E. K., and Freeman, G., Inorg. Synth., 1980,20,108. Vogel's Textbook of Practical Organic Chemistry, eds. Furniss, B. S., Hannaford, A. J., Rogers, V., Smith, P. W. G., and Tatchell, A. R.. Longman, London, 4th edn., 1978, p. 317. Wudl, F., and Heeger, A. J., PCT Int. Appl. WO 87 05,914; Chem. Abstr., 1988, 109, 38478~. Wagner, F., and Barefield, E. K., Inorg. Chem., 1976,15,408. De Santis, G., Di Casa, M., Mariani, M., Seghi, B.. and Fabbrizzi, L., J. Am. Chem. SOC., 1989, 111,2422. Hamnett, A., and Hillman, A. R., J. Electrochem. Soc., 1988, 135,2517. Beer, P. D., Kocian, O., and Mortimer, R. J., J. Chem. SOC., Dalton Trans., 1990, 3283. Paper 21005 77H Received February 3, 1992 Accepted February 17, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701243

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, AUGUST 1992, VOL. 117 1247 Surface Modification With Macrocycle-containing Redox-active Polymers: Towards the Design of Novel Spectroelectrochemical Group IA/IIA Metal Cation Sensors* Paul D. Beer Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, Oxford, OX I 3QR, UK Oldrich Kocian Department of Chemistry, University of Birmingham, P.O. Box 363, Birmingham, B15 ZTT, UK Roger J. Mortimer and Christopher Ridgway Department of Chemistry, L oug h boro ug h University of Tech nolog y, L oug h boroug h, L eicesters hire LEI1 3TU, UK The application of transition metal-polypyridyl complexes to chemical sensor technology is demonstrated. Using fluorescence emission spectroscopy the recognition of Na+ and Mg2+ metal cations by vinyl linked benzo- and aza- crown ether-bipyridyl ruthenium(ii) complexes is shown.Such complexes can be electropolymerized onto platinum and optically transparent conducting glass electrodes. Although electrochemical recognition by such modified electrodes is not observed, such systems show promise as novel spectrochemical sensors. Keywords: Modified electrode; sensor; spectrochemical; ruthenium tris-bip yridyl; crown ether Since their discovery 25 years ago1 crown ethers,2 as a new generation of complexing agents, have found many applica- tions in analytical chemistry.3 It is well known that crown ethers can bind Group IA and IIA metal cations, selectivity being determined by the compatibility of the cation radius and the size of the crown ether macrocyclic cavity. This effect has been exploited in the construction of ion-selective elec- trodes4.5 and recently in the development of selective electro- chemical recognition systems697 based on, for example, ferrocenyl-substituted crown ethers.Our aim is to extend such electrochemical recognition to the development of surface- modified electrode sensor systems. Variation of redox poten- tial with metal cation identity would allow development of such systems as amperometric sensors for the non-electro- active Group IA and IIA metal cations in flow injection analysis .* The surface modification approach described here utilizes the electropolymerization technique pioneered by Abruna et aZ.9 for metal complexes of vinyl-substituted bipyridyl (bipy) ligands. Experimental Synthesis The new trans vinyl-linked benzo-crown ether-, aza-crown ether- and bismethoxyphenyl-bipyridyl ligands L1 and L2 were prepared10 giving excellent yields (70-95%) via mono- or dilithiation of 4,4'-dimethyl-2,2'-bipyridine, addition of the appropriate 4-formyl substituted compound to give the alcohols L3 and L4, followed by dehydration.The structures of all these new compounds were characterized on the basis of spectroscopic and analytical evidence. The tris(1igand) ruthenium( 11) complexes [ RuL1-2,4a,b3] [PI?&, [R~(bipy)~][PF~]~ and the corresponding monoligand ruthenium(I1) complexes [ R~L~-~v~~*~(bipy)~][PF& were obtained by refluxing the appropriate ligand in dimethyl- formamide with RuC13.3H20 and [RuCl2(bipy)2]-2H20, respectively, followed by purification on Sephadex LH-20 and precipitation of the complexes on addition of ammonium hexafluorophosphate.10 * Presented at the Meeting on Analytical Applications of Chemi- cally Modified Electrodes, Bristol, UK, January 7-8. 1992. N T'R N T R L1a-C L2S-c L3" L4- c02 0 > a; R = * o d o J ,OMe c; R = *OM(?1248 ANALYST, AUGUST 1992, VOL. 117 Electrochemistry Electrochemical studies were performed using an EG & G Princeton Applied Research Model 273 potentiostat. A three-electrode system was employed with platinum flag (1 cm2 surface area) or indium-tin oxide optically transparent glass (30 G? 'per square' from Balzers, cut to a 9 X 50 mm surface area) working electrodes. The reference electrode [a Radiometer sodium chloride saturated calomel electrode (SSCE)] and counter electrode (platinum-mesh) were each separated from the working electrode compartment of the electrochemical cell by glass frits.Measurements were carried out in de-oxygenated acetonitrile (freshly distilled over calcium hydride) solutions containing a 0.1 mol dm-3 solution of the supporting electrolyte. Fluorescence Emission and Visible Absorption Spectrometry A Perkin-Elmer Model 3000 fluorescence spectrophotometer controlled by an Elonex 386 computer was used for recording fluorescence emission spectra. An interference filter provided a 450 nm excitation wavelength, with an absorption filter being used to prevent radiation below 490 nm from reaching the emission monochromator. The Perkin-Elmer front surface accessory (5212-3130) was used for measurements on the polymer films.A Hewlett-Packard HP 8451A diode-array spectro- photometer controlled by a Vectra QS/16S computer was employed for recording the visible absorption spectra. All measurements were conducted at 25 "C using a 1 x 1 cm rectangular quartz cuvette and de-oxygenated solutions. Results and Discussion Electropolymerization Following the earlier precedent,g it was anticipated that sequential potential scanning of solutions of the [Ru- L1-23][PF6]2 and [R~L~-~(bipy)~][pF~]~ complexes to the series of ligand-centred reductions would activate the vinylic lin- kages and initiate electropolymerization. All [RuLI3][PF6l2 complexes were indeed electropolymerized onto platinum electrodes to form smooth, adherent orange films, as exempli- fied by Fig. 1, which shows the steady increase in current attributable to the combined electroactivity of the polymeric film and that of the inward-diffusing complex.Complexes [ RuL23][ PF& were also electropolymerized although less efficiently (lower rate of current increase on sequential scanning), owing to steric crowding of the 4,4' substituents, and complexes [ R~L~-~(bipy)2][PF6]2 only exhibited solution redox processes analogous to the prototype [R~(bipy)3][PF6]~. 0 -0.4 -0.8 -1.2 -1.6 E N versus SSCE Fig. 1 Sequential cyclic voltammo rams for an acetonitrile solution containing 1.33 mmol dm-3 [RULla3f[PF& in 0.1 mol dm-3 Bu4NBF4 at 100 mV s-1; arrows indicate current increase Electrochemical Recognition Analysis of heteropolymetallic ruthenium(I1)-sodium com- plexes (isolated from the reaction of crown ether-containing ruthenium(I1) complexes and sodium hexafluorophosphate) and 13C nuclear magnetic resonance (NMR) titration studies have established10 that each crown ether moiety in the plexes is able to bind one metal cation.In order to discover whether the polymer films were electrochemically responsive to such binding, cyclic voltammograms of the modified electrodes were recorded in supporting electrolytes with a variety of cations. For illustration, a set of cyclic voltammo- grams for a poly[RuL1a3]2+ modified electrode in an aceto- nitrile solution containing 0.1 rnol dm-3 tetrabutylammonium tetrafluoroborate is shown in Fig. 2. The dependence of peak currents on scan rates was indicative of a combination of diffusional (peak current proportional to square root of scan rate) and surface control (peak current proportional to scan rate) for the RU'~"" redox response.Table 1 shows that Ru""" redox potentials, Epoly, for benzo-crown ether-containing poly[ RuL1a3]2+ and bismethoxyphenyl-containing poly- [RuL1c3]2+ modified electrodes are the same for a given supporting electrolyte , indicating that crown ether binding of metal cations in the former does not perturb the redox response of the Ru1I1"I wave. Unfortunately, although more likely to show electrochemical recognition due to the prox- imity of the crown ether macrocycle binding sites, it was not possible to monitor the ligand-based redox responses in the metal cation supporting electrolytes owing to low solvation energies of Group IA and IIA metal cations in acetonitrile, resulting in electroreduction to the metals at relatively positive potentials.11.12 [ R u L ~ ~ , ~ ~ ~ , ~ ~ ] [ P F ~ ] ~ and [ R u L ~ ~ , ~ ~ ~ , ~ (bipy) 21 fPF61 2 corn- Fluorescence Emission Spectrometry Analysis by fluorescence emission spectrometry has the advantage of high sensitivity, which allows the measurement of low analyte concentrations.That [R~(bipy)~][PF6]~ and related systems exhibit metal-to-ligand charge transfer (MLCT) emission maximal3 offers an alternative method for probing the binding of Group IA and IIA metal cations. The fluorescence emission spectra of all complexes in acetonitrile were fairly broad and featureless and of approximately the AE I 2.00 1 .oo 0 EN versus SSCE Fig. 2 Cyclic voltammograms, at A 20, B 50, C 100, D 200, and E 500 mV s-1 for a poly[RuL1a3]2+ modified electrode in a pure 0.1 mol dm-3 Bu4NBF4-acetonitrile solution after transfer from an electropolymerization solution containing 1.33 mmol dm-3 [RuLla3][PF6I2 in 0.1 rnol dm-3 Bu4NBF4-acetonitrileANALYST, AUGUST 1992, VOL.117 1249 Table 1 Ru””” redox potentials, Epoly (volts) versus SSCE, for polymer modified electrodes with variation in the supporting electrolyte. Values quoted were (f0.02) at 100 mV s-l in acetonitrile solutions containing 0.1 mol dm-3 supporting electrolyte. Redox potentials were calculated as EPoly = (& + Ep:,)/2 where E,,= = cathodic peak potential, EP,, = anodic peak potential Polymer Bu4NBF4 Bu4NC104 NaC104 Mg(C104)2 POly( RuLla3)2+ +1.20 +1.15 +1.10 +1.05 POly( RUL’C3)2+ +1.24 +1.15 +1.10 +1.07 Table 2 Metal-to-ligand charge transfer fluorescence emission wavelengths, h,,,(nm) for O.OOOO1 mol dm-3 solutions of complexes in pure acetonitrile and in acetonitrile with addition to individual solutions of excess amounts of each of the salts indicated Pure 605 606 605 672 659 670 673 696 69 1 686 Bu4NC104 605 608 606 672 659 670 673 696 69 1 686 NaC104 605 608 607 669 658 668 673 682 677 670 Mg(C104)2 605 608 607 665 657 666 673 694 689 683 same width.Since the spectra are all similar in shape and width, the emission maxima were used to assess the effect of ligand variation and metal-cation binding on excited-state energies. The emission maxima for the complexes of the alcohols [ R u L ~ ” ~ ] [ P F ~ ] ~ and [R~L~~(bbipy)~][PF~]~, shown in Table 2, are as for the prototype, [R~(bipy)~][PF&. In contrast, the presence of the vinyl linkage in the analogous complexes [ R u L ~ ~ ~ ] [ P F ~ ] ~ and [R~L~~(bipy)2][PF& causes a red shift owing to conjugation to the electron-donating benzo-crown ether.Addition of excess amounts of the salts, shown in Table 2, then provides evidence for the spectrochemical recognition of Group IA and IIA metal cations by the benzo-crown ether and aza-crown ether complexes. The MLCT emission maxima for these complexes are shifted to significantly lower wavelengths in the presence of Na+ or Mg2+. The largest shift was obtained with [R~L~~(bipy)z][PF&; addition of an excess amount of sodium ions giving a 16 nm blue shift.No corresponding shifts are observed for the bismethoxyphenyl-containing ‘model’ complex [ R u L ~ ~ ~ ] [ P F ~ ] ~ or the prototype [Ru(bipy)3][PF6]2, suggesting that metal cation coordination, causing lowering of the electron-donating strength of the crown ethers, is responsible for the effect. That the vinyl linkage is necessary for such spectrochemical recognition is shown by the data for the complexes of the alcohols [ R u L ~ ~ ~ ] [ P F ~ ] ~ and Spectrochemical recognition was also observed for polymer films prepared by the electropolymerization technique. For such measurements optically transparent conducting glass was chosen as the electrode material to allow visible absorption spectra to be recorded additionally. Poly[ RuL*a3]2+ films exhibited an MLCT absorption band (h,,, = 470 nm) corresponding to that of the monomeric precursor [ R u L ~ ~ ~ ] [ P F ~ ] ~ (h,,, = 475 nm).As for the monomeric complex, the MLCT emission maximum of the polymer was shifted to lower wavelengths in the presence of Mg*+, as shown in Fig. 3. The emission intensity also increases owing to the extra rigidity of the MLCT chromophore on metal-cation binding. Furthermore, the observation of a low wavelength [RUL4a(bipy)2] [PF612* E w o.2 I 0 1 500 550 600 650 700 750 Wavelengthhm Fig. 3 A, Fluorescence emission spectrum of a film of poly- [RuLla3I2+ coated onto an optically transparent conducting glass electrode and immersed in acetonitrile. B, Spectrum after addition of excess of Mg2+ to 10 mmol dm-3 [as Mg(C104)2-acetonitrile solution] shoulder in Fig.3, in addition to the main peak, is evidence that both saturated and vinyl linkages between the bipyridyls and the benzo-crown ethers are present in the polymer. Confirmation of spectrochemical recognition and indeed the retention of the benzo-crown ether after electropolymeriza- tion came from the visible absorption and fluorescence emission spectra of the bismethoxyphenyl-containing ‘model’ polymer. Poly[RuL2c3]2+ films displayed no shifts in the presence of Group IA and IIA metal cations. In conclusion these results have shown that the novel crown ether polymer film materials prepared by electropolymeriza- tion represent a new class of spectrochemical sensing devices for Group IA and IIA metal cations. 0. K. thanks the Science and Engineering Research Council (Molecular Sensors Initiative) for a postdoctoral research fellowship (GWE 71624). 1 2 3 4 5 6 7 8 9 10 11 12 13 References Pedersen, C. J., J. Am. Chem. SOC., 1967, 89,7017. Vogtle, F., Supramolecular Chemistry, Wiley, New York, 1991, ch. 2, pp. 27-83. Blasius, E., and Janzen, K.-P., Top. Curr. Chem., 1981, 98, 163. Covington, A. K., Grey, H., Kelly, P. M., Kinnear, K. I., and Lockhart, J. C., Analyst, 1988, 113, 895. Moody, G. J., Saad, B. B., and Thomas, J. D. R., Analyst, 1989, 114, 15. Beer, P. D., in Adv. Inorg. Chem., ed. Sykes, A. G., Academic Press, New York, 1992, vol. 39, in the press. Beer, P. D., Chem. SOC. Rev., 1989, 18,409. Thomsen, K. N., and Baldwin, R. P., Electroanalysis, 1990, 2, 263. Abrufia, H. D., Denisevich, P., Umaiia, M., Meyer, T. J., and Murray, R. W., J. Am. Chem. SOC., 1981, 103, 1. Beer, P. D., Kocian, O., Mortimer, R. J., and Ridgway, C., J. Chem. SOC., Chem. Commun., 1991, 1460. Kolthoff, I. M., and Coetzee, J. F., J. Am. Chem. SOC., 1957, 79, 870. Kolthoff, I. M., and Coetzee. J. F., J. Am. Chem. SOC., 1957, 79, 1852. Juris, A., Balzani, V., Barigelletti, F., Campagna, S., Belser, P., and von Zelewsky, A., Coord. Chem. Rev., 1988, 84,85. Paper 2/01 201 D Received March 5, 1992 Accepted April 23, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701247

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, AUGUST 1992, VOL. 117 1251 Mobile Species Uptake by Polymer-modified Electrodes* A. Robert HillmanJ David C. Loveday and Marcus J. Swann School o f Chemistry, University o f Bristol, Bristol BS8 ITS, UK Stanley Bruckenstein Department of Chemistry, State University o f New York at Buffalo, Buffalo, NY 14214, USA C. Paul Wilde Department of Chemistry, University of Ottawa, Ottawa, Ontario, Canada KIN 6N5 The use of mobile species uptake by a polymer-modified electrode as a probe of solution composition is discussed. The over-all mobile species exchange process between the polymer and solution phases is monitored gravimetrically, using the electrochemical quartz-crystal microbalance (EQCM). The influence of electroneutrality and activity constraints is considered under both equilibrium and transient conditions.The generality of detection by mass and the high sensitivity of the EQCM have analytical advantages. Selectivity requires separation of the total mass change into single species' components. The thermodynamic and kinetic approaches to the problem of selectivity are explored. Keywords: Polymer-modified electrodes; mobile species; permselectivity; quartz crystal microbalance; redox polymer Introduction The Problem The electrochemical quartz-crystal microbalance (EQCM) is presently finding considerable application as a probe of interfacial processes.1-4 In this paper its use for monitoring the exchange of mobile species between a polymer-modified electrode and its bathing solution is discussed. The key problem of deconvoluting the observed signal into its com- ponents is addressed, each being associated with a single mobile species.This problem exists as a consequence of the generality of detection by mass and the application of the EQCM to analytical problems requires its solution. Initially, the nature of the problem is illustrated by showing how a variety of phenomena alter the simple pattern of exchange predicted by electroneutrality arguments. Secondly, two methods are proposed for the separation of mobile species transfer processes. One approach involves the application of thermodynamic arguments to (pseudo-) equilibrium" data; the other exploits kinetic measurements. Then the principles of these methods are demonstrated by applying them to mobile species transfers for model systems. Background The modification of electrode surfaces with polymer films allows us to tailor chemically the properties of the electrode/ electrolyte interface.5.6 This facility has been exploited in a number of ways to give electrodes with analytical utility.Potentiometric responses have been reported for H+ using poly( 1,2-diaminobenzene) ,7 K+ using poly( acrylic acid) ,8 and SCN - using pol y [ trialk yl (vin ylbenz yl) ammonium] .8 Polymer film immobilization of suitable mediators allows ampero- metric detection of solution species. Choice of mediator is based upon the selectivity and kinetics of its reaction with the target species. The numerous cases (see refs. 5 and 6 for reviews) of amperometric detection are exemplified by the use of immobilized iridium mediators for nitrite9 and ascorbatelo detection.Selectivity and sensitivity of amperometric sensors are strongly related to partition of the target species into the polymer film. Strictly, the solution species need only interact with the film at the polymer/solution interface; however, the advantages of increasing film thickness576 are then not realized. In this paper, in situ nanogravimetric measurements are used to examine the solution/polymer phase exchange process. Insight into this process should be helpful both in the rational design of modified electrodes, and the development of alternative analytical strategies to existing potentiometric and amperometric approaches. Electroanalytical Strategy Generally, the solution composition determines the uptake by the polymer of mobile species, the target species and possible interferents.The basis of a sensor is inherent in this uptake if (i) there exists a means for monitoring its extent, and (ii) there is a known relationship between it and the analyte composi- tion. In the first instance, the response of the EQCM is related directly to the uptake process. In the second instance, two general interpretational issues arise: that of separating the total transfer process into its components and that of exploiting target species preconcentration. In the latter respect, polymer-modified electrodes appear promising, e. g . , partition coefficients in excess of 106 have been reported for ion-exchange into Nafion.11 Simple strategies for mobile species uptake are typically based on a single type of interaction.One approach, ion exchange into charged polymer films,12 has been utilized for the determination of cationic11 and anionic13 metal species, and neurotransmitters.14 Selectivity is primarily (see below) governed by ion charge-type.13 Another approach builds on the fact that coordination complexes have been electropoly- merized onto electrodes. 15 Electrodeposition of a metal-free polymer yields an electrode capable of complexing metal ions from the bathing solution.16.17 Here, selectivity is (primarily) governed by complexation chemistry. The chemical process * Presented at the meeting on Analytical Applications of Chemi- T Present address: Department of Chemistry, University of cally Modified Electrodes, Bristol, UK, January 7-8, 1992.Leicester, University Road, Leicester LE1 7RH, UK. a In some situations, a variety of metastable states exist. By (pseudo-)equilibrium we mean that on the timescale of the measure- ment one of these metastable states predominates, so that the system appears to be in equilibrium. On a different timescale another metastable state might predominate.1252 ANALYST, AUGUST 1992, VOL. 117 may be as simple as protonation7 for a pH sensor, or as sophisticated and specific as antibody interactions.18 Usually more than one type of interaction will contribute to selectivity patterns.11.14 In this paper, the ion-exchange and coordination uptake strategies are examined and it is shown how a single polymer may exhibit both characteristics. Partitioning of ion and neutral species is not independent, as there is a general thermodynamic requirement19 that neutral species (notably solvent) transfer accompanies ion transfer into a solvent-containing polymer film.Electrochemical Quartz Crystal Microbalance (EQCM) The quartz crystal microbalance (QCM) technique measures the variation (Af) in resonant frequency of a quartz crystal oscillator from its base value cfo) that accompanies a change (AM) in the mass attached to the crystal. When the additional mass is small and rigidly coupled20 (Af/Hz) = -(2/pv)f02(AM/g cm-2) where p is the density of the quartz and v is the wave velocity in the quartz. Characterization21722 of crystal oscillation in a liquid prompted in situ electrochemical application of the QCM. The EQCM technique has since been applied to a variety of electrochemical problems.1-4 Here we focus on its ability to follow quantitatively mobile species exchange between a polymer film and its bathing electrolyte.23-37 We draw attention to four attractive characteristics of the EQCM.The first is high sensitivity. For the 10 MHz AT-cut quartz crystals used, eqn. (1) shows that a frequency change of 1 Hz (routinely measurable) corresponds to a mass change of 4.4 ng cm-2. This areal density corresponds to 10 pmol for exchange of a solution species of molar mass 100 into the electrodes used in this work (area 0.23 cm2). The second is known sensitivity. Conversion of measured signal (An to moles of partitioning species only requires eqn. (1) and molar mass; the latter conversion is free of matrix effects unlike, for example , calculations based upon molar absorption coeffi- cients or cross-sections in spectroscopic methods.The third is in situ applicability, which allows analytical measurements to be made under optimum electrochemical control conditions. Certain powerful structural techniques, such as Auger,38 are inherently ex situ methods. This restricts their application to problems where removal from solution is acceptable. The fourth is the time resolution of the EQCM, which enables dynamic measurements to be made. This relatively unex- plored capability is crucial to one of the strategies described to extract a single species contribution from the over-all mass response. All the EQCM data analyses used are based on eqn. (l), which is appropriate for rigid films.If the polymer is extensively solvent-swollen, this rigidity requirement may not be satisfied.39 The data presented here are for rigid films. Although the QCM has been used for sensing biologically relevant species such as glucose40 and proteins,*s the full capabilities of the EQCM for biosensors have not yet been realized. A primary reason for this is the need to identify that component of the total response that is caused by the target species. This problem is addressed here. Theory The electrochemical potential, pj, of a species j in a single phase is given by: p, = p0, + RT In(.,) + z,F@ = p, + z,F@ pi is the chemical potential component, which is all that need be considered for neutral species (z, = 0). At equilibrium, the (2) - - p, values for each species in polymer and solution phases will be equal.p, imposes two constraints on each mobile species. They are associated with the two terms on the right-hand side of eqn. (2), and are the activity and electroneutrality constraints, respectively. Generally, these two constraints are fulfilled on different timescales. The EQCM responses on long timescales (typically >lo0 s) reflect satisfaction of both constraints by populations of all mobile species. The EQCM responses on shorter timescales (typically (1 s) are determined not by equilibrium paramet- ers, but rather by the rate(s) of mobile species transfer(s). These considerations suggest two procedures, one thermody- namic the other kinetic, for separating the total mass response into the components associated with individual transfer processes.Thermodynamic Aspects of the Exchange Process The problem The objective is to determine the population change of an individual species (the target species) within the polymer film. AM necessarily pertains to a sum of all mobile species population changes. We now describe a procedure for extraction of individual species contributions from the over-all gravimetric response for a modified electrode immersed in a solution of a single electrolyte (C+A-). This case involves three species [counter ion, co-ion and solvent (solv)], and is a first step towards solving the general problem. The analysis is presented for reduction of a polymer containing unipositively charged redox sites (the ‘redox’ couple): [(ox+A-).a(C+A-).P(s~lv)]~ + e- e [red.(cx - G)(C+A-).(B - E)(so~v)]~ + 6C,+ + (6 + l)As- + Esolv, (3) Subscripts p and s denote polymer and solution phases, respectively. The stoichiometric coefficients (Y and P b 0, and 6 and E may be of either sign or zero; none of these quantities need be integral.19 It is emphasized that the analysis is thermodynamic and pertains to over-all changes between (pseudo-)equilibrium states.The total mass change is the sum of three components, associated with counter ion, solvent19 and salt6 transfer: (4) Extraction of the three unknowns (population changes of counter ion, solvent and salt) requires three pieces of information. These are provided by AM (the experimental data), the electroneutrality constraint (commonly recog- nized5.6) and the concept of ‘constant solvent transfer’.Recently the background to this latter concept was pre- sented,41 and its application is demonstrated here for the first time. Counter ion contribution Electroneutrality dictates that the difference in counter- and co-ion population changes6 must balance the electronic charge injected. The counter ion contribution, AMcounter, to the mass change is given by where Q (/C cm-2) is the charge passed per unit area, and m and z represent the molar mass and charge number of the designated species (counter ion in this instance). b Electroneutrality demands that transferred co-ion be ac- companied by an equivalent amount of counter ion; these can be considered as ‘salt’.ANALYST, AUGUST 1992, VOL. 117 1253 Solvent contribution Partition coefficients (K) for solvent between the polymer (in its fully reduced and oxidized states) and the solution are defined in terms of activities (a) by - SOIV,o~n T, S0IVred Ksolv, red - asolv, red/asolv, soh (6) SOIVs,~n 2 soIvox (7) - Ksolv, ox - asolv, ox/aso~v, soln We have shown41 that - - - (Ksolv, ox/Ysolv. ox)} (8) where V is the volume of the polymer phase and y is the activity coefficient for the designated species and phase.For dilute solutions, when the activity coefficients are unity, the limiting expression for the change in solvent concentration in the polymer film upon reduction is AMsolv = msolv csolv, soh (&oh, red - Ksolv, ox) (9) Regardless of its oxidation state, the polymer is equilibrated with the same solution phase, so we can combine eqns. (6) and (7): soIvox * S0IVred Ksolv, red/ox = asolv, red/asolv, ox (10) This contains the key idea that the ratio of solvent activities in the oxidized and reduced states of the polymer is constant and does not depend on the solvent activity in the bathing solution.The solvent activity coefficient in the polymer phase cannot change in the permselective regime [a = 6 = 0 in eqn. (3)], so the redox-induced change in the polymer phase solvent population is independent of electrolyte concentration in the bathing solution .c As the electrolyte concentration is increased, salt will be partitioned into the film (permselectiv- ity failure: a, 6 # 0) and the activity coefficient of solvent in all phases will change. However, as the polymer phase (whatever its redox state) is equilibrated with the same solution phase [see eqn.(lo)], the ratio of the activity coefficients will not change significantly. This leads to the conclusion that the amount of water transferred is the same in the permselective and non-permselective regimes. We refer to this concept as ‘constant solvent transfer’. The constant solvent transfer concept will fail predictably at very high c,, typically in excess of 5 mol dm-3 (ref. 41) where the activity coefficient of water in the bathing solutions starts to decrease significantly. Salt contribution At sufficiently high electrolyte concentration, permselectivity fails: co-ion (accompanied by an equivalent amount of counter ion, and designated ‘salt’) partitions into the polymer film.We describe AMsalt, the extent of permselectivity failure, as a function of bathing solution and polymer compositions for the case where salt partitions only into one redox form (here, ox). The partition of anions and cations into a polymer contain- ing fixed (cationic) sites M+ (concentration c ~ + , ~ ) must satisfy two conditions. The first is electroneutrality cA-,p = cM+,p + cC++p (11) The second is the activity constraint,*g described by the salt partition coefficient (at a specified polymer oxidation state) Under these conditions the activity of water in the bathing Modulated by the thermodynamic function T [eqn. (15)]. solution hardly differs from unity. Combination of eqns. (11) and (12) yields a quadratic in CC+ +. The solution is expressed in terms of three dimensionless parameters R = CC+,dCM++p (13) s = cs/cM+,p (14) T = Ksalt4[Y k ,sly k ,pl (15) The extent of salt partitioning is described by R , which we term the ‘permselectivity index’.We have predicted41 that the departure of R from zero (when ST + 0) is initially quadratic (for ST <0.1), and eventually linear with electrolyte concen- tration (for ST >lo). Consequently AMsalt will vary in a quadratic mannerd as c, is increased from zero. Separation of contributions to AM The general expression for the mass change at any electrolyte concentration is given by eqn. (4). It is convenient to rewrite this as where AMperm describes the mass change at low electrolyte concentrations. For a single electrolyte, the ‘constant solvent transfer’ concept allows separation of experimental values of AM into three components, as follows.AM is measured over the widest range of c, available. At low c,, AM is independent of cs and becomes AMperm, defined by eqn. (16). We calculate AMcounter from eqn. (5). Subtrac- tion of AMcOunter from AMperm yields AMsolv. AMcounter and AMsolv are independent of c,. Consequently, their sum, the experimental quantity AMperm, can be subtracted from AM in the non-permselective regime [eqn. (4)] to yield AMsalt. This strategy is illustrated schematically in Fig. 1. Note that AMsalt and AMs0lV may each be of either sign. The sign of AMs01, is determined from the magnitude of AM in dilute solution (ST +- 0), and that of AMsalt from the behaviour of AM at high concentration (ST >0.1).When AMsalt and AMsolv are of opposite sign, interesting compensa- tory mass change effects can result (see below). Kinetics of Mobile Species Transfer We also discuss EQCM measurements on shorter timescales, following the application of a potential step. Starting with eqn. (2) the motion of charged and neutral species can be 8 6 w m C 5 4 rn s 2 0 -2.0 - 1 .o Fig. 1 Schematic diagram showing the separation of the total mass change, A M , into its components [see eqn. (4) . Solid line shows AM as a function of ST [see eqns. (14) and ( 1 5 i . Diagonally shaded, cross-hatched and open regions show (in cumulative manner) the single species contributions to A M . For illustrative purposes, contri- butions are normalized with the counter ion molar mass, and the (normalized) solvent contribution is taken to be 1.51254 ANALYST, AUGUST 1992, VOL.117 Table 1 Modified electrode deposition conditions (electrode area 0.23 cm-2) Deposition potential/ Concentration/ Species Electrode V versus SCE mmol dm-3 Electrolyte Solvent Thionine Au Poly(viny1ferrocene) Au 2,2‘-Bithiophene Pt 1.10 ca. 0.04 0.05 rnol dm-3 H20 (monomer) HC104 CH2C12 0.70 1 .o 0.1 mol dm-3 (pol ymer-reduced form) (C4H9)4 NC104 1.225 2.0 0.1 rnol dm-3 CH&N (monomer) (C2H.514 NBF4 described. Charged species are influenced by both activity and potential gradients while neutral species are influenced by activity gradients, but not potential gradients. This leads to the general expectation that ion transfers will ordinarily be more rapid than net neutral species transfers.Transient EQCM responses are therefore likely to be dominated by the transfer of charged species at short times; neutral species transfers will dominate the longer time responses. One manifestation of this field-induced transport rate enhancement is an apparently permselective response (no salt, or indeed any other neutral species, transfer) on short timescales. This will be true regardless of whether the thermodynamic behaviour of the film is permselective. In a preliminary paper,42 this effect was observed and termed ‘kinetic permselectivity’. Within the broad ‘ion’ and ‘net neutral’ classes of species, size effects and other more specific chemical interactions can lead to an observable dispersion of transfer rates. These observations suggest that individual transfers might be re- solved in the time domain.Now the analytical utility of this approach is discussed by considering the EQCM response of a polymer-modified electrode under transient conditions. A significant extension of the ‘constant solvent transfer’ concept is possible provided that salt transfer does not influence the transfer rates of other mobile species.e When this is the case we may differentiate eqn. (16) with respect to time to obtain dMldt = dM,,,,~,,/dt + dM,,,,/dt + dM,,ltldt = dMperm/dt + dM,,l,/dt (17) Rearranging eqn. (17) yields dM,,,t/dt = dM/dt - dM,,,,/dt (18) This result reveals that differences between kinetically con- trolled mass changes obtained in the non-permselective and permselective regions will provide quantitative information about instantaneous salt populations in the polymer.Experimental The instrumentation has been described previously.21~2+31 All films were deposited potentiostatically . Poly(thionine) (PTh) and poly(bithiophene) (PBT) films were deposited by electro- polymerization of the respective monomers. Electrodeposi- tion of poly(viny1ferrocene) (PVF) films exploited the lower solubility of the polymer upon oxidation.43 Details of solutions and deposition conditions are summarized in Table 1. For the experiments described here, all the modified electrodes were transferred to monomer/polymer-free solutions (see figure legends for composition). Polymer coverages were deter- mined by integration of slow scan rate (4 mV s-1) voltammetric current responses and are reported in terms of moles of electroactive sites, I‘Irnol cm-2.Coverages were in the ranges 3-10 nmol cm-2 (PTh), 5-12 nmol cm-2 (PVF), e This includes, but is not restricted to, the situation where all mobile species transfers are independent of each other. 20 - 10 im8 0 . . . . . . . . . . . 10 5 (v -- $ 1 8 m 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I .. ..... c 0 -” 0 2 4 PH 6 Fig. 2 Experimental mass change data for complete redox switching of poly(thionine) films immersed in aqueous acid solutions: U, HC104, pH varied via c,; and 0,O.l mol dm-3 total acetate, pH varied by addition of NaOH. Data taken from cyclic voltammetric scans at 5 mV s-1 and 16-38 nmol cm-2 (PBT). Comparison of gravimetric data for films with different coverages is made by dividing experimental mass changes, AMlng cm-2, by r (=Q/nF, where n is the number of electrons passed per redox site converted).Values of nAMFIQ correspond to the mass change per mole of redox sites converted. Results and Discussion Thermodynamics of Mobile Species Uptake Contributory phenomena Operationally, one seeks a single valued (ideally linear) relationship between the film mass change (AM) and the target species concentration in solution. In this section we illustrate some of the phenomena which contribute to the value of AM, emphasizing circumstances under which the desired behaviour will not prevail. Fig. 2 shows data for the redox switching of PTh films in aqueous solutions of a strong and a weak acid. The thionine redox couple in solution is a 2e--3H+ system: TH+ + 2e- + 3H+ e TH42+ where T represents the free base (see Scheme 1). Electroneut- rality requires that each redox site in a PTh film in the oxidized (reduced) state be associated with one (two) anions.The consequence of this is that 2AM,,,,t,,F/Q will be numerically equal to the anion mass. Including the protons, these electroneutrality considerations alone predict 2AMFIQ values of 102.5 and 62 g mol-1 for HC104 and CH3C02H electro- lytes, independent of solution composition (cs). These predic- tions fail, even at a qualitative level. The value of 2AMFIQ is dependent on solution composition, as required of a sensor, under some but not all circumstances (see acetic acid data at low pH). The variations of AM with electrolyte composition (19)ANALYST, AUGUST 1992, VOL.117 100 H - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . lPerchlorateJ I I I 1255 Scheme 1 are qualitatively different for the two acids; even the sign of AM is dependent upon HC104 concentration. These results pertain to complete redox switching of the modified electrode, under slow-scan voltammetric conditions and, therefore, relate to changes in (pseudo-)equilibrium populations of mobile species. We cannot invoke kinetic effects to rationalize these observations. The effects are of thermodynamic origin, but are different in the two examples. In HC104 solutions, the activity constraint requires transfer of two net neutral species: water and hydronium perchlorate.Equilibrium hydronium perchlorate populations in the film, for each polymer redox state, are dependent upon the acid concentration in solution. The extent of transfer (the differ- ence between these two equilibrium populations) is conse- quently dependent upon solution concentration. It is not possible a priori to calculate the (variation of) activity coefficients in the film. We are, therefore, required to determine the relationship between film and solution compo- sitions empirically. As different partitioning species would be characterized by different activity coefficients, it would be necessary to produce a calibration plot (of the form of Fig. 2) for each species of interest. The CH3C02H data represent a special case, of particular relevance to sensors employing specific reagents.Neutral acetic acid interacts very strongly with the polymer film,30 to the extent that a coordination model applies. Provided a minimum concentration of CH3C02H is present in the solution, all redox sites in the film are saturated with CH3C02H. Upon redox switching, one of these coordinated acetic acid molecules provides the counter ion required by electroneutrality and one of the protons [see eqn. (19)]. The result is that the only ionic species required to move during switching are two protons. The over-all 2AMFlQ value of -16 g mol-1 indicates expulsion of a species of molar mass 18, namely water. On this basis, the suggested30 half-reaction in acetic acid solutions of low pH is [(ThH+A-).(H20)-X.(HA)1, + 2e- + 2H3O+, = where X is permanently present in the film and may be either acetic acid or water.Attention is drawn to three points. Firstly, the counter ion-conjugate acid (acetate-acetic acid) is always resident in the film, i.e., does not transfer. Secondly, different species dominate the charge and mass responses, proton and solvent, respectively. Thirdly, the invariance of EQCM response at low pH (high CH3C02H concentration) is a consequence of the strong interaction between the acetic acid and the polymer. It is common for a sensor to exploit a strong, specific interaction between the target species and the polymer. If this interaction is too strong, saturation of all the available sites may result, even at fairly moderate target species concentra- tions. In the PTh example here, a composition-dependent response is seen only for pH > pK, when the solution concentration of neutral acetic acid is low and the polymer sites are no longer saturated with acetic acid.{ [ThH42+(A-)2]X}, + 3H20, (20) Separation of single species contributions We now apply the ‘constant solvent transfer’ concept to separate over-all (thermodynamic) mass change data into 500 I m rn Fig. 3 AMFIQ data for PVF oxidation as a function of NaC104 concentration. Symbols represent experimental data. Counter ion and solvent contributions (determined as in text) are shown cumulatively their components. We consider PVF films exposed to aqueous NaC104 solutions, for which the electroneutrality and activity constraints are sufficient to describe the mass changes. Fig.3 shows AMFIQ data for PVF oxidation as a function of NaC104 concentration, c,. At low concentration, AMFIQ is independent of c,, and given by AMperm [see eqn. (16)J. Eqn. (5) shows that AMcounterF/Q corresponds to +99.5 g mol-1. By difference, AMSoIvF/Q is 110 g mol-1. The magnitudes of the counter ion and solvent contributions to AM (experimen- tal points) are represented additively by the two marked regions in Fig. 3. In the light of the constant solvent transfer concept we are now able to extend our deconvolution of AM to the non-permselective region, here c, >1 mol dm-3. The low c, values of AM,,u,,erF/Q and AM,,IvF/Q in Fig. 3 have been extrapolated to high c,, using the strategy of Fig. 1. By difference [see eqn. (16)], we obtain AMSaltF/Q. Our theory predicts that the salt contribution should rise quadratically with NaC104 activity in solution.Figs. 1 and 3 are not combined, as the experimental data have a ‘concentration’ abscissa, whereas the theoretical prediction is in terms of ‘ST. Future work will involve reconciliation of theory and experi- ment: according to eqn. (15) this will require fitting activity coefficients and a partition coefficient. The central point is that we now have the theoretical machinery to effect complete separation of each single species contribution to the total mass change. Compensatory mass changes Previous PBT studies31 illustrate a more subtle problem. As for PVF, counter ion, salt and solvent are transferred upon PBT redox switching. The difference is that, for PBT, the transfers of the two net neutral species (salt and solvent) are in opposite directions. In terms of eqn.(3), 6 = -0.36 and 0.11 < E < 0.32 for a range of CH3CN-tetraalkylammonium salt solutions. Reduction of the PBT+ is accompanied by solvent exit and salt entry, in addition to counter-ion exit. The values of molar masses, 6 and E, are such that AMsolv = -AM,,It. This has the totally fortuitous result that AM = AMcounter. We term this set of circumstances ‘apparent permselectivity’. Detailed experiments31 for different electrolytes over a range1256 ANALYST, AUGUST 1992, VOL. 117 Y a1 2000 1 I x "3 .. I 1600 - N 1200 - m C 5 800 - 400 - 0 400 800 #pC cm-2 1200 Fig. 4 AM versus Q plots for the oxidation and subsequent reduction of a PVF film in 0.1 mol dm-3 (0) and 3 mol dm-3 (x) NaC104 solutions.The potential was stepped from 0 to +0.7 V, held at +0.7 V for 15 s, then stepped back to 0 V. The origin corresponds to reduced PVF at 0 V. The small intercept on the abscissa is attributed to double layer charging. The lines correspond to the predicted AM/Q relation- ship for anion and anion + solvent transfer, as marked of concentrations revealed deviations from AM = AMcounter, due to the concentration dependence of AMsalt (see above). We suspect that compensatory mass changes are fairly common. Analytical measurements made without due atten- tion to this process could lead to serious errors. Kinetic Aspects of the Exchange Process Kinetic permselectivity : the strategy In the previous section, we considered the over-all mass changes associated with complete oxidation-reduction of a film, which was then allowed to come to equilibrium with the solution phase.In this section we consider the time course of the redox transformation, i.e., the transfer of species during the oxidation-reduction process. When permselectivity does not exist in a thermodynamic sense (for example at high electrolyte concentration), we propose that it may be possible to achieve it transiently. We aim to exploit the differing rates of mobile ion and neutral species transfers. In a transient experiment the response on a short timescale will be dominated by the fastest moving species. The converse will be true on a longer timescale. We suggest that this approach might be exploited at two levels. Firstly, field assistance of ion transfer (migration) will lead to their being more rapid than neutral species transfers. Secondly, size effects (for ions or neutral species) will lead to a diversity of transport rates.The latter effect is likely to be more pronounced in the confined geometry of polymer films than for the same species in solution. Also, the extent to which transfer of a given species dominates the net transfer process (on a given timescale) will depend on its availability, i.e., the difference between its film and solution concentrations. Kinetic permselectivity: an example We first observed this effect during rapid scan voltammetry of PVF films in aqueous NaC104 solutions,42 for which the over-all mass changes accompanying redox switching have been described above (see Fig.3). The EQCM data acquired during chronoamperometric experiments on a PVF film exposed to 0.1 and 3 mol dm-3 NaC104 are shown in Fig. 4. Two straight lines are drawn corresponding to the mass- charge relationships that would result if (i) only counter ions and (ii) counter ion and solvent (but not salt) were trans- ferred. These correspond to AMcounter and AMperm, respec- tively. 2500 2000 ," 1500 c I 1000 m . 500 P m . -1000 I I I I I I -0.2 0 0.2 0.4 0.6 0.8 1.0 tls Fig. 5 Salt mass flux (dM,,,,ldt) as a function of time for the oxidation stage of the data in Fig. 4. Points were generated by taking the time differential of the difference in the two sets of data in Fig. 4 [see eqn. C18)l During the initial stage of the oxidation process, the experimental data in both bathing solutions superimpose on the line for AMcounter.This result implies that only C104- transfers on that timescale. Neutral species transfers are rather slower and progressively become more predominant at longer timescales. Fig. 5 gives the result of applying eqn. (18) to the first (oxidative) stage of the two data sets of Fig. 4. It shows the variation of AMsalt with time in 3 mol dm-3 sodium perchlor- ate (non-permselective conditions). As generally expected, salt (a net neutral species) transfer lags counter-ion transfer. For the particular case of our PVF films in this medium, salt transfer is the faster of the two neutral species transfers. Conclusions We have considered the transfers of mobile species accom- panying redox switching of a redox polymer film immersed in a solution of a single electrolyte.The problem has been cast in terms of the associated mass changes, which can be deter- mined by using the EQCM. It is possible to separate the total mass change observed during film redox switching into the components associated with counter ion, solvent and salt transfer under both equilibrium and transient conditions. This separation exploits the concept of constant solvent transfer, which asserts that the extent of solvent transfer is independent of electrolyte concentration in the bathing solution. Transient measurements allow separation of charged and the various net neutral species transfers in time. We suggest that this will have analytical utility when determining mixtures of mobile species via their uptake by a polymer film.The SERC (GFUE132946 and GR/E/78104), NATO (86/0830) and the National Science Foundation (CHE-9115462) are thanked for financial support. M. J. S. thanks the SERC for a research studen tship. References 1 Deakin, M. R., and Buttry, D. A . , Anal. Chem., 1989, 61, 1147A. 2 Schumacher, R., Angew. Chem., Znt. Ed. Engl., 1990,29,329. 3 Ward, M. D . , and Buttry, D. A . , Science, 1990,249, 1OOO. 4 Buttry, D. A., in Electroanalytical Chemistry, ed. Bard, A. J., Marcel Dekker, New York, 1991, vol. 17, p. 1. 5 Murray, R. W., in Electroanalytical Chemistry, ed. Bard, A. J., Marcel Dekker, New York, 1984, vol. 13, p. 192. 6 Hillman, A. R., in Electrochemical Technology of Polymers, ed. Linford, R . , Elsevier, London, 1987, p.241. 7 Heineman, W. R., Wieck, H. J., and Yacynych, A . M., Anal. Chem., 1980, 52,345.ANALYST, AUGUST 1992, VOL. 117 1257 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Lawton, R. S., and Yacynych, A. M., Anal. Chim. Acta, 1984, 160, 149. Cox, J. A., and Kulesza, P. J., J. Electroanal. Chem., 1984,175, 105. Facci, J. S., and Murray, R. W., Anal. Chem., 1982,54, 772. Szentirmay, M. R., and Martin, C. R., Anal. Chem., 1984, 56, 1898. Guadalupe, A. R., and Abruna, H. D., Anal. Chem., 1985,57, 142. Cox, J. A., and Kulesza, P. J., Anal. Chim. Acta, 1983,154,71. Nagy, G., Gerhardt, G. A., Oke, A. F., Rice, M. E., Adams, R. N., Moore, R. B., Szentirmay, M. N., and Martin, C. R., J. Electroanal. Chem., 1985, 188, 85. Denisevich, P., Abruna, H.D., Leidner, C. R., Meyer, T. J., and Murray, R. W., Inorg. Chem., 1982,21,2153. Wier, L. M., Guadalupe, A. R., and Abruna, H. D., Anal. Chem., 1985,57,2009. Itaya, K., Akahoshi, H., and Toshima, S., J. Electrochem. Soc., 1982, 129,762. Ebersole, R. C., and Ward, M. D., J. Am. Chem. Soc., 1988, 110, 8623. Bruckenstein, S., and Hillman, A. R., J. Phys. Chem., 1988,92, 4837. Sauerbrey, G. Z., 2. Phys., 1959,155,206. Bruckenstein, S., and Shay, M., Electrochim. Acta, 1985, 30, 1295. Kanazawa, K. K., and Gordon, J. G., Anal. Chim. Acta, 1985, 175, 99. Orata, D., and Buttry, D. A., J. Am. Chem. SOC., 1987, 109, 3574. Baker, C. K., and Reynolds, J. R., J. Electroanal. Chem., 1988, 251, 307. Reynolds, J. R., Sundaresan, N. S., Pomerantz, M., Basak, S., and Baker, C. K., J. Electroanal. Chem., 1988, 250, 355. Baker, C. K., and Reynolds, J. R., Synrh. Met., 1989, 28, C21. Hillman, A. R., Loveday, D. C., and Bruckenstein, S., J. Electroanal. Chem., 1989, 274, 157. 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 Hillman, A. R., Loveday, D. C., Swann, M. J., Eales, R. M., Hamnett, A., Higgins, S. J., Bruckenstein, S., and Wilde, C. P., Faraday Dkc. Chem. SOC., 1989, 88, 151. Bruckenstein, S., Wilde, C. P., Shay, M., and Hillman, A. R., J. Phys. Chem., 1990,94, 787. Bruckenstein, S., Wilde, C. P., and Hillman, A. R., J. Phys. Chem., 1990,94,6458. Hillman, A. R., Swann, M. J., and Bruckenstein, S., J. Electroanal. Chem., 1990, 291, 147. Kelly, A. J., Ohsaka, T., Oyama, N., Forster, R. J., and Vos, J. G., J. Electroanal. Chem., 1990,287, 185. Inzelt, G., J. Electroanal. Chem., 1990, 287, 171. Naoi, K., Lien, M., and Smyrl, W. H., J. Electrochem. SOC., 1991, 138,440. Borjas, R., and Buttry, D. A., Chem. Muter., 1991,3,872. Hillman, A. R., Swann, M. J., and Bruckenstein, S., J. Phys. Chem., 1991,95,3271. Hillman, A. R., Loveday, D. C., and Bruckenstein, S., J. Electroanal. Chem., 1991,300, 67. Bruce, J. A., and Wrighton, M. S., J. Am. Chem. SOC., 1982, 104,74. Borjas, R., and Buttry, D. A., J. Electroanal. Chem., 1990,280, 73. Lasky, S. J., and Buttry, D. A., ACSSymp. Ser., 1989,403,237. Bruckenstein, S., and Hillman, A. R., J. Phys. Chem., 1991,95, 10748. Hillman, A. R., Loveday, D. C., Bruckenstein, S., and Wilde, C. P., J. Chem. Soc., Faraday Trans., 1990,86, 437. Hillman, A. R., Loveday, D. C., and Bruckenstein, S., Langmuir, 1991, 7, 191. Paper 2102236 B Received April 30, 1992 Accepted May 14, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701251

出版商:RSC    

年代:1992

数据来源: RSC

摘要:

ANALYST, AUGUST 1992, VOL. 117 1259 Flow Injection Amperometric Determination of Nitrite at a Carbon Fibre Electrode Modified With the Polymer [O~(bipy)~(PVP)~&l]Cl* Michael M. Malone, Andrew P. Doherty, Malcolm R. Smytht and Johannes G. Vost School of Chemical Sciences, Dublin City University, Dublin 9, Ireland The development of carbon fibre electrodes modified with the polymer [Os( bipy)2(PVP)20CI]CI for the flow amperometric determination of nitrite is described. This osmium polymer modifier greatly enhances the kinetics of nitrite reduction compared with the reaction at bare carbon electrodes. Various electrode characteristics were optimized using both cyclic voltammetry and flow injection. The Cali bration graph yielded a slope of 0.197 nA cm3 pg-1 over the linear range 0-400 pg cm-3 with a limit of detection of 0.1 pg cm-3.The modified electrode was shown to exhibit good short-term reproducibility yielding a relative standard deviation of 2.1 5% ( n = 20). After a 3 week period of monitoring, involving 240 standard injections and 30 meat extract injections, the electrode continued to function with no significant change in sensitivity. The electrode was used to analyse a processed meat sample for nitrite content and the results compared favourably with those obtained using a standard reference spectrophotometric method. Keywords: Carbon fibre electrode; polymer modification; nitrite determination; flow injection amperometric detection; meat analysis The potential hazard of nitrite to human health has been well documented. 1 Conventional techniques for the determination of nitrite are based on spectrophotometric procedures using azo dye formation reactions.2 These methods however have limited sensitivity and dynamic range, and frequently suffer from interferences such as ascorbic acid.Several polarographic methods for the sensitive determina- tion of the nitrite ion have been reported in the literature.3" However, such polarographic methods have obvious disad- vantages for routine sensing and flow applications, and in recent years more emphasis has been placed on the use of solid electrodes for such applications. A large number of methods have been developed for the voltammetric determination of nitrite by oxidation at solid electrodes. Cox and Kulesza7 modified a platinum electrode by chemisorption of iodine, which was found to improve the reproducibility and decrease the peak width in the oxidation of nitrite by linear scan voltammetry.Nitrite oxidation at a bare glassy carbon electrode was reported by Newbery and Lopez de Haddad,g but this method suffered interferences from both ascorbate and chloride ions. The determination of nitrite following oxidation at both electrochemically pre-treated9 and polymer modified1@-12 glassy carbon electrodes at lower operational potentials has therefore been investigated. With the ruthen- ium polymer modified electrodes reported by Barisci et aZ.11 and Wallace et aZ.,1* the major problem associated with the electrodes was their long-term stability. Reduction of nitrite at Au, Pt and carbon electrodes is known to cause severe surface fouling.13 Reductive techniques are also limited by the negative potentials that are required for detection where interferences from metal cations, hydrogen peroxide and oxygen may be problematic.A modified electrode for the reduction of nitrite at moderate potentials, based on modification of glassy carbon macro-electrodes with the electrocatalytic polymer [ O S ( ~ ~ ~ ~ ) ~ ( P V P ) ~ ~ C I ] C I , has recently been reported. 14 Such modification offered several advantages over existing electrochemical and spectropho- tometric procedures. 14 In recent years there has also been much interest in the application of microelectrodes. These electrodes have several advantageous features when compared with macroelectrodes. * Presented at the meeting on Analytical Applications of Chemi- cally Modified Electrodes, Bristol, UK, January 7-8, 1992.Authors to whom correspondence should be addressed. For example, they enable time-independent currents to be monitored, are virtually non-destructive of the analyte and can be used in solutions of very high resistance.15 This paper describes the flow detection of nitrite in meat samples using a carbon fibre microelectrode that has been modified by chemisorption of the electrocatalytic polymer [0~(bipy)~(PVP)~~Cl]Cl, where bipy = 2,2'-bipyridyl and PVP = poly(4-vinylpyridine). The modified microelectrode was applied to the determination of nitrite in processed meat and the results are compared with those obtained using a standard spectrophotometric method.16 Experimental Reagents and Materials All reagents were of analytical-reagent grade.All aqueous solutions were prepared using de-ionized water obtained by passing distilled water through a Milli-Q water purification system. The electrolyte used throughout was 0.1 mol dm-3 HzSO4. Nitrite standard solutions were prepared daily using sodium nitrite [BDH (now Merck)], by appropriate dilutions in 0.1 mol dm-3 Na2S04 (Riedel-de Haen), as nitrite solutions are known to be unstable at low pH.16 The synthesis of the polymers has been reported elsewhere.17,18 Carbon fibres (14 pm diameter) were obtained from Avco. The surface of these fibres had no external coating. The meat sample analysed was Denny processed ham. Instrumentation Cyclic voltammetry was performed using an EG & G Princeton Applied Research (PAR) polarographic analyser/ stripping voltammeter in conjunction with a JJ Lloyd Instru- ments x-y recorder (Model PL4).A 20 cm3 laboratory-built three-electrode cell was employed for batch studies incorpor- ating the modified electrode, an Ag-AgCI reference electrode and a platinum auxiliary electrode. The working electrode used for the batch studies was prepared by inserting a carbon fibre into the narrow end of a plastic micropipette tip and gently heat-sealing it. The electrical connection was made by back-filling with mercury and dipping in a copper wire. The flow injection apparatus consisted of a Gilson Minipuls- 3 peristaltic pump, and a six-port Rheodyne injector with a 0.020 cm3 fixed-volume sample loop which was connected to the three-electrode system described below.The electrode1260 ANALYST, AUGUST 1992. VOL. 117 terminals were connected to an EG & G PAR Model 400 EC detector which was linked to a Philips PM8251 single-pen recorder to record the amperometric signals. A Shimadzu ultraviolet/visible recording spectrophotometer (Model UV- 240) was used in the comparison method for the determination of nitrite in meat. Construction of the Carbon Fibre Flow Cell The preparation of the carbon fibre working electrode was carried out using a method reported previously.19 The Ag-AgC1 reference electrode was prepared by firstly connect- ing a silver wire (0.1 mm diameter) to the anode and a platinum electrode to the cathode of a 1.5 V battery, after which the assembly was immersed in a solution of 1 rnol dm-3 HCl for 2 min.The wire was then inserted in a polyethylene tube (15 x 1 mm i.d.), one end of which was plugged with a ceramic porous rod (2 x 1 mm id.). The tube was then filled with the 1 rnol dm-3 HC1 internal reference solution and closed by heating. A 2 cm piece of stainless-steel tubing (1 x 0.2 mm id.) served as a counter electrode. Throughout this paper, potentials are quoted, after numerical conversion, versus the standard calomel electrode (SCE). After modification of the working electrode (described below) the working, reference and auxiliary electrodes were mounted in a T-tube arrangement reported previously,20 so that the electrolyte passed first through the working electrode and then via the auxiliary electrode to waste.Modification of the Working Electrode The electrode was connected to the tubing of a peristaltic pump via silicon tubing. The carbon fibre was cleaned by pumping 25 cm3 of 5 mol dm-3 HCI, 25 cm3 of de-ionized water and 25 cm3 of methanol, respectively, through the electrode at a rate of 1 cm3 min-1. The electrode was then air-dried by pumping air through for 1 h. When dry, the electrode was modified with the polymer solution of desired concentration in methanol. A 2 cm plug of 0.1% m/v polymer solution was first drawn into the peristaltic tubing, followed by air. The plug was pumped slowly towards the electrode. The pump was stopped for 40 s when the plug surrounded the electrode surface, allowing the polymer to chemisorb onto the fibre surface, after which time the pump was restarted so that air flowed past the electrode surface.In this way the electrode was dried for 3 h before use. Meat Sample Analysis The determination of nitrite in meat was carried out on a sample of processed ham (Denny). This involved the prior extraction of the nitrite from a 5 g sample of meat using a standard Association of Official Analytical Chemists (AOAC) method.21 The procedure involves extraction with hot water for 2 h. The extract was then filtered and made up to 50 cm3 in a calibrated flask. Prior to injection, the extract was diluted 1 + 1 with 0.2 mol dm-3 Na2S04, so that samples and standards were 0.1 mol dm-3 in Na2S04. Injections (0.020 cm3) of standards and samples were made. Results and Discussion Cyclic Voltammetry Studies were first carried out to investigate the retention of the osmium polymer on the carbon fibre surface and its electro- catalytic activity towards nitrite reduction.Well defined oxidation and reduction responses associated with the surface- bound OS~~--OS"~ couple were observed in 0.1 rnol dm-3 H2S04. This supporting electrolyte had previously been shown to be optimum for a macro glassy carbon electrode I I I I I E N 0.8 0.3 -0.2 Fig. 1 T pica1 cyclic voltammogram obtained for [Os(bipyh- (PVP)&$Cl. Scan rate = 0.10 V s-l; electrolyte = 0.1 rnol dm-3 H2S04 1 I I 1 0.6 0.2 -0.2 EN Fig. 2 Cyclic voltammograms for A, modified electrode in blank 0.1 rnol dm-3 H2SO4 electrolyte; B, same electrode as A in 5 x 10-3 rnol dm-3 NO2-; and C, same as B at higher sensitivity outlining the typical electrocatalytic shaped curve obtained on addition of nitrite.Scan rate = 0.010 V s-1; electrolyte = 0.1 rnol dm-3 H2SO4 using a similar polymer.14 Fig. 1 shows a typical cyclic voltammogram obtained from a modified microelectrode in 0.1 rnol dm-3 H2SO4. The peak-to-peak separation (AEp) of these waves was found to be 0.045 V, indicating the favourable kinetics associated with the OS~~-OS~I~ redox couple. A linear dependence of the cathodic and anodic peak currents on scan rate was observed at lower scan rates (less than 0.010 V s-*), indicating the predominantly surface behaviour of the modi- fier, as noted with some other polymer modified electrodes.22 The modified microelectrode stabilized very rapidly, and repetitive cycling over a 2 h period produced no significant change in response, indicating strong adsorption of the polymer.ANALYST, AUGUST 1992.VOL. 117 1261 The effect of nitrite reduction on this redox couple is shown in Fig. 2. On addition of nitrite to the solution, an electro- catalytic reduction occurred at the same potential as that observed for the reduction of Osl*' to 0s". When the voltammogram obtained after addition of nitrite (Fig. 2, B and C) is compared with that obtained from the same electrode in blank electrolyte solution (Fig. 2, A), the nitrite response can clearly be seen. The cyclic voltammogram is significantly elongated along the current axis, which is indicative of an electrocatalytic reaction. The electron transport kinetics at the bare glassy carbon electrode surface were too slow to yield a useful analytical signal.Cyclic voltammetry was used to determine the optimum length of time needed for optimum coating of the microelec- trode with the polymer. This was examined by dipping the fibres in a polymer solution for different lengths of time varying from 10 to 360 s and performing cyclic voltammetry on the resulting modified electrodes. The cyclic voltammetric peak heights were then measured, and 40 s was found to be the coating time giving rise to the best sensor response. At coating times longer than 40 s , the methanol solvent appeared to redissolve the polymer from the carbon fibre surface. The effect of the concentration of polymer solution in modifying the carbon fibres was also investigated.It was found that thicker polymer films produced smaller responses to nitrite, which is probably a result of hindered analyte and counter ion mass transport within the film.23324 An optimum polymer concentration of 0.1% m/v was found and used in further studies. As the osmium response is always present, the response for nitrite must be recorded on top of a substantial background current. However, by using amperometric detection with flow injection the background current can readily be offset, so that the subsequent responses will be due to addition of nitrite. Flow Injection When utilizing the polymer modified carbon fibre electrode in acidic flowing streams, the potential applied ensures that all the osmium is in the 0s" form, so that when the analyte reaches the surface it is reduced at the osmium centres within the polymer film according to the following proposed cross- exchange reaction mechanism14 N02- + H+ HNO2 (1) HN02 + H+ NO+ + HzO (2) NO+ + 0s" -+ NO(g) + 0s"' (3) The potential was varied from +0.195 to -0.250 V in 0.050 V increments, and injections of 0.020 cm3 of a 50 vg cm-3 NO2- standard solution were made at each poten- tial.Decreasing the potential resulted in increased sensitivity for nitrite reduction, but below -0.150 V the background noise also increased. Therefore, the detection potential was set at -0.150 V. The flow rate of the 0.1 mol dm-3 H2S04 electrolyte was kept low (0.2 cm3 min-I), as the kinetics of nitrite reduction are fairly slow. In fact, a useful signal cannot be obtained using these conditions at a bare carbon fibre electrode.However, under the same conditions using the polymer modified electrode, well defined and reproducible responses were obtained. The linearity of the method was determined by injecting a series of nitrite standards in the concentration range WOO pg cm-3 and constructing a calibration graph. The method was linear over this range with a correlation coefficient, r , = 0.999 and the regression equation was: y(nA) = 0.197~ (pg cm-3) - 1.9845. This linear range is better than that obtained using spectrophotometric procedures.25 The limit of detection was determined to be 0.1 pg cm-3 using a signal-to-noise ratio of 3: 1. The precision of the method was evaluated in terms of the variability between 20 replicate injections of various concen- trations of NO2- solution. This resulted in a relative standard deviation of 2.15% for all concentrations examined.Fig. 3 shows some typical amperometric responses obtained from the modified fibre flow cell at different nitrite concentrations, where the reproducibility of response is clearly evident. The long-term stability of the modified microelectrode was monitored over a 3 week period. After 3 weeks of continuous use, involving over 240 standard injections and 30 meat extract injections, the electrode showed no appreciable change in sensitivity. No surface-fouling effects were observed over this period of time. These results demonstrate the long-term stability of the modified microelectrode. This level of stability is rare for modified electrodes.Similar work has been carried out using the ruthenium polymer modified carbon fibre microelectrode analogue as a continuation of the study of ruthenium polymer modified glassy carbon electrodes.11.12 In this instance the oxidation of nitrite was investigated, which is a kinetically faster reaction than that of nitrite reduction. The use of the ruthenium polymer modifier lowered the overpotential for nitrite oxida- tion compared with that at a bare carbon fibre electrode. This method was approximately seven times more sensitive for the determination of nitrite with a slope of 1.41 nA cm3 pg-1 compared with the method reported here. It had a shorter analysis time and comparable short-term reproducibility. However, the main problem associated with this electrode was its long-term stability.When monitored over a 9 d period, a gradual decrease in response was observed owing to stripping of the polymer from the carbon fibre surface. Severe interference from ascorbic acid was also evident. The long- term stability of the osmium polymer modified electrode reported here was shown to be far superior, and deemed to be of more analytical importance in the design of an operational sensor for nitrite. Determination of Nitrite in Meat For both the electrochemical method and the spectropho- tometric reference method, nitrite was extracted from the meat sample using a standard AOAC method.21 For the analysis using the modified microelectrode the extract was diluted 1 + 1 with 0.2 mol dm-3 Na2S04. Using a range of nitrite standards (0-12 pg cm-3 prepared from NaN02), a calibration graph was constructed, which had a correlation coefficient, r , = 0.999 and a regression equation as follows: y (nA) = 0.098~ (pg (3121-3) - 0.016.Electrochemical analysis of the meat sample yielded a result of 88.0 k 1.0 pg g-1 of NO2- (n = 2), which compares favourably with the value of 84.0 k 1.2 pg g-1 ( n = 2) obtained using the reference spectropho- tometric method.16 Thirty consecutive injections of the meat extract were made, without the need for pre-treatment of the electrode, resulting in no deterioration of the response. Hence, in addition to exhibiting electrocatalytic properties, the polymer acts as a protective membrane by inhibition of adsorption of matrix compounds such as proteins on the carbon fibre surface, and thus preventing surface fouling.The polymer also prevented the severe surface fouling encountered at solid electrodes on nitrite reduction.I3 The method has a moderate analytical throughput and is capable of handling 30 samples h-1. Interferences For electrocatalysis at redox polymers such as those described here, the Gibbs free energy for the cross-exchange reaction must be negative. This requirement imposes certain limita- tions on the types of reaction that can occur at electrodes modified with these materials. By exploiting this thermo-ANALYST, AUGUST 1992, VOL. 117 a) 2 min H t I b) T0.79 nA 2 min - Time Fig. 3 Flow amperometric responses of the osmium polymer modified electrode to consecutive injections of (a) 50 and ( b ) 25 pg cm-3 of N02-.Constant potential operation at -0.150 V versus SCE; electrolyte = 0.1 mol dm-3 H2S04; flow rate = 0.2 cm3 min-1 2 min H acid. This feature is of considerable advantage for the determination of nitrite. Considering the limitations imposed by the thermodynamic requirements for cross-exchange reac- tions, then only reduction reactions are feasible with the electrocatalyst in the 0s" redox state. The thiocyanate anion, another potential interferent, was also found to be electro-inactive at the redox polymer. This is because the formal potential for thiocyanate oxidation is more positive than the ElI2 of the electrocatalytic centre, thus eliminating possible surface-fouling effects caused by the oxidation of this anion. The only major interferent in the electrochemical method was found to be iron(n1).However, standard procedures are available to remove iron(1ir) prior to the determination of nitrite.16 Interference from iron(ii1) can also be eliminated by the addition of ethylenediaminetetraacetic acid (EDTA) to the sample and carrier electrolyte. The [ Fe( EDTA)] - com- plex formed has a formal potential more negative than the of the osmium polymer and consequently is thermodynami- cally removed from the reaction at the modified electrode. Conclusion The osmium polymer modified carbon fibre flow electrode has been demonstrated to be sensitive , accurate and reproducible for the determination of the nitrite ion in a meat sample. The electrode was also shown to be very stable over extended periods of use without changes in sensor sensitivity.The sensor can be operated under conditions where common interferents such as ascorbic acid and thiocyanate are non- electroactive. The lack of interference and high stability of the sensor are major advantages of this device. When compared with the standard spectrophotometric techniques, the modified microelectrode has a comparable analysis time; however; considerably wider linear ranges, increased sensitivity and lower limits of detection are charac- teristic of the modified microelectrode. Also, the proposed method does not require the use of potentially unstable complexing or colour-forming reactions. The microelectrode flow cell is also simple and inexpensive to produce and is easy to operate. -Time Fig. 4 Effect of ascorbic acid on the response of the modified electrode to nitrite.A, 500 pg cm-3 ascorbic acid; B, 50 pg cm-3 NO2-; and C, 500 pg cm-3 ascorbic acid plus 50 pg ~ m - ~ NO2-. Operational conditions as for Fig. 3 dynamic limitation, the sensitivity and selectivity of sensors constructed from these materials can be controlled by the synthetic control of the E1/2 of the electrocatalytic centre and by the manipulation of the formal potential of possible analytes and interferents. As the osmium polymer has an Ellz of 0.250 V versus SCE," it therefore has wide scope for mediating reduction reactions. This provides a material that can be used in a range of analytical applications, with sensitivity and selectivity being controlled by the of the redox centre. The ascorbate anion is a common interferent when determining NO2- in processed meats, as sodium ascorbate is frequently added as an antioxidant.This com- pound does not interfere using the osmium polymer modified electrode, as the ascorbate ion cannot undergo oxidation at the modified electrode when the sensor is operating in the 0s" state. This can be seen in Fig. 4, A, when an injection of 500 pg cm-3 ascorbic acid produces a background response. In Fig. 4, B and C the response for 50 pg cm-3 of nitrite and 50 pg cm-3 of nitrite plus 500 pg (3111-3 of ascorbic acid solutions is shown. It can be seen that interference in the determination of nitrite does not occur even with a 10-fold excess of ascorbic 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 References Lijinsky, W., and Epstein, S.S., Nature (London), 1970, 223, 21. Barnes, H., and Folkard, A. R., Analyst, 1951, 76, 599. Chang, S., Kozeniauskas, R., and Harrington, G. W., Anal. Chem., 1977,49, 14. Zhao, Z., and Cai, X., J. Electroanal. Chem.. 1988, 252,361. Gao, Z . , Qing, G., and Zhao, Z., Anal. Chim. Acta, 1990,230, 105. Markusova, K., and Fedurco, M., Anal. Chim. Acta, 1991,248, 109. Cox, J. A., and Kulesza, P. J., J. Electroanal. Chem.. 1984,175, 105. Newbery, J. E., and Lopez de Haddad, M. P., Analyst, 1985, 110, 81. Chamsi, A. Y., and Fogg, A. G., Analyst, 1988, 113, 1723. Cox, J. A., and Kulkarni, K. R., Analyst, 1986, 111, 1219. Barisci, J. N., Wallace, G. G., Wilke, E. A., Meaney, M., Smyth, M. R., and Vos, J. G., Electroanalysis, 1989, 1,245. Wallace, G. G., Meaney, M., Smyth, M. R., and Vos, J. G., Electroanalysis, 1989, 1, 357. Mengoli, G., and Musiani, M. M., J. Electroanal. Chem., 1989, 269,99. Doherty, A. P., Forster, R. J., Smyth, M. R., and Vos, J. G., Anal. Chim. Acta, 1991, 255,45. Pons, S., and Fleischmann, M., Anal. Chem., l987,59,1391A. Vogels Textbook of Quantitative Inorganic Analysis, Longman, New York, 4th edn., 1978, pp. 97, 158 and 755. Forster, R. J., and Vos, J. G., Macromolecules, 1990,59,4372.ANALYST, AUGUST 1992, VOL. 117 1263 18 Clear, J. M., Kelly, J. K., O’Connell, C. M., and Vos, J. G., 1. Chem. Res. ( M ) , 1981, 3039. 19 Hua, C., Yunping, W., Chunghan, J., and Tonghui, Z., Anal. Chim. Acta, 1990,235,273. 20 Hua, C., Sagar, K. A., McLaughlin, K., Jorge, M., Meaney, M. P., and Smyth, M. R., Analyst, 1991, 116, 1117. 21 Official Metho& of Analysis of the Association of Official Analytical Chemists, AOAC, Washington, DC, 13th edn. , 1980. 22 Murray, R. W., in Electroanalytical Chemistry, ed. Bard, A. J., Marcel Dekker, New York, 1984, vol. 13, p. 240. 23 Denisevich, P., Abruna, H. D., Leidner, C. R., Meyer, T. J., and Murray, R. W., Inorg. Chem., 1982,21,2153. 24 Ikeda, T., Shmehl, R., Denisevich, P., Willman, K., and Murray, R. W., J. Am. Chem. SOC., 1982,104,2683. 25 Silva, M., Gallego, M., and Valchrcel, M., Anal. Chim. Acta, 1986,179, 341. Paper 2f00413E Received January 27, 1992 Accepted April 23, 1992

ISSN:0003-2654    

DOI:10.1039/AN9921701259

出版商:RSC    

年代:1992

数据来源: RSC