ANALYST NOVEMBER 1986 VOL. 111 1235 Modified Platinum Wire Glucose Oxidase Amperometric Electrode G. J. Moody G. S. Sanghera and J. D. R. Thomas Department of Applied Chemistry Redwood Building UWIST PO Box 13 Cardiff CFI 3XF UK Aflow injection system incorporating a modified platinum wire amperometric enzyme electrode is described. Glucose oxidase was covalently attached to an activated platinum surface to form an enzyme electrode which was then incorporated in a laboratory-built three-electrode flow-through cell. The system exhibited good linearity (for glucose concentrations of 0.1-20 mM) where log (current/A) = 0.992 log ([glucose]/~) -3.94 with a correlation coefficient of 0.999. Response times (<25 s) and wash times (<30 s) were short and a lifetime of 9 h was obtained for continuous exposure to 2.5 and 10 mM glucose.Normal use during the flow injection analysis of glucose gave lifetimes of 10 d. Keywords Amperometric glucose sensor; enzyme electrode; flow injection analysis; modified platinum electrode An enzyme electrode essentially consists of a layer of immobilised enzyme held over a suitable electrochemical sensor such as the Clark electrode.’ This assembly combines the specificity of an enzyme with the sensitivity of potentio-metric2.3 and amperometric3 electrodes for a wide range of substrates. Over the past decade the chemical modification of elec-trodes in order to provide an electrode surface more selective than bare metal has become well established. The attachment of suitable redox centres to the bare metal can be achieved by direct covalent bonding,4 coating with a polymer film5 and the covalent attachment of an enzyme directly on to a chemically modified electrode surface.6 The last of these is a logical development as the thinner membrane accelerates the diffu-sion of substrate to and the diffusion of product(s) from the active sensor zones.The direct covalent attachment of an enzyme on to a chemically modified electrode surface has good prospects, especially for the microfabrication of simple implantable sensors. The use of glucose oxidase in the presence of oxygen for the direct electrochemical detection of glucose is usually based on the amperometric measurement of hydrogen per-oxide at a platinum - enzyme electrode. However the use of mediators7 and conducting organic salts8 can lead to a sensor that is potentially independent of oxygen concentration in the vicinity of the probe.Thus ideally an enzyme sensor for implantation would be small in size and by the agency of redox mediators would be independent of oxygen concen-tration. Although this goal remains elusive present develop-ments with oxygen independent sensors and miniature devices are a step in this direction. This paper describes a miniature enzyme electrode consist-ing of glucose oxidase covalently attached to a silanised and anodised platinum wire surface via the bifunctional glutar-aldehyde enzyme immobilising reagent. Its response charac-teristics were determined in a three-electrode amperometric mode by monitoring the anodic decomposition of hydrogen peroxide.Experimental Reagents Glucose oxidase (E.C. 1.1.3.4 100 IU mg-1 purified from Aspergillus niger) glutaraldehyde (25% aqueous solution), p a ( +)-glucose and 3-aminopropyltriethoxysilane were obtained from Sigma Chemical (Poole Dorset UK) and platinum wire from Goodfellow Metals (Cambridge UK). All other materials were of the best analytical grade available. Glucose standards were prepared from a fresh stock solution of P-D(+)-glucose (1 M) in 0.1 M orthophosphate buffer (pH 6.0). The carrier stream phosphate buffer was sodium di-hydrogen orthophosphate (0.1 M) adjusted with sodium hydroxide solution. Immobilisation of Enzyme Essentially the method incorporated a simplification of the chemical modification of platinum outlined by Yao ,9 followed by the attachment of glucose oxidase via the bifunctional reagent glutaraldehyde using a procedure similar to that described for controlled porosity glass by Masoom and Townshend.10 This approach avoided the more complicated thin film approach of Ya0,9 which involved the use of bovine albumin.The platinum wire electrodes were cleaned by successive bathing for 10 min in hot concentrated nitric chromic and hydrochloric acids respectively followed by electrochemical treatment in sulphuric acid (0.5 M) with voltammetric cycling between 0 and +1.3 V (relative to Ag - AgC1) for 2 h or alternatively overnight. Cyclic voltammetry was then used to indicate “clean” platinum wires in each instance. Prior to silanisation the clean platinum wire was anodised at +2.50 V (relative to Ag - AgCI) in sulphuric acid (0.5 M) for 5 min.After thorough washing with de-ionised water and drying the anodised platinum wire was refluxed for 45 min in anhydrous 3-aminopropyltriethoxysilane in toluene (10% V/V). Glutaraldehyde (2.5% VIV) was prepared by a 1+9 dilution of the stock solution with phosphate buffer (0.1 M pH 7.0). The platinum wire was then placed in nitrogen-saturated glutaraldehyde in a stoppered flask for 1 h while nitrogen was bubbled through at 10-min intervals to maintain anaerobic conditions. After thorough washing in fresh phosphate buffer the wire was placed in glucose oxidase (10 mg cm-3) in the phosphate buffer and nitrogen deoxygenation continued as for the first hour. Before use the platinum wire was immersed in the enzyme solution overnight at 4 “C.The scheme for the platinum wire anodisation and subsequent glucose oxidase immobilisation is shown in Scheme 1. Apparatus The cyclic voltammetric approach to activating the platinum wire was carried out with a potentiostat (Bruker E130M). The detector cell (Metrohm EA1102) was based on a three-electrode assembly incorporating a silver - silver chloride reference electrode a glassy carbon auxiliary electrode and a platinum wire working electrode (length 2.5 cm diameter 0.1 cm). All voltammograms were recorded on an X - Y recorder (Omnigraphic Model 2000) 1236 Y lu a 20 ANALYST NOVEMBER 1986 VOL. 111 -OEt Activation 3-Aminopropyl I Pt - Pt-0 Pt-0- Si -0Et I Anodisation -triethoxysilane Pt - 0-Si-OEt I I II I I (CH2)3 N CH (CH2)3 CH-"Enzyme Scheme 1 Glucose -(CH-213 I i" Glutaraldehyde OEt 1 I I 11 I I Pt- 0 -Si-OEt (CH2)J N CH (CH2)3 CHO The anodic decomposition of hydrogen peroxide a product of the enzyme catalysis was monitored by setting the platinum wire enzyme electrode at +600 mV (relative to Ag - AgC1) and measuring the change in current.The electrode was used in a three-electrode system in a laboratory-built flow-through cell. The cell design was such that the reference (silver wire coated with silver chloride) and auxiliary (platinum wire) electrodes were placed in a stationary solution of saturated potassium chloride and contacted with a flowing buffer stream by means of a T-junction (Fig.1). A slight back-pressure from a potassium chloride reservoir maintained a stationary phase for the reference and auxiliary electrodes whereas the working electrode was positioned in the flowing stream. When not in use the electrode was stored at 4°C in the pH 6.0 phosphate buffer. Electrode potentials were controlled and the currents monitored with a potentiostat (Metrohm VA-detector E611). A Linear y - t chart recorder (Model 500) was used to record the flow injection signals. Sample propulsion was achieved with a four-channel peristaltic pump (Ismatec Model IP-4) and sample injections were made with a manual (PTFE) valve (Tecator). All connecting tubing was of PTFE (nominal i.d. 1.27 mm). Pump pulsation was reduced with a suppressor A-I H /D ? J Fig.1. Amperometric flow-through cell. A Perspex block; B, reference electrode chamber; C auxiliary electrode chamber; D, enzyme electrode chamber; E sample inlet; F sample outlet; G from saturated potassium chloride reservoir; H platinum wire enzyme electrode; I electrode connector; and J silicone-rubber seal situated immediately after the pump. Static noise was con-trolled by earthing the flowing stream immediately after the injection valve. Results Response to Glucose Prior to the calibration of the platinum wire enzyme electrode its optimum pH range and the effect of flow-rate were investigated. For assessing the optimum pH a glucose standard (1 mM) was injected (500 mm3 samples) over the electrode at a buffer flow-rate of 2.5 cm3 min-1.The pH effect was investigated over the range 4.5-8.0 with sample injections at 0.5 pH unit intervals. The resulting peak height versus pH profile (Fig. 2) exhibits a plateau region between ca. pH 5.8 and 6.5. Hence all further work was carried out at pH 6.0. The effect of the flow-rate of the carrier stream at pH 6.0 over the range 0.5-4 cm3 min-1 was investigated relative to the peak height of a glucose standard (1 mM). The resulting peak height versus flow-rate profile (Fig. 3) increased almost linearly until a limiting value was reached for flow-rates above 3.2 cm3 min-1. A more detailed investigation over the flow-rate range 2-4 cm3 min-1 showed a clear optimum response at around 3.2 cm3 min-1 and there was little change for higher flow-rates.Hence a flow-rate of 3.5 cm3 min-1 was adopted in this study. The electrode was calibrated at the optimised pH and flow-rate with glucose standards over the range 0.1-30 mM. Fig. 4 illustrates a typical chart recorder output and Fig. 5 shows the corresponding calibration. The calibration was linear over the range 0.1-10 mM glucose with excellent response times (<25 s) and wash times (<30 s). The linear portion of the calibration graph corresponds to log(current/A) = 0.992 log ([glucose]/~) - 3.94 with a correlation coefficient of 0.999. 2 100 & a E 7f) 120 u) 1 Y lu a 80 I I I 5 6 7 8 Phosphate buffer (100 mM) pH Fig. 2. samples Effect of pH on glucose (1 mM) response for 500-mm3 140 2 $ 100 Y $.- 01 I I I I 1.0 2.0 3.0 4.0 Flow-rateicm3 min-l Fig.3. Effect of flow-rate on glucose (1 mM) response for 500-mm3 samples. 0 Electrode 1 d old; A electrode 4 d old; . electrode 8 d ol ANALYST NOVEMBER 1986 VOL. 111 1237 5 rnin Scan Fig. 4. electrode Chart recorder output for glucose calibration for 3 d old 0-J 4- -1 L 0, a L Y m a -.--2 Log([glucoselimM) Fig. 5. Calibration graph for glucose Discussion Activation of Platinum Wire Platinum electrodes may be rendered more active by pulsing the electrode between anodic and cathodic potentials. 11.12 It has been suggested11 that the pulsing technique roughens the platinum surface. Roughening has been attributed to a re-distribution of surface metal ions as a direct result of the formation and breakdown of platinum - oxygen bonds.Essentially anodic - cathodic treatment results in metal from the electrode surface dissolving on the anodic sweep and a fraction of it being re-deposited on the cathodic sweep. This process is equivalent to surface evaporation and selective condensation to produce a clean fresh metal surface. Cyclic voltammetry was used to monitor the activation of the platinum wire electrode. The changing shape of the voltammogram during activation and a typical voltammetric profile for such a modified platinum electrode are shown in Fig. 6(a). As the fractional coverage of impurities is reduced ~~ 0 0.65 1.3 VoltageN Fig. 6. Cyclic voltammograms illustrating the effect of activation and silanisation of the platinum electrode.(a) Change in shape of the curve during activation; (b) activated platinum wire and (c) silanised platinum wire during activation there is a corresponding increase in oxygen adsorption (beyond +1.1 V vs. Ag - AgCl for the anodic branch) and desorption (ca. +0.6 V vs. Ag - AgCl for the cathodic branch) as depicted in Fig. 6(a) and (b). Attempts to immobilise enzyme on an inactivated platinum wire proved unsuccessful further emphasising the importance of electrode surface modification by activation. The anodisation time recommended by Ya09 (1 h at +2.5 V vs. SCE) was reduced in this work to 5 min without any decrease in the over-all performance of the enzyme electrode. Biegler and Woods13 demonstrated that the maximum oxygen coverage was attained within ca.1 s for an electrode set at >2.2 V (vs. SHE) in sulphuric acid (1 M). Consequently the reduction of anodisation time reported here still provides for the maximum oxygen coverage of the platinum wire. Following anodisation the platinum wire was silanised as described and immediately after silanisation it was examined by cyclic voltammetry [Fig. 6(c)]. The disappearance of the oxygen desorption previously noted at ca. +0.6 V [see Fig. 6(b) and 6(c)] clearly indicates the extensive coverage of the platinum surface by the action of the silanising agent. Response Characteristics The optimum pH for the platinum enzyme electrode lies between 5.8 and 6.5. Other workers have reported a broad pH range of 4.0-7.0 with a maximum response around pH 5.5 for solubilised glucose oxidase.14.15 The extent of such pH shifts is a direct result of the change in the microenvironment of the enzyme and is related to the immobilisation technique and the nature of the support material.The increasing current response to glucose with increasing flow-rate is unexpected (Fig. 3) because as the flow-rate is increased the residence time of the substrate over the electrode decreases and consequently a decrease in current response would be expected. However the observed phe-nomenon may be associated with the proximity of the enzym 1238 ANALYST NOVEMBER 1986 VOL. 111 to the platinum. As the flow-rate is increased there is turbulence and the degree of substrate diffusion to the electrode and of products away from the electrode is increased.An optimum response is reached at a flow-rate of 3.2 cm3 min-1 but a further increase produces no change in the glucose response. The flow-rate might need to be adjusted for real samples such as blood serum as in flow injection systems there can be differences arising from viscosity considerations. Important parameters when considering immobilised en-zymes are the lifetime durability and storage suitability and stability. The modified platinum wire enzyme electrode exhibited a good and reproducible response to variable glucose levels for daily use with flow injection analysis samples over a period of 10 d after which there was a sudden inactivation of the electrode. For a continuous exposure to glucose achieved by pumping glucose (10 and 2.5 mM) over the electrode at 3.5 cm3 min-1 the electrode functioned well for 9 h before inactivation.The relatively short over-all lifetimes compared with those observed for an electrode with glucose oxidase immobilised on nylon mesh16 (loading 22 nmol cm-2 min-1) may relate to the enzyme loading on the wire (5-10 nmol cm-2 min-I) small surface area and/or weakness in the nature of the Pt - 0 bonding. Masoom and Townshendl0 reported glucose oxidase activity of up to 1 year for 3-aminopropyltriethoxysilane - glutaraldehyde - glucose oxidase on controlled-porosity glass (CPG) which has a much larger usable surface area. The large surface area of CPG used in a reactor provides a highly active immobilised system so that some loss of activity has a negligible effect on glucose response.The modified platinum wire carries a thin layer of immobilised enzyme but there is the advantage that the sensor is in dwelling. However an additional factor in reducing electrode lifetime is the setting of the electrode at an anodic potential that may facilitate desorption of the oxygen, resulting in the premature loss of enzyme activity by the breaking of the Pt - 0 bonds which are however weaker than the Si - 0 bonds of controlled porosity glass. The authors thank the Department of Trade and Industry (Laboratory of the Government Chemist) for financial sup-port. Thanks are also extended to Mrs. Geraldine Alliston of the Laboratory of the Government Chemist for very helpful discussions. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Clark L. C. and Lyons C. Ann. N . Y. Acad. Sci. 1962 102, 29. Guilbault G. G. Ion-Sel. Electrode Rev. 1982 4 187. Karube I. and Suzuki S . Ion-Sel. Electrode Rev. 1984,6,15. Lehhart J. R. and Murray R. W. J. Electroanal. Chem., 1977 78 195. Murthy A. S. N. and Reddy S . Electrochim. Acta 1983,28, 473. Ianniello R. M. Lindsay T. J. and Yacynych M. Anal. Chem. 1982 54 1980. Cass A. E. G. Davis G. Francis G. D. Hill H. A. O., Aston W. J. Higgins I. J. Plotkin E. V. Scott L. D. L. and Turner A. P. F. Anal. Chem. 1984 56 667. Albery W. J. and Bartlett P. N. J. Chem. SOC. Chem. Commun. 1984 234. Yao T. Anal. Chim. Acta 1983 148 27. Masoom M. and Townshend A. Anal. Chim. Acta 1984, 166 111. Woods R. Electroanal. Chem. 1976 9 9. Gilman S . Electroanal. Chem. 1967 2 111. Biegler T. and Woods R. J. Electroanal. Chem. 1969 20, 73. Bright H. J. and Appleby M. J. Biol. Chem. 1969 224, 3625. Weibel M. K. and Bright H. J. J. Biol. Chem. 1971 246, 2734. Moody G. J. Sanghera G. S . and Thomas J. D. R. Analyst, 1986 111 605. Paper A61141 Received May 12th 1986 Accepted June 16th 198