Anal. Chem. 1994,66, 566-571
Activated Carbon Paste Electrodes for Biosensors Noel Motta and Ana R. Guadalupe' Department of Chemistty, University of Puerto Rico, Rib Piedras Campus, P.0. Box 23346, Rib Piedras, Puetio Rico 0093 1
A reagentlessamperometric glucose biosensor was constructed coimmobilized using glucose oxidase and hydroquinone (HzQ) in carbon paste. The sensor response was studied by amperometry and cyclic voltammetry in quiescent solutions, as well as in a flow injection apparatus. Studies were conducted as a function of surface activation and sensor working conditions such as glucose concentration, storage, aging, and reusability. Surface activation proved to be useful to improve the electrochemical reversibility of the mediator and the analytical characteristics of the sensor. The harsh conditions of the surface pretreatment did not deactivate the enzyme. Results from the assay of a clinical samplegave a glucose concentration value of 76 f 6 mg/dL, which compared favorably with the expected range of 72-88 mg/dL. It has been demonstrated that carbon paste can be used to immobilize enzymes for mediated biosensors. 1-12 They can be prepared in two ways: by mixing the mediator and the paste alone, with the enzyme onto the electrode surface, or with all the components within the paste. These sensors exhibit low background current and faster response time, they are inexpensive and relatively easy to fabricate in different configurations and sizes, and the amount of modifier can be easily controlled. Despite of these advantages, there are various drawbacks associated with CP enzymatic e1ectr0des.l~ Because most of these enzymes prefer a more hydrophilic environment than that actually present in the paste, surface preconditioning is usually necessary. This preconditioning is done by soaking the electrode for several hours, in a buffered solution containing the enzyme substrate, at open circuit or under an applied potential. In addition, although it is possible to expose a fresh surface for each experiment by simply squeezing the paste out of the electrode, this complicates its use in flow systems. The paste is prepared by hand mixing, and there is no guarantee that a homogeneous and identical surface will be exposed every time the surface is renewed. (1) Dicks, J. M.;Aston, W. J.; Davis, G.;Turner, A. P. F. Anal. Chim. Acra 1986, 182, 103-112. (2) Bonakdar, M.; Vilchez, J. L.; Mottola, H. A. J . Elecrroanal. Chem. 1989,266, 47-55. (3) Gorton, L.; Karan, H. I.; Hale, P. D.; Inagaki, T., Okamoto, Y.;Skotheim, T. A. Anal. Chim. Acta 1990, 228, 23-30. (4) Schuhmann, W.; Schmidt, H. L.; Kulys, J. Anal. Lett. 1992,25,1011-1024. ( 5 ) Nader, P. A,; Vives, S.S.; Mottola, H . A. J . Elecrroanal. Chem. 1990, 284,
323-333.
(6) Martin, G. B.; Rechnitz, G.A. A n d . Chim. Acta 1990, 237, 91-98. (7) SQnchez,J.;Ruiliang, Li; Ziling, lu; Li-Huey, Wu; Wang, J. Anal. Chim.Acta
1990, 228, 251-257.
(8) Amine, A.; Kauffmann, J. M. Bioelectrochem. Bioenerg. 1992.28, 117-125.
(9) Sakura, S.;Buck, R. P. Bioelectrochem. Bioenerg. 1992, 28, 387-400. (10) Ikeda, T.; Shibata, T.; Todoriki, S.;Senda, M.; Kiroshita, H. Anal. Chim. Acta 1990, 230, 75-82. (1 1) Ikeda, T.; Shibata, T.; Senda, M. J. EIecrroanaL Chem. 1989,261,351-362. (12) Kubiak, W. W.; Wang, J. Anal. Chim. Acta 1989, 221, 43-51. (1 3) Amine, A.; Patriarche, G. J.; Kauffman, J.-M.; Kaifer, A. E. Anal. Lett. 1991, 24, 1293-1315.
566
AnalytlcalChemistry, Vol. 66,No. 4, February 15, 1994
Electrode stability is also limited because the enzyme or the mediator can be lost during both storage or working time. The advantages of using carbon paste as a matrix for the immobilization of enzymes can surpass its drawbacks if the sensors can be made reproducible, reusable, and stable with regard to the surface preparation and conditioning. As pointed out recently by Alvarez-Icaza and Bilitewski,14 this is a necessary condition for the mass production of any kind of biosensor. With this idea in mind, we undertook a study of the effects of surface activation on a CPE biosensor response as a way to achieve reproducibility and reusability. The wellknown system of glucose oxidase (GOx)/glucose with p-hydroquinone (HzQ) as a redox mediator was used as the biosensor. This system was chosen for two reasons: first, previous studies have demonstrated that the electrochemical reversibility of quinones is improved by surface pretreatment of carbon electrodes.15J6 Second, Ikeda and Senda17J8 published a study using p-benzoquinone (BQ) with GOx immobilized behind a dialysis membrane onto the surface of a CPE containing the BQ (30%). The operational potential of this sensor was 500 mV vs SCE in the absence of any surface activation. This sensor exhibited a good linear response with a short response time (20 s); however, it might suffer from more easily oxidized interferences including hydrogen peroxide. The activation of the CPE was done by electrochemical moderately fast scanning in basic solutions (e.g., NaHCO3 or NaOH, 0.5 M) within a given potential interval. The response of the GOx and the H2Q was tested before and after surface activation, either in solution or coimmobilized on the CPE. The studies were conducted in quiescent solutions and under hydrodynamic conditions. The results demonstrate an improvement in the biosensor response, as determined by its dynamic linear range and sensitivity at low redox potential, after surface activation.
EXPERIMENTAL SECTION A. Reagents. GOx (EC 1.1.3.4, Aspergillus niger) and &D-(+)-glUCOSe were from Sigma Chemicals. Spectroscopic graphite powder, (SP-1 grade, Union Carbide) and mineral oil were used for the carbon paste. All other reagents were at least analytical reagent grade. Nanopure water (18 MS2) was obtained through a Barnstead purification train. Monitrol, ES Level I Chemistry Control (B5103-75, Baxter Laboratories) was used to test the biosensor response. (14) Alvarez-Icaza, M.; Bilitewski, U. Anal. Chem. 1993, 65, 525A. (15) Lindquist, J. J. Electroanal. Chem. 1974, 52, 37-46. (16) Urbaniczky, C.; Lundstron, K. J. ElectroaMI. Chem. 1984, 176, 169-182. (17) Ikeda, T.; Hamada, H.; Miki, K.; Senda, M. Agric. Biol. Chem. 1985, 49, 541-543. (18) Ikeda, T.; Hamada, H.; Senda, M. Agric. B i d . Chem. 1986, 50, 883-890.
0003-2700/94/0366-0566$04.50/0
0 1994 American Chemical Society
+ E-
Flgure 1. Carbon paste electrodes: (A) carbon paste; (B) plastic tip; (C) glass tube; (D) brass rod; (E) Teflon holder; d = 0.8 mm; d' = 0.6 mm.
B. Instruments. The electrochemical experiments were done with an EG&G PAR Universal programmer (Model 175) coupled to a potentiostat/galvanostat (Model 173) and XYT-HP 7090A plotter. In order to suppress signal interference coming from the 60-Hz power line and light transformers, a Butterworth low-pass filter was incorporated to the XYT-HP plotting system. Besides noise reduction, this filter amplifies the current by a factor of 2.53; consequently, all currents reported were corrected to reflect the true value without filtering. Spectrophotometric measurements were done in a HP 8452A diode array spectrophotometer interfaced to a HP89500 UV/Vis Chem Station and 7475A HP plotter. An Orion pH/ISE (Model 720) meter was used for pH measurements. A Waters Associates solvent delivery system (Model M-45) and a Rheodyne 7010 injector with a 50-pL sample loop were used to build a conventional flow injection apparatus. C. Experimental Procecures. 1. CPE Preparation. Bare or modified CPE were prepared by the following procedure: for the bare electrodes, the graphite powder was mixed with mineral oil in a ratio of 5 mg of graphite/pL of oil. If a modified electrode was desired, the H2Q was added first, followed by a mixture of the enzyme and oil. The H2Q was added from a methanolic solution in a fixed volume-to-mass ratio given by 100 pL of x M H2Q/50 mg of graphite, where x was the molarity of the H2Q solution,whosevalue depended on the desired H2Q loading in the paste. The graphite and H2Q solution were thoroughly mixed for several minutes, and the methanol was allowed to evaporate in air. The graphiteto-oil ratio was always constant and equal to the bare CPE. Two electrode configurations were used (see Figure 1). The tip configuration was used in a conventional cell under quiescent and convective conditions while the planar configuration was used in the flow system. The corresponding materials and dimensions are illustrated in the diagram. The plastic tips were cut from a Bio-Rad Prot. Elec. tip (BR-42 type) with a razor blade. In both cases, the paste was pressed
into the holder using a wood applicator until it became well compacted without distorting its end circular geometry. The resulting surface was polished over weighing paper until a smooth surface was achieved. Electrical contact was done with a brass rod. The electrodes were stored in 1-5 mM glucose in 0.1 M phosphate buffer (PBS), pH 7.0 at 6 "Cor under dry and dark conditions at 0 OC. Ag/AgCl (3.0 M NaCl) was used as the reference electrode, and a platinum wire as the auxiliary. Potentials are quoted without regard for the liquid junction potential. 2. Activationof the CPE. The CP electrodeswere activated by applying a linearly varying potential between 0.6 and 2.0 V at 5 V/s for a given time in an unstirred solution of 0.5 M NaOH or NaHC03. 3. Determination of Enzyme Activity. The concentration of GOx was expressed as molar units by using the method of Weibel and Bright.19 In this method, the molarity of the catalyticallyactive FAD is determined spectrophotometrically with a differential molar extinction coefficient between the oxidized and reduced bound FAD, C(FAD/FADH~),of 1.31 X lo4M-l cm-l at 450 nm. This molarity is related to the amount of catalytic active GOx by the use of eq 1, where A is the
absorbance difference of the sample without added glucose and with glucose, corrected by dilution, and the C G Ois~ the concentrationof the GOx in thecuvette expressedas milligrams per milliliter. 4. Electrochemical Experiments. The CPE area was determined by chronocoulometry using a 4.0 mM &Fe(CN)6=3H20solution containing 0.1 M KCl. The diffusion coefficient for ferrocyanide under those conditions is 6.50 X cm2/s.20 The potential was stepped from 0 to +500 mV for 250 ms and the area calculated from the integrated form of the Cottrell equation. The same protocol was followed for different activation medium, activation time, and consecutive treatments. The performance of the biosensor was tested by measuring the current from the mediated enzymatic oxidation of glucose at a fixed applied potential. This was accomplished by using either a conventional cell or a flow cell. A fixed potential was applied to a pH 7.0 (PBS, 0.1 M) solution under aerobic conditions. When the background current reached a constant value, successive additions of 100 pL (50 pL in the FIA) of glucose were added from a 0.5-0.1 M glucose stock, also in 0.1 M PBS, pH 7.0. This stock was prepared and allowed to stand 24 h before use to allow equilibration between the a! and @ anomers. For each glucose addition, the catalytic current was determined by taking the differencebetween the measured faradaic current and the background current. 5. Determination of Glucose in Monitrol. The CPE was prepared by mixing 250 nmol of H2Q, 0.16 mg of GOx, and 0.2 pL of mineral oil per milligram of graphite. The sensor geometrical area was 3 X 1P3cm2. The surface was activated in 0.5 M NaOH for 10 min and rinsed extensively with Nanopure water right before use. The assays were performed (19) Weibel, M. K.; Bright, H. J. J. Biol. Chem. 1971, 246, 2734-2744. (20) Sawyer, D. L.; Roberts, J. L., Jr. Experimental Electrochemistry for Chemists; John & Wiley Sons, Inc.: New York, 1974; p 77.
Anal'icalChemistry,
Vol. 66, No. 4, February 15, 1994
567
in a conventional cell under aerobic conditions. The protocol for glucose measurement was as follows: the CPE was potentiostated at 150 mV in 1.O mL of NaHCO3 0.1 M, pH 7.4. The current was monitored as a function of time until the background reached a steady-state value, after which, 1.O mL of Monitrol, reconstituted with 0.1 M NaHC03 (pH 7.4) for 10 min, was added while stirring. The stirring was stopped and the current monitored until it reached a constant value. This procedure was repeated for three 10-pL successiveglucose additions from a 0.5 M stock solution. The analytical signal for the sample and the standard additions was taken as the difference between the current at the plateau and the background signal. Theconcentration ofglucose in Monitrolzl was determined by eq 2, where icatis the catalytic current
(nA), V, is the volume of Monitrol (mL), Vsis the volume of the added glucose stock solution (mL), K is the slope of the calibration plot (nA/mM), C, is the concentration of glucose in diluted Monitrol (mM), and C, is the concentration of the glucose stock (mM). The actual concentration of glucose in Monitrol was obtained from a linear fit of the data resulting from a plot of icat(Vx+ Vs)vs V,.
RESULTS AND DISCUSSION A. Activation of the CPE Surfaces. Previous studies have demonstrated that H2Q is an electrochemically irreversible redox couple at most solid electrodes, including CPE/Nujol, with Up= 200-300 mV between pH 2 and 9.l5 Urbaniczky and Lundstrom16 observed that thermal oxidation of a previously reduced graphite surface improved its electrochemical reversibility while the reduction makes it slower. A closely related compound, commonly known as DOPAC, also improves its electron-transfer rate when the CPE was electrochemically oxidized under neutral pH conditions.22 These authors concluded that a distinct interaction between the oxidized CPE and the H2Q (or DOPAC) might be responsible for this improvement. Few studies on the effect of surface pretreatment on CPE have been published. A significant one was that by Adams and co-workers,22where the effects of chemical and electrochemical oxidation (in neutral media) in the electron-transfer rate of [Fe(CN)613-/" and DOPAC were studied. For both systems the electrochemical reversibility improved upon surface activation, which was attributed to oil depletion from the surface with a concomitant increase in the surface hydrophilicity. As pointed out by McCreery and others,23 several surface variables may be operative in improving the electrochemical reversibility of redox couples. However, it is generally agreed that the removal of adsorbed impurities, the increase of active edge-plane sites and, in some cases, the presence of surface oxides are responsible for the observed surface activation. In view of the current interest in carbon paste as a matrix for the construction of reagentless biosensors, (21) Bader, M. J . Chem. Educ. 1980,57, 703-706. (22) Rice, M. E.; Galus, Z.; Adams, R. N. J . Elecrroanol. Chem. 1983, 143, 89102. (23) McCrecry, R. L. In Electroanalytical Chemistry; Bard, A. J., Ed.;Marcel Dekker: New York, 1991, Vol. 17, pp 221-375.
568
Ana&tical Ghembtry, Vol. 66, No. 4, February 15, 1994
I
I
I
+800
I
I
I
I
I
I
+200
+400
+GOO
0
I
1.55 pA
I
400
I
I
+600
I
I
+400
I
I
+200
I
I
0
E (mV vs Ag/AgCI) Figure 2. Cycllc voltammogram of 4.0 mM K,Fe(CN)d.3H20 In 0.1 M KCI at a CFE (A) before surface actlvatlon; (8)after 5 min of surface activatkn;(C)afteranaddltknal30mhofsurfaceacthratlon.Acthratlon medium, 0.5 M NaHCO,; scan rate, 50 mV/s.
a study was carried out to elucidate the effect of electrochemical pretreatment in the analytical characteristics of a reagentless CPE biosensor for glucose. Initial experiments were conducted to determine the CPE area using &Fe(CN)6 as the electrochemical probe. Figure 2A-C shows the cyclic voltammograms of [Fe(CN),+ at a freshly prepared CPE surface, before and after being treated for 5 and 30 min in a 0.5 M NaHCO3 solution. Several features areevident: (1) therewasasubstantialincreasein thecathodic and anodic peak current, (2) the electrochemical reversibility of [Fe(CN)# was improved after surface activation, and (3) as the pretreatment time was increased, theelectrochemical behavior of [Fe(CN)# approached that for a chemical and electrochemically reversible redox couple, e.g., ipc/ipe= 1.06 and Up= 68 mV compared to the theoretical values of ip,c/ i , , = 1 and Up= 59 mV24for le- redox process. The formal redox potential, taken as Eo' = l / 2 (Ep,a + EpJ, of +214 mV compared favorably with that published in the literature ( E O ' = +0.240 mV, 0.1 M KCl).2 Chronocoulometry was used to determine the electrode area and how this was affected by the surface pretreatment as a function of activation time and medium. These results are summarized in Table 1. A freshly prepared electrode was used for each determination. The activation time is additive; Le., total activation time was 30 min. There was a slight increase in the surface area as the activation time was increased. (24) Bard, A. J.; Faulkner, L. R.Electrochemical Methods: Fundamentals and Applications; John Wiley & Sons, Inc.: New York, 1980; Chapter 6.
I
Tabk 1. Ekctrock Aroa and Ekctrodmnlcal Paramotm of I
