Biosensors & Bioelectronics 7 (1992) 361-365

Glucose biosensor based on carbon black strips Plamen Atanasop,

Anastasia

Kaishevaa, llia Ilk@, J uozas Ku I&

Valdemaras

Rawma&*

81

Wentral Laboratory of Electrochemical Power Sources ofthe Bulgarian Academy of Sciences, Sofia 1113, Bulgaria ‘Institute of Biochemistry of the Lithuanian Academy of Sciences, Mokslininku 12, 2600 Vilnius, Lithuania (Received 20 June 1991; revised version received 24 September

1991; accepted 22 October 1991)

Abstract: Amperometric biosensors for the determination of /&o-glucose have been constructed. They were based on a porous matrix of carbon blacks ‘Ketjenblack’ (KB) and ‘Shawinigan black’ (SB) wet-proofed with polytetrafluorethylene. Glucose-sensitive elements were prepared by subsequent adsorptional immobilization of l,l’-dimethylferrocene (DMFc) and nickelocene (NC) on ‘Shawinigan black’ or tetracyanoquinodimethane (TCNQ) on ‘Ketjenblack’ together with Penicillium chtysogenum glucose oxidase. Maximum surface concentrations of DMFc, NC and TCNQ on carbon black electrodes were 95, 116 and 151 nmol cmm2. The biosensor based on KB and TCNQ (KB-TCNQ biosensor) could be used at a potential of O-5 V (vs. Ag/AgCl reference electrode) in the concentration range up to 7 mM. This biosensor possessed an approximately ten times higher sensitivity than the ones based on SB and DMFc (SB-DMFc biosensor) and on SB and NC (SB-NC biosensor) which acted at 0.3 V and 0.05 V, respectively. The biosensors were suitable for practical use longer than one week. Keywords: enzyme electrode, glucose, chemically modified dimethylferrocene, nickelocene, tetracyanoquinodimethane.

INTRODUCTION

More than two thousand publications concerning glucose biosensors are known. The first generation of electrochemical sensors of this type was based on glucose oxidase and an oxygen electrode (Updike & Hicks, 1967). In the second generation hydrogen peroxide the product of the enzymatic reaction - was detected amperometrically on platinum or other

*To whom correspondence

should be addressed.

0965-5663/92/$05.00 @ 1992 Elsevier Science Publishers

carbon

black,

noble metal electrodes (Clark, 1970; Guilbault & Lubrano, 1973). To achieve higher sensitivities special multi-layer membranes were used in these biosensors. The most famous commercializations of these biosensors were done by Miles Laboratories (Clemens & Chang, 1977) and YellowSprings, Inc. (Dray et al., 1977). Glucose biosensors of the third generation were prepared by screen-printing technology using chemically modified graphite as electrode material (Matthews et al., 1987). Inexpensive materials (graphite, organic mediators) and modem printing technologies allowed the preparLtd.

361

I? Atanasov

ation of cheap single-use biosensors for blood glucose monitoring. Carbon black is another carbon material with very interesting physicochemical properties (Besenhard & Fritz, 1983; Razumas et al., 1984). Carbon black powder was used for the construction ofa different kind ofbiosensor based on dehydrogenases and cytochrome bz (Kulys etal., 1987; StaSkeviEiene et al., 1991). The aim of our work was: (1) creation of electrocatalytically active surfaces comprising hydrophobic chemically modified carbon black and glucose oxidase (GOD); (2) preparation of an Ag/AgCl reference electrode based on carbon black; (3) development and investigation of macro-kinetic properties of strip glucose biosensors. Common mediators - tetracyanoquinodimethane (Cenas & Kulys, 1981) and l,l’-dimethylferrocene (Cass et al 1984) - and a new mediator - nickelocene - were used as modifiers. A new type of GOD from Penicillium chlysogenum was also used for the preparation of biosensors.

EXPERIMENTAL, Chemicals and materials 1.1.3.4) isolated from Penicillium (Pharmachim-CBS, Sofia, Bulgaria) and with an activity of 200 U mg-’ (purification degree 95% molecular weight 150 kDa). l,l’-Dimethylferrocene (DMFc), dicyclopenthadienylnickel - nickelocene (NC) - and tetracyanoquinodimethane (TCNQ) were obtained from Aldrich Chemical Co., Inc. (Milwaukee, WI, USA). Glucose, ethanol, acetonitrile, argon, helium and other reagents were of ‘reagent grade’ class. For the preparation of enzyme and reference electrodes carbon blacks ‘Ketjenblack’ (KB) (ACKO Chemie, Amsterdam, Netherlands), ‘Shawinigan black’ (SB) (Gulf oil Chem. Co., Shawinigan Products, Englewood Cliffs, NJ, USA) and acetylene black (AB) (V.V.B. Agrochemie und Zwichenproducte, Halle, Germany) were used after the hydrophobization by polytetrafluorethylene (Iliev er al., 1977). The hydrophobization degree expressed in terms of the weight/weight ratio of polymer: carbon black was 19%, 7% and 35% for KB, SB and AB, respectively. GOD

(EC

chrysogenum

362

Biosensors & Bioelectronics

et al.

1

/ hl8uhtbf’

cuw&m

WE \\

(B)

/

mped

+=%wc

plate

Fig. 1. Schematic representation of glucose biosensors: (A) top view and (B) side view. WE = working electrode, RE = reference electrode, Cu = copper, AB = acetylene black, PVC = polyvinyl chloride.

Enzyme electrode preparation The enzyme electrode (Fig. 1) comprised two layers of carbon blacks, consistently piled up and pressed at 216 MPa pressure and 613 K. ‘Ketjenblack and ‘Shawinigan black were used for the upper (working) layer, and acetylene black for the bottom (non-working) layer. Weight ratio AB/KB (or SB) was 5 : 1. After the compression the plates obtained (thickness 1 mm) were cut into strips (4 X O-5cm). Copper wires were fixed to the electrodes with silver epoxy. These junctions were covered with an isolating varnish resulting in an electrode area of approximately 1 X 0.5 cm (Fig. 1). The electrodes were modified by DMFc, NC or TCNQ by applying 30~1 of 8 mM solutions of DMFc or NC in ethanol or TCNQ in acetonitrile on the working electrode. The solvent was evaporated at room temperature. Afterwards GOD was adsorbed on the modified surface by applying 20 ~1 of a 10 PM enzyme solution in 0.1 M phosphate buffer (pH 7.0, 0.1 M KCl). After drying at room temperature electrodes were washed by the same buffer solution for 30 min and stored in a refrigerator. Reference electrode Five grams of unhydrophobized AB were wetted in a solution of 4 g AgNO3 in 250 ml distilled water by heating at 363 K for 1 h. After cooling and filtration the carbon black with adsorbed AgNOs was dried at 393 K. The dry product was heated at 573 K for 1 h. The hot material was poured out in a hot solution of log NaCl in 250 ml distilled water and after filtration the

Biosensors & Bioelectmnics product was dried at 393 K The dried product was mixed with AB, which was hydrophobized by polytetrafluorethylene to a hydrophobization degree of 35%, in a ratio of 1: I (w/w) and the mixture was pressed at 49 MPa and 293 K. After compression the plates obtained (thickness 1 mm) were cut into strips (4 X 0.5 cm). The copper wire was fixed in the same way as described for the enzyme electrodes. The junction was covered with isolating varnish, resulting in an electrode area of 1 X 0.5 cm. The prepared carbon black (AB) Ag/AgCl electrodes had a potential (E) of 286 (+2) mV versus a normal hydrogen reference electrode in 0 1 M phosphate buffer solution pH 7.0, containing 0.1 M KCl.

Biosensor construction and electrochemical measurements The prepared enzyme electrode and the reference electrode were stuck to a polyvinyl chloride (PVC) plate (Fig. 1). The distance between the enzyme and reference electrode was 1 mm. Electrochemical measurements were done in a two-electrode circuit using a glass cell (10 ml) thermostated at 298 K(+O- 1 K) and a potentiostat EP 22 with E reamer generator EG 20 (Eplan, Lubawa, Poland) or polarograph OH-105 (Radelkis, Budapest, Hungary). The electrochemical experiments were performed in 0.1 M phosphate buffer pH 7.0, containing 0.1 M KCl. All potentials are referred to the carbon black Ag/ AgCl electrode. To measure the dependence of the steady-state current (I,,) on glucose concentration (cs) the biosensor was immersed into the buffer solution and a fixed potential was applied to the electrodes for 5-10 min until a constant residual current, Iss, was established. After this pretreatment a 3 M solution of glucose in buffer was introduced into the cell and the current increase was recorded. The solution was mixed with a magnetic disc stirrer (rotation speed 300 rev min-‘).

RESULTS AND DISCUSSION Figure 2 presents cyclic voltammograms of SB-DMFc, SB-NC and KB-TCNQ-based electrodes after ten potential scans in pure buffer solution. As it is seen the 1-E curve of the SB-DMFc electrode has two maxima in the anodic and cathodic

Glucose biosensor

-1.5 1

-0.3

0

POTENTIAL(Vvs.

0.3

0.6

Ag/AgCl)

Fig. 2. Cyclic voltammograms of SB-NC (curve I), SB-DA4Fc (curve 2) and KB-TCNQ (curve 3) electrodes in 0.1 M phosphate buff@ pH 7.0, containing 01 M KCI. Potential scan rate = 30 m Vs? anaerobic conditions, 298 K.

region at 0.27 V (Ei) and 0.18 V (Ez), respectively. On the basis of these results the calculated (E” = (Ei + EC,)/2) formal redox potential of DMFc is 0.225 V, which is 167 mV more positive than the value known for DMFc+/DMFc (Cass etal., 1984). The difference between these two values can be explained by multi-layer tilling of the surface and by higher solubility of DMFc+ (Bard &Faulkner, 1980). The amount of electricity Q necessary for the oxidation of DMFc adsorbed on electrode was 4.6 mC (determined from the area under the anodic wave). From these data the apparent surface concentration I of DMFc was found to be 95 nmol cm-* using l? = QInFA where

n = 1

(1)

(the

number

of

electrons),

F = 96485 C mol-’ (Faraday constant) and A = 0.5 cm* (the apparent surface area of electrode).

From the values of the peak potentials of the SB-NC electrode (Ei -35 mV, Ei -0.2 V) the calculated E” of a pair of Nc+/Nc is found to be -0.118 V. Q necessary for the oxidation of adsorbed NC was 5.6 mC and I was calculated to be 116 nmol cm-*. The voltamperometric curve of the KB-TCNQ electrode possessed two anodic and two cathodic maxima at -0.03 V, 0.59V and at -0*39V, -O+tV, respectively. According to these values the more anodic process is characterized by E’: = O-315 V and reflects redox conversions of the TCNQ”/I’CNQ-• pair. The more cathodic process had a formal potential of Eq = -0.20 V and can be attributed to the electrode reaction of the TCNQ-•/TCNQHpair. According to Sharp 363

P. Atanasov et al.

Biosensors & Bioelectronics

(1976) E’: was 74 mV and Eq = -176 mV. The observed differences are most probably caused by the structure of the surface modifier layer and the slow flux of counterions and protons (Inzelt, 1986) in the electrochemical process TCNQ’ + e- + K+ = K+TCNQ-•

(2)

K+TCNQ-•

(3)

+ e- + H+ = K+TCNQH-

The reduction process of reaction (2) at the KB-TCNQ electrode is characterized by Q = 7.3 mC and P was calculated to be 151 nmol cm-*. The given value and P for SB-DMFc and SB-NC electrodes are undoubtedly raised because the real A value for carbon blacks considerably exceeds O-5 cm*. The adsorption of GOD at the surfaces of the modified electrodes did not lead to a considerable change in their cyclic voltammograms. However, in contrast to electrodes, containing no enzyme, they generated steady-state anodic currents at 03, 0.05 and 05 V, respectively, for SB-DMFc, SB-NC and KB-TCNQ biosensors after the introduction of glucose into the buffer solution. The time for reading 95% of the steady-state current did not exceed 20 s. Figure 3 presents the dependence of Zss (with subtraction of Zssin buffer solution) for different biosensors on glucose concentration. The elimination of dissolved oxygen from the solutions by bubbling with argon or helium for 30 min increased Z,, no more than 10% in the whole investigated interval of glucose concentrations. This fact indicates that adsorbed GOD is mainly

oxidized TCNQ’:

by DMFc+,

NC+ or TCNQ-•

and

,_ k$ S-GOD & P + GOD,,’

S+GGD

I

GODXd + M,, k? GOD-M 5 3

+

(4)

GOD,,

Mred

where S = glucose, P = gluconolactone, indices ‘ox’ and ‘red are used for the oxidized and reduced state of the active centre of GOD and of the mediator (M), and M,, = 2DMFc+, 2Nc+, 2TCNQ-• or TCNQ’. Mred was oxidized again at the carbon black electrodes. The data of Fig. 3 were linearized in the coordinates I,‘-& and resultant dependences are expressed by the following equations for SBDMFc, SB-NC and KB-TCNQ biosensors, respectively: I,-,’ @A-‘)

= 0.12 (+0.02) + 0.68 (fO.12) c;’ (mM-‘), r = O-975

Z,-,’@A-‘)

(6)

= 2.3 (f2*0)10-* + 0.61 (f0.01) ci’

(rnM_‘),

r = 0.938

Z,’ (PA-‘)

(7)

= 3.36 (f4*33)10-3 + 0.08 (kO.02) c;’

(rnM_‘),

r = O-953

(8)

where r = correlation coeff’cient. The linearization of the experimental data in reverse coordinates enables the conclusion that the action of enzyme electrodes was limited by the rate of processes (4) and (5). Following the GOD action scheme (eqns. (4) and (5)) Z;’ can be calculated (Ikeda et al 1984): GLUCOSE-CONCENTliATION(miij Fig. 3. Dependences of steady-state currents of SB-NC-GOD (curve I). SB-DMFc-WD (curve 2) and KB-TCNQ-GOD (curve 3) sensors on glucose concentration in 0.1 h4 phosphate bufler pH 7.0. containing O-1MKCI. Medium values offour newly prepared sensors at 0.05 (curve I), 0.3 (curve2) and 0.5 V (curve 3). Background I, = 0.5 /.tA (curves 1 and 2) and 6 pA (curve 3). Disc stirrer rotation speed = 300 rev min-t, 298 K. 364

G’

=

+

K2 + zgaXCM

(GYT’

K’

ZFXCS

(9)

where

cx =

nFAcGODk2k4 k2 + k4

(10)

Biosensors & Bioektronics

k4 K, = ~ k,+k,

k_, + kz kl

k2 k_3 + k4 K2 = k2 + k4 k3

Glucose biosensor

(11) (12)

According to eqns (6) to (8) and (9) the parameters cF/Kl for the DMFc-, NC- and TCNQ-modified enzyme electrodes were close to 1.5, 1.6 and 12.5 PA mM_‘, respectively. The difference between these values can be accounted for exceptionally by surface area and enzyme surface concentration, because ex/Kt = nFAcGoDkl kzl (k, + k2). The hydrophobization of the SB-carbon black was 2.7 times lower than that of the KB-carbon black and it is unlikely that the real surface area of the KB-TCNQ electrode considerably exceeds the active surface area of the electrode based on SB-carbon black. Hence, it is concluded that the increase in eXIK, value for the KB-TCNQ-based biosensor was defined by a higher adsorption of GOD on KB-carbon black. As the two values of csaXIKt for the SB-DMFc and SB-NC biosensors were very similar it can be concluded that prepared biosensors acted in the kinetic regime because the reduction constant calculated from the expression kred = klk2/(k-I + k2) does not depend on the nature of the mediator. Among the investigated biosensors the SB-Ncbased biosensors acted at the lowest potential. This has some advantages for such type of biosensors for the elimination of influences of interfering species in the real glucose assay. All three enzyme electrodes are sufficiently stable while kept in a refrigerator. The response was decreased no more than 49% in the period of the first 48 h. Later they remain stable for a week. Most likely the higher solubility of the oxidized mediators and GOD desorption account for the poor stability of the biosensors. REFERENCES Bard, A. J. & Faulkner, L. R (1980). Electrochemical Methd, Fundamentals and Applications, John Wiley, New York Besenhard, J. 0. & Fritz, H. P. (1983). The electrochemistry of black carbons. Angew. Chem. Znt. Ed Engl., 22, 950-15.

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. & Turner, A. P. F. (1984). Ferrocenemediated enzyme electrodes for amperomettic determination of glucose. Anal. Chem.. 56,667-7 1. &as, N. K. & Kulys, J. J. (198 1). Biocatalytic oxidation of glucose on the conductive charge transfer complexes. Bioelectrochem. BioeneG, 8, 103-13. Clark, L. C., Jr (1970). Method and apparatus for polarographic analysis. US Patent No. 3539,455. Clemens, A H. & Chang, P. H. (1977). Continuous blood glucose analysis and feedback control dynamics for glucose controlled insulin fusion (artificial beta-cell). Paper presented at the XIVth International Congress of Therapeutics, Montpellier, France, L’Expansion Scientitique Francaise (publisher), pp. 45-58. Dray, D. N., Keyes, M. H. & Watson, B. (1977). Immobilized enzymes in analytical chemistry. Anal. Chem., 49, 1067A-78A. Guilbault, G. G. & Lubrano, G. J. (1973). An enzyme electrode for the amperometric determination of glucose. Anal. Chim. Acta. 64,439-55. Ikeda. T.. Katasho, I., Kamei. M. & Senda, M. (1984). Electrocatalysis with a glucose-ox&se-immobilized graphite electrode. Agric. Biol. Chem., 48,1969-76. Iliev, I., Kaisheva, A, Gamburzev, S., Vakanova, E. & Budevski, E. (1977). Method for producing powdered wet proofed material useful in making gas-diffusion electrodes. US Patent No. 4,031,033. Inzelt, G. (1986). The influence of experimental conditions on the cyclic voltammetric response to multilayer surface modified electrodes. I. The effect of film thickness, temperature, and sweep rate. Acta Chim. Hung., 122, 187-202. Kulys, J. J., C&as, N. K. & Kanapieniene, J. J. (1987). Electrochemical regeneration of NAD+ on chemically modified electrodes and its application to biosensors. Studia Biophys., 119, 175-8. Matthews, D. R, Holman, R R, Bown, E., Steemson, A, Watson, A, Hughes, S. & Scott, D. (1987). Pensized digital 30-second blood glucose meter. L.ancef April, 778-9. Razumas, V. J., Jasaitis, J. J. & Kulys, J. J. (1984). Electrocatalysis on enzyme-modified carbon materials. Bioekchvchem. Bioenerg., 12,297-322. Sharp, M. (1976). Studies of solid state ion-selective electrodes prepared from semiconducting organic radical-ion salts. Anal. Chim. Acta, 85, 17-30. StaSkeviBent, S. I., Cdnas, N. K. & Kulys, J. J. (1991). Reagentless lactate electrodes based on electrocatalytic oxidation of flavocytochrome br. Anal. Chim. Acta, 243, 167-71. Updike, S. J. & Hicks, G. P. (1%7). The enzyme electrode. Nature, 214, 986-8.

365

Glucose biosensor based on carbon black strips.

Amperometric biosensors for the determination of beta-D-glucose have been constructed. They were based on a porous matrix of carbon blacks--'Ketjenbla...
489KB Sizes 0 Downloads 0 Views