Anal. Chem. 1996, 68, 2705-2708
Screen-Printable Sol-Gel Enzyme-Containing Carbon Inks Joseph Wang,* Prasad V. A. Pamidi, and Deog Su Park
Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, New Mexico 88003
Enzymes usually cannot withstand the high-temperature curing associated with the thick-film fabrication process and require a separate immobilization step in connection with the production of single-use biosensors. We report on the development of sol-gel-derived enzyme-containing carbon inks that display compatibility with the screenprinting process. Such coupling of sol-gel and thick-film technologies offers a one-step fabrication of disposable enzyme electrodes, as it obviates the need for thermal curing. The enzyme-containing sol-gel carbon ink, prepared by dispersing the biocatalyst, along with the graphite powder and a binder, within the sol-gel precursors, is cured very rapidly (10 min) at low temperature (4 °C). The influence of the ink preparation conditions is explored, and the sensor performance is evaluated in connection with the incorporation of glucose oxidase or horseradish peroxidase. The resulting strips are stable for at least 3 months. Such sol-gel-derived carbon inks should serve as hosts for other heat-sensitive biomaterials in connection with the microfabrication of various thickfilm biosensors. The microfabrication of enzyme-based electrochemical biosensors commonly relies on thick-film technology.1,2 For example, disposable screen-printed enzyme strips are widely used by diabetic patients for self-monitoring of their blood glucose levels.3 Such a microfabrication route offers high-volume production of extremely inexpensive and yet highly reproducible single-use enzyme electrodes.4,5 It involves printing of different electrode patterns (through a predesigned mask) on the surface of ceramic or plastic substrates. While various conducting inks are viable for this task, carbon-based ones are commonly used due to electrochemical and economical considerations. The printed films are commonly baked at elevated temperatures to drive off the solvent and cure the patterned ink deposit. Because of the high temperatures used in this curing step, enzymes are commonly immobilized after the printing and firing processes.2 A UVpolymerizable material has been suggested recently for facilitating the room-temperature fabrication of enzyme-containing inks in connection with a photocuring step.6 The present note describes a novel enzyme-containing carbon ink, derived from sol-gel materials, for the low-temperature mass production of thick-film amperometric biosensors. Sol-gel is a (1) (2) (3) (4) (5) (6)
Prodenziuati, M., Ed., Thick-Film Sensors; Elsevier, Amsterdam, 1994. Alvarez-Icaza, M.; Bilitewski, U. Anal. Chem. 1993, 65, 525A. Green, M.; Hildrich, P. Anal. Proc. 1991, 28, 374. Craston, D.; Jones, D.; Williams, D.; El Murr, N. Talanta 1991, 38, 17. Liu, C. C. Appl. Biochem. Biotechnol. 1993, 41, 99. Rohm, I.; Kunnecke, W.; Bilitewski, U. Anal. Chem. 1995, 67, 2304.
S0003-2700(96)00159-X CCC: $12.00
© 1996 American Chemical Society
low-temperature technology for the production of ceramic materials through the formation of colloidal suspensions of metal oxides.7,8 Because of its low-temperature preparation, the solgel process represents an attractive route for the immobilization of biomolecules.9-12 Sol-gel-derived carbon composite13 and carbon biocomposite14 working electrodes have been introduced recently by Lev and co-workers. Yet, to our knowledge, enzymecontaining screen-printable inks based on sol-gel matrices have not been described. The newly developed biogel carbon ink obviates the need for high-temperature curing and time-consuming immobilization steps. It is prepared by mixing the sol-gel precursors, with the enzyme, graphite powder, and a binder. The characteristics and advantages of such a coupling of sol-gel and screen-printing processes are reported in the following sections. EXPERIMENTAL SECTION Apparatus. Experiments were performed with a Bioanalytical Systems (BAS) Model CV-27 voltammetric analyzer, in connection to a BAS X-Y-t recorder. Amperometric measurements were carried out in a 10-mL electrochemical cell (Model VC-2, BAS). The screen-printed carbon electrode, reference electrode (Ag/ AgCl (3 M NaCl), Model RE-1, BAS), and platinum wire auxiliary electrode joined the cell through the holes in its Teflon cover. A magnetic stirrer and a stirring bar provided convective transport during amperometric measurements. Surface characterization of the screen-printed electrodes was performed using a Hitachi Model S-3200N variable-pressure scanning electron microscope. Reagents. Potassium ferrocyanide, hydroxypropyl cellulose and hydroxypropylmethyl cellulose, cobalt phthalocyanine (CoPC), and acetonitrile (99.9%) were purchased from Aldrich. Glucose oxidase (GOx, EC 1.1.3.4, 150 units/mg) and tetraethoxysilane (TEOS) were purchased from Fluka. Glucose, o-phenylenediamine, hydrogen peroxide, 2-butanone peroxide, horseradish peroxidase (HRP, EC 1.11.1.7, 120 units/mg), uric acid, acetaminophen, and ascorbic acid were from Sigma. Graphite powder (No. 38) was purchased from Fischer Scientific. Palladium (1%)on-graphite was purchased from Johnson Matthey Electronics. Ethanol (200 proof) was obtained from Quantum Chemical Co., while hydrochloric acid (reagent grade) was from J. T. Baker. All solutions were prepared using double-distilled deionized water. (7) (8) (9) (10) (11) (12)
(13) (14)
Brinker, C.; Scherer, G. Sol-Gel Science; Academic Press: New York, 1990. Hench, L.; West, J. Chem. Rev. 1990, 90, 33. Avnir, D.; Braun, S.; Lev, O.; Ottolenghi, Chem. Mater. 1994, 6, 1605. Lev, O.; Tsionsky, M.; Rabinovich, L.; Glezer, V.; Sampath, S.; Pankratov, I.; Gun, J. Anal. Chem. 1995, 67, 22A. Dave, B.; Dunn, B.; Valentine, J. S.; Zink, J. I. Anal. Chem. 1994, 66, 1120A. Narang, U.; Prasad, P. N.; Bright, F. V.; Ramanathan, K.; Kumar, N.; Malhotra, B. D.; Kamalasanan, M. N.; Chandra, S. Anal. Chem. 1994, 66, 3139. Tsionsky, M.; Gun, G.; Glezer, V.; Lev, O. Anal. Chem. 1994, 66, 1747. Pankratov, I.; Lev, O. J. Electroanal. Chem. 1993, 393, 35.
Analytical Chemistry, Vol. 68, No. 15, August 1, 1996 2705
Figure 1. Current versus time recordings for the successive additions of 2 mM glucose at sol-gel screen-printed carbon electrodes doped with (b) and without (a) glucose oxidase. Applied potential, +1.0 V; electrolyte, phosphate buffer (pH 7.4, 0.05 M); stirring rate, 400 rpm. Inset shows calibration plots for glucose strips containing different enzyme loadings: 0 (A), 0.5 (B). 1.0 (C) and 1.5 (D) wt %.
A phosphate buffer solution (0.05 M, pH 7.4) served as the supporting electrolyte. Ink Preparation. Three milliliters of tetraethoxysilane, 0.66 mL of water, 0.66 mL of ethanol, and 50 µL of 1.0 M HCl were mixed thoroughly using a magnetic stirrer for about 1 h to make a clear sol-gel (of pH 5.0). A 300-µL portion of the enzyme solution (containing 30 mg of GOx or HRP) was added to 1.0 mL of the sol-gel, and the mixture was sonicated for 1 min. Subsequently, 0.80 g of graphite powder was added to this mixture, and the new mixture was hand-mixed and then sonicated for 1 min. To this mixture was added 500 µL of 4% hydroxypropyl cellulose to increase the viscosity, and this mixture was sonicated for another 5 min. The entire composite was cooled for 1 h prior to the screen-printing. A similar preparation scheme was employed for the Pd-doped sol-gel ink, in connection with Pd-ongraphite particles. The CoPC-based enzyme sol-gel carbon ink was also prepared in a similar fashion by adding 100 mg of CoPC prior to the graphite addition. Electrode Fabrication. A semiautomatic screen-printer (Model TF-100, MPM Inc., Franklin, MA) was used for fabricating the working electrodes. The ink was printed onto alumina ceramic substrates (3.4 × 10.0 cm) to yield 10 strips of 3.4 cm × 1.0 cm with 0.15- × 2.8-cm working electrode structures. The thickness of the resulting electrode was 25 µm. The electrodes were subsequently dried in the refrigerator (at 4 °C) for 10 min. A portion of their surface was covered with a 20-µm-thick nail-polish insulating layer, leaving 0.05-cm2 areas on both ends for defining the working electrode and the electrical contact. Unlike most commercial insulators, this type of insulator is compatible with the low-temperature curing. RESULTS AND DISCUSSION The characteristics and advantages of the new biogel carbon inks are illustrated below in connection to the thick-film fabrication of amperometric biosensors for glucose and peroxide species. Figure 1 displays typical amperometric responses for successive additions of 2 mM glucose, as obtained at sol-gel derived strips, fabricated in the absence (a) and presence (b) of GOx. No response is observed in the absence of biocatalytic activity. In 2706 Analytical Chemistry, Vol. 68, No. 15, August 1, 1996
Figure 2. (A) Eight biogel-based strip electrodes produced by screen-printing process. (B) SEM image of such strip electrodes.
contrast, the screen-printed enzyme electrode responds very rapidly to the change in the substrate concentration. Steady-state currents are attained within ∼10 s, and linearity prevails up to 16 mM. Such a fast response is attributed to the absence of supporting membrane and the porous character of the printed surface. The favorable signal-to-noise characteristics allow convenient quantitation over this range. Also shown in Figure 1 (inset) are calibration plots for sol-gel inks containing different enzyme loadings. The sensitivity increases dramatically upon raising the GOx content (119, 158, and 310 nA/mM for 0.5, 1.0, and 1.5 wt %, respectively). Higher enzyme loadings resulted in the formation of nonuniform solidified biocomposites that were not compatible with the thick-film process. The sensitivity of the new strips compares favorably with those (∼15 nA/mM) of other strips6 or carbon biocomposites15 of similar size, indicating a favorable biocatalytic activity. Electrodes fabricated from biogel-derived inks are mechanically strong and possess an excellent adhesion to the ceramic substrate. Their well-defined appearance is similar to those of conventional screen-printed electrodes (Figure 2A). A closer microscopic visualization revealed distinct, smoothed conductor edges with minimal defects (not shown). Figure 2B shows a typical scanning electron micrograph of the biogel screen-printed electrode. The surface is characterized by a microporous structure of nonuniform particle size and distribution. A similar appearance is common to strips based on conventional carbon inks. Dipping the biogelbased strips in an organic solvent (e.g., acetonitrile) for an extended period (e.g., 3 days) had no effect on their appearance or performance. Various parameters of the sol-gel process, known to influence the biosensor reponse, were examined and optimized. Particular attention was given to the presence of acid or alcohol that may (15) Amine, A.; Kauffmann, J. M.; Patriarche, G. J. Talanta 1991, 38, 107.
Figure 4. Hydrodynamic voltammograms for 2 mM glucose at biogel-derived screen-printed carbon glucose sensors. (A) Plain gel carbon surface; (B) GOx-containing gel carbon surface; (C) GOxcontaining gel Pd-on-graphite surface. Voltammograms were recorded by applying the appropriate potential and allowing the transient current to decay prior to the glucose addition. Other conditions as in Figure 1.
Figure 3. Effect of ethanol (A) and hydrochloric acid (B) concentration during the ink preparation on the amperometric response of 2 mM glucose at screen-printed sol-gel carbon glucose sensors. Other conditions as in Figure 1.
be “hostile” to the incorporated biocatalyst. Changing the ethanol content over the 5-25% (v/v) range had minimal effect on the response (Figure 3A). In contrast, the response decreases slowly upon raising the level of hydrochloric acid between 2.3 and 11.5 mM, and then more rapidly (Figure 3B). Such change is attributed to the loss of enzyme activity. Films prepared using acid levels lower than 2.3 mM exhibit poor adhesion to the surface. Different binders, common to screen-printed inks, were tested, with hyroxypropyl cellulose and hydroxypropylmethyl cellulose offering the most favorable results. A 10-min period (at 4 °C) yielded complete curing and sensitivity similar to that observed with longer periods. Calculation of the apparent MichaelisMenten constant (Km) offers additional insights into the activity of GOx within the screen-printed sol-gel carbon network. A Km value of 38 mM was estimated from the Eadie-Hofstsee plot for a calibration experiment involving 10 4-mM increments in the glucose level. This is higher than the value (26 mM) reported for GOx in solution,16 reflecting mass-transport limitations associated with the entrapment of the enzyme within the growing oxide network. A Km value of 50 mM was reported by Yamanaka et al. for GOx entrapped in a different sol-gel network.17 Changing the stirring rate between 200 and 800 rpm had a negligible effect on the glucose response, reflecting the composite/microelectrode character, associated with the relatively low (∼30 wt %) carbon dispersion. Figure 4 shows the dependence of the response of various solgel-based printed strips upon the operating potential. Steady-state hydrodynamic voltammetry was employed for this task due to large background current in analogous cyclic voltammetric experiments. No glucose response is observed in the absence of biocatalytic activity (A). With the GOx-containing strip, the response to glucose starts at +0.6 V and approaches a maximum value at potentials higher than +0.9 V (B). The dispersion of metalized carbon particles in the sol-gel-derived ink (C) further (16) Bright, H. J.; Gibson, Q. J. Biol. Chem. 1967, 242, 994. (17) Yamanaka, S.; Nishida, F.; Ellerby, L.; Nishida, C.; Dunn, B.; Valentine, J. S.; Zink, J. I. Chem. Mater. 1992, 4, 495.
Figure 5. Current-time recordings for the successive additions of 2 mM hydrogen peroxide (A) and 2 mM 2-butanone peroxide (B) at plain (a) and HRP-containing (b) screen-printed sol-gel carbon electrodes. Applied potential, 0.0 V; electrolyte, phosphate buffer (pH 7.4, 0.05 M) containing 2 mM o-phenylenediamine.
facilitates the detection of glucose, as expected from their catalytic detection toward the peroxide product.18 At this catalytic sensor, the anodic response of glucose starts above +0.2 V, with cathodic detection occurring at lower potentials. The low operating potentials greatly minimize interferences from coexisting electroactive species. For example, the response to 2 mM glucose at +0.3 V was not affected by the addition of 0.2 mM uric acid and suffered from a very small (∼5%) interference from 0.2 mM acetaminophen; ascorbic acid, in contrast, still displayed a large interference (not shown). Similar improvements were observed upon incorporating the CoPC catalyst within the GOx/sol-gel carbon strip (not shown). The preparation of the sol-derived biosensor strips results in good electrode-to-electrode reproducibility. Measurements of 2 mM glucose at 10 different bioelectrodes yielded a mean peak current of 588 nA and a relative standard deviation of 11%. (18) Wang, J.; Pedrero, M.; Pamidi, P. V. A.; Cai, X. Electroanalysis 1995, 8, 1032.
Analytical Chemistry, Vol. 68, No. 15, August 1, 1996
2707
Improving or replacing the manual (hand-mixing) ink preparation should lead to more homogeneous inks and hence to better precision. In addition to single-use applications, the sol-gelderived biochips hold great promise as reusable devices. Indeed, these biosensors displayed a remarkably stable response. The long-term stability was tested over a 90-day period, using the same strip, with intermittent usage (every 2-3 days) and dry storage at 4 °C. No apparent change in the response to 2 mM glucose was observed over this period. Improved thermal stability of enzymes upon entrapment in sol-gel networks was reported by Avnir and co-workers.19 We tested also the incorporation of horseradish peroxidase (HRP) within the sol-gel-derived carbon ink. Figure 5 compares current-time recording at the ordinary (a) and HRP-containing (b) sol-gel carbon strips on successive additions of 2 mM hydrogen peroxide (A) and 2-butanone peroxide (B). While the HRP sensor offers convenient quantitation of its peroxide substrates, no response is observed in analogous measurements at the plain sol-gel carbon surface. A similar repsonse was obtained after dipping the HRP strip in acetonitrile for a 48-h period. (19) Braun, S.; Rappoport, S.; Zusman, R.; Avnir, D.; Ottolenghi, M. Mater. Lett. 1990, 10, 1.
2708
Analytical Chemistry, Vol. 68, No. 15, August 1, 1996
In conclusion, we have demonstrated that biogel-based carbon inks are very compatible with the screen-printing microfabrication process. Such coupling of sol-gel and screen-printing technologies offers a one-step mass production of enzyme-containing strips and addresses deactivation problems encountered during the thermal curing of thick-film sensors. The performance of the resulting biosensors compares favorably to that of strips fabricated with conventional carbon inks. This ink preparation strategy can be readily expanded to the incorporation of other heat-sensitive biomaterials. ACKNOWLEDGMENT The authors acknowledge Prof. F. V. Bright for his valuable suggestions. D.S.P. acknowledges a fellowship from the Korea Science and Engineering Foundation for the postdoctoral programs (1995). Received for review February 19, 1996. Accepted May 13, 1996.X AC960159N X
Abstract published in Advance ACS Abstracts, June 15, 1996.