Available online at www.sciencedirect.com

Biosensors and Bioelectronics 23 (2008) 771–779

A novel glucose biosensor based on immobilization of glucose oxidase into multiwall carbon nanotubes–polyelectrolyte-loaded electrospun nanofibrous membrane K.M. Manesh a , Hyun Tae Kim b , P. Santhosh a , A.I. Gopalan a,c , Kwang-Pill Lee a,c,∗ a

b

Department of Chemistry Graduate School, Kyungpook National University, Daegu 702-701, South Korea Department of Sensor and Display Engineering, Kyungpook National University, Daegu 702-701, South Korea c Nano Practical Application Center, Daegu 704-801, South Korea Received 26 April 2007; received in revised form 29 July 2007; accepted 17 August 2007 Available online 30 August 2007

Abstract Nanofibrous glucose electrodes were fabricated by the immobilization of glucose oxidase (GOx) into an electrospun composite membrane consisting of polymethylmethacrylate (PMMA) dispersed with multiwall carbon nanotubes (MWCNTs) wrapped by a cationic polymer (poly(diallyldimethylammonium chloride) (PDDA)) and this nanofibrous electrode (NFE) is abbreviated as PMMA–MWCNT(PDDA)/GOx–NFE. The NFE was characterized for morphology and electroactivity by using electron microscopy and cyclic voltammetry, respectively. Field emission transmission electron microscopy (FETEM) image reveals the dispersion of MWCNT(PDDA) within the matrix of PMMA. Cyclic voltammetry informs that NFE is suitable for performing surface-confined electrochemical reactions. PMMA–MWCNT(PDDA)/GOx–NFE exhibits excellent electrocatalytic activity towards hydrogen peroxide (H2 O2 ) with a pronounced oxidation current at +100 mV. Glucose is amperometrically detected at +100 mV (vs. Ag/AgCl) in 0.1 M phosphate buffer solution (PBS, pH 7). The linear response for glucose detection is in the range of 20 ␮M to 15 mM with a detection limit of 1 ␮M and a shorter response time of ∼4 s. The superior performance of PMMA–MWCNT(PDDA)/GOx–NFE is due to the wrapping of PDDA over MWCNTs that binds GOx through electrostatic interactions. As a result, an effective electron mediation is achieved. A layer of nafion is made over PMMA–MWCNT(PDDA)/GOx–NFE that significantly suppressed the electrochemical interference from ascorbic acid or uric acid. In all, PMMA–MWCNT(PDDA)/GOx–nafion–NFE has exhibited excellent properties for the sensitive determination of glucose like high selectivity, good reproducibility, remarkable stability and without interference from other co-existing electroactive species. © 2007 Elsevier B.V. All rights reserved. Keywords: Composite; Nanofibrous membrane electrode; Carbon nanotubes; Polymethylmethacrylate; Poly(diallyldimethylammonium chloride); Glucose oxidase; Biosensor; Amperometry

1. Introduction Tailoring of the electrodes and fabrication of enzyme-based biosensors has been the active subject over the past few decades (Bakker, 2004; Prodromidis and Karayannis, 2002). Amperometric enzyme electrodes have the combined advantages of specificity of the enzyme for recognizing a particular target molecule and direct transduction rate of reaction into current

∗ Corresponding author at: Department of Chemistry Graduate School, Kyungpook National University, Daegu 702-701, South Korea. Tel.: +82 539505901; fax: +82 539528104. E-mail address: [email protected] (K.-P. Lee).

0956-5663/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2007.08.016

(Amine et al., 1993). Among the amperometric biosensors, the development of glucose biosensor has been given much attention because of its importance in monitoring the blood glucose. The important aspect to consider in the fabrication of amperometric biosensors is to keep the enzyme on the surface of electrode for detecting the active species electrochemically. This has been achieved by immobilization of enzyme on suitable substrate. The immobilization of enzyme could be done in gels, cross-linked polymers, conductive salts or carbon pastes or carbon–organic polymer hosts (Campuzano et al., 2003; Chang et al., 2002; Christophe et al., 2003). For the fabrication of an efficient biosensor, the selection of substrate for dispersing the sensing material decides the sensor performance. It is highly desirable to use the substrates that

772

K.M. Manesh et al. / Biosensors and Bioelectronics 23 (2008) 771–779

would favor effective loading of enzymes to have high sensitivity towards the analyte (D´ıaz and Balkus, 1996). The host material (substrate) having high-surface area, optimum porosity, highthermal stability, chemical inertness and minimum or negligible swelling in aqueous and non-aqueous solutions would be an ideal choice to obtain high performance for the biosensors. Among the host materials, electrospun fibrous membrane meets many of the requirements for achieving improved performance for a sensor electrode (Manesh et al., 2007). Main benefits of fibrous materials include design flexibility, dimensional stability for the flow of gas and liquids through the bundle of fibers, high-surface area, safer operations, easy scaling and reusability. The present study utilized an electrospun polymer membrane having these requirements for the fabrication of a biosensor. Electrospinning has been proved to be an efficient method for generating nanofibrous materials with high surface-to-volume ratios. Electrospun nanofibrous membrane could immobilize biological molecules with higher loading (Lee and Belcher, 2004; Li et al., 2003; Patel et al., 2006). Electrospinning involves drawing of fibers from polymer solutions. In the electrospinning process, an electric field is applied to surpass the surface tension of the polymer and to generate fine jets of the polymer solution through the capillary tip. As a result, ultrathin fibers are formed. The main features of electrospun fibers include large surface area, reusability of the fiber mats due to easy separation from the reaction mixture and retention of catalytic activities (Deitzel et al., 2001; Bognitzki et al., 2001). Biological molecules could be immobilized onto the surface of the electrospun membranes. However, the molecules on the surface of electrospun fibers tend to leach out from the surface when the fibrous mat is placed in a solution. It is therefore essential to have an additional functional material that would bind the biological molecule/enzyme within the fibrous mat and minimize the leaching of the biomolecules/enzyme. The present study focuses on this aspect. The present study utilizes electrospun nanofibrous membrane to induct functionality to have stronger binding with biomolecules/enzymes. Carbon nanotubes (CNTs) with their mechanical properties, high-aspect ratio, electrical conductivity, chemical stability and ability to promote electron transfer reaction find wide range of applications that include fabrication of electrochemical sensors and biosensors (Yao and Shiu, 2007). The electrocatalytic role of CNTs for the electron transfer reaction of biomolecules such as cytochrome c (Wang et al., 2002), ascorbic acid (Lee et al., 2006), dopamine (Zhang et al., 2005), NADH (Musameh et al., 2002), H2 O2 (Santhosh et al., 2006), glucose (Withey et al., 2006) and nucleic acid (Wang et al., 2003) has been well documented and utilized for the development of electrochemical sensors. CNTs have been reported to act as nanowires for the electrical contact of enzyme with the electrode (Yao and Shiu, 2007). The immobilization of biomolecules onto the surface of CNTs results in a new class of biosensors with a few improved performance characteristics (Joshi et al., 2005). Enzymes were effectively immobilized onto the surfaces of CNTs through either physical entrapment or by covalent binding (Lin et al., 2004; Guan et al., 2005; Salimi et al., 2004). Most of the amperometric glucose sensors utilize the specific recognition

of glucose by the enzyme, glucose oxidase (GOx) (Ohnuki et al., 2007). GOx (from Aspergillus or Penicillium) is a homodimer with a molecular weight of about 150–180 kDa. GOx contains two tightly bound flavine adenine dinucleotide (FAD) cofactors and can catalyze the electron transfer from glucose to oxygen with an accompanying production of gluconolactone and hydrogen peroxide. GOx

glucose + O2 −→gluconolactone + H2 O2 GOx (from Aspergillus niger) has an isoelectric point of 4.2 (Swoboda and Massey, 1965), and carries a net positive charge at pHs lower than 4.2. GOx is negatively charged in neutral pHs. Electrostatic interactions between a cationic surfactant/polymer and CNTs were utilized for the fabrication of biosensors. The enzyme, GOx, was bound to CNTs through the electrostatic interaction between the negative charges in the enzyme and the positive charges in CNTs (Cai and Chen, 2004; Liu et al., 2005; Wen et al., 2007). Recently, a layer-by-layer technique, based on the alternate electrostatic adsorption of the negatively/positively charged individual components, has been adopted for the fabrication of CNTs multilayer film on the surface of working electrode. Multiwall carbon nanotubes (MWCNTs) wrapped by positively charged poly(diallyldimethylammonium chloride), PDDA, were assembled by the layer-by-layer deposition of negatively charged GOx. A detection limit of 58 ␮M with the response time of 8 s have been noticed towards glucose (Zhao and Ju, 2006). A gradual increase in the sensitivity toward glucose was noticed up to five bilayers. Zhang et al. (2007) observed the direct electrochemical response of GOx for the electrode fabricated with alternative layers of SWCNTs and PDDA. A sandwich-like structure comprising of PDDA, GOx and MWCNTs has been reported through a self-assembly approach (Liu and Lin, 2006). Importantly, it must be noted that the present study involves the use of electronspun nanofibrous membrane (NFM) for the fabrication of biosensor electrode. The membrane material used for the fabrication of sensor electrode is an entirely different one from the previously known composite systems. We have aimed to utilize the chief characteristics of electrospun membrane such as large surface area, dimensional stabilities upon flow of gases and liquids, reusability, ease of scaling up and reproducible catalytic activities, etc. for achieving the high performance to the biosensor. The porous nature of NFM suits for loading higher extent of immobilization of enzymes. Added to these characteristics, the extent of leaching of enzyme during storage can be minimized and/or biomolecules can be bound to NFM through interaction between functional groups in NFM and MWCNTs. The present work involves the fabrication of a novel amperometric enzyme electrode through immobilizing GOx into an electrospun membrane of polymethamethylacrylate (PMMA) containing a dispersion of cationic polymer-wrapped MWCNTs. PDDA, the cationic polymer, was selected as the cationic polymer to wrap the surface of MWCNTs. The steps involved in the fabrication of the modified electrode are presented in Scheme 1. Firstly, electrospun fibrous membrane of PMMA was prepared with a dispersion of PDDA-wrapped MWCNTs,

K.M. Manesh et al. / Biosensors and Bioelectronics 23 (2008) 771–779

773

Scheme 1. Schematic representation of interaction between PMMA–MWNT(PDDA) and GOx.

MWCNT (PDDA), into the PMMA matrix. The positive charges in PDDA were effectively utilized to immobilize GOx into PMMA–MWCNT(PDDA) fibrous matrix in the fabrication of the modified electrode, PMMA–MWCNT(PDDA)/GOx–NFE (Scheme 1). The PMMA–MWCNT(PDDA)/GOx–NFE was characterized for its electroactivity by cyclic voltammetry and tested for its utility as amperometric glucose biosensor. Thus, the present investigation details the fabrication of PMMA–MWCNT(PDDA)/GOx–NFE and the capability to perform as glucose biosensor. The NFE exhibits low-detection limit, fast response time, high sensitivity for glucose and has a good storage stability. 2. Experimental 2.1. Materials MWCNTs (10–50 nm in diameter, CNT Co. Ltd. Incheon, Korea) were rinsed with double-distilled water and dried. Polymethamethylacrylate (PMMA, MW: 350,000 g/mol), 20% poly(diallyldimethylammonium chloride) (PDDA, MW: 400,000–500,000 g/mol) aqueous solution, d(+)-glucose, glucose oxidase (GOx) (256 U mg−1 ; U: enzyme units), nafion, ascorbic acid (AA) and uric acid (UA) were of analytical grades and used as received. Double-distilled water was used throughout the experiments. Aqueous solutions of glucose were prepared in 0.1 M phosphate buffer (PBS, pH 7) afresh at the time of experiments. Indium-doped tin oxide (ITO)-coated glass plate of 2 cm × 4 cm area (specific surface resistance of about 10 ) was used for making sensor electrode. Before performing each of the experiment, the ITO plate was rinsed with acetone and washed with distilled water. 2.2. Preparation of PMMA–MWCNT(PDDA)/GOx–NFE MWCNTs (50 mg) were dispersed in a 20 mL aqueous solution of 10% PDDA by ultrasonication for 30 min. The solution

was refluxed at 50 ◦ C for 12 h and filtered. The residue was washed with distilled water and dried to get MWCNTs wrapped PDDA, MWCNT(PDDA). Electrospun membrane electrode was fabricated using the methodology, as described elsewhere (Manesh et al., 2007). Typically, an adequate amount of PMMA and MWCNT(PDDA) were dissolved in DMF/acetone mixture (7:3, v/v). Electrospinning of the composite solution was performed at a flow rate of 10 mL/h with a potential difference of 25 kV. A distance of 15 cm was kept between the syringe tip and collector. PMMA–MWCNT(PDDA)–NFE was accumulated over the ITO-coated glass plate. The area of the working electrode was 2 cm × 2 cm. This area was maintained in all measurements. PMMA–MWCNT(PDDA)–NFE was soaked in a GOx solution (10 mg mL−1 ) and incubated overnight at 4 ◦ C for the immobilization of enzyme. The GOx immobilized PMMA–MWCNT(PDDA)–NFE was washed thoroughly with a buffer solution (pH 7) and dried in a vacuum oven. 2.3. Morphological characterization Field emission scanning electron microscopy (FESEM; Hitach-530) and field emission transmission microscopy (FETEM; JOEL JEM-2000EX) were used to study the morphology of the fibrous PMMA–MWCNT(PDDA) membrane. 2.4. Electrochemical characterization of PMMA–MWCNT(PDDA)/GOx–NFE as an amperometric glucose sensor Electrochemical measurements were performed using EG&G PAR Electrochemical Analyzer. A conventional threeelectrode cell assembly was used. PMMA–MWCNT(PDDA)/ GOx–NFE was used as the working electrode. Ag/AgCl and platinum wire were used as reference and counter electrodes, respectively. The electrocatalytic activity of PMMA– MWCNT(PDDA)/GOx–NFE was evaluated using cyclic voltammetry in 0.1 M PBS (pH 7). Amperometric measure-

774

K.M. Manesh et al. / Biosensors and Bioelectronics 23 (2008) 771–779

ments at PMMA–MWCNT(PDDA)/GOx–NFE were carried out at an applied potential of +100 mV (vs. Ag/AgCl) (desired working potential in the case of hydrodynamic voltammetric measurements) for various glucose concentrations in an electrochemical cell containing a magnetically stirred PBS electrolyte. The current response was recorded as follows: after applying the potential (+100 mV vs. Ag/AgCl), the current was allowed to stabilize (after 30 s). Subsequently, aliquots of glucose solutions were added incrementally and the current response was recorded for each an incremental addition of glucose.

3.2. Electroactivity of PMMA–MWCNT(PDDA)/GOx–NFE Electroactivity of PMMA–MWCNT(PDDA)/GOx–NFE was tested by cyclic voltammetry for various scan rates (Fig. 2a). Cyclic voltammograms (CVs) were recorded by scanning the potentials between 0 mV and −600 mV at differ-

3. Results and discussion 3.1. Morphology of PMMA–MWCNT(PDDA) and PMMA–MWCNT(PDDA)/GOx–NFE The electrospun membrane obtained from electrospinning a solution of PMMA-containing MWCNT(PDDA) has a fibrous morphology, as evident from FESEM micrograph (Fig. 1a). The diameter of the fibers is in the range of 200–400 nm. And, there is no abnormal morphology such as beads in the fibers. The fibers are randomly oriented and of several micrometer length. FETEM image of a portion of the fiber (Fig. 1b) clearly reveals that MWCNT(PDDA) is finely distributed within the fibrous PMMA matrix. Such a nanofibrous morphology for PMMA/MWCNT (PDDA) provides a high-surface area for the immobilization of GOx. The wrapping of PDDA over the surface of MWCNTs was done with two main objectives: Firstly, the positive charges on the protonated nitrogen in PDDA chains provide electrostatic repulsions between MWCNT(PDDA) units and hence prevent aggregation of MWCNTs. Thus, a good dispersion of MWCNT(PDDA) within the matrix of PMMA was achieved. Secondly, the positive charges (from PDDA) on the surface of MWCNTs electrostatically bind the negatively charged GOx. After loading the enzyme into PMMA–MWCNT(PDDA) membrane (Fig. 1c), the fiber like morphology of PMMA–MWCNT(PDDA) changed. A layer of GOx was found to cover over the porous matrix of PMMA–MWCNT(PDDA). It is to be remembered that the porous and high-surface area PMMA enables diffusion of GOx to reach the surface of MWCNT(PDDA). The role of PDDA wrapping over MWCNTs on the immobilization of GOx was ascertained by checking the GOx immobilization into an electrospun mat of PMMA loaded with a pristine MWCNTs. The morphology of PMMA/MWCNT(pristine)/GOx (figure not shown) is different from PMMA–MWCNT(PDDA)/GOx. A layer of GOx is not found to be covered over PMMA–MWCNT(pristine) in contrast to the morphology noticed in the case of PMMA–MWCNT(PDDA). Rather, GOx was found at scattered locations in the fiber. Thus, we conclude that wrapping of MWCNTs with PDDA causes higher loading of GOx in the fiber mat. The ITO plate with PMMA–MWCNT(PDDA)/GOx over its surface is fabricated as nanofibrous electrode (NFE). The NFE was characterized for electroactivity and used as a glucose biosensor.

Fig. 1. FESEM (a) and FETEM (b) images of PMMA–MWNT(PDDA) fibrous membrane and FESEM image of PMMA–MWNT(PDDA)/GOx (c).

K.M. Manesh et al. / Biosensors and Bioelectronics 23 (2008) 771–779

ent scan rates in 0.1 M PBS (pH 7). In this potential region, the PMMA–MWCNT(PDDA)/GOx–NFE displays a pair of redox peaks at −386 mV (Epc ) and −351 mV (Epa ) (scan rate: 100 mV s−1 ). The formal potential of the redox process was found to be −365 mV with a peak separation (Ep ) of 35 mV. These features are the characteristics of the reversible redox processes in GOx molecules at the electrode. It is clear that MWCNT(PDDA) acts as a mediator for the electron transfer reaction of GOx. CNTs exhibit electrocatalytic activity due to their inherent features like topological defects in lattices, high-surface active sites and the nanostructure (Li et al., 2002). The anodic and cathodic peak current values were found to show linearity with the scan rates between 10 mV s−1 and 175 mV s−1 (Fig. 2a inset) and the observation suggested that the reaction is a surface-controlled one as expected for immobi-

Fig. 2. (a) Cyclic voltammograms of PMMA–MWCNT(PDDA)/GOx–NFE in 0.1 M PBS (pH 7) at various scan rates, mV s−1 : 10 (i); 20 (ii); 40 (iii); 60(iv); 80 (v); 100 (vi); 125 (vii); 150 (viii) and 175 (ix) (Inset: anodic and cathodic peak currents vs. scan rate). (b) Plot of anodic and cathodic peak potential vs. ln υ.

775

lized systems. Interestingly, when the scan rates were higher than 700 mV s−1 , an irreversible electrode reaction was evident from the shifting of Epa and Epc to the positive and negative directions. For an irreversible electrode reaction, the relationship between peak potential and the scan rate follows the Laviron’s equation (Laviron, 1974, 1979): RT RTks RT ln − ln υ, αnF αnF αnF RT RTks RT Epa = E + ln − ln υ (1 − α)nF (1 − α)nF (1 − α)nF

Epc = E +

where α is the electron transfer coefficient, ν the scan rate and E is the formal potential. E (−365 mV) is estimated as the midpoint between the cathodic and the anodic peak potentials at the low-scan rate. Fig. 2b shows the plot of anodic and cathodic peak potential vs. ln ν. The plots of Ep vs. ln ν were found to be linear at higher scan rates. The electron transfer coefficient (α) was evaluated to be 0.53. The electron transfer rate constant, ks was estimated to be 13.3 ± 0.2 s−1 by substituting the value of α using the Laviron theory and the fitting the line at high-scan rates, The ks for PMMA–MWCNT(PDDA)–G0x NFE is higher than at GOx immobilized on SWCNTs electrode (ks = 1.7 s−1 ) (Elie et al., 2002), Nafion–GOx–MWCNTs composite electrode (ks = 1.53 ± 0.45 s−1 ) (Cai and Chen, 2004) and GOx immobilized on MWCNTs/chitosan composite electrode (ks = 7.73 s−1 ) (Liu et al., 2005). The surface coverage (Γ ) of GOx was estimated to be 8.43 × 10−10 mol cm−2 through the integration of area under cathodic peak at scan rates less than 100 mV s−1 and using the equation, Γ = Q/nFA (Bard and Faulkner, 2001), where Q is the charge, n the electron transfer number, F the Faraday constant and A is the geometric area of the working electrode. The value of Γ is higher than the theoretically predicted value (2.86 × 10−12 mol cm−2 ) for a monolayer of GOx on the bare electrode surface (Xu et al., 2003), GOx immobilized on PDDA-wrapped MWCNTs (4.34 × 10−10 mol cm−2 ) (Wen et al., 2007), GOx immobilized on SWCNTs (2.35 × 10−11 mol cm−2 ) (Zhang et al., 2007), GOx immobilized on silica nanoparticles (0.6 × 10−12 mol cm−2 ) (Sun et al., 2006) and GOx adsorbed on a colloidal gold-modified carbon paste electrode (9.8 × 10−12 mol cm−2 ) (Liu and Ju, 2003). The higher Γ value for GOx loading clearly indicates that electrospun membrane provides a very large active area for enzyme immobilization. The effect of solution pH on the response of immobilized GOx was monitored (Fig. 3). Stable and well-defined CVs were obtained for the pH values from 3 to 9 in N2 saturated PBS at a scan rate of 100 mV s−1 . The decrease of GOx response at higher pHs (pH 8 and 9) is due to the depletion of proton concentration. The decrease in peak current at low pH (pH 3–6) corresponds to the decrease in bioactivity of immobilized GOx. Maximum current was noticed at pH ∼ 7. An increase in solution pH causes a negative shift in E with a slope of −41 mV/pH (Fig. 3 inset), which is comparable to the previous reports, such as −44.5 mV/pH (Zhang et al., 2004) and −43.7 mV/pH (Liu and Ju, 2003). The results are in accordance with two

776

K.M. Manesh et al. / Biosensors and Bioelectronics 23 (2008) 771–779

Fig. 3. CVs of PMMA–MWCNT(PDDA)/GOx–NFE in N2 saturated PBS of various pHs: (a) 8.0, (b) 7.0, (c) 6.0, (d) 5.0 and (e) 4.0 at a scan rate of 100 mV s−1 . Inset shows the dependence of E and peak currents on the solution pH.

protons and two electrons mechanism for the electron transfer processes. GOx–FAD + 2e− + 2H+  GOx–FADH2 3.3. Sensor characteristics of PMMA–MWCNT(PDDA)/GOx–NFE The mechanism of glucose sensing at PMMA–MWCNT (PDDA)/GOx–NFE is based on the amperometric detection of hydrogen peroxide, H2 O2 , by the GOx-catalyzed oxidation of glucose in the presence of dissolved oxygen. Cyclic voltammograms were recorded at PMMA–MWCNT(PDDA)/GOx–NFE in the presence of various concentrations of H2 O2 (figure not shown). PMMA–MWCNT(PDDA)/GOx–NFE exhibits significant electrocatalysis to the oxidation and reduction of H2 O2 starting around 0 mV. In order to authenticate this and to determine the optimum potential for the biosensor operation, hydrodynamic voltammetric studies were carried out with 15 mM H2 O2 in 0.1 M PBS (pH 7). The potential at the working electrode was varied between −400 mV and 400 mV and the transduced current values were noted. Fig. 4 inset shows the amperometric response signals of 15 mM H2 O2 at PMMA–MWCNT(PDDA)/GOx–NFE under different applied potentials in 0.1 M PBS (pH 7). As is seen from Fig. 4 (inset), the oxidation of H2 O2 starts at potentials much more positive than 0 mV, while the reduction starts at potentials more negative than 0 mV. Also, it can be seen that oxidation is more predominant over reduction as inferred from the response current. An operating potential of +100 mV (vs. Ag/AgCl) was chosen in the further experiments to demonstrate the applicability of the biosensor electrode towards the detection of glucose. CVs were recorded at PMMA–MWCNT(PDDA)/GOx–NFE (Fig. 4a) and PMMA/MWCNT(pristine)/GOx electrode

Fig. 4. Cyclic voltammograms of PMMA–MWCNT(PDDA)/GOx–NFE (a) and PMMA–MWCNT(pristine)/GOx electrode (b) in 0.1 M PBS (pH 7) containing 15 mM glucose; scan rate: 50 mV s−1 . Inset shows the hydrodynamic voltammogram for 15 mM H2 O2 at the PMMA–MWCNT(PDDA)/GOx–NFE electrode in PBS (pH 7).

(Fig. 4b) for a solution of 15 mM of glucose in 0.1 M PBS (pH 7) at a scan rate of 50 mV s−1 . It can be seen that the oxidation of the enzymatically formed H2 O2 starts from more positive potential than 0 mV at PMMA–MWCNT(PDDA)/ GOx–NFE. Also, the oxidation current at PMMA–MWCNT (PDDA)/GOx–NFE (Fig. 4a) is found to be higher than at PMMA/MWCNT(pristine)/GOx electrode (Fig. 4b). The oxidation current is maximum at +100 mV at PMMA– MWCNT(PDDA)/GOx–NFE and it is nearly five times higher than at PMMA–MWCNT(pristine)/GOx electrode.The higher current response for glucose at PMMA– MWCNT(PDDA)/GOx–NFE in comparison to PMMA– MWCNT(pristine)/GOx signifies the importance of PDDA wrapping. The existence of PDDA over MWCNTs augments binding of GOx with MWCNTs and hence mediates the electron transfer effectively. 3.4. Amperometric determination of glucose at PMMA– MWCNT(PDDA)/GOx–NFE Fig. 5 shows the amperometric responses of PMMA– MWCNT(PDDA)/GOx–NFE at +100 mV (vs. Ag/AgCl) for the successive addition 1 mM of glucose (amperogram recorded for 1–10 mM glucose concentration is shown). The electrode exhibits a rapid and sensitive current response for the changes of glucose concentration and indicates the excellent electrocatalytic behavior of the electrode. Fig. 5 (bottom inset) shows the plot of current vs. glucose concentration recorded at the PMMA–MWCNT(PDDA)/GOx–NFE. A linear response to current is noticed for a wider concentration range of glucose (20 ␮M to 15 mM) at PMMA–MWCNT(PDDA)/GOx–NFE. The linear least-squares calibration plot over the range 20 ␮M to 15 mM had a slope of 3.7048 ␮A mM−1 with a correlation coefficient

K.M. Manesh et al. / Biosensors and Bioelectronics 23 (2008) 771–779

777

PMMA–MWCNT(PDDA)/GOx–NFE possesses higher enzymatic activity. Thus NFE exhibits a higher affinity for glucose than the other glucose biosensors (Manso et al., 2007; Zhu et al., 2006). 3.5. Response time of glucose detection at PMMA–MWCNT(PDDA)/GOx–NFE PMMA–MWCNT(PDDA)/GOx–NFE shows a response time of ∼4 s for the detection of glucose. We believe that higher loading of GOx, larger active surface area provided by the NFE and porous nature of PMMA fiber are the reasons for the effective mediation of electron transfer to result such a lower response time for glucose detection. 3.6. Interference study

Fig. 5. Amperometric responses of PMMA–MWCNT(PDDA)/GOx–NFE to successive additions of 1 mM glucose; electrolyte: 0.1 M PBS (pH 7). Potential: +100 mV (vs. Ag/AgCl). Inset (bottom) shows the calibration plot of the concentration of glucose with current at PMMA–MWCNT(PDDA)/GOx–NFE (20–900 ␮M (line a) and 1–20 mM (line b); inset (top) shows the Lineweaver–Burk plot.

of 0.9983. The lower detection limit, 1 ␮M, was calculated as the glucose concentration giving a signal equal to the blank signal yB (intercept) plus three standard deviations of y-residuals sy/x . This detection limit is much lower than noticed for the other CNTs based glucose biosensors (Zhao and Ju, 2006; Tsai et al., 2005). A plateau in current response was observed for a glucose concentration beyond 15 mM. This signifies the operation of the Michaelis–Menten kinetic mechanism for the enzymecatalyzed process. The apparent Michaelis–Menten constant (KM ), a parameter of importance in enzyme–substrate kinetics, is obtained from the Lineweaver–Burk equation (Li et al., 1996): 1 1 KM 1 = + ISS Imax Imax C where ISS is the steady-state current after the addition of substrate, C the bulk concentration of substrate and Imax is the maximum current measured under saturated substrate solution. A plot of I−1 vs. [glucose]−1 (Fig. 5—top inset) is a straight line (linearity correlation coefficient of 0.9921) showing the Lineweaver–Burk-like relationship with a slope of 0.3137 mM ␮A−1 (KM /Imax ) and an intercept of 0.031 ␮A−1 (1/Imax ). Typical KM and Imax values were calculated to be 10.12 mM and 32.25 ␮A, respectively. The value of KM for GOx at PMMA–MWCNT(PDDA)/GOx–NFE is lower in comparison to other glucose biosensors based on GOx immobilized prussian blue/MWCNT nanocomposites (KM = 18 mM) (Zhu et al., 2006), GOx immobilized on colloidal gold-MWCNT/Teflon (KM = 14.9 mM) and MWCNT/Teflon composite electrodes (KM = 30 mM) (Manso et al., 2007). The smaller value of KM validates that the immobilized GOx on

One of the most important analytical factors for an amperometric biosensor is the ability of the sensor to discriminate the interfering species having electroactivities similar to the target analyte. Ascorbic acid (AA) and uric acid (UA) are the most common interfering electroactive species for the amperometric detection of glucose. We have recorded the current response at the PMMA–MWCNT(PDDA)/GOx–NFE (Fig. 6(i)a) for an addition of AA and UA of 5 mM each in the presence of 5 mM of glucose. There was an increase in the response (current signal) at PMMA–MWCNT(PDDA)/GOx–NFE upon addition for AA and UA, when the potential was held at +100 mV (vs. Ag/AgCl). Hence, a layer of nafion was coated over the surface of PMMA–MWCNT(PDDA)/GOx–NFE to obtain PMMA–MWCNT(PDDA)/GOx–nafion–NFE. Interestingly, there is no interference at PMMA–MWCNT(PDDA)/ GOx–nafion– NFE for AA and UA as there were no current responses for the addition of AA or UA (Fig. 6(i)b). We have ascertained the sensitivity of PMMA–MWCNT(PDDA)/GOx–nafion–NFE for the glucose detection. The current responses for 5 mM at PMMA–MWCNT(PDDA)/GOx–nafion–NFE for glucose is nearly the same as that at PMMA–MWCNT(PDDA)/ GOx–NFE. Hence, PMMA–MWCNT(PDDA)/GOx–nafion– NFE provides minimum interference for AA and UA and high selectivity towards glucose detection. The role of nafion is justified as follows. Nafion is a negatively charged polyelectrolyte and can electrostatically repel the negatively charged AA or UA. As a result, there could not be any pre-concentration of AA or UA at the electrode surface with glucose. The use of a thin layer of nafion over the electrode surface has been earlier reported to decrease the interferences caused by anions present in biological media (Ying et al., 2002; Lim et al., 2005; Choi et al., 2005). 3.7. Real sample analysis In order to test the applicability of PMMA– MWCNT(PDDA)/GOx–NFE for the practical purpose, the glucose in the serum sample was assayed. Serum sample was diluted to its half concentration using 0.1 M PBS (pH 7) and

778

K.M. Manesh et al. / Biosensors and Bioelectronics 23 (2008) 771–779

solution of desired pH after each of the measurement. A reproducible current with a R.S.D. value of 2% was noticed for eight successive assays. Repeatability of the fabrication procedure was also assessed. Four PMMA–MWCNT(PDDA)/GOx electrodes were fabricated concurrently and the response of the electrodes towards 5 mM glucose were measured. A R.S.D. value of 5 ± 0.5% was obtained. The stability of the electrode stored at 4 ◦ C was investigated by recording periodically the current response for 5 mM glucose (Fig. 6(ii)). The electrode shows negligible change in current response for 15 days. The electrode activity of electrode remained constant indicating a good long-term stability for the biosensor. There is a decrease in the current response up to 15% only after 35 days. The good storage stability of PMMA–MWCNT(PDDA)/GOx–NFE is due to the porous structure of the fibrous matrix that could preserve GOx molecules without loss of activities. 4. Conclusions We have developed an electrospun-based nanofibrous composite electrode, PMMA–MWCNT(PDDA)/GOx–nafion–NFE, for the detection of glucose. Fibrous morphology and wrapping of PDDA over MWCNTs result in a high loading of GOx into the electrospun matrix. The NFE showed an excellent detection limit, wide linear range response, operational stability and free from electrochemical interferes of co-existing species. These features provide scope for utilizing the methodology proposed in the present study to immobilize other biomolecules in the process of fabricating novel biosensors. Acknowledgments

Fig. 6. (i) Effects of interfering signals of 5 mM uric acid (UA) and 5 mM ascorbic acid (AA) on the performance of PMMA–MWCNT(PDDA)/GOx–NFE (a) and PMMA–MWCNT(PDDA)/GOx–nafion–NFE (b) towards glucose (5 mM) at +100 mV (vs. Ag/AgCl) in 0.1 M PBS (pH 7). (ii) Stability of the PMMA–MWCNT(PDDA)/GOx–NFE to glucose stored in PBS (pH 7); (E = +100 mV vs. Ag/AgCl; [glucose] = 5 mM).

was analyzed without any pretreatment. The glucose concentration level was determined to be 7.92 mM which is in agreement to a value of 8.04 mM determined by the spectrophotometric method. Hence, the accuracy of glucose detection in serum sample at PMMA–MWCNT(PDDA)/GOx–NFE is estimated as 98% (R.S.D. = 2%, n = 4). 3.8. Repeatability, reproducibility and stability Repeatability and reproducibility of glucose detection at PMMA–MWCNT(PDDA)/GOx–NFE for 5 mM glucose were checked by amperometric measurements performed under the optimum conditions (E = +100 mV vs. Ag/AgCl, 0.1 M PBS, pH 7). The electrode was regenerated by immersing in a buffer

This work was supported by Korean Research Foundation Grant (KRF-2006-J02402). The authors acknowledge the Korea Basic Science Institute (Daegu) and Kyungpook National University Center for Scientific Instrument and KRF-2006-C00001. We thank the anonymous reviewers for many insightful comments and suggestions for useful experiments. References Amine, A., Kauffmann, J.M., Guilbault, G.G., Bacha, S., 1993. Anal. Lett. 26, 1281–1299. Bakker, E., 2004. Anal. Chem. 76, 3285–3298. Bard, A.J., Faulkner, L.R., 2001. Electrochemical Methods: Fundamentals and Applications, 2nd ed. John Wiley & Sons, New York. Bognitzki, M., Czado, W., Frese, T., Schaper, A., Hellwig, M., Steinhart, M., Greiner, A., Wendorff, J.H., 2001. Adv. Mater. 13, 70–72. Cai, C., Chen, J., 2004. Anal. Biochem. 332, 75–83. Campuzano, S., Serra, B., Pedrero, M., de Villena, F.J.M., Pingarr´on, J.M., 2003. Anal. Chim. Acta 494, 187–197. Chang, S.C., Rawson, K., McNeil, C.J., 2002. Biosens. Bioelectron. 17, 1015–1023. Choi, H.N., Kim, M.A., Lee, W.Y., 2005. Anal. Chim. Acta 537, 179–187. Christophe, V., Silvia, F., Minh, C.T., 2003. Talanta 59, 535–544. D´ıaz, J.P., Balkus, K.J., 1996. J. Mol. Catal. B 2, 115–126. Deitzel, J.M., Kleinmeyer, J., Harris, D., Beck, T.N.C., 2001. Polymer 42, 261–272. Elie, A.G., Lei, C., Baughman, R.H., 2002. Nanotechnology 13, 559–564.

K.M. Manesh et al. / Biosensors and Bioelectronics 23 (2008) 771–779 Guan, W.J., Li, Y., Chen, Y.Q., Zhang, X.B., Hu, G.Q., 2005. Biosens. Bioelectron. 21, 508–512. Joshi, P.P., Merchant, S.A., Wang, Y., Schmidtke, D.W., 2005. Anal. Chem. 77, 3183–3188. Laviron, E., 1974. J. Electroanal. Chem. 52, 355–393. Laviron, E., 1979. J. Electroanal. Chem. 101, 19–28. Lee, K.P., Gopalan, A., Santhosh, P., Manesh, K.M., Kim, J.H., Kim, K.S., 2006. J. Nanosci. Nanotechnol. 6, 1575–1583. Lee, S.W., Belcher, A.M., 2004. Nano Lett. 4, 387–390. Li, D., Wang, Y., Xia, Y., 2003. Nano Lett. 3, 1167–1171. Li, J., Cassell, A., Delzeit, L., Han, J., Meyyappan, M., 2002. J. Phys. Chem. B 106, 9299–9305. Li, J., Tan, S.N., Ge, H., 1996. Anal. Chim. Acta 335, 137–145. Lim, S., Wei, J., Lin, J., Li, Q., You, J., 2005. Biosens. Bioelectron. 20, 2341–2346. Lin, Y., Lu, F., Tu, Y., Ren, Z., 2004. Nano Lett. 4, 191–195. Liu, G., Lin, Y., 2006. Electrochem. Commun. 8, 251–256. Liu, S., Ju, H., 2003. Biosens. Bioelectron. 19, 177–183. Liu, Y., Wang, M., Zhao, F., Xu, Z., Dong, S., 2005. Biosens. Bioelectron. 21, 984–988. Manesh, K.M., Santhosh, P., Gopalan, A., Lee, K.P., 2007. Anal. Biochem. 360, 189–195. Manso, J., Mena, M.L., Sede˜no, P.Y., Pingarr´on, J., 2007. J. Electroanal. Chem. 603, 1–7. Musameh, M., Wang, J., Merkoci, A., Lin, Y., 2002. Electrochem. Commun. 4, 743–746. Ohnuki, H., Saiki, T., Kusakari, A., Endo, H., Ichihara, M., Izumi, M., 2007. Langmuir 23, 4675–4681.

779

Patel, A.C., Li, S., Yuan, J.M., Wei, Y., 2006. Nano Lett. 6, 1042–1046. Prodromidis, M.I., Karayannis, M.I., 2002. Electroanalysis 14, 241–261. Salimi, A., Compton, R.G., Hallaj, R., 2004. Anal. Biochem. 333, 49–56. Santhosh, P., Manesh, K.M., Gopalan, A., Lee, K.P., 2006. Anal. Chim. Acta 575, 32–38. Sun, Y., Yan, F., Yang, W., Sun, C., 2006. Biomaterials 27, 4042–4049. Swoboda, B.E.P., Massey, V., 1965. J. Biol. Chem. 240, 2209–2215. Tsai, Y.C., Li, S.C., Chen, J.M., 2005. Langmuir 21, 3653–3658. Wang, J., Li, M., Shi, Z., Li, N., Gu, Z., 2002. Anal. Chem. 74, 1993–1997. Wang, J., Kawde, A.N., Musameh, M., 2003. Analyst, 912–916. Wen, D., Liu, Y., Yang, G., Dong, S., 2007. Electrochim. Acta 52, 5312– 5317. Withey, G.D., Lazareck, A.D., Tzolov, M.B., Yina, A., Aich, P., Yeh, J.I., Xu, J.M., 2006. Biosens. Bioelectron. 21, 1560– 1565. Xu, J.Z., Zhu, J.J., Wu, Q., Hu, Z., Chen, H.Y., 2003. Chin. J. Chem. 21, 1088–1091. Yao, Y., Shiu, K.K., 2007. Anal. Bioanal. Chem. 387, 303–309. Ying, L., Kang, E.T., Neoh, K.G., 2002. J. Membr. Sci. 208, 361–374. Zhang, J., Feng, M., Tachikawa, H., 2007. Biosens. Bioelectron. 22, 3036– 3041. Zhang, P., Wu, F.H., Zhao, G.C., Wei, X.W., 2005. Bioelectrochemistry 67, 109–114. Zhang, W., Huang, Y., Dai, H., Wang, X., Fan, C., Li, G., 2004. Anal. Biochem. 329, 85–90. Zhao, H., Ju, H., 2006. Anal. Biochem. 350, 138–144. Zhu, L., Zhai, J., Guo, Y., Tian, C., Yang, R., 2006. Electroanalysis 18, 1842–1846.