Microchemical Journal 119 (2015) 66–74

Contents lists available at ScienceDirect

Microchemical Journal

j ourna l homepage: www.e lsev ie r .com/ locate /mic roc

Improvement of the detection limit for biosensors: Advances on theoptimization of biocomposite composition

R. Montes, J. Bartrolí, M. Baeza ⁎, F. CéspedesGrup de Sensors i Biosensors, Departament de Química, Facultat de Ciències, Edifici C-Nord, Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallès (Bellaterra), Spain

⁎ Corresponding author. Tel.: +34 935814927.E-mail address: mariadelmar.baeza@uab.cat (M. Baeza

http://dx.doi.org/10.1016/j.microc.2014.11.0040026-265X/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 28 July 2014Received in revised form 18 September 2014Accepted 11 November 2014Available online 14 November 2014

Keywords:Amperometric biosensorEISCVGraphite–epoxyGOD(Bio)composite

In this work the application of advanced characterization techniques in the development of amperometric bio-sensors based on biocomposites is described. The optimization of the conductive particle distribution and theamount of the biological material inside the biomaterial have allowed an improvement of the electrochemicalproperties, regarding the electroanalytical properties such as signal stability and limit of detection. The highsignal-to-noise ratio obtained in the electrochemical transduction has allowed enhancing the limit of detectionof the biosensor. In the present study, it has been demonstrated the feasibility of electrochemical impedancespectroscopy (EIS) and cyclic voltammetry (CV) for the characterization and optimization of biosensors basedon graphite–epoxy–enzyme, using an enzyme model. The optimum biocomposite proportion based on graph-ite–epoxy which incorporates the enzyme glucose oxidase (GOD) on the matrix ranges between 16% and 17%of graphite using 1% and 2% of enzyme. This range provides the optimal electroanalytical properties. Low limitof detection and good sensitivity have been achieved. Furthermore, confocal laser scanning microscopy wasused to visualize the enzyme distribution onto the surface electrode.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

The need of detecting small amounts of different kinds of com-pounds, which are usually present in a complex matrices, has led tothe development of new sensitive, economical and robust (bio)sensordevices based on composite materials that allow to perform in situand real analysis [1,2]. Moreover, the use of conductive materialsas transducers or conductive phase based on carbon materials(i.e. graphite, carbon nanotubes, etc.) and dispersed on a polymericmatrix (i.e. epoxy, methacrylate, Teflon, etc.) has been opened a newgeneration of rigid conducting composites that have been applied inthe realization of (bio)sensors. Electrodes obtained by using a mixtureof particulate conductive carbon phase and an insulating matrix repre-sent an attractive approach for the fabrication of electrochemical (bio)sensors, whose surfaces can be renewed by polishing [3–6]. Biocompat-ibility and capability to incorporate chemical species without their lossin operating medium are of utmost importance [7]. A biosensor basedon a biocomposite is defined as a rigid material made by combiningtwo or more materials of different nature (phases) where at least oneof them has a biological origin [8]. An important aspect of the develop-ment of biosensors is the method of immobilization of the enzyme [9].The biological component of biosensors has traditionally been placedon the surface of the transducer, either by direct adsorption [10],

).

cross-linking [11,12] or covalent attachment [13,14] or immobilizedpreviously on a membrane [15]. Another alternative strategy proposedin the development of biocomposites is the immobilization of thebiological component inside the matrix of the composite by physicalentrapment [16–20]. This immobilization that allows forming a rigidand renewable sensing surface showed excellent performance for enzy-matic determinations.

An important feature of composite electrodes is that their overallanalytical performance is strongly influenced by the carbon loadingwithin polymeric matrix. It is due to carbon loading that influencesdirectly on the electrochemical surface and inner structure (bulkresistance) of the composite electrode [21,22]. Both parameters strong-ly affect on the overall electroanalytical performance of such compositeelectrodes [23]. The characterization and optimization of compositesbased graphite–epoxy have been widely studied using different strate-gies based on several techniques as well as percolation theory[24–26], atomic force microscopy (AFM) [3,25] or chronoamperometry[3,24]. Up to now the principle applied to the optimization of the com-posite proportions has been done using the percolation theory, underthe criteria of maximizing the conductive particle loading, withoutlosing its physical and mechanical stability, but without taking care ifthe composite provides the best electroanalytical characteristics ofresponse [24,27–29]. Recently, it has established new alternativestrategies of characterization which demonstrates that if the compositeproportions are optimized the response of the electrode is improved[30] in terms of the signal-to-noise ratio which has a direct relationshipwith the limit of detection.

67R. Montes et al. / Microchemical Journal 119 (2015) 66–74

These techniques are electrochemical impedance spectroscopy(EIS) and cyclic voltammetry (CV). EIS measurements provide, inan easy way, information about the electron-transfer rate, double-layercapacitance, contact resistance and resistance of the solution [31,32].The electrochemical properties required by a transducer are highelectron-transfer rate, the lowest double-layer capacitance and ohmic re-sistance in order to guarantee the optimal electroanalytical characteristicsof the electrode response as high sensitivity, a high signal/noise ratio, andlow detection limits. Consequently, by EIS technique it is possible to de-termine the optimal composite composition that exhibits these improvedelectroanalytical properties. These results can also be complementedwithvoltammetric measurements. This is the first time that these strategiesare applied in the characterization and optimization of biocomposite elec-trodes in order to develop more efficient biosensors for determining lowanalyte concentrations in a final analytical application.

The study of the electrochemical properties of the biocomposites asa function of the conductive particle and biological chargematerial pres-ent on the transducer matrix has been performed. In the developmentof amperometric biosensors based on biocomposites, the incorporationof a biological compound produces a modification in the spatial separa-tion and inner distribution of the conductive particles. The electrochem-ical response of a biosensor based on a biocomposite depends onthe electron-transfer rate and also the active area of the electrode.Moreover, the electronic-transfer on the surface electrode dependson the conductive particle loading and hence on their distribution[33,34]. The improvement of the electrochemical properties of the bio-sensor is due to an appropriate distribution of the conductive particlesinside the biocomposite and, in consequence, in the biosensor surface.

In the literature some references have been found which realize astudy of the influence of the enzyme load. On one side, Pérez et al. [35]reported a study of L-lactate biosensors based on polysulfone/carbonnanotubes membranes where the criterion of optimization of theamount of enzyme presented on the biosensor is the amount that pro-vides the wider linear range according to the analytical requirements.On the other side, Wang et al. [36] reported a study based on enzymedispersed on carbon nanotubes for monitoring glucose where the crite-rion of optimizing the enzyme loading followed is the amount that pro-vides the best electroanalytical signal. Moreover, under the samecriterion, Caro-Jara et al. [37] optimize the GOD/HRP ratio in abienzymatic amperometric biosensor. However, in any case the optimi-zation of the transducer has been considered.

The main goal of this study is the use of alternative strategiesof characterization [30], in order to characterize and optimize thebiocomposite composition based on graphite–epoxy which incorpo-rates different amounts of an enzyme model on the polymeric matrix.Enzymatic amperometric glucose biosensors have been widely studiedin the last four decades [38] because of the relatively high durability ofthe enzyme and low cost, so glucose oxidase (GOD) has been chosen asan enzyme model for this study.

Firstly,we have constructed two series of graphite–epoxy–GODwithdifferent graphite loadings between the percolation threshold zone andamounts of GOD. We have applied the electrochemical strategies ofcharacterization in order to obtain the optimized biocomposite compo-sition. The analytical response of the optimized biocomposite has beenevaluated with synthetic samples of glucose. Besides the electroanalyt-ical characterization, the surface of the biosensors has been also charac-terized by optical techniques such as fluorescence microscopy.

2. Experimental

2.1. Apparatus

Electrochemical impedance spectroscopy and voltammetric measure-ments were performed using a computer controlled Autolab PGSTAT12potentiostat/galvanostat (Eco Chemie, Utrecht, The Netherlands) witha three-electrode configuration. A platinum-based electrode 53-671

(Crison instruments, Alella, Barcelona, Spain), an AgCl covered silverwire and the constructed graphite biocomposite electrodes were usedas a counter, reference, and working electrodes, respectively.

Amperometric measurements were done using an amperimeter LC-4C (Bio analytical Systems Inc., West Lafayette, IN, USA), connected to apersonal computer by data acquisition card ADC-42 Pico Technology(St. Neots, Cambridgeshire, UK) for data registering and visualization.Electroanalytical experiments were carried out in 20 mL glass cell, atroom temperature (25 °C), using three-electrode configuration. A singlejunction reference electrode Ag/AgCl Orion 900100 (Thermo ElectronCorporation, Beverly, MA, USA) and platinum-based electrode wereused as reference and auxiliary, respectively. The graphite biocompositeelectrodes were used as working electrode. A magnetic stirrer providedthe convective transport during the amperometric measurements.

Confocal laser scanning microscopy microphotographs were takenwith a LEICA TCS SP2 microscope.

2.2. Chemical reagents

Graphite powder (particle size 50 μm) was received from Merck(Merck Millipore, Darmstadt, Germany). Epoxy resin Epotek H77A andhardener Epotek H77B were obtained from Epoxy Technology (EpoxyTechnology, Billerica, MA, USA). Potassium ferricyanide/ferrocyanide(99.8%), potassium chloride (99.5%), potassium phosphate monoba-sic (99.5%), nitric acid (65%), potassium dibasic-anhydrous (98%),D-(+)-glucose (99.5%) and glucose oxidase type VII from Aspergillusniger (174,400 units/g) were supplied from Sigma-Aldrich (St. Louis,MO, USA) and used without further purification. EZ-Link Sulfo-NHS-LC-Biotin was purchased from Thermo Scientific (Rockford, USA). Allthe dissolutions were prepared using deionized water from Milli-Qsystem (Millipore, Billerica, MA, USA).

2.3. Fabrication of the electrodes

2.3.1. Working electrodesHandmade graphite–epoxy compositeswere prepared following the

conventionalmethodology previously established in the research group[39]. A resin Epotek H77 and their corresponding hardener compoundwere mixed in the ratio 20:3 (w/w). The graphite composite was pre-pared by loading different amounts of graphite (13, 14, 15, 16, 17, 18,19 and 20% (w/w)) into the epoxy resin before hardening. The compos-ite was homogenized for 30 min. After the homogenization time, theglucose oxidase amount (1% and 2% for each series, respectively)was in-troduced to the composite paste andhomogenized for 15minmore. Thefinal biocomposite paste electrodewas allowed to harden during 5 daysat 40 °C and when not in use it was stored at 4 °C [16]. Finally the elec-trode surface was polished with different sandpapers of decreasinggrain size. The final electrode dimensions were 28 mm2 and 3 mm forits geometric area and thickness, respectively.

2.3.2. Graphite–epoxy–GOD electrodes for fluorescence measurementsIn order to perform the fluorescencemeasurements, graphite–epoxy–

GOD biocomposite was prepared as discussed in Section 2.3.1, but it wasintroduced into a special support. Hand-made biocomposite was placedinto a PVC disk with an external diameter of 35 mm, internal diameterof 15 mm, and thickness of 3 mm. The process of hardening was 5 daysat 40 °C and when not in use it was stored at 4 °C. The electrode surfacewas polished using different sandpapers of decreasing grain size.

2.4. Procedure

2.4.1. Electrochemical characterizationEIS and voltammetricmeasurementsweremade in a 0.1Mpotassium

chloride solution containing 0.01 M potassium ferricyanide/ferrocyanideunder quiescent condition. Amperometric detection was made underforce convection by stirring the solution with magnetic stirrer.

68 R. Montes et al. / Microchemical Journal 119 (2015) 66–74

2.4.2. Microscopic characterization of biosensor surfaceThe distribution of the glucose oxidase enzyme on the electrode sur-

face was studied by confocal laser scanning fluorescence microscopy.In this instance a fluorescent conjugate was used for the labeling of

the graphite–epoxy–GOD electrode surface by covalent bond betweenthe amino groups of GOD and the carboxyl groups of biotin. The back-ground control was performed using non-biotinylated fluorescein sur-face of graphite–epoxy–GOD electrode.

The biocomposite surface was labeled in two steps. An incubationstep with 100 μL of biotin/streptavidin–Cy5 solution (10 mM of biotinand 0.02 mg/mL of streptavidin–Cy5 fluorescent dye) in PBS (0.1 Mand pH = 7.6) at 25 °C during 60 min was performed. After that,three washing steps with PBS buffer (0.1 M and pH = 7.6) at 25 °Cduring 5 min were realized.

3. Results and discussion

3.1. Electrochemical impedance spectroscopy measurements

By means of electrochemical impedance spectroscopy we could ob-tain general trends in the electrochemical parameters of the graphite–epoxy–enzyme biocomposites such as the solution resistance or theohmic resistance (RΩ) (which comprises the resistance of the solution(Rs) plus any contact resistance (Rc)), the charge-transfer resistance(Rct) and the double-layer capacitance (Cdl) for composites with differ-ent graphite loadings and amount of GOD. These parameters were ob-tained by fitting the impedance spectra to an equivalent circuit. Thiscircuit was sufficiently suitable to interpret the RΩ, Rct and Cdl valuesin terms of interfacial phenomena that occur at the electrochemicalcell [30]. Such electrochemical characterization was performed in thepresence of benchmark redox species such as Fe(CN)63−/Fe(CN)64−

which is very sensitive to the electrode surface characteristics [40].The impedance spectra were recorded in the frequency range 0.1 Hzto 100 kHz at the redox equilibrium potential. The signal amplitude toperturb the system was 10 mV.

Two series of eight different compositions of graphite–epoxy com-posite electrodes were constructed by varying the graphite loadingfrom 13% to 20% and the amount of glucose oxidase was 1% and 2%,respectively. The interval of graphite composition was selected basedon the percolation curve for the graphite–epoxy composite previouslyestablished in an earlier work [25]. This interval of composition corre-sponds to the low resistivity region, where it is important to highlightthat the graphite loading has to be high enough to guarantee aconducting pathway and sufficiently low bulk resistivity. For eachgraphite–epoxy–GOD composition, three equal electrodes were fabri-cated and evaluated. The corresponding parameters of EIS, the ohmicresistance (RΩ), charge-transfer resistance (Rct) and the double-layercapacitance (Cdl) were obtained by fitting the impedance spectra to asimple equivalent circuit proposed (see inset Fig. 1). Using such ap-proach, we attempted to obtain a general trend of these physical param-eters as a function of the graphite loading, using constant amount ofglucose oxidase for each series.

The impedance response of biocomposites with 1% and 2% of GOD isdepicted in Fig. 1A and 1B, respectively. As we can see, the impedancebehavior is dominated by small diameter semicircle representingkinetic-controlled electrode process, though in some cases the diffusioncontrol starts to be discerned at low frequencies, when the amount ofgraphite is high corresponding to the low resistivity area (from 20% to16% of graphite loading). On the other hand, the impedance plot forcomposites with low graphite loading (from 15% to 13% of graphiteloading) appears to be dominated by a big diameter semicircle andonly the kinetic-controlled electrode process is present in the recordedfrequency range. We could observe that when the amount of GOD in-creases from 1% (Fig. 1A) to 2% (Fig. 1B) the diameters of the impedanceresponse for each composition are higher. So, an increase of the GODamount produces an increase of the kinetics-controlled electrode

process and the electron-transfer is more limited. This can be associatedwith the separation of the conductive particles due to the increase ofenzyme load.

Fig. 2A shows the variations of the ohmic resistance as a function ofthe graphite composition for both GOD amounts. Ionic concentration,type of ion, electrode area and dry resistance of the composite are fac-tors that influence on this resistance. At low graphite loads, the ohmicresistance is dominated by composite resistance whereas at highergraphite loads the ohmic resistance decreases to low values and it isdominated by the solution resistance [25,26]. In general trends, theohmic resistance values decrease when the graphite loading increasesfor both GOD concentrations. The increase of the ohmic resistance be-comes more significant when the enzyme load increases. Therefore,the modification of the matrix by a third element has an influence onthe active area of the electrode because the conductive material is sep-arated by the enzyme particles, decreasing the active area of the elec-trode and increasing the value of the ohmic resistance. Consequently,in order to assure the sensitivity and low response time it is importantthat ohmic resistance is low.

The quantitative values of Rct are depicted in Fig. 2B. This parameteris inversely proportional to the heterogeneous charge-transfer rate andalso affects the sensitivity and response time of the electrode. A de-crease of the Rct is observed with the increase of the graphite loadwhich indicates the strong relation between electrochemical reactivityand the surface characteristics of the conducting material. As carbonload increases, the probability of having more electroactive sites in-creases and hence the electrode kinetics. Therefore low values in thecharge-transfer resistance allow increasing the electroanalytical appli-cations of the electrode. As well as in ohmic resistance, the values ofRct are higher when the GOD amount increases. The normalization ofthis parameter by the electrochemical electroactive area should be con-stant in metal electrode surface (see Table 1(A) and 1(B)). Graphite as acarbonmaterial exhibits electrochemical anisotropy aswell as it was re-ported for CNTs [30]. Low values on the charge-transfer resistance of thebiocomposites will guarantee fast electron-transfer and their appropri-ate use in electroanalytical applications. According to the results obtain-ed biocomposites between 20% and 16% of graphite loading and bothamounts of GOD presented lower charge-transfer resistance values.

Finally, it is important to consider the double-layer capacitancevalue which is directly related to the charging or background current.Fig. 2C depicts the decrease of the double-layer capacitance valueswith the decrease of the graphite loadings with constant GOD amount.An increase of the values of Cdl is observed when the amount of GODincreases on biocomposite electrodes with the same load of graphite.High proportions of graphite increase remarkably the background cur-rent because the proportion of conductingmaterial exposed on the sur-face area becomes higher and that not enhances the signal-to-noiseratio and consequently increased the detection limit of the analyte.Therefore, an increase on the amount of GOD increases the Cdl valuedue to the presence of the enzyme on the matrix that reduces thegraphite portion on the surface.

Based on the impedance results obtained for the biosensors andtaking into account the properties required by an electrode for electro-analytical purposes, such as rapid response time, low limit of detectionand high sensitivity, the interval between 16% and 17% of graphite load-ing seems to fulfill these requirements, for both GOD concentrationsevaluated. The fabricated composites in such interval of compositiondepicted similar electrochemical behavior, an increase of the electro-chemical reproducibility and analytical properties improved.

3.2. Cyclic voltammetry characterization

In order to complement the results obtained by EIS technique wehave also performed cyclic voltammetry (CV) measurements. Cyclicvoltammogramswere taken for the different biocomposite compositionelectrodes in the presence of the benchmark Fe(CN)63−/Fe(CN)64− redox

(A)

Z (Real)

-500 0 500 1000 1500 2000 2500 3000 3500

Z (Im

ag)

0

200

400

600

800

1000

1200

1400

1600

1800

200013% graphite14% graphite15% graphite16% graphite17% graphite18% graphite19% graphite20% graphite

RsRct

Rc

Cdl

(B)

Z (Real)

-500 0 500 1000 1500 2000 2500 3000 3500

Z (Im

ag)

0

200

400

600

800

1000

1200

1400

1600

1800

200013% graphite14% graphite15% graphite16% graphite17% graphite18% graphite19% graphite20% graphite

Fig. 1. Nyquist plots for different graphite loading electrodes with (A) 1% and (B) 2% of GOD in the presence of Fe(CN)63−/Fe(CN)64−. The equivalent circuit used for the impedance spectrafitting is shown in (A).

69R. Montes et al. / Microchemical Journal 119 (2015) 66–74

couple under the same experimental condition (10 mV/s of scan). Forboth GOD concentrations, it is important to highlight that none of thecompositions exhibit sigmoidal shaped voltagramms so there is no elec-trochemical behavior as a microelectrode array where radial diffusiondominates mass transport; at least in the interval of graphite studied.Biocomposite electrodes exhibit the typical peak-shaped [4,6,41] profilecorresponding to more massive electrodes with planar diffusion wherelinear diffusion controlsmass transport (see Fig. 3A and 3B). Different pa-rameters were extracted from the cyclic voltagramms such as the peakseparation potential (ΔE) and peak current (Ip) as shown in Table 1 (Aand B) for each series of graphite–epoxy–GOD biocomposite. As we cansee in Table 1, there is an increase of the peak current with the graphiteloading increases associated with an increase of the electroactive area,together with a decrease of peak separation related to an enhancement

Fig. 2. Values of (A) ohmic resistance, (B) charge transfer resistance and (C) double-layer capacloading electrodes with 1% and 2% of GOD, using the redox probe Fe(CN)63−/Fe(CN)64−.

of the electron-transfer rate. This behavior is observed for both GOD con-centrations. The relative electroactive area was estimated from the peak-shaped voltammograms by quantifying the peak current with the use ofthis relationship, Ip = 3.01 × 105 n3/2 (α Dred υ)1/2 A C⁎red [31], which isappropriate for electron-transfer controlled process. In this equation αcorresponds to the transfer coefficient which was considered to be ap-proximately 0.5, Dred = 6.32 · 10−6 cm2s−1 corresponds to the diffusioncoefficient of the reduced species, υ=0.01 Vs−1 represents the scan rate,A is the electroactive area and C⁎red= 0.01M is the bulk concentration ofthe electroactive species.We also evaluate the exchange current (io) fromTafel plots (log current vs. potential), a parameter which provides infor-mation about the reversibility of the process. From the value of the ex-change current we can also evaluate the charge-transfer resistancethrough the relation io = RT/nFRct. The charge-transfer resistance values

itances, with their corresponding standard deviation (n = 3) for the different graphite

(A)

(B)

% graphite12 13 14 15 16 17 18 19 20 21

R(o

hms)

0

2000

4000

6000

8000

100001% GOD2% GOD

% graphite12 13 14 15 16 17 18 19 20 21

Rct

Ω

(ohm

s)

0

1000

2000

3000

4000

50001% GOD2% GOD

(C)

% graphite12 13 14 15 16 17 18 19 20 21

Cdl

(F)

0.00

2.00e-5

4.00e-5

6.00e-5

8.00e-5

1.00e-4

1.20e-4

1.40e-4

1.60e-4

1.80e-41% GOD2% GOD

70 R. Montes et al. / Microchemical Journal 119 (2015) 66–74

71R. Montes et al. / Microchemical Journal 119 (2015) 66–74

obtained by EIS are following the same trend like the results obtainedby CV technique. We have normalized the Rct with respect to theelectroactive area and it is observed that there is a decrease on thevalue when the active area increases showing the evident influence ofthe electrochemical anisotropy of the graphite which can bemore notice-able as the graphite loading is increased. From the comparison of the nor-malized Rct values extracted from cyclic voltammetry with the onesobtained by electrochemical impedance spectroscopy measurements itcan be observed that the values agree quite well being more similar inthe optimal range of composition.

3.3. Electroanalytical performance

Glucose was used as analyte for evaluating the electroanalyticalproperties. When GOD and glucose interact, hydrogen peroxide is pro-duced at the biosensor surface. In this case, hydrogen peroxide pro-duced by the biocatalytic reaction can be amperometrically measuredby direct oxidation on the surface of the biosensor. Measurementswere carried out at 1050 mV polarization potential fixed. The responseof biosensor to changes in glucose concentration was evaluated byhydrodynamic amperometric measurements for all biocomposite com-positions studied.

The analytical parameters such as detection limit, sensitivity andlinear range were evaluated for 16% and 17% of graphite compositewith 1% and 2% of GOD and compared to those obtained with 20% ofgraphite loading with the same amount of GOD. For each compositionthree electrodes were evaluated.

Aswe can see fromTable 2, the sensitivity increaseswhen the graph-ite loading increases and the limit of detection increases remarkably,using the same amount of glucose oxidase. The experimental resultsshow that in the optimal composite composition range (16–17% ofgraphite loadingwith 1–2% of GOD), there are no significant differencesregarding the sensitivity and limit of detection for both amounts of GOD.However, we can observe that the biocomposite with 16% of graphiteloading and 1% of GOD presents a wider linear range.When the amountof GODdecreases from2% to 1%, the sensitivity is slightly inferior but thedetection limit remains constant. Nevertheless, when it is comparedwith the 20% of graphite loading, in spite of the sensitivity and the linearrange increased, the limit of detection also increases in one order ofmagnitude. These results show that the biocomposites in the optimalcomposition range provide quite good sensitivity, low limit of detectionand a wide linear range.

Table 1Cyclic voltammetry parameters for the different composite compositions with (A) 2% of GOD atance, Ip to peak current, A to active area andΔE to the peak separation potencial. Rct · A and Rc

E

and normalized with respect to the active area.

Electrodes io (A) Rct (Ω) Ip (A)

(A)13% 1.93 · 10−6 12873.2 7.80 · 10−5

14% 3.03 · 10−6 8327.2 8.21 · 10−5

15% 5.93 · 10−6 4254.9 1.01 · 10−4

16% 1.51 · 10−5 1671.0 1.23 · 10−4

17% 1.79 · 10−5 1409.6 1.37 · 10−4

18% 2.20 · 10−5 1146.9 1.43 · 10−4

19% 5.27 · 10−5 478.8 1.73 · 10−4

20% 8.46 · 10−5 298.2 2.03 · 10−4

(B)13% 2.60 · 10−6 9704.7 8.26 · 10−5

14% 5.97 · 10−6 4226.36 1.00 · 10−4

15% 1.08 · 10−5 2336.2 1.23 · 10−4

16% 1.85 · 10−5 1363.9 1.35 · 10−4

17% 2.00 · 10−5 1261.6 1.35 · 10−4

18% 2.34 · 10−5 1078.3 1.46 · 10−4

19% 6.01 · 10−5 419.8 1.82 · 10−4

20% 7.76 · 10−5 325.2 2.06 · 10−4

In the literature it has been reported that non-optimized electrodesbased on graphite–epoxy–GOD (18:80:2 (w/w)) which compared tothe results obtained with the optimized biocomposite electrodes hadlower sensitivity and higher limit of detection [18]. Hence, this factreaffirms that complementary strategies to optimize the electrochemi-cal properties of biosensors, both conducting material and enzymema-terial load, improve the electroanalytical properties of the biosensorsdeveloped.

3.4. Microscopic characterization of biosensor surface

Confocal laser scanning fluorescence microscopy was used to evalu-ate the distribution of the biological compound in theworking electrodesurface. This study was performed with the biotinylation and markingwith a fluorescein dye of the surface electrode (see Experimentalsection 2.4.2). The compositions evaluated in this study are 16% and20% of graphite with both amounts of GOD. The surface of differentbiocomposite electrodes with different amounts of graphite load (16%and 20%) and glucose oxidase (1% and 2%) is shown in Fig. 4. The intensefluorescence area corresponds to the enzyme bidingwith the fluorescentmarker. After this study fluorescence microscopy confirms that the glu-cose oxidase is distributed homogenously in the electrode surface.

4. Conclusion

The proposed electrochemical techniques, such as electrochemicalimpedance spectroscopy and cyclic voltammetry, as well as hydrody-namic amperometry are suitable for the characterization and optimiza-tion of biocomposites based on graphite–epoxy–enzyme. By thesealternative techniques, and especially EIS, provide a versatile tool to op-timize the biocomposite composition. Some relevant parameters can beextracted from EIS, as ohmic resistance which is related to the percola-tion theory, charge-transfer resistance which can be related to the het-erogeneous electron-transfer and which depends on the surfaceelectrochemical reactivity (active area) and finally the electrode capac-itance which can be correlated to the background current and conse-quently with the signal/noise ratio. As a result, all these parametersare relevant to the biosensor analytical response and the evaluation ofthese parameters allows choosing the biocomposite compositionwhich fulfills the electroanalytical requirements of high sensitivity,fast response and low limits of detection. In the present work the opti-mized composition ranges between 16% and 17% of graphite loading

nd (B) 1% of GOD. Io corresponds to the exchange current, Rct to the charge transfer resis-tIS · A correspond to the Rct obtained by voltammetric and EIS measurements, respectively,

A (cm2) ΔE (V) Rct · A (Ω cm−2) RctEIS · A (Ω cm−2)

0.15 1.5666 1931 4690.15 1.3245 1249 4000.19 0.9720 789 4410.23 0.6622 384 2760.26 0.5732 366 2640.27 0.5091 310 2500.32 0.2563 153 970.38 0.1780 113 77

0.15 1.5595 1456 4560.19 1.0895 803 3370.23 0.7868 537 3480.25 0.5803 341 1930.25 0.4949 315 1750.27 0.4733 291 1730.34 0.2314 143 440.38 0.1923 124 60

(A)

(B)

Potential (V)-1,5 -1,0 -0,5 0,0 0,5 1,0 1,5 2,0 2,5

Inte

nsity

(A)

-3,00e-4

-2,00e-4

-1,00e-4

0,00

1,00e-4

2,00e-4

3,00e-4

13% graphite14% graphite15% graphite16% graphite17% graphite18% graphite19% graphite20% graphite

% graphite12 13 14 15 16 17 18 19 20 21

Rct

(ohm

s)

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

Potential (V)-1,5 -1,0 -0,5 0,0 0,5 1,0 1,5 2,0

Inte

nsity

(A)

-3,00e-4

-2,00e-4

-1,00e-4

0,00

1,00e-4

2,00e-4

3,00e-413 % grafito14 % grafito15 % grafito16 % grafito17 % grafito18 % grafito19 % grafito20 % grafito

% graphite12 13 14 15 16 17 18 19 20 21

Rct

(ohm

s)

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

Fig. 3. Cyclic voltammogram for biocompositeswith (A) 1% of GOD and (B) 2% of GOD using 0.01M ferricyanide/ferrocyanide and 0.1M KCl. Scan rate 10mV/s. The inset figure shows thetrend of charge transfer resistance for different biocomposite compositions.

72 R. Montes et al. / Microchemical Journal 119 (2015) 66–74

for graphite–epoxy–GOD model electrodes for both GOD proportions.Small variations in compositions around this optimal range do not pro-duce high changes in the electrochemical behavior. To confirm these

Table 2The calibration parameters for 20%, 17% and 16% of graphite biocomposite electrode with1% and 2% of GOD using amperometric measurements with glucose as analyte and PBS0.1 M at pH = 7.0 as background electrolyte.

% Graphite % GOD Sensitivitya

(μA·L·mmol−1)(% RSD)

LOD(mmol·L−1)

Linear range(mmol·L−1)

16 1 0.185 ± 0.003 (2%) 0.037 ± 0.001 0.037–1.282 0.21 ± 0.01 (7%) 0.046 ± 0.008 0.046–0.70

17 1 0.214 ± 0.005 (3%) 0.036 ± 0.007 0.036–0.702 0.231 ± 0.009 (4%) 0.035 ± 0.009 0.035–0.70

20 1 0.74 ± 0.01 (2%) 0.265 ± 0.006 0.265–0.812 1.35 ± 0.05 (4%) 0.24 ± 0.01 0.24–1.80

a n = 3, 95% confidence level.

predictions, we showed the electroanalytical detection of glucosewhich demonstrates that although the sensitivity decreases, comparedwith 20% graphite; we have obtained a better limit of detection (oneorder of magnitude lower) and wider linear range. It is important tohighlight that these results showed that to obtain low limit of detectionthe amount of graphite loading it is more critical than the enzyme load-ing in biocomposites. Moreover, there are no significant differences inthe electroanalytical parameters between 1% and 2% of GOD, when itused the same graphite loading. Finally, in the present work it hasbeen demonstrated the applicability of these techniques in the charac-terization and optimization of biosensors based on biocompositeswhich will allow us to apply them to other biological compounds fortheir application in the determination of pesticides, DNA, immunoassay,food safety, etc., where it is very important to achieved low analyteconcentrations. In addition, the homogeneous and stable immobiliza-tion of the enzyme in the biocomposite surface by entrapment hasbeen demonstrated by confocal laser microscopy.

A1 A2

B1 B2

C1 C2

D1 D2

Fig. 4. Confocal laser scanning fluorescence microphotograph for biocomposites with 16% with (A1) 1% and (C1) 2% of GOD and 20% of graphite load and (B1) 1% and (D2) 2% of GOD sub-mitted to biotinylated fluorescein. Microphotographs (A2), (B2), (C2) and (D2) correspond to non-biotinylated fluorescein. Laser excitation was 620 nm. Voltage was 352 V.

73R. Montes et al. / Microchemical Journal 119 (2015) 66–74

74 R. Montes et al. / Microchemical Journal 119 (2015) 66–74

Acknowledgments

R. Montes thanks Universitat Autònoma de Barcelona (UAB) for theaward of PIF studentship.

References

[1] R. Olivé-Monllau, A. Pereira, J. Bartrolí, M. Baeza, F. Céspedes, Highly sensitive CNTcomposite amperometric sensors integrated in an automated flow system for thedetermination of free chlorine in waters, Talanta 81 (2010) 1593–1598, http://dx.doi.org/10.1016/j.talanta.2010.03.008.

[2] A.B. Ibáñez, A. Gutés, M. Baeza, F. Céspedes, Electronic tongue applied to phenoliccompounds analysis, Anal. Lett. 41 (2008) 1419–1429.

[3] D.O. Hare, J.V. Macpherson, A. Willows, On the microelectrode behaviour ofgraphite–epoxy composite electrodes, Electrochem. Commun. 4 (2002) 245–250.

[4] S. Ramírez-García, S. Alegret, F. Cespedes, R.J. Forster, Carbon composite electrodes:surface and electrochemical properties, Analyst 127 (2002) 1512–1519, http://dx.doi.org/10.1039/b206201a.

[5] L. Rassaei, M. Sillanpää, M.J. Bonné, F. Marken, Carbon nanofiber–polystyrenecomposite electrodes for electroanalytical processes, Electroanalysis 19 (2007)1461–1466, http://dx.doi.org/10.1002/elan.200703887.

[6] J. Wang, Practical considerations, Anal. Electrochem.Wiley-VCH, New York, 2000,pp. 100–139.

[7] D. Bellido-Milla, L.M. Cubillana-Aguilera, M. El Kaoutit, M.P. Hernández-Artiga, J.L.Hidalgo-Hidalgo de Cisneros, I. Naranjo-Rodríguez, et al., Recent advances ingraphite powder-based electrodes, Anal. Bioanal. Chem. 405 (2013) 3525–3539,http://dx.doi.org/10.1007/s00216-013-6816-2.

[8] F. Cespedes, S. Alegret, New materials for electrochemical sensing II. Rigid carbon–polymer biocomposites, Trends Anal. Chem. 19 (2000) 276–284.

[9] A. Numnuam, P. Thavarungkul, An amperometric uric acid biosensor based onchitosan–carbon nanotubes electrospun nanofiber on silver nanoparticles, Anal.Bioanal. Chem. 406 (2014) 3763–3772, http://dx.doi.org/10.1007/s00216-014-7770-3.

[10] A. Parra, E. Casero, L. Vázquez, F. Pariente, E. Lorenzo, Design and characterization ofa lactate biosensor based on immobilized lactate oxidase onto gold surfaces, Anal.Chim. Acta. 555 (2006) 308–315, http://dx.doi.org/10.1016/j.aca.2005.09.025.

[11] S.A. Marzouk, V.V. Cosofret, R.P. Buck, H. Yang,W.E. Cascio, S.S. Hassen, A conductingsalt-based amperometric biosensor for measurement of extracellular lactate accu-mulation in ischemic myocardium, Anal. Chem. 69 (1997) 2646–2652 http://www.ncbi.nlm.nih.gov/pubmed/9230678.

[12] M.M. Barsan, C.M.A. Brett, A new modified conducting carbon composite electrodeas sensor for ascorbate and biosensor for glucose. Bioelectrochemistry 76 (2009)135–140, http://dx.doi.org/10.1016/j.bioelechem.2009.03.004.

[13] F. Palmisano, R. Rizzi, D. Centonze, P.G. Zambonin, Simultaneous monitoring ofglucose and lactate by an interference and cross-talk free dual electrode ampero-metric biosensor based on electropolymerized thin films, Biosens. Bioelectron. 15(2000) 531–539 http://www.ncbi.nlm.nih.gov/pubmed/11419650.

[14] J. Haccoun, B. Piro, L.D. Tran, L.A. Dang, M.C. Pham, Reagentless amperometric detec-tion of l-lactate on an enzyme-modified conducting copolymer poly(5-hydroxy-1,4-naphthoquinone-co-5-hydroxy-3-thioacetic acid-1,4-naphthoquinone), Biosens.Bioelectron. 19 (2004) 1325–1329, http://dx.doi.org/10.1016/j.bios.2003.11.006.

[15] S. Sánchez, M. Pumera, E. Fàbregas, Carbon nanotube/polysulfone screen-printedelectrochemical immunosensor, Biosens. Bioelectron. 23 (2007) 332–340, http://dx.doi.org/10.1016/j.bios.2007.04.021.

[16] F. Céspedes, E. Martínez-Fàbrega, J. Bartrolí, S. Alegret, Amperometric enzymaticglucose electrode based on an epoxy graphite composite, Anal. Chim. Acta. 273(1993) 409–417.

[17] J. Wang, L. Fang, D. Lopez, H. Tobias, Highly selective and sensitive amperometricbiosensing of glucose at ruthenium-dispersed carbon-paste enzyme electrodes,Anal. Lett. 26 (1993) 1819–1830.

[18] B. Perez, M. Pumera, M. del Valle, A. Merkoci, S. Alegret, Glucose biosensor based oncarbon nanotube epoxy composites, J. Nanosci. Nanotechnol. 5 (2005) 1694–1698.

[19] E. Zacco, M.I. Pividori, X. Llopis, M. Del Valle, S. Alegret, Renewable Protein Amodified graphite–epoxy composite for electrochemical immunosensing, J.

Imunol. Methods 286 (2004) 35–46, http://dx.doi.org/10.1016/j.jim.2003.11.014.

[20] M.D. Rubianes, G. Rivas, Enzymatic biosensors based on carbon nanotubes pasteelectrodes, Electroanalysis 17 (2005) 73–78, http://dx.doi.org/10.1002/elan.200403121.

[21] M. Baeza, R. Olivé-Monllau, M.J. Esplandiu, Advances on rigid conducting compos-ites for electroanalytical applications, in: Deborah B. Son (Ed.), Resin Compos.Prop. Prod. Appl.Nova Science Publishers, Inc., 2011, pp. 153–211.

[22] M.J. Esplandiu, M. Baeza, R. Olivé-Monllau, F. Céspedes, Chapter 5: Development ofTunable Nanocomposites Made from Carbon Nanotubes for ElectrochemicalApplications, 2010.

[23] H. Zhao, D.O. Hare, Characterisation and modeling of conducting composite elec-trodes, J. Phys. Chem. C 112 (2008) 9351–9357, http://dx.doi.org/10.1021/jp711366u.

[24] S. Ramírez-García, S. Alegret, F. Céspedes, R.J. Forster, Carbon composite microelec-trodes: charge percolation and electroanalytical performance, Anal. Chem. 76(2004) 503–512, http://dx.doi.org/10.1021/ac034536p.

[25] R. Olivé-Monllau, M. Baeza, J. Bartrolí, F. Céspedes, Novel amperometric sensorbased on rigid near-percolation composite, Electroanalysis 21 (2009) 931–938,http://dx.doi.org/10.1002/elan.200804494.

[26] J. Trijueque, J.J. Garcı́a-Jareño, J. Navarro-Laboulais, A. Sanmatı́as, F. Vicente, Ohmicdrop of Prussian-blue/graphite + epoxy electrodes, Electrochim. Acta 45 (1999)789–795, http://dx.doi.org/10.1016/S0013-4686(99)00257-1.

[27] M. Pumera, A. Merkoçi, S. Alegret, Carbon nanotube–epoxy composites for electro-chemical sensing, Sensors Actuators B Chem. 113 (2006) 617–622, http://dx.doi.org/10.1016/j.snb.2005.07.010.

[28] S. Ramírez-García, F. Cespedes, S. Alegret, Development of conducting compositematerials for electrochemical sensing in organic media, Electroanalysis 13 (2001)529–535.

[29] F. Cespedes, E. Martínez-Fàbregas, S. Alegret, New materials for electrochemicalsensing I. Rigid conducting composites, Trends Anal. Chem. 15 (1996) 296–304.

[30] R. Olivé-Monllau, M.J. Esplandiu, J. Bartrolí, M. Baeza, F. Céspedes, Strategies for theoptimization of carbon nanotube/polymer ratio in composite materials: applicationsas voltammetric sensors, Sensors Actuators B Chem. 146 (2010) 353–360, http://dx.doi.org/10.1016/j.snb.2010.02.017.

[31] M. Pacios, M. del Valle, J. Bartroli, M.J. Esplandiu, Electrochemical behavior of rigidcarbon nanotube composite electrodes, J. Electroanal. Chem. 619–620 (2008)117–124, http://dx.doi.org/10.1016/j.jelechem.2008.03.019.

[32] M.J. Esplandiu, M. Pacios, L. Cyganek, J. Bartroli, M. del Valle, Enhancing the electro-chemical response of myoglobin with carbon nanotube electrodes, Nanotechnology20 (2009) 355–502, http://dx.doi.org/10.1088/0957-4484/20/35/355502.

[33] N. Bojorge, E. Alhadeff, Graphite-composites alternatives for electrochemical biosensor,in: J. Cuppoletti (Ed.), Met. Ceram. Polym. Compos. Var. UsesInTech, 2011, pp. 597–620.

[34] A. Sadana, N. Sadana, Handbook of Biosensors and Biosensor Kinetics, Elsevier, B.V.,2011

[35] S. Pérez, S. Sánchez, E. Fàbregas, Enzymatic strategies to construct L-lactate biosen-sors based on polysulfone/carbon nanotubes membranes, Electroanalysis 24(2012) 967–974, http://dx.doi.org/10.1002/elan.201100628.

[36] J. Wang, M. Musameh, Enzyme-dispersed carbon-nanotube electrodes: a needle mi-crosensor for monitoring glucose, Analyst 128 (2003) 1382, http://dx.doi.org/10.1039/b309928h.

[37] N. Caro-Jara, R. Mundaca-Uribe, C. Zaror-Zaror, J. Carpinelli-Pavisic, M. Aranda-Bustos, C. Peña-Farfal, Development of a bienzymatic amperometric glucosebiosensor using mesoporous silica (MCM-41) for enzyme immobilization and itsapplication on liquid pharmaceutical formulations, Electroanalysis 25 (2013)308–315, http://dx.doi.org/10.1002/elan.201200391.

[38] J.Wang, Glucose biosensors: 40 years of advances and challenges, Electroanalysis 13(2001) 983–988, http://dx.doi.org/10.1002/1521-4109(200108)13:12b983::AID-ELAN983N3.0.CO;2-#.

[39] A. Morales, F. Cespedes, J. Muñoz, S. Alegret, E. Martinez-Fàbregas, Hydrogen perox-ide amperometric biosensor based on a biocomposite, Anal. Chim. Acta. 332 (1996)131–136.

[40] R.L. McCreery, Advanced carbon electrode materials for molecular electrochemistry,Chem. Rev. 108 (2008) 2646–2687, http://dx.doi.org/10.1021/cr068076m.

[41] K. Stulik, C. Amatore, K. Holub, V. Marecek, W. Kutner, Microelectrodes. Definitions,characterization and applications, Pure Appl. Chem. 72 (2000) 1483–1492.