Materials Science and Engineering C 29 (2009) 442–447

Contents lists available at ScienceDirect

Materials Science and Engineering C

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

Immobilization of cholesterol oxidase in LbL films and detection of cholesterol usingac measurements

Marli L. Moraes a,b,⁎, Nara C. de Souza c, Caio O. Hayasaka b, Marystela Ferreira d,Ubirajara P. Rodrigues Filho a, Antonio Riul Jr. d, Valtencir Zucolotto b, Osvaldo N. Oliveira Jr. b

a Instituto de Química de São Carlos, Universidade de São Paulo, CP 780, 13560-970, São Carlos, SP, Brazilb Instituto de Física de São Carlos, Universidade de São Paulo, CP 369, 13560-970, São Carlos, SP, Brazilc Instituto Universitário do Araguaia, Universidade Federal de Mato Grosso, 78698-000, Pontal do Araguaia, MT, Brazild Universidade Federal de São Carlos, Campus Sorocaba, CP 3031, 18043-970, Sorocaba, SP, Brazil

⁎ Corresponding author. Instituto de Química de SãPaulo, CP 780, 13560-970, São Carlos, SP, Brazil.

E-mail address: moraes@ifsc.usp.br (M.L. Moraes).

0928-4931/$ – see front matter © 2008 Published by Edoi:10.1016/j.msec.2008.08.040

a b s t r a c t

a r t i c l e i n f o

Article history:

The preserved activity of im Received 27 May 2008Received in revised form 15 July 2008Accepted 29 August 2008Available online 15 September 2008

Keywords:Layer-by-layerCholesterol oxidaseCholesterolBiosensorEggPGAtomic force microscopy

mobilized biomolecules in layer-by-layer (LbL) films can be exploited in variousapplications, including biosensing. In this study, cholesterol oxidase (COX) layers were alternated with layersof poly(allylamine hydrochloride) (PAH) in LbL films whose morphology was investigated with atomic forcemicroscopy (AFM). The adsorption kinetics of COX layers comprised two regimes, a fast, first-order kineticsprocess followed by a slow process fitted with a Johnson–Mehl–Avrami (JMA) function, with exponent ~2characteristic of aggregates growing as disks. The concept based on the use of sensor arrays to increasesensitivity, widely employed in electronic tongues, was extended to biosensing with impedance spectroscopymeasurements. Using three sensing units, made of LbL films of PAH/COX and PAH/PVS (polyvinyl sulfonicacid) and a bare gold interdigitated electrode, we were able to detect cholesterol in aqueous solutions downto the 10−6 M level. This high sensitivity is attributed to the molecular-recognition interaction between COXand cholesterol, and opens the way for clinical tests to be made with low cost, fast experimental procedures.

© 2008 Published by Elsevier B.V.

1. Introduction

The fabrication of nanostructured films from biomolecules hasbecome an important tool for investigating fundamental adsorptionmechanisms and biological activity, in addition to serving for immobi-lizing materials in different supramolecular architectures for a numberof biotechnological applications [1]. In this context, one can benefit fromthe LbL (layer-by-layer) technique [2] that allows efficient immobiliza-tion of enzymes due to the mild conditions under which the filmfabrication takes place, in addition to the choice of template materialsthatmay help preserve enzyme activity [3]. Significantly, recentwork byLourenço et al. [4] and Silva et al. [5] has proven that entrained water ispresent in LbL films, in spite of the drying procedures, which isimportant for the preservation of activity of biomolecules. There is nowan extensive list of biosensorsmadewith LbL films (for a review see [6]).

Biosensors obtained with immobilized cholesterol oxidase may beadvantageous in comparison with standard methods of cholesteroldetection, e.g. spectrophotometry [7,8], gas–liquid chromatography andHPLC [9,10], owing to the simplicity and low cost involved. This isrelevant to thedevelopmentof reliablemethods of cholesterol detection

o Carlos, Universidade de São

lsevier B.V.

in blood, a fundamental parameter to identify disorders such ashypercholesterolemia [11,12]. Furthermore, it is important to controlthe cholesterol level of foodstuffs for human intake.

Cholesterol is a sterol found in eggs, meats, yellow cheese, andderivatives. Biosensors for cholesterol have been used in biochemicalanalysis owing to their good selectivity, fast response, low cost, smallsize and long term stability [8]. Most of the literature on cholesterolbiosensors has focused on diagnosing disorders [13–15] and the intakecontrol of foodstuffs [16,17]. For example, Basu et al. [18] developed acholesterol biosensor based on immobilized cholesterol esterase andcholesterol oxidase to determine the total cholesterol content infoodstuffs, and electrochemical measurements were performed in thecholesterol analysis of food samples.

In this paper we report the immobilization of cholesterol oxidase(COX) in LbL films, with alternating layers of poly(allylamine hydro-chloride) (PAH). In addition to studying the adsorption characteristics ofthe LbL films, including an analysis of the film morphology, we employan array of sensors to detect cholesterol in chicken eggs at µMconcentrations. Combining the analysis of adsorption experimentswith investigation of performance of sensing units is important becausefilmmorphologymay be a key parameter to determine the suitability ofa sensing unit. Indeed, while in some types of devices films with lowroughness may be preferred, rough films may be advantageous forsensors as a high surface area may be convenient [19].

443M.L. Moraes et al. / Materials Science and Engineering C 29 (2009) 442–447

2. Experimental details

2.1. Materials

Cholesterol oxidase (COX) from Cellulomonas species (EC 1.1.3.6)(100 units), cholesterol from egg, poly(allylamine hydrochloride)(PAH) and poly(vinylsulfonic acid) (PVS) were purchased from Sigma-Aldrich. The natural phospholipid L-α-phosphatidyl-DL-glycerol fromchicken egg (EggPG) was obtained from Avanti Polar Lipids. Allreagents were used without further purification. The structure ofsome of the materials is shown in Fig. 1. Yolk of natural egg chickenwas separated from the white part and diluted in Tris–HCl buffer untilthe formation of a 10−3 M cholesterol solution, calculated from the eggnutrition information.

2.2. Enzyme immobilization

Cholesterol oxidase enzyme (COX) was alternately immobilizedwith PAH onto glass substrates using the LbL technique [2]. Thesubstrates were previously cleaned in HCl/H2O2/H2O (1:1:6) andNH4OH/H2O2/H2O (1:1:5) solutions, both for 10 min at 80 °C. Beforethe PAH/COX film fabrication, 2-bilayers of (PAH/PVS) were depositedonto the solid substrates to reduce substrate effects [20]. All solutionswere prepared in 10 mM Tris–HCl buffer pH=7.5. At this pH COX canbe used as a polyanion, as its isoelectric point lies between 4.6 and 5.2[8]. Both PAH and PVS concentrations were 1 mg/mL, while COXconcentration was 0.2 mg/mL. The substrate was immersed in PAHsolution for 3 min, being further rinsed with 10 mM Tris–HCl buffersolution to remove loosely adsorbed molecules. Subsequently, it wasimmersed in the COX solution at 4 °C (controlled temperature) for10min for enzyme adsorption, being once againwashedwith Tris–HClbuffer to remove inadequately adsorbed molecules. This procedurewas repeated until the desired number of PAH/COX bilayers.

The adsorption time profile of COX multilayers was monitored ateach deposition step by measuring the absorbance at 280 nm with aHitachi U-2001 spectrophotometer, which corresponds to the max-imum in absorption. Film morphology was studied by atomic forcemicroscopy (AFM). Fig. 2A shows the scheme of the multilayers byconsecutive adsorption of PAH and COX onto a substrate previouslymodified with two bilayers of PAH/PVS [8].

Fig. 1. Structures of the materials employed.

2.3. Preparation of cholesterol in liposomes

Since cholesterol is insoluble inwater and occurs in cell membranesbetween phospholipids a biomimetic environment was provided usingunillamelar vesicles (SUVs) (liposomes). Hence, 1 mM of EggPG(phospholipid) and cholesterol at various concentrations were homo-geneously dissolved in chloroform to prepare the SUVs. A thin filmwasformedbyevaporating the solventwith a gentle streamof nitrogen, thenthe dry lipid film was hydrated overnight in 10 mM Tris–HCl bufferpH=7.5 resulting in multilamellar vesicles (MLVs). To obtain the SUVsthe eggPG/cholesterolmixturewas sonicated in ten cycles of 30 s,with adelay time of 1 min between the sonication period.

2.4. Electrical measurements

Impedance measurements were carried out with a Solartron 1260Aimpedance/gainphase analyzer in a frequency range from100 to 1000Hz.All measurements were performed using three interdigitated (IDE)electrodes: a bare IDE electrode, one coated with seven PAH/PVS bilayers,and another with two PAH/PVS bilayers covered with five PAH/COXbilayers. The experiments were conducted by soaking the electrodesduring 20 min in liposome/cholesterol solutions containing differentmolar concentrations of cholesterol (0, 10−6, 10−5, 10−4 and 10−3 M), asschematically shown in Fig. 2B. The samemeasurementswere carried outwithyolk egg fromchicken indifferent concentrationsof cholesterol (10−5,10−4 and 10−3M). The pHof the solutionswas buffered at 7.5 using 10mMTris–HCl. The capacitance values obtained by modeling the impedanceresponse with an equivalent circuit [21] were treated with PrincipalComponent Analysis (PCA [22], which is a statistical method correlatingsamples, decreasing the dimensionality of the original datawithout losinginformation. Briefly, it makes a linear combination of the data, creatinganother pair of orthogonal axis into the directions where the highercorrelations exist.

2.5. Morphological characterization

Surface morphology was investigated using a Nanoscope IIIa(Digital Instruments) microscope with 512×512 pixels per imageobtained under ambient conditions in the tapping mode. The averageheight and average diameter of the aggregates were determined usingthe software from Digital Instruments.

3. Results and discussions

3.1. Film growth and adsorption mechanisms for PAH/cholesterol oxidase(COX) LbL films

In order to characterize the COX-containing LbL films, adsorptionwascarried out onto quartz substrates, which are amenable to opticalmeasurements. As alreadymentioned, thequartz substrateswere initiallycoated with two PAH/PVS bilayers to minimize substrate effects. Thekinetics of adsorptionofPAH/COXontoPAH/PVSLbLfilm is represented inFig. 3, which displays the absorption at 280 nm versus time. Thesaturationobservedat ca. 200 s indicates that a complete layer is achievedwithin3min. This adsorptionkineticsmaybemodeledwith the Johnson–Mehl–Avrami (JMA) equation [23–25], which allows one to infer aboutgrowth processes without resorting to molecular-level information [25].The fitting of the data is shown in the solid line, obtained with theparameters of Table 1. The plateau reached in the absorption is related tothe saturation of available sites for COX molecules adsorption fromsolution, repelled by those already adsorbed with the same charge. Thissecond process can be explained with a Johnson–Mehl–Avrami (JMA)function (Eq. (1)) with n~2 and characteristic times of 180 s (see Table 1).

A ¼ k1 1− exp −t1τ1

� �� �þ k2 1− exp −

t2τ2

� �n� �� �ð1Þ

Fig. 2. Schematic representation of COX immobilization in LbL films (A) and a schematic representation of the enzymatic oxidation of cholesterol in liposomes (B).

444 M.L. Moraes et al. / Materials Science and Engineering C 29 (2009) 442–447

where A is the absorbance, k and n are constants and τ (1 and 2) are thecharacteristic times [23]. Avrami's model gives the phenomenologicaldescription of adsorption kinetics and has been successfully used toexplain the growth of polymer films [24,25]. The values of τ1 and τ2 aretypical for adsorption of LbL films. The first process, represented by τ1 isvery fast, usually within a few seconds or less, while the second processhas characteristic times of hundreds of seconds. A value of n≈2corresponds to a 2D adsorption, without forming new nuclei during

Fig. 3. Absorbance at 280 nm versus time of adsorption for the COX layer. The solid linerepresents the fitting with the JMA equation.

the adsorption process. In other words, adsorption takes place with thegrowth of nuclei already formed in the beginning of the process, whichactually corresponds to the first, fast process represented by τ1. This is infact consistent with the morphological analysis presented below.

Fig. 4(A) shows an AFM image for a scanwindow of 10×10 µm2 of a10-bilayer PAH/COX film. Here, we employ information from surfacemorphology to analyze growthmechanisms [24]. For instance, from theheight profiles in Fig. 4B one may infer about the direction of filmgrowth. Fig. 5 shows the average diameter of the aggregates increasinglinearly with the average height, indicating no preferential direction forthe aggregate growth. Note that the average diameter increases muchfaster than the average height. Therefore, the aggregates grow as discs,consistentwith n=2 obtained in the kinetics of adsorption (Table 1). Theerror bars included in the plots of Fig. 5 are average values from fivescans.

We have also used the dynamic scale theory to obtain roughnessexponents. Fig. 6 shows the log–log plot of roughness, W, againstscanning window size, L, from AFM images on PAH/COX films and ascanned area of 10×10 µm2. The roughness is typical of LbL films of non-conventional polyelectrolytes. It is possible to apply scaling laws [26] asW saturates at a given L, which allows one to investigate themorphologyand dynamics of film formation. The parameters involved are theroughness, fractal dimension and the critical exponents, namely, the

Table 1Parameters n, τ andKemployed in the JMA equation for fitting the data for kinetics ofPAH/COX layers

K1 τ1 K2 τ2 n

0.041±0.002 0.8±0.2 0.014±0.003 157±7 1.8±0.1

Fig. 4. (A) AFM image of a 10-bilayer (PAH/COX)n LbL film adsorbed on a quartz substrate with scanning window of 10×10 µm2. (B) Height profile of the film.

445M.L. Moraes et al. / Materials Science and Engineering C 29 (2009) 442–447

exponents of roughness (α) and growth (β), and the dynamic exponent(z). Scaling laws, therefore, help to describe how film morphology,characterized by rms (rootmean square) roughness, evolves during filmgrowth, based on two factors: the time formation of the surface and thesize of the scanningwindow [26,27]. The rms roughnessW is taken as afunction of the scanningwindow size, L, in the formW (L)∝Lα, where αis independent of the units employed for the lateral dimension L. The αshould be calculated when the roughness has already reachedsaturation, as in Fig. 6.

The slopes in the logW×log L plots correspond to the roughnessexponents (α) given inTable2alongwith the fractal dimension (DF=3−α).Lx, the average diameter, corresponds to the crossover length defining

Fig. 5. Log–log plot of average diameter of the domains versus height for a 10-bilayerPAH/COX LbL film.

the transition between the two growth regimes, while L̄ is the averagediameter estimated from the AFM images and calculated with theNanoscope III software (see Fig. 4B). As Table 2 indicates, the dynamicscale analysis provides a length scale that is in good agreementwith theaverage size of domains obtained from AFM.

For length scales larger than the average diameter of the domains,LNLx, the surfaces are characteristic of self-affine fractals that may bedescribed with the Kardar–Parisi–Zhang (KPZ) equation [26,28]. KPZis one of the models used to describe stochastic formation at aninterface, which can be related to the growth of a film surface [29–32].The α exponent in the KPZ model is close to 0.4 [19,26,33–35]. It ispossible that this roughness could decrease if additional bilayers wereadsorbed, as new molecules could occupy empty spaces [36]. For LbLx

Fig. 6. Log–log plot of rms roughness (W) versus size of the scan window (L).

Table 2Lx; root mean square roughness (rms); exponents and fractal dimension (α1 and DF1correspond to the region inside the domains) and L̄

Lx (nm) rms (nm) α1 α2 DF1 DF2 L̄ (nm)

418 8.7 0.61±0.01 0.35±0.05 2.39 2.65 410

Fig. 8. Principal components analysis (PCA) plots using three electrodes: bare, PAH/PVSand PAH/COX. The capacitance values at 1 kHz for: Milli-Q water (●) and egg yolknatural at concentrations of 10−5 (O), 10−4 (X) and 10−3 (+) M.

446 M.L. Moraes et al. / Materials Science and Engineering C 29 (2009) 442–447

(inside the domains) the measured α≈0.6 is characteristic of agrowing kinetics governed by diffusion along the surface [26,37].Indeed, the marked oscillations on top of the domains observed in theheight profiles, see Fig. 4B, are characteristic of a rough surface, incontrast to other LbL films [36,38] for which the surface domains areclose to a Euclidian surface (DF≈2).

3.2. Detection of cholesterol using impedance measurements

Impedance measurements were used to detect cholesterol ineggPG liposome and natural egg yolk using the 3-electrode array withLbL films deposited onto gold interdigitated electrodes (IDEs). Thedata takenwith the 3 electrodes at 1 kHz, which is the frequency mostsensitive to changes in film properties [39–41], were treated withPrincipal Component Analysis (PCA). The PCA plots in Figs. 7 and 8display the results for the electrical measurements in aqueoussolutions with cholesterol in eggPG liposomes or in natural egg yolk,respectively. A clear distinction is made between water and thecholesterol containing solutions at different molar concentrations. InFig. 7 there is a high correlation between the Second Principal Com-ponent (PC2) and the cholesterol concentration. The sensor arraycould also distinguish water from cholesterol egg yolk, as illustrated inFig. 8. In this case a good correlation between PC1 and naturalcholesterol concentration was obtained, with a left hand trend to themost diluted samples. It is worth mentioning that the data presentedin Figs. 7 and 8 show similar results in different sets of independentmeasurements, taken also during different days of analysis. Such highcorrelation was attributed to the presence of specific interactionsbetween COX and cholesterol, increasing the “chemical orthogonality”of the data, as reported in e-tongue analysis of beverages [40].

In PCA plots considering results acquired only with the bareelectrodes and PAH/PVS, i.e. without the cholesterol oxidase film,there is no clear distinction between distinct cholesterol concentra-tions and no correspondence in the data that could be related to thePrincipal Components (results not shown). Possibly, this is due to thelack of specific interactions due to the absence of COX for cholesteroldetection. Nevertheless, in both cases there was a clear distinction

Fig. 7. Principal components analysis (PCA) plots using three electrodes: bare, PAH/PVSand PAH/COX. The capacitance values at 1 kHz for: Milli-Q water (●) and cholesterolinto eggC liposome at concentrations of 0 (O), 10−6 (X), 10−5 (+), 10−4 (⁎) and 10−3 (□) M.

betweenwater and liposomes containing cholesterol. In the measure-ments presented here LbL films with a fixed number of bilayers(seven) were used, and no attempt was made to investigate possibleeffects of the number of layers on the sensitivity. Previous studieshave, however, demonstrated that optimal performance is usuallyattained with 5 to 10-bilayer films because they are sufficiently thin toincrease sensitivity while providing full coverage of the electrodes,which may not occur for very thin films [42].

4. Conclusions

Layer-by-layer (LbL) films of cholesterol oxidase (COX) could beproduced, in which COX layers were alternated with poly(allylaminehydrochloride), and the COX activity was preserved. This demonstratesthe suitability of the film-formingmethod for biomolecules, whichmaybe largely attributed to the mild conditions under which the films arefabricated. Adsorption of a layer of COX was complete within ca. 180 s,with a kinetics that comprises two steps. The nuclei are formed right inthe beginning of the adsorption, and then aggregates grow as disks,according to atomic force microscopy data that were analyzed usingdynamic scaling theories. LbL films adsorbed onto gold electrodes wereused as sensing units, with impedance spectroscopy as the method ofdetection. Owing to the specific interaction with COX in the LbL film,cholesterol could be detected in aqueous solutions with a sensitivityreaching 10−6 M. Because the methodology combining ultrathin,nanostructured films and a.c. electrical measurements can be extendedto many other molecular-recognition pairs, one may envisage itsapplication in clinical diagnosis.

Acknowledgements

This work was supported by FAPESP, CNPq and CAPES (Brazil).

References

[1] K. Ariga, in: S.K. Tripathy, J. Kumar, H.S. Nalwa (Eds.), Handbook of Polyelectrolytesand Their Applications, 2002.

[2] G. Decher, J.D. Schmitt, Thin Solid Films 210 (1992) 831.[3] Y. Lvov, K. Ariga, I. Ichinose, T. Kunitake, J. Am. Chem. Soc. 117 (1995) 6117.[4] J.M.C. Lourenço, P.A. Ribeiro, A.M.B. do Rego, M. Raposo, J. Colloid Interface Sci.

313 (2007) 26.[5] H.S. Silva, T.M. Uehara, K. Bergamaski, P.B. Miranda, J. Nanosci. Nanotech. 8 (2008)

3399.[6] K. Ariga, T. Nakanishi, T. Michinobu, J. Nanosci. Nanotechnol. 6 (2006) 2278.[7] B. Zak, Am. J. Clin. Pathol. 27 (1957) 583.[8] M.K. Ram,P. Bertoncello, H.Ding, S. Paddeu, C.Nicolini, Biosens. Bioelectron.16 (2001)

849.

447M.L. Moraes et al. / Materials Science and Engineering C 29 (2009) 442–447

[9] J.P. Blomhoff, Clin. Chim. Acta 43 (1973) 257.[10] E. Agullo, B. Susna, Food Res. Int. 29 (1996) 77.[11] J. Zicha, J. Kunes, M.A. Devynek, Am. J. Hypertens. 12 (1999) 315.[12] R. Foster, J. Cassidy, E. O'Donoghue, Electroanalysis 12 (2000) 716.[13] Y. Minghui, Y. Yang, H. Yang, G. Shen, R. Yu, Biomaterials 27 (2006) 246.[14] J.C. Vidal, E. Garcia-Ruiz, J. Espuelas, T. Aramendia, J.R. Castillo, Anal. Bioanal. Chem.

377 (2003) 273.[15] C. Bongiovanni, T. Ferri, A. Poscia, M. Varalli, R. Santucci, A. Desideri, Bioelec-

trochemistry 54 (2001) 17.[16] N. Peña, G. Ruiz, A.J. Reviejo, J.M. Pingarrón, Anal. Chem. 73 (2001) 1190.[17] X.J. Wu, M.M.F. Choi, Anal. Chem. 75 (2003) 4019.[18] A.K. Basu, P. Chattopadhyay, U. Roychoudhuri, R. Chakraborty, Bioelectrochemistry

70 (2007) 375.[19] N.C. de Souza, M. Ferreira, K. Wohnrath, J.R. Silva, O.N. Oliveira Jr., J.A. Giacometti,

Nanotechnology 18 (2007) 75713.[20] R.F.M. Lobo, M.A. Pereira-da-Silva, M. Raposo, R.M. Faria, O.N. Oliveira Jr.,

Nanotechnology 14 (2003) 101.[21] D.M. Taylor, A.G. MacDonald, J. Phys. D Appl. Phys. 20 (1987) 1227.[22] K.R. Beebe, R.J. Peel, M.B. Seasholtz, Chemometrics: A Pratical Guide, John Wiley

and Sons, New York, 1998.[23] J.W. Christian, The Theory of Transformation inMetals and Alloys, Pergamon,1963.[24] J.N. Hay, Br. Polym. J. 3 (1971) 74.[25] N.C. de Souza, J.R. Silva, C.A. Rodrigues, A.C. Hernandes, L.D. Costa, J.A. Giacometti,

O.N. Oliveira Jr., Synth. Met. 135 (2003) 121.[26] A.L. Barabási, H.E. Stanley, Fractal Concepts in Surface Growth, Cambridge Univ.

Press, Cambrige, UK, 1995.

[27] N.C. de Souza, J.R. Silva, R.D. Thommazo, M. Raposo, D.T. Balogh, J.A. Giacometti, O.N.Oliveira Jr., J. Phys. Chem. B 36 (2004) 108.

[28] M. Kardar, G. Parisi, Y.C. Zhang, Phys. Rev. Lett. 56 (1986) 889.[29] D.M. Aurongzeb, Solid State Commun. 134 (2005) 165.[30] H.j. Qi, L.H. Huang, J.D. Shao, Z.X. Fan, Acta Phys. Sin. 52 (2003) 2743.[31] C. Oguey, N. Rivier, J. Phys. A Math. Gen. 34 (2001) 6225.[32] P. Neogi, J. Electroanal. Chem. 595 (2006) 1.[33] N. Happo, M. Fujiwara, M. Iwamatsu, K. Horii, Jpn. J. Appl. Phys. 37 (1998) 3951.[34] R.C. Salvarezza, L. Vázquez, P. Herrasti, P. Ocón, J.M. Vara, A.J. Arvia, Europhys. Lett.

20 (1992) 727.[35] A.E. Lita, J.E. Sanchez Jr., Phys. Rev. B 61 (2000) 7692.[36] N.C. de Souza, V. Zucolotto, J.R. Silva, F.R. Santos, D.S. dos Santos Jr., D.T. Balogh, O.N.

Oliveira Jr., J.A. Giacometti, J. Colloid Interface Sci. 285 (2005) 544.[37] Z.W. Lai, S. Das Sarma, Phys. Rev. Lett. 66 (1991) 2348.[38] M. Losche, J. Schmitt, G. Decher, W.G. Bouwman, K. Kjaer, Macromolecules 31 (1998)

8893.[39] B.A. da Silva, P.A. Antunes, D. Pasquini, A.A.S. Curvelo, R.F. Aroca, A. Riul, C.J.L.

Constantino, J. Nanosci. Nanotechnol. 7 (2007) 510.[40] P. Ciosek, W. Wróblewski, Talanta 71 (2007) 738.[41] V. Zucolotto, A.P.A. Pinto, T. Tumolo, M.L. Moraes, M.S. Baptista, A. Riul Jr., A.P.U.

Araújo, O.N. Oliveira Jr., Biosens. Bioelectron. 21 (2006) 1320.[42] M. Ferreira, A. Riul Jr., K. Wohnrath, F.J. Fonseca, O.N. Oliveira Jr., L.H. Mattoso, Anal.

Chem. 75 (2003) 953.