Controlled Electrochemically-Assisted Deposition of Sol− Gel

May 4, 2011 - Department of Analytical Chemistry, Taras Shevchenko National University of Kyiv, Volodymyrska str., 62a, 01033 Kyiv, Ukraine. bS Suppor...
0 downloads 0 Views 3MB Size
ARTICLE pubs.acs.org/Langmuir

Controlled Electrochemically-Assisted Deposition of SolGel Biocomposite on Electrospun Platinum Nanofibers Ievgen Mazurenko,†,§ Mathieu Etienne,*,† Rainer Ostermann,‡ Bernd M. Smarsly,‡ Oksana Tananaiko,§ Vladimir Zaitsev,§ and Alain Walcarius† †

Laboratoire de Chimie Physique et Microbiologie pour l’Environnement, UMR 7564, CNRS  Institut Jean Barriol  Nancy-Universite, 405 rue de Vand!uvre, 54600 Villers-les-Nancy, France ‡ Physikalisch-Chemisches Institut, Justus-Liebig-Universit€at Giessen, Heinrich-Buff-Ring 58, Giessen D-35392, Germany § Department of Analytical Chemistry, Taras Shevchenko National University of Kyiv, Volodymyrska str., 62a, 01033 Kyiv, Ukraine

bS Supporting Information ABSTRACT: The modification of platinum nanofibers by silica using the electrochemically-assisted deposition is reported here. Pt nanofibers are obtained by electrospinning and deposited on a glass substrate. The electrochemicallyassisted deposition of the solgel material then gives the unique possibility to finely tune the silica film thickness around these nanofibers. It also allows the successful encapsulation of a biomolecule (glucose oxidase was chosen here as a model) while retaining its biological activity, as pointed out via the electrochemical monitoring of H2O2 produced upon addition of glucose in the medium. This silicaglucose oxidase composite offers the possibility of comparing systematically the influence of the deposition time on the bioelectrode response and to compare it with the particular features of the deposits. It was found that the film first grew uniformly around the nanofibers and then started to deposit between them, covering the whole sample (fibers and glass substrate), and tended to fully embed the nanofibers for prolonged deposition. The thickness of the silica film is critical for the electroactivity of the biocomposite, the best response being obtained for a silica layer thickness in the range of the fiber diameter (∼50 nm).

’ INTRODUCTION Nanostructuration of electrode surfaces has recently captured a great deal of attention in view of enhancing the performance of electrochemical devices, notably in the field of electroanalysis.1 A major challenge is to increase the electroactive surface area of conventional electrodes in comparison to the geometric ones. This is usually achieved by using molecular templates, supramolecular assemblies, macromolecules, micro- and nano-objects, or even macroscopic porous materials, as templates to generate continuous porous structures onto the surface of solid electrodes, mainly via bottom-up approaches.2 Another way to reach this goal of enhanced performance is resorting to assemblies of conducting nano-objects such as carbon nanotubes,3 metal nanoparticles,4 and nanowires.5 Individually, these ultramicro- or nanoelectrodes offer the advantage of low background charging currents and high faradaic current densities due to convergent diffusion, as well as the possibility of handling small sample size or enabling in vivo analysis, contributing thereby to the miniaturization of devices and systems one can observe in modern analytical chemistry.6,7 In addition, these nano-objects can be arranged into ensembles of nanoelectrodes to circumvent the intrinsic limitation of the individual ones (i.e., ultralow currents).8 Assemblies of metal nanofibers combine the intrinsic electrochemically-chemical features of the bulk metal electrode, with r 2011 American Chemical Society

much larger electroactive areas for the same amount of material, and possibly offer the advantage to be designed into a porous texture and to exhibit catalytic properties. Several methods have been reported to fabricate such metallic networks, including electrospinning,911 template synthesis,1214 and electrodeposition.15 To date, most of the developed 1-D metallic structures were used as such (i.e., not chemically modified), with applications in the fields of hydrogen sensors,16,17 fuel cells,11,1820 or catalysis for fine chemicals.21,22 Of related interest is the application to biosensors of platinum nanowires,23 ensembles of 1-D nanoelectrodes vertically aligned,24,25 and porous 3-D electrodes.26,27 On the other hand, electrodes modified with solgel-derived silica-based materials or related organicinorganic hybrids are increasingly attractive for applications in various fields, including electrochemical sensors and biosensors, power sources, electrochemically-chromism, gas sensors, or permselective membranes.28 The versatility of the solgel process enables quite easy deposition of (organo)silica thin films onto solid surfaces, which is based on the transfer of a sol solution onto a solid support (by dip-coating, spin-coating, or dropping) and its subsequent evaporation with Received: January 7, 2011 Revised: March 11, 2011 Published: May 04, 2011 7140

dx.doi.org/10.1021/la200069z | Langmuir 2011, 27, 7140–7147

Langmuir

ARTICLE

Scheme 1. Pt-Nanofiber Connection and Electro-Assisted Sol-Gel Deposition

concomitant condensation.29 This method can be also applied to solgel film deposition on electrodes,30 but getting deposits of uniform thickness is restricted to basically flat surfaces. An elegant way to overcome this limitation is the electrochemically-assisted generation method,3133 involving the application of a suitable negative potential to the working electrode immersed in a hydrolyzed sol to produce the necessary OH ions that catalyze the precursor condensation in the near vicinity of the electrode surface. In doing so, it is possible to finely control the film thickness (by tuning the applied potential or the deposition time) and to get uniformly deposited solgel layers onto electrode surfaces of complex geometry or complex conductive patterns, such as gold CD-trodes,33,34 macroporous electrodes,35,36 or printed circuits.37 The method is also compatible with encapsulation of biomolecules without preventing their activity,38,39 in agreement with the excellent biocompatibility provided by the solgel silica matrix.40 Except for one example dealing with the electrodeposition of silica films onto single-walled carbon nanotube surfaces,41 the method has not yet been evaluated for covering selectively conductive nanoobjects deposited on insulating supports. We have thus examined in this work the potential interest of the electrochemically-assisted deposition method to coat the surface of platinum nanofibers (Pt-Nfb) networks prepared by electrospinning with thin solgel layers encapsulating a model redox protein (glucose oxidase). As the basic electrochemical properties of such electrospun Pt-Nfb networks are not known, the work started with some voltammetric and scanning electrochemical microscopic experiments, along with the necessary microscopic examinations. Then, discussion is made on the way to control solgel coating of each individual fiber and/or all the fibers together, underlying the critical role of the electrodeposition time on both the film thickness and its resulting response to glucose biosensing.

’ EXPERIMENTAL SECTION Chemicals and Reagents. Tetraethoxysilane (98%) was purchased from Alfa Aesar, hydrogen peroxide, H2O2 (35%) from Acros, and glucose oxidase (GOx, EC 1.1.3.4, from Aspergillus niger, Mw 15 00050 000) from Sigma. Solutions of glucose (Acros) were left to mutorotate at least for 24 h. Phosphate buffer solutions (PBS) adjusted at necessary pH were prepared using Na2HPO4 3 2H2O (99.5%, Merck) and KH2PO4 (99.9%, Prolabo). 1,10 -Ferrocenedimethanol (98%) was from Aldrich. All other reagents were of analytical grade. All solutions were prepared with high purity water (18 MΩ cm) from a Purelab Option water purification system. Apparatus. All electrochemical experiments were carried out using a PalmSens potentiostat and a three-electrode cell, including a Pt-wire auxiliary electrode, an Ag/AgCl reference electrode (3 M KCl, Metrohm) and electrospun-Pt-nanofiber assemblies as working electrode. Surface observations were made using atomic force microscopy (AFM, Asylum Research MFP-3D-Bio) in contact mode using V-shaped

silicon nitride tips (ref MLCT-EXMT-BF, Veeco Instruments) with a spring constant of 0.1 N m1 (manufacturer specifications), and scanning electron microscopy (SEM, Hitachi FEG S4800). Scanning electrochemical microscopy (SECM) measurements were performed with a home-built setup operating with a shear force control of the ultramicroelectrode-to-sample distance.42 Fabrication of the Pt-Nanofiber Assembly (Pt-Nfb). The nanofiber assemblies have been prepared according to a protocol adapted from the literature.11 In a typical procedure, 55 mg Pt(NO3)2 was dissolved in 150 mg DMF (N,N-dimethylformamide) and 75 mg acetic acid under heating, then 150 mg methanol and 150 mg of 12 wt % PVP (poly(vinylpyrrolidone), MW 1 300 000) in methanol were added. The electrospinning was carried out in a homemade setup consisting of a high voltage power supply (Spellman CZE-1000R) and a syringe pump (KDS Scientific 100-CE) with which the solution was fed at a rate of 0.25 mL h1. The applied electrical field was about 0.9 kV cm1, and the distance from the needle tip to the collector was 7 cm. The relative humidity was controlled to 25%. Samples were collected on glass plates for times from 40 to 160 s and subsequently heated with a ramp of 10 °C min1 to 425 °C and maintained at this temperature for 5 min. It is wellknown that during electrospinning residual charges build up on the deposited fiber mats. This slows down the rate of fiber deposition, and therefore, it can be difficult to precisely control the amount of nanofibers. For conducting nanofibers, the fiber density can be estimated from the conductivity of the fiber mat.43 Electrochemical Experiments. Each glass plate with Pt-Nfb was cleaned by a gentle flow of compressed air before use. Scheme 1 shows the simple protocol that was followed to connect the Pt-Nfbs before using them as working electrodes for electrochemical characterization and solgel deposition. A part of its surface was restricted by an O-ring to define a known geometric working area, and the plate was placed at the bottom of a Teflon cell designed with an underlying flat support and an upper Teflon container with a hole of the same size as the O-ring (6 mm diameter) at the bottom. The electroactive surface area was estimated by cyclic voltammetry (CV) using 0.5 M sulfuric acid as electrolyte, at a scan rate of 100 mV s1. 1 mM ferrocenedimethanol in phosphate buffer (pH 6.0) was used for further characterization of PtNfb by CV. A solution of 1 mM ferrocenedimethanol in 0.05 M potassium hydrogen phthalate was used for SECM measurements. Electrochemically-assisted deposition of solgel biocomposite films and electrochemical detection of glucose was performed in a cell designed with a 10-mm-diameter O-ring. The sol composition was selected and adapted from a previously described procedure.38 Briefly, 2.28 mL of TEOS, 2.0 mL of H2O, and 2.5 mL of 0.01 M HCl were mixed with a magnetic stirrer for 16 h. Then, prior to introduction of the enzyme in the medium, 1.66 mL of 0.1 M NaOH was added to neutralize the sol (to avoid possible enzyme denaturation in acidic medium). The enzyme solution (50 μL of PBS (0.067 M, pH 6.0) and 100 μL of 10 mg mL1 GOx solution) was added to 0.5 mL of the hydrolyzed sol and left to stay for 1 h. The electrochemically-assisted deposition was performed in the above-described mixture, typically by applying a potential of 1.2 V for a deposition time ranging from 1 to 10 s. After film formation, the modified Pt-Nfb electrodes were left in the sol for 5 min, softly rinsed with water and dried for 1 h. 7141

dx.doi.org/10.1021/la200069z |Langmuir 2011, 27, 7140–7147

Langmuir

ARTICLE

Figure 1. Characterization of platinum nanofibers. (A) SEM picture; (B) AFM pictures of fibers with scratch and cross-section profile.

’ RESULTS AND DISCUSSION Characterization of the Pt-Nfb Assemblies. Figure 1 depicts typical microscopic characterization data of the Pt-Nfb assemblies. SEM (Figure 1A) and AFM (Figure 1B) imaging reveals that the diameter of individual fibers is in the range 3060 nm. Under the conditions used here, the overall thickness of the twolayers assembly remains limited to ca. 100 nm, as shown in the AFM profile (Figure 1B). Platinum nanofibers deposited by electrospinning are highly interconnected, forming a 2D network of Pt nanoelectrodes on the glass substrate, suggesting good conductivity. This is indeed the case, as the conductivity of the assembly varied typically from 0.5 to 10 kΩ cm1 (two points measurement), depending on the film density (which can be tuned by adjusting either the deposition time and/or the number of deposited layers). As also shown, individual Pt nanofibers can be easily evidenced by AFM, and their diameter can be evaluated quite accurately when they are deposited directly onto the glass support (see the two first features around X = 1 μm in the line scan in Figure 1B). The mechanical stability of the assembly is good enough in solution to perform electrochemical measurements. As shown in Figure 2A, the fiber density affects the electroactive surface area. Samples displaying various densities of Pt nanofibers have been prepared and characterized by cyclic voltammetry in sulfuric acid solution. The estimation made from the integration of the hydrogen desorption peak shows that one fiber layer exhibits an electroactive surface area (0.29 cm2) similar to the geometric surface area defined by the O-ring of the electrochemical cell (0.28 cm2). Comparing to the SEM image (showing much incomplete coverage of the underlying glass support; see Figure 1A), this result confirms the interest of nanoobjects/ nanostructuration to improve the electrochemical performance.2 The electroactive surface area increases proportionally to the number of layers, reaching almost 1.2 cm2 for a four-layers sample. As a consequence, the background current measured in 1 mM KCl solution was also increasing with the number of layers (See Figure S2 in Supporting Information). By contrast, CV curves recorded in the presence of 1 mM ferrocene-dimethanol resulted in peak-shaped amperometric signals, the intensity of which being independent of the fibers density/thickness (Figure 2B). Peak separation was about 0.08 V, independent from the number of Pt-Nfbs layer when using 5 mV s1 potential scan rate suggesting a quasi-reversible redox process (Figure 2B). At higher potential scan rate (100 mV s1), the peak

Figure 2. (A) Cyclic voltammogram of Pt-Nfbs in 0.5 M H2SO4 solution; scan rate 100 mV s1 (inset: dependence of estimated electroactive surface area of Pt-Nfbs as a function of the number of nanofibers layers). (B) Cyclic voltammograms recorded with Pt bare, 1-layer Pt-Nfbs, 2-layers Pt-Nfbs, and 4-layers Pt-Nfbs in 0.1 M KCl containing 1 mM ferrocenedimethanol. (C) Cyclic voltammogram of Pt-Nfbs in 0.1 M H2SO4 solution containing 50 mM H2O2 for (a) Pt bare electrode, (b) 1-layer Pt-Nfbs, (c) 2-layers Pt-Nfbs, (d) 4-layers Pt-Nfbs, scan rate 10 mV s1. (D) SECM characterization: variation of the ultramicroelectrode (UME) current versus the UME-to-sample distance (the experiment has been done successively on (a) Pt bare electrode, (b) platinum nanofibers assembly, (c) platinum nanofibers assembly covered by an electrodeposited thin silica film, and (d) a glass substrate).

separation increased slightly when decreasing the number of PtNfb layers from four (0.155 V) to one (0.114 V). In the same time, peak current intensities were not significantly affected by the number of Pt-Nfbs layers (data for both ferrocene-dimethanol and Fe(CN)63 are reported in Figures S3 to S5 in Supporting Information). Moreover, the linear dependence of the peak current intensity to the square root of potential scan rate 7142

dx.doi.org/10.1021/la200069z |Langmuir 2011, 27, 7140–7147

Langmuir confirms that the electrochemical response to ferrocene-dimethanol was controlled by the diffusion of the probe in solution (Figures S6 to S8 in Supporting Information). These results correspond to the proposed theory44 considering the behavior of Pt-nanofibers as partially blocked electrode and the impossibility of observing any changes in peak current for the redox probes with fast electron transfer rate.45 Despite the nanometric dimension of Pt fibers, the assembly behaves as a macroelectrode with the voltammetric signals governed by linear diffusion of the redox probe, indicating that such nanoelectrode ensembles are characterized by a total overlap regime.46 The situation is very different for the probes with slow electron transfer rate, for example H2O2 (Figure 2C). As predicted by Amatore et al.45 and confirmed by De Leo et al.,25 the reduction and oxidation currents of hydrogen peroxide are strongly affected by the number of nanofiber layers at slow scan rates (compare curves b, c, and d of Figure 2C that correspond, respectively, to one, two, and four Pt-Nfbs layers). Moreover, Pt-Nfbs display in these conditions a better electrochemical detection than the bare Pt electrode (curve a of Figure 2C). The platinum nanofibers show behavior typical of ensemble of 1-D nanoelectrodes, having advantages only for the detection of substances with low electron transfer rates. SECM has also been used here for the characterization of the metallic nanofiber assembly. Figure 2D shows the typical current variation when the ultramicroelectrode-to-sample distance is varied from 1 to 80 μm over a flat/bulky platinum surface (curve a), a platinum nanofibers assembly (curve b), the same platinum nanofibers assembly covered by a thin silica layer (curve c, see discussion in the following section for more details), and a glass substrate (curve d). The measurement on glass and bulk platinum substrates allows one to see the typical current variations for the two limit cases that are called, respectively, negative and positive feedback.47 The positive feedback indicates a facile regeneration of the ferrocene species at the platinum surface, while the negative feedback corresponds to the total absence of regeneration on the insulating glass surface. Pt nanofiber assembly leads to a rather good regeneration of the ferrocene species (i.e., positive feedback), but not as efficient as on bulk Pt. For the point of view of SECM, the assembly of Pt nanofibers does not perform as well as the bulk Pt electrode and a kinetic limitation affects the regeneration process. This sort of kinetic limitation can be found in the literature for ensembles of gold nanoparticles48 or polyaniline filaments.49 This observation is also consistent with the peak separation observed previously in CV (i.e., larger than the expected value for a reversible process). The situation becomes even worse when a thin silica layer has been electrodeposited on the surface of Pt fibers (see curve c in Figure 2D), but in this case, the limitation is mainly due to the insulating character of the silica films (yet porous; see discussion below) which contributes to restrict the access to some platinum fibers and/or to limit the diffusion of the probe to the Pt surface, making the regeneration of ferrocene species more difficult. Such limitation was already reported for silica films electrogenerated onto electrodes of macroscopic dimensions.50 Controlled Deposition of Protein-Containing Silica on Pt Nanofibers. In an attempt to cover only the surface of Pt-Nfb, solgel electrochemically-assisted deposition33 was preferred over the classical evaporation method,29 which is basically restricted to film deposition onto flat surfaces.34,37 The present approach involves the use of a sol solution containing the silica precursors and the protein (here GOx) and the production of OH species

ARTICLE

at the electrode/solution interface by cathodic electrolysis. The so-generated local pH increase leads to a fast (within a few seconds) gelification of the sol into a silica matrix thus growing onto the electrode surface. In such a way, a thin silica film is expected to be produced and the protein can be physically entrapped in the film, as we previously showed for bulk electrodes.38 Here, the particular size and morphology/topography of the Pt-Nfb assembly is likely to strongly affect the process. In addition, the properties of the modified nanofibers (i.e., after silica deposition) are expected to be significantly related to the electrolysis time as it affects the quantity of material deposited on the electrode. All data that are shown in the following sections— AFM, high resolution SEM, and electrochemical detection of glucose—have been obtained from the same batch of Pt nanofibers (two-layers) and using the same silica sol to avoid any deviations that might have occurred if changing such experimental conditions. Note that the electro-assisted solgel deposition method is reproducible, as repeating film formation (3 s electrolysis) from new sols with distinct GOx batches resulted in similar electrochemical responses (as checked for glucose detection). AFM allows monitoring, step-by-step, the electrochemicallyassisted deposition of silica on the platinum nanofiber assembly. Both the image (5  5 μm) and a selected profile permit the visualization of the unmodified fibers and their packing on the glass substrate (Figure 1B). Figure 3 reports the AFM analysis of Pt-Nfb assemblies after electrochemically-assisted deposition for 1 s (Figure 3A), 2 s (Figure 3B), 3 s (Figure 3C), 4 s (Figure 3D), 5 s (Figure 3E), and 10 s (Figure 3F). In addition, high resolution SEM imaging of selected samples (Figure 4) has been performed to confirm and complete AFM data. When applying deposition times less than 5 s, the Pt nanofibers remain visible on the AFM images and the corresponding height profiles. The SEM observation (top-view) shows that a thin layer of material is effectively deposited onto the whole surface of the fibers after 1 s (Figure 4B) and 3 s (Figure 4C). The thickness of such a silica layer cannot be determined accurately, but one can estimate values of about 78 nm and ∼15 nm, respectively, for 1 and 3 s deposits. No deposit was observed on the unmodified fiber (Figure 4A). The depth profile of the deposits is difficult to analyze from SEM pictures, but it can be evaluated from AFM data obtained on scratched samples, pointing out peculiar features as a function of the deposition time. Whereas no silica deposit can be observed in between the individual fibers for 1 and 2 s deposition (the film was only present on the Pt-Nfb surface; see Figure 3A,B), such an “inter-fiber” film started to grow on the glass support from 3 s deposition and beyond. The presence of this additional deposit for longer deposition times can be explained by the fact that the OH catalysts are generated not only in the close vicinity of the Pt-Nfb surface, but also in the diffusion layer located at the electrode/solution interface, which is growing with time. The thickness of the deposit increases with the deposition time from 20 nm (3 s) to 80 nm (4 s). It becomes thicker and starts to cover both the glass and the Pt nanofibers for 5 s deposition, the PtNfb becoming completely encapsulated with 10 s deposition (Figure 3F) or more (not shown). The corresponding SEM picture of the same sample prepared with 10 s electrolysis (Figure 4D) confirms that the Pt nanofibers are embedded within a textured material, the silica-GOx composite (note that a better magnification could not be obtained here due to the insulating property of this layer). To conclude this section, it can be stated that silica deposition occurs essentially on the Pt 7143

dx.doi.org/10.1021/la200069z |Langmuir 2011, 27, 7140–7147

Langmuir

ARTICLE

Figure 3. AFM images and cross-section profiles (along white lines shown in the images) of silica-modified Pt-Nfbs prepared with using increasing electrolysis times: (A) 1 s; (B) 2 s; (C) 3 s; (D) 4 s; (E) 5 s; (F) 10 s.

Figure 4. SEM images of silica-modified Pt-Nfbs prepared for various electrodeposition times: (A) 0 s (blank); (B) 1 s; (C) 3 s; (D) 10 s. 7144

dx.doi.org/10.1021/la200069z |Langmuir 2011, 27, 7140–7147

Langmuir

ARTICLE

Figure 5. (A) Amperometric currenttime response of Pt-Nfbs coated with a silica film encapsulating glucose oxidase, as prepared using increasing electrolysis times: (inset) clean Pt-Nfbs (blank); 1 s; 2 s; 3 s; 4 s; 5 s; 10 s, upon successive additions of glucose: 0.4 mM; 1.2 mM; 3.1 mM; 6.7 mM. Applied potential: þ0.6 V. (B) Dependence of electrode sensitivity on glucose as a function of the electrolysis time used for the solgel electrodeposition of the silica-GOx biocomposite.

nanofibers for very short deposition times (