Journal of Nanoparticle Research

, Volume 11, Issue 6, pp 1311–1319

Nanofabrication by electrochemical routes of Ni-coated ordered arrays of carbon nanotubes

Authors

    • Dip. di Scienze e Tecnologie Chimiche, MINASlabUniversità di Roma “Tor Vergata”
  • Francesco Toschi
    • Dip. di Scienze e Tecnologie Chimiche, MINASlabUniversità di Roma “Tor Vergata”
  • Valeria Guglielmotti
    • Dip. di Scienze e Tecnologie Chimiche, MINASlabUniversità di Roma “Tor Vergata”
  • Elisa Scatena
    • Dip. di Scienze e Tecnologie Chimiche, MINASlabUniversità di Roma “Tor Vergata”
  • Silvia Orlanducci
    • Dip. di Scienze e Tecnologie Chimiche, MINASlabUniversità di Roma “Tor Vergata”
  • Maria Letizia Terranova
    • Dip. di Scienze e Tecnologie Chimiche, MINASlabUniversità di Roma “Tor Vergata”
Research Paper

DOI: 10.1007/s11051-008-9520-y

Cite this article as:
Tamburri, E., Toschi, F., Guglielmotti, V. et al. J Nanopart Res (2009) 11: 1311. doi:10.1007/s11051-008-9520-y

Abstract

Ordered arrays of carbon nanotubes (CNT) have been coated by Ni nanoparticles and Ni thin films by using the chronoamperometry technique for nickel reduction. Two different kinds of nanotube arrays have been used: aligned bundles of CNT grown on Si substrates by chemical vapour deposition (CVD) and networks of CNT bundles positioned via a dielectrophoretic post-synthesis process between the electrodes of a multifinger device. The morphology and structure of the Ni-coated CNT bundles have been characterized by field emission scanning electron microscopy (FE-SEM) and X-ray diffraction (XRD). By changing the parameters of the electrochemical process, it is possible to modulate the morphological characteristics of the Ni deposits, which can be obtained in form of nanoparticles uniformly distributed along the whole length of the CNT bundles or of Ni thin films. A qualitative study of the nucleation and growth mechanism of Ni onto CNT has been performed using the theoretical model for diffusion-controlled electrocrystallization, and a correlation between growth mechanism and samples morphology is presented and discussed. The possibility to maintain the architecture of the pristine nanotube deposits after the Ni coating process opens new perspectives for integration of CNT/Ni systems in magnetic and spintronics devices.

Keywords

Nickel nanoparticlesElectrodepositionCarbon nanotubesOrdered arraysSWCNTNanomanufacturing

Introduction

Ni/CNT materials are emerging as promising systems for many technological applications. Many studies have been performed in the last few years to prepare Ni/CNT systems and some interesting results are here summarized. Nickel nanoparticles have been grown on CNT both by electrochemical and by electroless deposition methods (Arai et al. 2004; Fan et al. 2005; Chen et al. 2002; Jin et al. 2007; Shi et al. 2006; Cheng et al. 2006; Ang et al. 2000; Ayala et al. 2006; Bittencourt et al. 2007). These Ni-decorated CNTs are used as catalyst for oxidation of ethanol (Jin et al. 2007), for electrochemical detection of traditionally ‘non-electroactive’ amino acids (Deo et al. 2004) and are proposed to achieve low-resistance ohmic metal-nanotube contacts thus improving CNT integration in new nanodevices (Tersoff 2003; Mann et al. 2003).

Metal composite powders and films have been prepared by electrodeposition techniques (Chen et al. 2002; Jin et al. 2007; Shi et al. 2006) and tested as protective coating against corrosion (Chen et al. 2005). Electrodeposition methods present some advantages to respect electroless techniques. In fact, it is an economic and easy to perform method that allows the quantitative control of the deposited metal and does not need drastic process conditions.

As first observed by Arai et al. (2004), the Ni is preferentially electrodeposited on CNT defects which have highest chemical reactivity. By using a proper potential, a selective electrochemical deposition can be performed and this method has been used to detect the point defects in nanotubes (Fan et al. 2005). Lowering the deposition potential, Ni nucleates without selectivity and a uniform coverage of CNT can be obtained (Fan et al. 2005).

Ni electrodepositions on CNT are usually performed by adding CNT directly in the Ni-plating bath. Following this procedure, the nanotubes are codeposited together with Ni on the cathode during the electrochemical process (Chen et al. 2002; Jin et al. 2007; Shi et al. 2006). Alternatively, Ni is deposited directly on CNT-modified electrode, but the electrode preparation is long, and the use of several complex chemical reactions is often necessary (Jin et al. 2007; Deo et al. 2004). In general, until now the coated or decorated Ni/CNT systems are characterized by a random distribution of CNT and not ordered arrays of Ni/CNT are obtained. Conversely, the use of CNT/Ni for advanced technological applications, i.e. electronics, magnetics, spintronics, needs to satisfy strict geometrical requirements for the integration of the material in the devices.

In this article we describe an electrochemical approach which enables to control the architecture of the CNT/Ni systems. The electrodeposition of Ni nanoparticles has been performed on CNT arrays either deposited on Si plates by CVD synthesis or aligned on a multifinger device using dielectrophoresis. An analysis of deposition mechanism is also reported and discussed.

Experiments

Preparation of nanotube arrays

Two kinds of CNT samples have been used in the present study as templates for fabrication of hybrid nanotube–Ni nanowires. The first type of samples consists of CNT arrays grown by hot-filament chemical vapour deposition (HFCVD) and named CVDCNT; the second type is CNT aligned by dielectrophoretic process and named DEPCNT.

The quality and the general features of the nanotube samples prepared in our laboratories are routinely checked by field emission scanning electron microscopy (FE-SEM), electron diffraction (RHEED) and Raman spectroscopy (Terranova et al. 2000; Orlanducci et al. 2003). The structural characterizations of the Ni-coated CVDCNT and DEPCNT samples used in the present experiments have been performed by X-ray diffraction (XRD).

CVDCNT

The nanotubes are grown in a HFCVD reactor using carbon nanopowders as carbon source and in situ produced atomic H on Si substrate. Fe is used as catalyst. The synthesis procedure and parameters are described in Terranova et al. 2000. In our modified HFCVD apparatus (Terranova et al. 1999), a carrier gas is used to transport the powders in the reactive area of atomic H plasma. The direction of carrier flux can be controlled in such a way that it allows us to select the growth direction of nanotubes with respect to the substrate surface.

Some relevant parameters are substrate temperature 900 °C, Ar as carrier gas, filament temperature 2100 °C. A solution of Fe(NO3)3 · 9H2O is deposited on the substrate by drop casting and then reduced to the catalyst metal form by atomic hydrogen.

The obtained CNT are prevalently single wall (SWCNT) with a diameter ranging from 1.2 to 2.0 nm about and can be grown with various alignments to respect the substrate and on selected area using substrates patterned by lithography (Orlanducci et al. 2003).

DEPCNT

Under the influence of an alternate electric field, the nanotubes dispersed in a proper solvent are found to move and align by means of a dielectrophoretic process (Krupke et al. 2003; Terranova et al. 2007).

For the preparation of DEPCNT samples, commercial nanotube materials obtained by arc discharge (SWCNT content: 50–70%) have been used. The as-received material was purified using HNO3 solution and following the protocols described in Terranova et al. (2007).

A controlled amount of purified SWCNT was dispersed in CHCl3 by ultrasonic bath for 30 min. A controlled volume of this dispersion was deposited onto an interdigitated electrode platform (gold electrodes, with 40 μm spacing, evaporated on SiO2 layer grown on Si substrate). Efficient conditions for alignment of SWCNT between the interdigitated electrodes were achieved using an AC field with frequency of 1 MHz and 10 Vpp. More details about the optimization of dielectrophoretic parameters are reported in Terranova et al. (2007).

Ni electrodeposition process

Electrodeposition processes were performed with a Palm Sens Instrument at 25 °C using a standard three electrode cell: an Ag/AgCl/Cl electrode and a Pt wire were, respectively, used as reference and counter, whereas the CVDCNT-on-Si and the DEPCNT-on-multifinger samples were adopted as working electrodes.

The nickel-plating solution was constituted of 1 M Ni(NH2SO3)2 · 4H2O, 0.17 M NiCl2 6H2O and 0.24 M HBO3. The pH of the solution was adjusted to 3.99.

The Ni electroreduction was carried out by means of chronoamperometry, which is a current transient technique. A potential step was applied from an initial potential at which the nickel reduction is negligible up to a final fixed potential at which the formation of Ni nuclei and their growth can be observed directly by monitoring the current. The applied potentials were selected from the corresponding cyclic voltammetry.

Results and discussion

In Fig. 1 the cyclic voltammetry for the CVDCNT electrode in the nickel-plating solution is reported. In the first cycle, the onset of Ni2+ reduction occurs at −1 V (vs. Ag/AgCl/Cl), with a characteristic diffusion limited grow peak at −1.15 V (vs. Ag/AgCl/Cl). After the deposition peak, the current again increases, probably due to the reduction of water adsorbed on the electrode surface. In the reverse scan, no anodic peak can be observed and this confirms that the electrochemical stripping of the Ni coating is not significant.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9520-y/MediaObjects/11051_2008_9520_Fig1_HTML.gif
Fig. 1

Cyclic voltammetry for a CVDCNT electrode in the nickel-plating solution

Considering these results, in the chronoamperometric experiment the CVDCNT electrode was switched from −0.5 to −1.15 V, keeping the voltages constant for 2 and 10 s, respectively. In Fig. 2 the current versus time curve for the deposition of Ni on CVDCNT is reported. We can observe that after the switching of the electrode to −1.15 V, the chronoamperogram presents three transients of current density. The first transient (1) is characterized by a maximum of the current density at very short times and is followed by a fast decrease of the current density down to a minimum whose coordinates are t0, j0. This transient is known as the induction time τ and corresponds to the formation of the Ni first nuclei on the electrode surface after the charging of the double layer. The second transient (2) exhibits a current increase with formation of a peak whose coordinates are tmax, jmax, whereas the final transient (3) is characterized by a slight increase of the current density. The behaviour of these two last transients is better evidenced in the inset of the Fig. 2 where the chronoamperogram is rewritten with respect to the onset of the deposition current; therefore, t and j result corrected for the induction time τ.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9520-y/MediaObjects/11051_2008_9520_Fig2_HTML.gif
Fig. 2

Chronoamperogram for the Ni deposition on CVDCNT electrode. The voltage is kept constant at −0.5 V for 2 s and at −1.15 V for 10 s. The induction time τ, the three current transients and the maximum current density (tmax; jmax) are indicated. In the inset, the chronoamperogram corrected for the induction time τ is also shown

The mechanisms involved in the electrodeposition of a metal onto a substrate are subjected to several kinetic and thermodynamic factors which deal with the interaction energy between the adsorbed metal atom and the substrate, and with the difference in interatomic spacing between the metal phase and the substrate (Budevski et al. 1996). Even if much work has been done on investigating these mechanisms, there is still considerable controversy regarding the most basic principles involved in modelling such systems (Fletcher 2002a, b; Abyaneh and Fleischmann 2002a, b). Anyway, it is commonly accepted that electrodeposition occurs by a process of nucleation and growth characterized by an initial appearance of metal nuclei at active sites on the substrate and by the subsequent grow of such nuclei via the incorporation of further ions from the solution (Hyde and Compton 2003). The nucleation and growth processes can be controlled by charge transfer or mass transfer reactions depending on the experimental conditions adopted during the electrodeposition. In particular, high metal ion concentrations and low deposition overpotentials favour a deposition mechanism controlled by charge transfers, whereas low concentrations and high overpotentials promote diffusion-controlled reactions. In the early 1980s, nickel electrodeposition on vitreous carbon electrodes has been largely studied in the frame of these two kinds of limiting experimental conditions looking for theoretical models able to explain the experimental results. The most important models derived to determine the nucleation and growth mechanisms are the one elaborated by Abyaneh and Fleischmann (1981) for electrodeposition controlled by charge transfer reactions and the one elaborated by Scharifker and Hills (1981) for diffusion-controlled processes. Both the models allow one to classify the experimental current transients into the two limiting nucleation mechanisms: the instantaneous and the progressive ones. In the instantaneous nucleation, the number of nuclei is constant and the nuclei slowly grow on relatively few active sites, all activated at the same time. The progressive nucleation instead corresponds to a fast growth of the nuclei on many active sites, activated during the course of the electroreduction.

We performed a non-conventional nickel electrodeposition using electrolytes with high Ni2+ concentration, applying high overpotentials and most of all adopting a non-regular surface substrate (i.e. array of carbon nanotubes) as working electrode. As a consequence, the experimental conditions in which we performed the deposition strictly fall in neither of the two limiting experimental conditions above described. Anyway, we find that the model to determine the nickel nucleation mechanism which best explains our chronoamperometric measurements is the one that utilizes the coordinates of chronoamperometric peaks as described by Scharifker and Hills. In this approach, the current–time relations describing instantaneous and progressive mechanisms are expressed in terms of the maximum current density, jmax, and of the time at which the maximum current density is observed, tmax (Scharifker and Hills 1983). The relations can be written as the following equations for instantaneous and progressive nucleation, respectively:
$$ \frac{{i^{2} }}{{i_{\max }^{2} }} = 1.9542 \cdot \left( {\frac{{t_{\max } }}{t}} \right) \cdot \left[ {1 - \exp \left( { - 1.2564\frac{t}{{t_{\max } }}} \right)} \right]^{2} $$
(1)
$$ \frac{{i^{2} }}{{i_{\max }^{2} }} = 1.2254 \cdot \left( {\frac{{t_{\max } }}{t}} \right) \cdot \left[ {1 - \exp \left( { - 2.3367\frac{{t^{2} }}{{t_{\max }^{2} }}} \right)} \right]^{2} $$
(2)
Figure 3 shows the experimental chronoamperogram plotted in (j/jmax)2 versus t/tmax coordinates compared with the theoretical curves for progressive (dotted line) (Eq. 2) and instantaneous nucleation (solid line) (Eq. 1). Note that the plotted chronoamperogram is the one corrected for the induction time τ (shown in the inset of Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9520-y/MediaObjects/11051_2008_9520_Fig3_HTML.gif
Fig. 3

Corrected chronoamperogram plotted in dimensionless form compared with the theoretical curves for progressive (dotted line) and instantaneous (solid line) nucleation mechanism

We can see as the chronoamperometric data reveal behaviour intermediate between the two limiting cases; in fact, the peak current follows the progressive nucleation in the early stage of deposition, after that it switches to instantaneous nucleation with the continuation of the reaction. Anyway, due the complexity of our experimental system, it is difficult on the basis of our experimental data, to discriminate between the nucleation mechanisms, and therefore, we cannot exclude that processes controlled by charge transfer reactions may occur during the nickel nucleation and growth upon carbon nanotube-based electrode.

In any case, the morphological investigation by FE-SEM is compatible with the model for Ni deposition adopted for the analysis of the electrochemical data.

Figure 4a, b show the Ni coating after the first 2 s of deposition on the CVDCNT electrode. We can observe that at this time a lot of Ni nanoparticles have already been grown on the CNT bundles. The Ni appears in form of chains formed by distinct particles placed one next to the other along the tubes, with some particles starting to coalesce (Fig. 4b). Moreover, it is evidenced as Ni particles with different dimensions are found along the tubes likely as consequence of the progressive nucleation which characterizes the first step of deposition. At the conclusion of the process (10 s), the metal particles have all merged and all the CNT bundles in the sample are coated by the metal (Fig. 4c). It is interesting to note as the general organization of the pristine nanotube deposit can also be maintained during the deposition as one can see in Fig. 4d, where a perpendicularly aligned CNT array coated by nickel is reported.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9520-y/MediaObjects/11051_2008_9520_Fig4_HTML.jpg
Fig. 4

a, b FE-SEM images of the Ni coating after the first 2 s of deposition on the CVDCNT electrode. c FE-SEM image of the Ni coating after 10 s of deposition. d FE-SEM image of vertically aligned CNT arrays coated by Ni

For deposition times exceeding 10 s, a thin continuous film starts to form on the electrode surface, and it can be observed as the morphological details of the CNT bundles are partially lost (Fig. 5a).
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9520-y/MediaObjects/11051_2008_9520_Fig5_HTML.jpg
Fig. 5

a FE-SEM image of the Ni coating for 15 s deposition. b FE-SEM image of a Ni-coated dense CNT deposit. The inset shows the morphology of the CNT deposit before Ni electrodeposition

A different situation occurs in the presence of very dense nanotube arrays. As an example, in Fig. 5b one can observe the FE-SEM image of a CNT deposit coated by Ni. In the inset of Fig. 5b, the morphology of the same sample before the Ni electrodeposition is shown.

For all the samples characterized by high density of CNT bundles, it is found that the Ni electrodeposition produces modifications of the initial organization of the CNT. In particular, after 2–3 s of deposition the CNT bundles are seen to coalesce forming separated bumps, whereas the metal is deposited as a uniform coverage that encompasses the tops of the conical objects. This effect can be likely ascribed to the interaction between CNT bundles and the aqueous medium constituting the Ni-plating solution. Because of the hydrophobic nature of the nanotubes, densely packed structures could limit the surface available for the electrodeposition producing separate forelocks of CNT coated by the metal.

In the case of the DEPCNT samples, we did not perform an electrochemical study in view of the fact that the electrode configuration (interdigitated multifingers connected each other by means of the CNT) does not allow one to discriminate between the Ni electrodeposition on the CNT and on the Au multifingers. Anyway, the cyclic voltammetry performed by the DEPCNT electrode indicated a Ni reduction potential in the range of −1.1 and −1.2 V; therefore, we performed the Ni electrodeposition on this electrode with the same electrochemical parameters adopted for the CVDCNT electrode.

In Fig. 6a–c FE-SEM images of Ni-coated DEPCNT samples are reported. It is evident that the feature of the CNT deposits on the multifinger influences the deposition of the nickel. On the CNT directly connected to the finger electrode (Fig. 6c), the electron transfer probably is facilitated because the metal fingers are expected to increase the rate of charge transfer to nickel ions in solution. As a consequence, a greater amount of Ni is deposited. On the other hand, the inhomogeneity of the electric field lines crossing the multifinger seems to be reflected in the inhomogeneity of the Ni deposit along the tubes lengths.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9520-y/MediaObjects/11051_2008_9520_Fig6_HTML.jpg
Fig. 6

ac FE-SEM images of DEPCNT samples. The photos show some Ni-coated CNT bundles inside the track of the interdigitated electrode

Moreover, the amount of deposited Ni on closely packed CNT bundles is lower than the amount deposited on the CNT bundles that directly contact the metal finger. These results indicate that the charge transfer in systems formed by CNT bundles is not as efficient as in bundle/metal-finger systems.

The morphological differences between CVDCNT and DEPCNT samples can be likely rationalized taking in account the different electron transfer. With respect to the DEPCNT deposits on the multifinger device, the CVDCNT samples present in general a better uniformity of the coating, associated to a lower thickness of the deposit. This probably is due to different kinetics of the Ni reduction processes on the two samples, but at the moment no more definitive conclusions can be drawn, because of the many differences between the two working electrodes.

The crystalline characteristics of the Ni deposits have been studied by means of a Seifert-XRD 3003 instrument, using the Cu Kα radiation (E = 8041.3 eV). Figure 7 shows the θ–2θ XRD spectrum of a CVDCNT sample coated with Ni (time deposition: 10 s). The positions and the relative intensity of the two signals at 44.41° and 51.76° can be attributed to the cubic Ni (Fm3m) (Powder Diffraction File). The peak sharpness reveals a crystal good quality of the metal deposit. Both the spectra acquired on Ni-coated CVDCNT and on the DEPCNT samples give the same crystalline phase.
https://static-content.springer.com/image/art%3A10.1007%2Fs11051-008-9520-y/MediaObjects/11051_2008_9520_Fig7_HTML.gif
Fig. 7

Typical θ–2θ XRD spectrum of Ni nanoparticles generated by a 10 s deposition on a CVDCNT sample. The signal (*) relative to the silicon substrate is also reported

Conclusions

In this article, Ni deposition by electrochemical methods has been used for the fabrication of ordered SWCNT–Ni composite arrays.

Two kinds of samples with different architectures have been prepared and thereafter coated by Ni: vertically aligned SWCNT grown by the CVD technique and SWCNT aligned by dielectrophoretic method. A remarkable feature is, for both the type of samples, the preservation of the alignment of the nanotube arrays.

An analysis of the mechanism of Ni electrodeposition on the nanotube deposits has been carried out adopting the theoretical model for diffusion-controlled nucleation and growth of metal. The chronoamperometric data revealed behaviour intermediate between the progressive and instantaneous nucleation in the early stages of the deposition process. Anyway, the complexity of the experimental system does not allow us to draw exhaustive conclusions about the real growth mechanism of the metal on carbon nanotubes.

These ordered nanotube–Ni composite arrays hold the potential for further optimization by systematically adjusting the process parameters. A variety of promising nanotube–Ni systems are expected to be fabricated on patterned electrodes or on substrates with non-planar geometries (wires, tips). These hybrid nanomaterials demonstrate promise for a series of magnetic and spintronic applications: from quantum information processing to magnetic data storage, from nanomagnets to nanosensors, from biosystem probes to medical imaging.

Some preliminary results have moreover suggested a high catalytic activity of the nanotube/Ni systems for the electrooxidation of ethanol and glucose.

Copyright information

© Springer Science+Business Media B.V. 2008