Decorating carbon nanotubes with nanostructured nickel particles via chemical methods

Decorating carbon nanotubes with nanostructured nickel particles via chemical methods

Chemical Physics Letters 431 (2006) 104–109 www.elsevier.com/locate/cplett Decorating carbon nanotubes with nanostructured nickel particles via chemi...

900KB Sizes 0 Downloads 46 Views

Chemical Physics Letters 431 (2006) 104–109 www.elsevier.com/locate/cplett

Decorating carbon nanotubes with nanostructured nickel particles via chemical methods P. Ayala a

a,*

, F.L. Freire Jr. a, L. Gu b, David J. Smith b, I.G. Solo´rzano c, D.W. Macedo c, J.B. Vander Sande e, H. Terrones d, J. Rodriguez-Manzo d, M. Terrones d

Departamento de Fı´sica, Pontifı´cia Universidade Cato´lica do Rio de Janeiro, Caixa Postal 38071, 22453-970 Rio de Janeiro, RJ, Brazil b Center for Solid State Science, Az State University, Tempe, AZ 85287, USA c Departamento de Ciencia dos Materiais e Metalurgia, Pontifı´cia Universidade Cato´lica do Rio de Janeiro, Caixa Postal 38071, 22453-970 Rio de Janeiro, RJ, Brazil d Advanced Materials Department, IPICYT, Camino a la Presa San Jose´ 2055, 78216, San Luis Potosı´, Mexico e Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA Received 24 April 2006; in final form 11 September 2006 Available online 20 September 2006

Abstract An alternative method for decorating carbon nanotubes (CNTs) with Ni nanoparticles is described. In this work our purpose is to establish a novel and low cost alternative route for coating the outer walls of multiwalled CNTs with metals such as Ni. Using nanotubes during the dissociation of nickel nitrate upon dehydration, it is possible to obtain nanostructured nickel oxide. After reduction of this metal oxide in H2 nickel nanoparticles are obtained and randomly anchored to the surface of nanotubes. Comparative experiments were carried out using pure and nitrogen-containing CNTs. A detailed characterization of the products is also presented. Ó 2006 Elsevier B.V. All rights reserved.

1. Introduction Over the last decade, various publications have emphasized the outstanding physicochemical characteristics of Carbon Nanotubes (CNTs) [1]. The study of these nanostructures opened up an important field of research with enormous potential for developing new technologies [2]. Concerning mechanical properties, carbon nanotubes could exhibit high strength-to-weight ratio; their individual Young’s modulus has been measured to reach about 1 TPa [3–5]. In fact, the excellent elasticity and flexibility of CNTs make them attractive as alternative reinforcing filler in composite materials. Numerous authors have reported the incorporation of these tubular structures in diverse matrices, although the most reproducible and successful advances have been achieved when fabricating ceramic and polymer composites [6,7]. *

Corresponding author. Fax: +55 21 3527 1040. E-mail address: [email protected]fis.puc-rio.br (P. Ayala).

0009-2614/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.cplett.2006.09.039

Regarding the production of metal–nanotube composites, some attempts have been carried out. However, this is not a trivial problem because controlling the wetting properties of nanotubes becomes a crucial issue that could ensure strong interactions between the CNT surfaces and the metal matrix. As stated by Dujardin et al. [8], materials with high surface tension, such as metals, do not spontaneously wet CNTs. Therefore, alternative ways of processing metal–nanotube composites need to be explored in depth [9,10]. Most published work to date has focused on the coating and filling of carbon nanotubes with metals. However, the difficulty of obtaining a uniform nanotube coating imposes limitations. This uniformity is desirable and we envisage that it should be possible to form homogenous nanowires by coating the outer tube walls uniformly with different metals. Several publications have shown that, depending on the metal, the uniformity of the coating varies drastically depending on the metal coating desired. In this respect, some experimental work was reported on the decoration of CNTs with Ni and other metals [9–13], using

P. Ayala et al. / Chemical Physics Letters 431 (2006) 104–109

electron-beam deposition of metallic layers with different thicknesses. Theoretical as well as experimental investigations have also been carried out in order to explain the interactions of metal layers with graphitic surfaces, indicating that metals usually exhibit non-covalent interactions with hexagonal layered carbon. Although it is not yet certain, it is generally accepted that van der Waals interactions are responsible for this apparent interface adhesion in metal– CNTs, rather than chemical bonds [14]. This Letter aims to establish a novel and alternative route for coating the outer walls of multiwalled CNTs with metals such as Ni. In this context, we should emphasize that it has been proved that Ni is able to establish enhanced binding interaction with highly curved carbon layers [15]. We believe that these interactions could be enhanced if structural nanotube surface defects or oxide layers are present. The experimental method we report is based on the work of Brocchi et al. [16], who suggested that some compounds such as nickel nitrate could be easily dissociated into nanoscale particles (in this case leading to the formation of the metal oxide). In this Letter, we provide evidence that such chemical transformations occur in the presence of CNTs, consequently leading to Ni decorated tubes, after nickel oxide reduction in the presence of hydrogen. In particular, we report a simple and inexpensive way of decorating different types of carbon nanotubes with nanostructured nickel using wet chemical routes in ambient conditions. We envisage that this material could be used as a matrix precursor to fabricate homogenous metal– nanotube composites with enhanced elastic properties. Scanning electron microscopy (SEM) was used for a general morphology study of the resulting material. A careful characterization was performed using high resolution electron microscopy (HREM) and electron diffraction techniques. Chemical analysis was carried out using energy-dispersive X-ray spectroscopy (EDX) and electron-energy loss spectroscopy (EELS) in conjunction with scanning transmission electron microscopy (STEM). 2. Experimental

105

Table 1 Nanotube growth conditions and morphologies Solvent

Reaction temperature (°C)

Concentration of Fe(C5H5)2 (wt%)

Overall morphology

C7H8 C7H9N

800 800–850

1.5–2% 1.5–2%

Long MWCNTs N-doped MWCNTs exhibiting bamboo-like morphologies

in N-containing carbon nanotubes that usually exhibit a bamboo-like morphology [18], as previously reported by Terrones et al. [19]. All nanotube materials were collected by scratching the quartz tube walls. The powder samples were dispersed in isopropanol, and a few droplets of the suspension were transferred onto a holey carbon grid for TEM and HREM observation and analysis. For this purpose a JEOL 2010 and a JEOL4000EX Transmission Electron Microscopes were used. The latter instrument, operating at 400 keV, was employed for a detailed structural characterization of the tubes as well as the material described in the next section. 2.2. Formation of nanostructured nickel and decoration of nanotubes Ni oxide is obtained by the dehydration of the metal nitrate in solution with deionized water. Preparation of the nickel oxide is carried out through the reaction of 99% pure nickel nitrate Ni(NO3)2 dissolved in H2O, and dissociating it at 500 °C according to the following reaction: 2Ni(NO3 )2ðsÞ ! 2NiOðsÞ + 4NO2ðgÞ + O2ðgÞ For the nanotube decoration, we carried out the same Ni(NO3)2 dissociation but this time in the presence of our synthesized nanotubes. The NiO + nanotubes synthesis is carried out by homogeneously dispersing nickel nitrate together with 1%wt of nanotube powder (pure CNTs and CNx were used in separate experiments) in deionized water. Mild sonication is initially required in order to avoid nanotube agglomeration caused by their low solubility in water. nanotubes

2.1. Nanotube synthesis and characterization

2NiðNO3 Þ2ðsÞ

Experiments were carried out using pure multi-walled CNTs and N-doped multi-walled CNTs synthesized by the spray pyrolysis method reported by Kamalakaran et al. [17]. This method is able to produce high yields of nanotubes (g/h), and involves the atomization of a solution released from a glass liquid container through a quartz tube preheated at 800–850 °C in an Ar atmosphere. In our experiments, Ferrocene (Fe(C5H5)2(s)) was dissolved in Toluene (C7H8) or Benzylamine (C7H9N). These two solutions were then pyrolyzed in order to obtain two different types of tubular structures as summarized in Table 1. The experiments using benzylamine as the solvent resulted

The solution was then dehydrated at 500 °C and later annealed in ambient conditions at the same temperature. After obtaining the nanotube–metal oxide product, an additional step is required to reduce the oxide to metallic nickel. For this purpose it is possible to propose the following reaction when hydrogen is present in a reducing atmosphere.

!

2NiOðsÞ þ 4NO2ðgÞ þ O2ðgÞ

NiO+H2ðgÞ !Ni+H2 O We used two different reducing atmospheres at 650 °C: extra high purity 99% H2 and a mixture of Ar 95% with H2 5%. The samples were placed into a quartz tube and heated

106

P. Ayala et al. / Chemical Physics Letters 431 (2006) 104–109

until attaining the reduction temperature in presence of Ar flow. Subsequently, the reduction was carried out for 1 h in a hydrogen flow of 100 sccm for the pure H2 and 150 sccm for the Ar/H2 gas mixture. The nickel oxide obtained in the presence of the tubes as well as the Ni-decorated tubes were carefully characterized by HREM. Annular dark field scanning transmission electron microscopy as well as energy dispersive X-ray spectroscopy elemental mapping were performed using a VG HB603 dedicated STEM equipped with a large X-ray detector operating at 250 kV. Energy-filtered imaging was performed in an FEI Tecnai F20 equipped with a post-column Gatan imaging filter (GIF). 3. Results and discussion In order to compare the surface reactivity and the metal decorating ability, we have analyzed the differences when using pure CNTs and N-doped CNTs. Thus, the morphological characterization of the nanotubes used as precursor material is fundamental, as significant differences were observed for the multiwalled nanotubes from toluene solutions and the N-containing tubes produced with benzylamine. We noted that the pure carbon MWCNTs obtained when pyrolyzing toluene:ferrocene solutions were heterogeneous regarding the tube diameter distribution but consisted of nested carbon cylinders (Fig. 1a and b). The encapsulation of Fe-metal inside the nanotube cores is a common feature observed in our samples. In some cases long Fe nanowires were witnessed, although the Fe material was most commonly found as small agglomerates inside the nanotube cores. The presence of straight and parallel graphite fringes suggests a good crystallinity within the produced tubes, prior to the decoration experiments (see Fig. 1b). It is also noteworthy that amorphous carbon

impurities were not usually evident and appeared in very low concentrations in rare cases. The material resulting from the pyrolysis experiments with benzylamine:ferrocene mixtures (N-containing nanotubes) were also tubular structures with a similar range of diameter distribution but exhibiting a characteristic bamboo-like morphology (Fig. 2a). They also displayed a certain amount of nitrogenated defects on their outermost surfaces, and occasionally an incomplete layer coating on the outer layers (Fig. 2b). The nitrogen is presumably found either encapsulated inside the tubes or incorporated within the walls in very low amounts (<2% confirmed by EELS measurements), thus introducing structural defects on the crystalline graphene walls. In addition, these tubes are rather pure and contained almost no foreign particles such as amorphous carbon or encapsulated metal particles. Following the nickel nitrate decomposition, the resulting NiO material was characterized prior to the reduction process. For the NiO synthesized in the presence of MWCNTs as well as CNx nanotubes, it was not possible to determine a nanotube concentration by means of electron microscopy, since TEM and STEM images do not reveal clearly their presence in the intermediate product. Identification of the carbon nanotubes by Analytical Electron Microscopy resulted unfeasible because the carbon nature of the supporting holey film used for the microscopy characterization combined with the strong signals from Ni and O, blurred out the tubular carbon features in the bulky NiO material (figure not shown here). The decorated nanotube structure is only revealed once the reduction of the NiO–nanotube material in a hydrogenrich atmosphere. We observed different particle sizes assembled on the tubes depending on reduction time, nanotube type, gas flow and reduction temperature. Thus, we succeeded to obtain a uniform distribution of metallic Ni

Fig. 1. (a) Bright field low magnification image of pure MWCNTs showing the overall distribution and morphology prior to nanotube decoration experiments and (b) high resolution image of a MWCNT.

P. Ayala et al. / Chemical Physics Letters 431 (2006) 104–109

107

Fig. 2. (a) Low magnification TEM image and (b) HREM of a N-doped carbon nanotube exhibiting the characteristic features of the bamboo-like structure.

nanoparticles attached along the surfaces of the tubes (pure CNTs and N-containing CNTs). For both types of nanotubes, the lower hydrogen content in the carrier gas, and longer reduction times resulted in the formation of Ni particles of uniform size (e.g. 4–15 nm). Both types of tubes have given similar results regarding the adhesion characteristics of Ni particles onto the superficial layers of the tubes, as observed by the SEM and TEM characterization in Figs. 3–5. Pure multiwalled CNTs tend to keep their crystalline structure, and exhibit mainly nickel particles ranging from

4 to 15 nm attached on their outer layer. However, for Ndoped tubes, which to some extent exhibit structural defects presumably induced by the nitrogen incorporation, we found less size dispersion of the Ni material attached on the surface (exhibiting a higher yield between 8 and 14 nm). The particles attached reveal higher average size when compared to the particles attached on pure MWCNTs (see Figs. 4d and 5d). Nevertheless, this does not seem to influence the contact angle observed between the Ni particles and the tube surfaces.

Fig. 3. (a) Bright field STEM image showing Ni particles on N-doped CNTs; (b) X-ray elemental mapping showing that the attached particles are Nickel and (c) high resolution TEM image of Ni particles attached to the outer walls of the CNx tubes.

108

P. Ayala et al. / Chemical Physics Letters 431 (2006) 104–109

a)

100nm

b) Relative Yield

CNx Nanotubes + nanostructured Ni

4

6

10 12 14 Particle size (nm)

24

40

Fig. 4. Characterization of N-doped nanotubes after decoration: (a) TEM image of a representative coated tube with Ni clusters and (b) size distribution of Ni nanostructured particles attached to the walls of Ncontaining nanotubes.

In addition, as part of the grid preparation for TEM observations, a mild sonication process was carried out. We observed no evident particle detachment from the sonicated tubes in contrast to the grids prepared just by pressing the nanotube processed material onto the holey carbon film. Another important fact to consider is whether the oxidation state of Ni remains with +5 oxidizing state when the nickel nitrate decomposes, which will somehow lead to determine the nature of the atomic interaction. Although it has been suggested that wetting of the nanotube surfaces

with high surface tension materials such as metals and transition metals is not possible, we have now corroborated the theoretical work carried out by Curtin and Sheldon [7] and the experimental observations performed by Zang et al. [9], which illustrate that the interaction of Ni with singlewalled CNTs is possible and could have certain ‘covalent’ bonding characteristics that make the tube decoration possible. We believe this might be the reason why the Ni particles get easily attached to the nanotube surfaces, even on carbon tubes without doping that are expected to be chemically inert and tend to disallow good metallic wetting while coating the surface. Since we started from the production of the metal oxide–nanotube mixture (by dissociating Ni(NO3)2 dissolved in H2O with both type of nanotubes at 500 °C), it highly likely that numerous carboxyl groups appear on the tube surface, thus promoting the formation of an oxide layer on the tube interface; this being responsible for an efficient metal anchoring. We performed HREM studies (Figs. 3c and 5d), and noted that Ni tends to adhere strongly to the nanotube surface. In this context, we cannot rule out the possibility of having some nickel oxide left on the tube–metal interface (see high contrast fringes at the tube interface; Fig. 5d). This interface arises from the fact that metal reacts with the oxygen and carbon, thus forming a few layers of an oxide, which facilitates the adhesion of metallic clusters. This explains why we did not observe clear differences in the particle attachment yield on both types of tubes. However, the size of the clusters varied between the two types of tubes. For pure multiwalled CNT samples, the synthesis process gave origin to encapsulated iron found in the hollow

Fig. 5. Characterization of pure carbon MWCNTs after decoration: (a) Dark field SEM Micrograph showing long MWCNTs decorated with Ni particles. The long tubular structures remain after the sonication process, (b) micrographs obtained using an energy filter (GIF) showing that it is possible to obtain Ni particles decorating the outer walls of the carbon nanotubes, (c) size distribution of Ni nanostructured particles attached to the walls of N containing nanotubes and (d) high resolution TEM image showing the possible formation of a nickel oxide layer between the tube and the nickel particle.

P. Ayala et al. / Chemical Physics Letters 431 (2006) 104–109

core. In this case, this feature arises due to use of ferrocene as catalyst during the nanotube synthesis. We also noted that these types of undoped tubes get decorated with Ni particles of similar diameters to those observed in the hollow cores. We did not notice any preferential adhesion on sites where the iron was encapsulated from other regions of the tube where the core was completely hollow. However, the ability that nanotubes possess to host gases as well as solid compounds, induces different structural properties. Therefore, further experiments such as magnetic interaction between filling and coating should be carried out in order to determine the further properties in these metal– carbon systems. 4. Conclusions We have demonstrated that CNTs (pure and N-doped) could be decorated with Ni nanoparticles (4–15 nm in diameter differently for each case). The advantage of our method lies in the fact that it represents a relatively simple and low cost process that could yield bulk quantities of metal-decorated nanotubes. The reduction of the oxide particles occurs when passing H2 gas on the NiOx–nanotube material at 650 °C. In our systems, the Ni particles are able to attach firmly to the nanotube surfaces due to the presence of a thin oxide layer established between the tube and the metal. Our materials could also provide valuable information related to the interactions between nanotubes and metal clusters, and could motivate further theoretical research in order to elucidate the anchoring mechanisms. Similar techniques could be extended to efficiently coat other metal compound nanostructrures using the method reported here. The electronic, thermal and mechanical properties of compounds based on Ni coated CNTs should be different when compared to composites based on uncoated dispersed tubes. We believe that these coated tubular materials could be used on a bulk scale to fabricate Ni–CNT composites that could exhibit outstanding mechanical, thermal and electrical properties. Moreover, we believe that modified methods could also yield even metal coatings on nanotubes (templates), thus generating 1D nanowires of different metals with interesting properties, taking advantage of the potential for industrial production owing its low costs.

109

Acknowledgements The authors acknowledge the financial support of the Brazilian Agencies: CNPq through CIAM and CTEnergia programs and FAPERJ, as well as the National Science Foundation (US) Grant DMR-0303429. We gratefully acknowledge the use of facilities within the John Cowley Center for High Resolution Electron Microscopy at Arizona State University (USA). This work was also sponsored by CONACYT-Mexico grants: 45762 (HT), 45772 (MT), 41464-Inter American Collaboration (MT), 42428Inter American Collaboration (HT), 2004-01-013/SALUD-CONACYT(MT), PUE-2004-CO2-9 Fondo Mixto de Puebla (MT) and Ph.D. Scholarship (JARM). P.A. thanks T.Pichler from IFW-Dresden for fruitful discussions. References [1] M. Terrones, Int. Mater. Rev. 49 (2005) 325. [2] M.S. Dreselhaus, G. Dresselhaus, Ph. Avouris (Eds.), Carbon Nanotubes: Synthesis, Structure, Properties and Applications, Springer, Berlin, 2001. [3] E. Dujardin, T.W. Ebbesen, A. Krishnan, P.N. Yianilos, M.M.J. Treacy, Phys. Rev. B 58 (1998) 14013. [4] C. Li, T. Chou, Comp. Sci. Tech. 63 (2003) 1517. [5] J.P. Salvetat et al., Phys. Rev. Lett. 82 (1999) 944. [6] E.T. Thonstenson, Z. Ren, Tsu-Wei Chou, Comp. Sci. Tech. 61 (2001) 1899. [7] W.A. Curtin, B.W. Sheldon, Mater. Today 7 (2004) 44. [8] E. Dujardin, T.W. Ebbesen, H. Hiura, K. Tanigaki, Science 265 (1994) 1850. [9] Y. Zhang, N. Franklin, R. Chen, H. Dai, Chem. Phys. Lett. 331 (2000) 35. [10] Y. Zhang, H. Dai, Appl. Phys. Lett. 77 (2000) 3015. [11] B.C. Satishkumar, E.M. Vogl, A. Govindaraj, C.N.R. Rao, J. Phys. D: Appl. Phys. 29 (1996) 3173. [12] K. Jiang et al., Nano Lett. 3 (2003) 275. [13] X. Chen, J. Xia, J. Peng, W. Li, S. Xie, Comp. Sci. Tech. 60 (2000) 301. [14] M. Ba¨umer, J. Libuda, H.-J. Freund, Surf. Sci. 327 (1995) 321. [15] M. Menon, A. Andriotis, G. Froudakis, Chem. Phys. Lett. 320 (2000) 425. [16] E.A. Brocchi, M.S. Motta, I.G. Solo´rzano, P.K. Jena, F.J. Moura, Mater. Sci. Eng. B 112 (2004) 200. [17] R. Kamalakaran et al., Appl. Phys. Lett. 77 (2000) 3385. [18] R. Kurt, A. Karimi, Chemphyschem 2 (6) (2001) 388. [19] M. Terrones, R. Kamalakaran, T. Seeger, M. Ru¨hle, Chem. Commun. 23 (2000) 2335.