A molecular dynamics study of the graphitization ability of transition metals for catalysis of carbon nanotube growth via chemical vapor deposition
Graphical abstract
Graphitization ability of transition metals used as catalysts for nanotube growth via CVD was investigated on crystalline surfaces and nanoclusters.
Introduction
The growth mechanisms [1] of single-wall carbon nanotubes (SWCNTs) have been widely studied since they were first brought to the attention of the scientific community at large [2]. With the development of catalytic chemical vapor deposition (CCVD) techniques [3], [4], [5], [6] as a promising synthesis method for both single and multi-wall tubes, growth models to describe metal-catalyzed growth have been widely discussed [1], [7]. In 2003, Shibuta and Maruyama [8] first reported classical molecular dynamics (MD) simulations showing a nanotube-cap growth process occurring on a metal cluster, which have also been observed since by a number of other groups [9], [10], [11], [12], [13], [14], [15]. Combined with the support of in situ observations [16], these numerical simulations have given rise to a broad consensus that the initial cap structure of the carbon nanotube is nucleated by a transition metal catalyst nanocluster, followed by subsequent longitudinal growth either from tip or root of the tube. Furthermore, the effect of the substrate on CNT growth has also been investigated by classical MD simulation [17], [18]. Recently, not only the initial cap formation process, but also the subsequent longitudinal growth process has been captured by the present authors using a multi-scale modelling approach [19], [20].
One of the most significant unresolved issues related to the nanotube growth is to understand the factors determining the chirality of grown CNTs. Several previous reports have shown that epitaxy between the graphene sheet and crystallographic facets of catalytic metal clusters may be important in determining the chirality of the CNTs [21], [22]. Moreover, the surface crystallographic orientation of the catalytic metal clusters is sensitive to their bulk structural configuration [23], which is in turn related to the phase of the cluster. Since the phase of the cluster is difficult to define clearly, the effect of size on the phase of the nanoparticle has been widely studied [23], [24], [25], [26], [27], [28]. For example, Harutyunyan et al. [24] examined the phases of cobalt nanoparticles during CNT growth by calorimetric measurements and concluded that the liquid phase is most favorable for the growth of SWCNTs. Depression of the melting point of nanoparticles due to the Gibbs–Thomson effect was confirmed by MD simulation of phase transition process of metal nanoparticles [25], [26]. Moreover, discussion of the Fe–C phase diagram as a function of the nanoparticle size on the basis of MD simulation [27] and ab initio calculations [28] have demonstrated that formation of cementite arrests the formation process of nanotubes due to its different activity and diffusion properties.
Although the stability of the graphene in the metal cluster has been discussed at length in the above studies, the catalytic ability of the transition metals is still under debate. From previous experimental work, it is known that the best catalysts for SWCNT growth have a higher graphitization ability, low solubility for carbon and stable crystallographic orientation on graphite [29]. The graphitization ability, which is defined as the ability of a catalytic surface to stabilize the formation of a defective graphene sheet so that it can anneal to make a perfect sheet, has not yet been the explicit focus of any theoretical studies to date. Therefore, the graphitization ability of several transition metals commonly used as catalysts for carbon nanotube growth has been investigated here via classical molecular dynamics simulation. After briefly summarizing the simulation methodology in Section 2, the remainder of the paper is structured as follows: first, the stability of the graphene sheet on the surface of the bulk metal is examined in Section 3.1; then, the stability and the interaction energy between graphene sheet and metal cluster are discussed in Section 3.2.
Section snippets
Simulation methodology
A classical MD method was used to study the graphitization ability of transition metals for catalysis of carbon nanotube growth. A Brenner potential [30] in its simplified form [31] was used to describe covalent bonding between carbon atoms. A bond-order type potential [22], which was constructed for transition metal carbide cluster, was used to describe metal–metal and metal–carbon interactions in which the bonding energy between atom i and j is expressed as:
Stability of the graphene on the surface of the bulk metal
Before considering nanoparticles, which generally exhibit a variety of crystalline facets, the stability of graphene on various surfaces of the bulk crystalline metals was first examined. A schematic image of calculation process is shown in Fig. 1a. Three layers of the Ni(1 1 1)fcc, Co(0 0 0 1)hcp and Fe(0 0 1)bcc surfaces were prepared at the base of the cubic cell, with the close-packed surfaces comprising of 672 atoms, and the cubic-packed surfaces comprising of 484 atoms, respectively. A
Conclusions
In summary, the graphitization ability of several transition metals commonly used as catalysts for carbon nanotube growth (nickel, cobalt and iron) has been investigated via classical molecular dynamics simulation. First, the stability of the graphene on the surfaces of the bulk crystalline metals was examined. After some initial bond breakage in the graphene due to kinetic energy of impact with the metal surface, the hexagonal carbon network was rapidly reconstructed on the Ni(1 1 1)fcc and Co(0 0
Acknowledgements
J.A.E. gratefully acknowledges Japan Society for Promotion of Science (JSPS) for funding of a Long Term Invitation Fellowship (No. L08536), held at the University of Tokyo, and both authors would like to thank Fitzwilliam College, Cambridge for supporting a visiting Fellowship for Y.S. Part of this research was supported by the Grant-in-Aid for Young Scientists (a) (No. 18686017) from MEXT, Japan.
References (34)
- et al.
Chem. Phys. Lett.
(1996) - et al.
Chem. Phys. Lett.
(1999) - et al.
J. Catal.
(2001) - et al.
Chem. Phys. Lett.
(2002) Prog. Mater. Sci.
(2005)- et al.
Chem. Phys. Lett.
(2003) - et al.
Chem. Phys. Lett.
(2007) - et al.
Appl. Surf. Sci.
(2006) - et al.
Chem. Phys. Lett.
(2006) - et al.
Comp. Mater. Sci.
(2007)
Chem. Phys. Lett.
Chem. Phys. Lett.
Lect. Notes Phys.
Nature
J. Chem. Phys.
J. Nanosci. Nanotechnol.
Phys. Rev. Lett.
Cited by (30)
The microstructure evolution of graphene in nanoindentation G/WC-Co based on molecular dynamics simulation
2024, Diamond and Related MaterialsSingle catalyst particle growth modeling in thermocatalytic decomposition of methane
2021, Chemical Engineering JournalCitation Excerpt :The potential applications of carbon nanomaterials in semiconductors, additives to building materials, energy storage, and catalytic materials due to their unique physicochemical properties such as high conductivity, high tenacity and mechanical strength, high specific surface area and semiconductor properties make TCD more economically and environmentally appealing [1–6]. In the literature, including the work of Ashik some studies have been reported on modeling at the molecular scale [2,15,16], which helps scientists in catalyst evaluation and to develop a microscopic level of understanding of the complete chemical transformation. Although these models are helpful in the understanding of reaction mechanisms, a model that properly describes the behavior of a catalyst at the macro level has not been reported to the best of our knowledge.
Chiral vector and metal catalyst-dependent growth kinetics of single-wall carbon nanotubes
2018, CarbonCitation Excerpt :This is readily recognized in Fig. 7 where the Arrhenius lines of α and β for nickel (open squares and circles) are upshifted from those for cobalt (red squares and green circles). The CVD growth rates of carbon nanotubes or fibers were studied previously in relation with the type of metal catalyst [31,53,73–82]. The results were contradictory to one another due to different synthesis conditions.
Ab initio molecular dynamics simulation of dissociation of methane on nickel(1 1 1) surface: Unravelling initial stage of graphene growth via a CVD technique
2013, Chemical Physics LettersCitation Excerpt :The role of catalytic metals during the growth of low-dimensional carbon materials such as carbon nanotubes (CNTs) [1] and graphene [2] has been widely investigated since catalytic metals are essential to synthesize single-walled carbon nanotubes and monolayer graphene via a chemical vapor deposition (CVD) technique [3–5]. Especially, the interaction between the metal surface and graphene is widely studied focusing on commensurate orientational relationships (called an epitaxy) between graphene and the metal surface with close-packed facet such as fcc(1 1 1) and hcp(0 0 0 1) [3–9], although such epitaxy has been suggested for a long time in the field of surface science prior to the discovery of CNTs and graphene [10,11]. Recently, epitaxial growth of the graphene on the highly orientated metal surface via the CVD process becomes a reality [12–19] and the mechanism of epitaxial growth has been examined.
Growth of carbon clusters on a Ni(1 1 1) surface
2012, Computational Materials ScienceCitation Excerpt :With this aim in view, in addition to the C–Ni potential designed in this work, more realism is needed in the modeling of homonuclear Ni–Ni interactions in a carbon environment, as discussed in Section 2. Finally, we stress once more that the situation for the growth at high temperature can be substantially different: there, linear chains do cross-link as shown in previous work [25–28,50,51]. The purpose of this work is to predict the first steps of carbon clustering in the formation of graphene on a perfectly flat Ni(1 1 1) surface, simulating the experimental conditions of recent work [11].