Growth of carbon nanotubes over nonmetallic based catalysts

Growth of carbon nanotubes over nonmetallic based catalysts. Phát triển các ống nano carbon trên chất xúc tác phi kim loại. Growth of carbon nanotubes over nonmetallic based catalysts. Phát triển các ống nano carbon trên chất xúc tác phi kim loại.Growth of carbon nanotubes over nonmetallic based catalysts. Phát triển các ống nano carbon trên chất xúc tác phi kim loại.

G Model CATTOD-8288; No. of Pages 12 ARTICLE IN PRESS Catalysis Today xxx (2012) xxx–xxx Contents lists available at SciVerse ScienceDirect Catalysis Today journal homepage: www.elsevier.com/locate/cattod Review Growth of carbon nanotubes over non-metallic based catalysts: A review on the recent developments Lling-Lling Tan a , Wee-Jun Ong a , Siang-Piao Chai a,∗ , Abdul Rahman Mohamed b a b Chemical Engineering Discipline, School of Engineering, Monash University, Jalan Lagoon Selatan, Bandar Sunway, 46150 Selangor Darul Ehsan, Malaysia School of Chemical Engineering, Universiti Sains Malaysia, Engineering Campus, Seri Ampangan, 14300 Nibong Tebal, Pulau Pinang, Malaysia a r t i c l e i n f o Article history: Received 16 July 2012 Received in revised form 1 October 2012 Accepted 17 October 2012 Available online xxx Keywords: Metal-free-catalyst Carbon nanotubes Nanoparticles Chemical vapor deposition Methane decomposition Catalyst preparation a b s t r a c t The conventional method to synthesize carbon nanotubes (CNTs) requires the use of metallic catalysts. However, the metallic particle impurities usually contaminate the CNTs, and are difficult to remove without introducing defects and contaminations. An alternative and more desirable approach is to grow CNTs directly on a substrate without further treatment being required. The review presents and discusses the current development of the controlled synthesis of SWCNTs and/or MWCNTs using non-metallic catalysts, which comprises of various ceramics and semiconductors as catalytic particles. All-carbon systems for CNT growth without employing any catalysts are also discussed. Despite the enormous strides in the synthesis of CNTs, the precise atomistic mechanism explaining their nucleation and growth still remains unclear. Therefore, the different growth systems proposed by several authors are also examined in this article. The review concludes with a summary and an outlook on the challenges and future directions in the metal-free-catalyst growth of CNTs. © 2012 Elsevier B.V. All rights reserved. Contents 1. 2. 3. 4. 5. 6. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General background on the metal-free catalyst growth of CNTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Non-metallic catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Semiconductor–catalyst systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Nanoparticulate oxide catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Other catalyst systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Controlling the length, diameter and chirality of CNTs over non-metallic catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Growth mechanism discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Introduction Tubular form of carbon products known as carbon filaments was first observed using electron microscopes around 1950 [1]. Since the observation and detailed structural study of carbon nanotubes (CNTs) by Iijima of the NEC Corporation in 1991 [2], CNTs have stimulated extensive research activities in most areas of science and engineering due to their extraordinary physical and chemical properties, including high mechanical strength, high electron ∗ Corresponding author. Tel.: +60 355146234. E-mail address: chai.siang.piao@monash.edu (S.-P. Chai). 00 00 00 00 00 00 00 00 00 00 00 conductivity and superior surface property. In 1991, Iijima initially observed only multi-walled carbon nanotubes (MWCNTs) grown in an arc discharge process, and it was not until 2 years later when single-walled carbon nanotubes (SWCNTs) with diameters between 1.1 and 1.3 nm were synthesized using laser ablation [3,4]. In 1996, Smalley’s group successfully produced bundles of SWCNTs for the first time [5]. Currently, the vast majority of research is being carried out on SWCNTs, as they are known to possess remarkable electronic and mechanical properties. They represent the ultimate carbon fiber, with the highest thermal conductivity and the highest tensile strength of any material [6,7]. Huge efforts have been spent by the international scientific community in order to study their 0920-5861/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cattod.2012.10.023 Please cite this article in press as: L.-L. Tan, et al., Growth of carbon nanotubes over non-metallic based catalysts: A review on the recent developments, Catal. Today (2012), http://dx.doi.org/10.1016/j.cattod.2012.10.023 G Model CATTOD-8288; No. of Pages 12 2 ARTICLE IN PRESS L.-L. Tan et al. / Catalysis Today xxx (2012) xxx–xxx application in fields including medicine [8–10], catalyst support for electrode of fuel cells [11,12], high-performance adsorbent [13–15], field emitter for field-emission display [16,17], electrode material for energy storage device [18,19] and so on. The three major techniques for the fabrication of CNTs are arc discharge [3], laser ablation [20] and chemical vapor deposition (CVD) [21,22]. Among these techniques, CVD shows obvious advantages for the growth of CNTs in terms of low growth temperature, good controllability and easiness of scaling up, at a relatively low cost. It is possible to control the growth and the structure of CNTs by adjusting reaction parameters such as growth temperature, carbon source, and catalyst concentration [23]. The growth process typically requires the assistance of metallic catalysts, predominantly 3d valence transition metal nanoparticles (Fe, Co and Ni) in the decomposition process of hydrocarbons due to high solubility and diffusion rate of carbon at high temperatures. Generally, for supported metal catalysts, there are two proposed mechanisms for the growth of CNTs, namely tip-growth and base-growth model, both of which are based on the vapor–liquid–solid (VLS) theory described by Baker and co-workers [24,25]. Initially, it was believed that the application of these carbide-forming metals is indispensable for the growth of CNTs. However, recent experimental [26,27] and theoretical [28] studies demonstrated that the chiralities of the SWCNTs produced could be controlled to some extent by the selection of catalysts. Since then, many other metallic species such as Au, Ag, Cu, Pd, and Rh have been reported to yield SWCNTs [29,30]. Despite these advancements, the metal species remaining in the CNT samples would result in major drawbacks for their intrinsic property characterization. Due to the chemical, redox, and magnetic properties of the metal nanoparticles, interference with the corresponding tube properties cannot be avoided [31,32]. The performance of CNT-based materials as catalyst support and its application in molecular electronics, biology, and medicine are also obscured by the presence of metal catalyst particles [33,34]. In semiconductor electronics fabrication (CMOS) processes, the adaptation of CNTs onto electronic substrates will result in detrimental incompatibility issues due to the presence of metal particle impurities. The toxicity and effects of metal nanoparticles on human health have hindered the use of CNTs in biology and drug delivery applications. Furthermore, in many cases, the catalyst particles are covered by a carbon shell, which presents additional problems for the non-destructive purification of the CNTs e.g. by treatment with non-oxidizing acids [35]. It has been until now an intractable problem in completely eliminating metal catalysts from CNT samples without introducing defects and contaminations. Over the past 5 years, many non-metallic species have been reported to have the ability to catalyze CNT growth under specific conditions. The list encompasses various ceramics (e.g. Al2 O3 and ZrO2 ) [36,37] and semiconductors (e.g. Si, SiC and Ge) [38–40]. The fact that CNTs can be grown over non-metallic catalysts is a major breakthrough in nanotechnology. The growth of CNTs over non-metallic catalysts will play very important roles in facilitating the applications of CNT materials in the field of nanoelectronics, photonics, biomedicine, membrane technology and so forth. In addition, it enables simpler purification techniques and mitigates toxicity concerns. Membrane- and CNT-nanofluidics-based research can also benefit from the non-metallic CNTs in bypassing one or more manufacturing steps, thus leading to lower costs. Comprehensive reviews on the production and growth mechanism of CNTs have been published [41–49]. The growth process and mechanism for metallic catalysts have been adequately discussed [42], whereas a clear mechanism is still being sought for the proper understanding of CNT growth over non-metallic catalysts. As oppose to metallic catalysts, the synthesis of CNTs over the nonmetallic catalysts still suffers from low yield, making it unattractive for mass production at the moment. Nevertheless, since a number of important findings have been reported on the growth of CNTs over non-metallic catalysts, we believe that a review on this subject is timely to promote latest developments in this interesting area of research to shed a different light on previous experiments. The review begins with an overview of the non-metallic catalyst growth of CNTs, followed by an in-depth analysis of CNT growth based on semiconductors, ceramics, all-carbon systems, and other new uncommon catalysts. The different growth mechanisms, particularly on the graphitization of silica, are also examined in this article. Finally, we will present a summary of challenges and future directions for investigation. 2. General background on the metal-free catalyst growth of CNTs Using metallic catalysts has long been considered indispensable for the nucleation and growth of CNTs. Only very recently, several groups have demonstrated the possibility to grow CNTs from ceramic [36,37], semiconducting [38–40,50–53], and nanosized diamond particles [54], all of which were considered inactive in the growth of CNTs in the past. More interestingly, dense CNTs were also demonstrated to grow on porous carbon substrates without employing any catalysts [55,56]. These findings clearly show that the decomposition of hydrocarbons and CNT production are not bound to the functions of metallic particles. Instead, Takagi et al. [38] proposed that the essential role of catalysts is to provide a template for cap formation. It should be noted that the growth of CNTs over these newly found species does not always happen. The template for CNT nucleation requires the presence of porous structures or nanoparticles with the appropriate diameter and the right curvature. It is said that the nano-scale curvatures provide a platform on which carbon atoms can form a hemispherical cap, where CNTs are grown in a self-assembled function [36,38,57]. There are four factors for the growth of CNTs: (1) the catalyst size [29], (2) the catalyst/substrate pretreatment [38], (3) the growth temperature [58] and (4) the role of water [59]. The solubility of carbon in catalyst particles and the precipitation rates of carbon from catalyst particles both show great dependence on the catalyst size. When the size of particles is below 10 nm, the quantum size effects greatly influence the properties of the catalyst particles [60,61]. The synthesis of SWCNTs generally requires catalyst particles with several nanometers in diameters, preferably 3 nm or less. With an increase in particle size, the number of walls and diameters of the nanotubes would increase. Interestingly, SWCNTs have been reported to grow on large Al2 O3 particles, ranging from several tens of nanometer to hundreds of nanometers [36]. Scanning electron microscopy (SEM) images revealed that the SWCNTs were grown on nano-sized protrusions, once again indicating the importance of nano-scale curvatures in the nucleation and subsequent growth of CNTs. The second factor for the growth of CNTs is the pretreatment of the catalyst/substrate. It was reported that preheating the particles in air at 950–1000 ◦ C not only increases the yield of SWCNTs, but also improves the activity of the catalyst particles and removes hydrocarbon contaminants on the particles [58,62]. On the other hand, it is noted that the annealing duration (from 1 to 60 min) in air had no significant effect on the growth of CNTs. This directly implied that pretreatment of SiO2 substrates in air was merely aimed to clean the substrate surface [58]. The growth temperature is also considered to be one of the essential factors which affects the growth of CNTs. Liu et al. [58] investigated the effect of growth temperature from 800 ◦ C to 900 ◦ C on the yield of SWCNTs grown on a SiO2 substrate. They reported that a high temperature of 850 ◦ C was important to induce the pyrolysis of ethanol. However, when the temperature increased above 900 ◦ C, a higher thermal decomposition rate Please cite this article in press as: L.-L. Tan, et al., Growth of carbon nanotubes over non-metallic based catalysts: A review on the recent developments, Catal. Today (2012), http://dx.doi.org/10.1016/j.cattod.2012.10.023 G Model CATTOD-8288; No. of Pages 12 ARTICLE IN PRESS L.-L. Tan et al. / Catalysis Today xxx (2012) xxx–xxx of carbon source would take place, leading to the coverage of substrates by a thick encapsulated carbon layer in a short residence time, resulting in no CNT growth. A similar trend of results was reported in the study by Xu et al. [59]. They studied the temperature dependence of the CNT growth ranging from 750 ◦ C to 950 ◦ C on silica nanoparticles synthesized from the thermal decomposition of PSS-(2-(trans-3,4-Cyclohexanediol)ethyl)-Heptaisobutyl substituted (POSS). Apparently, no CNTs could be produced when the growth temperature was too low (at 750 ◦ C) or too high (at 950 ◦ C). They claimed that when the temperature was increased from 850 ◦ C to 900 ◦ C, the CNT content decreased drastically from 47% to 20.3% due to much more aggregation of SiO2 particles. Very recently, Xu et al. [59] demonstrated the effect of introducing water in the synthesis of CNTs on silica nanoparticles. They found that water plays a cleansing role by removing amorphous carbon and also plays an indirect role in promoting CNT growth. This observation is similar to that reported in the previous studies on the growth of CNTs over metal catalyst particles [63,64]. They studied the influence of water content from 0 to 10 vol%. The results showed that in the absence of water, CNTs with some amorphous carbon were observed whereas much less amorphous carbon with 2% water. However, by increasing the water content from 2% to 5% and 10%, respectively, much less CNTs were obtained. Most recently, Liu et al. [65] reported that the chemical composition of the catalyst particles is another crucial factor for the growth of SWCNTs, in addition to the well-known catalyst size effect. In the paper, the authors studied the CVD growth of SWCNTs from SiOx and Si nanoparticles. Atomic force microscopy (AFM) and SEM images showed the growth of dense SWCNTs on SiOx nanoparticles while no CNTs were observed on Si nanoparticles. Density functional theory (DFT) calculations have also indicated that oxygen atoms can improve the capture of carbon-bearing molecules. This indicates that oxygen may play an important role in promoting the formation of graphitic carbon structures and facilitating the growth of SWCNTs on oxygen-containing SiOx nanoparticles [65]. 3. Non-metallic catalysts 3.1. Semiconductor–catalyst systems The non-metallic synthesis of CNTs was demonstrated for the first time in 1997, in which a film of well-oriented CNTs were produced by the sublimation decomposition of SiC nanoparticles at high temperatures (>1700 ◦ C) [66]. However, the success of transition metal-based catalysts and the large number of researchers working on CVD at that time left this work largely ignored. With the recent developments in novel catalysts, interest in the use of non-metallic catalysts for CNT growth is gradually rising. One of the most widely used catalysts in the synthesis of CNTs is SiC. Growth of CNTs using this catalyst can be achieved by annealing SiC particles [38,53,66], amorphous SiC films [52], or hexagonal SiC (6H-SiC) [67] in vacuum. Kusunoki et al. [68] demonstrated that under low vacuum conditions, SiC oxidizes according to the following process: SiC(s) + CO(g) → SiO(g) + 2C(s) (1) The oxidation of SiC results in the formation of initial nucleation caps, which then enable the subsequent construction of CNTs. However, the mechanism behind their formation still remains unclear. Several groups claimed that the formation of the nanocaps followed the transformation process of graphene layers [69,70] or amorphous carbon clusters [71], while others argued that their formation is a result of convex structures on the SiC surface [72–75]. Most recently, Wang et al. [51] suggested the formation of SiO nanoclusters at the C/SiC interface. The authors claimed that the possible roles of these nanoclusters may be twofold. First, the 3 curvature of the molten SiO nanoclusters provides a platform for graphene lifting, thus leading to the formation of hemispherical carbon nanocaps and subsequent CNT growth. Second, the coordination-unsaturated Si species in the nanoclusters present high activity for the incorporation of carbon atoms, and facilitate the attachment of carbon atoms to the tube edges. In addition to SiC, the use of Ge catalysts in the synthesis of CNTs has also been extensively researched [40,76]. The pioneering work by Uchino et al. [76] employed carbon-doped SiGe islands on Si substrates to synthesize CNTs in CH4 -CVD. The authors suggested that the growth of CNTs occurred from Ge clusters, which were produced following the chemical oxidation and annealing treatment of the SiGe layers. Since Si has the greater thermodynamic tendency to be oxidized as compared to Ge, the oxidation treatment results in the formation of SiO2 and the segregation of Ge clusters. These clusters then serve as catalysts for the growth of CNTs. Other growth techniques associated with the use of Ge catalysts include those which are based on Ge Stranksi–Krastanow dots, Ge nanoparticles formed by ion implantation [77] and colloidal Ge nanoparticles [78,79]. Takagi et al. [38] also demonstrated the production of CNTs over semiconductor nanoparticles Si, Ge and SiC in ethanol-CVD. Singlewalled or double-walled CNTs, with diameters less than 5 nm were produced. The CNT yield from Ge nanoparticles was found to be higher than those of Si and SiC. Since the melting point of Ge is known to be lower compared to Si and SiC, it is plausible that the Ge nanoparticles are in molten state during the growth process, thus contributing to the higher CNT yield. Despite the advancements made in the development of these non-metallic catalysts, the yield was still much lower compared to the 3d valence transition metals of Fe, Ni and Co. Since Si, Ge and SiC should have little activity in this process, a higher growth temperature would be necessary to induce the pyrolysis of the carbon source. Another factor that might contribute to the deviation in CNT yield is the phase of the nanoparticles (solid or liquid) because it relates to the precipitation mechanism of carbon atoms on its surface. Other semiconductor catalysts that have been reported in literature include ZnO [80], TiO2 [81] and Te [82]. These findings provide more insights into the actual role of catalysts and are helpful for understanding the mechanism behind the growth of CNTs. 3.2. Nanoparticulate oxide catalysts In catalyst-supported CVD, pure oxides such as SiO2 , Al2 O3 , TiO2 , ZrO2 and MgO are basically employed as physical supports for the catalysts. However, recent experimental studies revealed that oxides have the ability to grow graphitic sheets under typical CVD conditions for CNT growth [83,84]. Rümmeli et al. [83] demonstrated that difficult-to-reduce nanoparticle oxides are extremely effective in promoting ordered carbon (graphene) growth. In contrast, no carbon formation was reported on bulk/film samples. The authors attributed this difference to the presence of surface defect sites on the nanoparticle oxides, and proposed that the interface between the catalyst particle and the substrate behaves as a circular defect site where nanotube growth can take place. Huang et al. [81] reported that many oxide nanoparticles including SiO2 , Al2 O3 , Er2 O3 and all lanthanide oxides except promethium oxide are active for the growth of SWCNTs. Among the nanoparticle oxides reported in literature, the use of SiO2 as a catalyst for CNT and graphene production is of particular interest due to their potential applications in silicon-based technology [85]. To date, different approaches to SiO2 catalyst preparation have been reported. In 2009, Liu et al. [86] reported that a 30-nmthick SiO2 film deposited onto a Si/SiO2 wafer could directly serve as a substrate for SWCNT growth in a CH4 -CVD process at 900 ◦ C. A dense and large-area of uniform SWCNTs were reproducibly Please cite this article in press as: L.-L. Tan, et al., Growth of carbon nanotubes over non-metallic based catalysts: A review on the recent developments, Catal. Today (2012), http://dx.doi.org/10.1016/j.cattod.2012.10.023 G Model CATTOD-8288; No. of Pages 12 ARTICLE IN PRESS L.-L. Tan et al. / Catalysis Today xxx (2012) xxx–xxx 4 Fig. 1. SEM images of the SWCNTs grown on Si/SiO2 wafer: (a) overview of the cross pattern, (b and c) enlarged images of the blue square area in image (a). Reprinted with permission from [86]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.) Copyright (2009) American Chemical Society. obtained on the substrate surface, indicating the effectiveness of the synthesis approach. The same group of researchers also developed a facile “scratching” method for the patterned growth of SWCNTs, in which a Si/SiO2 wafer was scratched using another one with a sharp tip to obtain a desired pattern, before being subjected to CH4 -CVD. From SEM observations (Fig. 1), CNTs can be found on scratched areas but not on nanoscopically smooth surfaces. It is said that the “scratching” between the two wafers will crack the thermally grown SiO2 layer and consequently generate some active sites for the growth of SWCNTs. Along the same lines, Huang et al. [81] scratched a clean Si/SiO2 wafer with a diamond blade without putting any catalyst on the surface. SWCNTs with a narrow diameter distribution ranging from 0.8 to 1.4 nm were observed. This indicates that only SiO2 with an appropriate catalyst size of less than 2 nm is active for SWCNT growth. The presence of oxidized Si was confirmed by the XPS analysis at the rich region of SWCNTs. The “scratching” approach is both simple and cheap, without requiring the need for complex patterning process to grow SWCNTs at a predefined position for device fabrication. However, since the structure of these nanoparticles fluctuate, it is difficult to precisely control the structure of the as-grown CNTs. In addition to the “scratching” approach, a wet chemical etching process has also been reported to generate SiO2 nanoparticles [81,87]. Huang et al. [81] etched a Si/SiO2 wafer with an aqueous solution of HF, followed by thermally annealing it in air at 1000 ◦ C for 1 h. A circular trace was formed on the Si wafer after drying as shown in Fig. 2a. It was proposed that HF dissolves the SiO2 layer which is hydrophilic and the aqueous solution shrinks onto the bottom Si layer which is hydrophobic. Parts of the SiO2 are said to be dissolved, leaving small SiO2 particles in the water drop on the Si wafer. These particles then serve as nucleation sites for the growth of CNTs in the subsequent CVD process. Based on SEM observations, highly dense random SWCNTs can be observed around the circular trace (Fig. 2b and c) while carbon filament or MWCNTs are observed inside the circle (Fig. 2d). There are other reports on the use of SiO2 catalysts for CNT growth. Most recently, Xu et al. [59] demonstrated the first example of the use of SiO2 nanoparticles for the continuous synthesis of MWCNTs by alcohol-CVD. In another paper, Liu et al. [58] successfully synthesized dense SWCNTs by simply annealing the SiO2 substrates in H2 at high temperatures (950–1000 ◦ C) before CVD. The authors proposed that the annealing treatment at high temperature locally evaporates SiO2 substrate surfaces resulting in the formation of defects on the surfaces. These defects provide nucleation sites for the production of carbon nanoparticles, and assist the formation of carbon nanocaps, thus resulting in the growth of SWCNTs. The growth mechanism of CNTs over SiO2 nanoparticles has yet to be clarified. Some argue that SiO2 undergoes carbothermal reduction to SiC, while others claim that it remains stable throughout the growth process. An in depth analysis of the graphitization mechanism of SiO2 catalysts is presented in Section 5. In the past, Al2 O3 ceramic nanoparticles were regarded as an inactive catalyst in the growth of CNTs and were only used as a buffer layer to disperse metallic catalyst particles and enhance their catalytic properties in CNT growth [88]. However, in a recent study by Liu et al. [36], dense SWCNT layers catalyzed by Al2 O3 nanoparticles were successfully grown using an alcohol-CVD. The morphologies of Al2 O3 particles were found to be different compared to metallic catalyst particles such as Fe, Co and Ni, which are typically spherical or semi-spherical with a smooth surface. The authors indicate that the Al2 O3 particles are likely to be in solid state during the CNT growth process. This signifies that a growth model other than the traditional VLS mechanism must be involved for these solid catalysts. The research also indicates the possibility of growing large area “catalyst-free” SWCNTs on flat Al2 O3 substrates by simply manipulating the nanostructures on their surfaces. Steiner et al. [37] successfully synthesized both SWCNTs and MWCNTs on ZrO2 nanoparticles by thermal CVD. ZrO2 offers several advantages as a catalyst for CNT growth due to its nonmagnetic nature, making it an interesting possible catalyst for electronics application. The authors demonstrated that solid ZrO2 nanoparticles are active catalysts, and neither reduce to Zr nor carbonize into ZrC during the growth process. This observation is substantiated by the understanding of ZrO2 chemistry at high temperatures in literature. ZrO2 is known to not be reduced by H2 , even at temperatures of 1500 ◦ C and higher [89]. Since this temperature is much higher than the growth temperature used in the study, reduction of ZrO2 to Zr or ZrC is highly unlikely. Both Al2 O3 and ZrO2 possess high melting points (>2000 ◦ C) in their bulk form. Hence, it is unlikely that they would be in molten state during the entire growth process, even after factoring in particle size effects. Moreover, given the low carbon diffusivity in bulk ZrO2 and Al2 O3 , the successful growth of CNTs is suggestive to occur via a surface-borne mechanism which will be discussed in detail in Section 5. 3.3. Other catalyst systems In 2009, Takagi et al. [54] demonstrated that nanosized diamond particles (4–5 nm) acted as CNT growth nuclei effectively in CVD. Interestingly, the diamond nanoparticles did not fuse with each other, even when agglomerated particles were used in the CVD process. Particle density was enhanced due to the non-fusion characteristic of nanodiamond particles. The findings suggest that the growth of CNTs occurred from solid carbon nanoparticles. Furthermore, since bulk diffusion of carbon into nanodiamond particles is unlikely, the surface of the diamond must play an important role in the synthesis of CNTs, in which it provides a template for the formation of CNT caps. In recent years, a broad array of growth routes using pure carbon systems without any addition of catalyst particles has been Please cite this article in press as: L.-L. Tan, et al., Growth of carbon nanotubes over non-metallic based catalysts: A review on the recent developments, Catal. Today (2012), http://dx.doi.org/10.1016/j.cattod.2012.10.023 G Model CATTOD-8288; No. of Pages 12 ARTICLE IN PRESS L.-L. Tan et al. / Catalysis Today xxx (2012) xxx–xxx 5 Fig. 2. SEM images of CNTs grown on HF-treated Si/SiO2 wafer: (a) overview of the circular trace, (b) indicated area B in image (a), (c) high magnification of image (b) and (d) inner area of the circle. Reprinted with permission from [81]. Copyright (2009) American Chemical Society. developed. The growth of CNTs on graphitic surfaces has been successfully demonstrated in various works by Lin et al. [55,56]. The authors employed carbon black, flake graphite powder and highly oriented pyrolytic graphite as substrates to grow CNTs by CVD. Previous studies indicate that dangling structures of carbons have the catalytic ability to decompose hydrocarbons [90,91]. Hence, the defective sites present on the substrates will result in the decomposition of the carbon feedstock and lead to the subsequent formation of amorphous nanobumps, as shown in Fig. 3. Once these nanobumps have formed, whether from a lift-off process or selfassembly mechanism, the growth of CNTs can occur. However, two major disadvantages still need to be overcome before catalyst-free CNT growth can be employed in macroscopic applications: the significantly low synthesis yield and the high growth temperature required. Two of these factors are unattractive for scaling up to mass production. Recent developments in nanomaterial synthesis and characterization have brought up many new catalysts for the production of CNTs. Apart from the non-metallic catalysts mentioned above, most recently, calcium silicate, CaSiO3 was shown to have the ability to catalyze CNT growth on a pyrolytic graphite paper tape [92]. The authors proposed that the role of silicate is similar to that of a transition metal and is based on the solubility of the carbon in the silicate. The electron energy loss spectra (EELS) also revealed that the carbon surrounding the catalyst is more amorphous than that in the tube, which supports the solid state transformation of carbon [92]. Recent reports on the growth of CNTs over non-metallic catalysts are summarized in Table 1. 4. Controlling the length, diameter and chirality of CNTs over non-metallic catalysts Almost all currently available technologies for CNT fabrication can only produce mixtures of CNTs with a range of (n, m) indices [48]. This poses a huge limitation for the application of CNTs. Hence, the growth of CNTs with well-controlled structures is highly desirable for both fundamental research and practical applications [44]. Liu et al. [95] reported the direct length-sorted growth of SWCNTs using SiO2 nanoparticles. The authors found that SiO2 catalyzed grown SWCNTs have an extremely low growth velocity of 8.3 nm/s, which is about 300 times slower than that of the commonly used Co catalyst for SWCNT growth at the same reaction condition. The slow growth velocity allows direct length-sorted growth of SWCNTs with average lengths of 149, 342 and 483 nm by simply adjusting the growth durations correspondingly [95]. Owing to their finite length effect, the short SWCNTs are believed to display intriguing physics and are attractive for various practical applications including scanning probes [96–98], catalyst supports [99], biological imaging [100,101], molecular sensing [100], electronic devices [100,102] and so on. Furthermore, comparative studies revealed that SiO2 catalyst exhibits considerably longer catalytic active time compared to the commonly used Co catalyst, which loses its catalytic activity after a short period of time. Recently, the growth of SWCNTs with controlled diameters using SiO2 nanoparticles were demonstrated by Chen and Zhang [103]. Various sizes of SiO2 nanoparticles were generated by the thermal oxidation of 3-aminopropyltriethoxysilane (APTES) layers with different thicknesses. It was shown that the size of SiO2 nanoparticles increased with the number of assembled APTES layers. These nanoparticles served as nucleation centers, where SWCNTs with diameters ranging from 0.90 to 1.82 nm were grown in ethanol-CVD. The findings clearly indicated a direct correlation between SWCNT diameter and SiO2 nanoparticle size. Fig. 4 presents the schematic of the preparation procedures from APTES layers to SiO2 nanoparticles with controlled sizes, followed by the growth of SWCNTs in CVD. Chirality-selective synthesis of SWCNTs is essential for their application in nanoelectronic devices [104] because electronic structures are defined by the chiral index (n, m). It is well known that the structures of SWCNTs are determined by the initial carbon structure as the growth commences. The structure of this “cap” is determined during the nucleation stage. Hence, it is possible to control the chirality of SWCNTs by controlling the process of cap formation. Yu et al. [105] demonstrated a rational approach to Fig. 3. SEM image of (a) carbon nanobumps initially formed on the surface of the acid-treated flake graphite, TEM images of (b) a typical nanobumps and (c) an as-grown MWCNT with the close-cap feature. Reprinted with permission from [56]. Copyright (2011) American Chemical Society. Please cite this article in press as: L.-L. Tan, et al., Growth of carbon nanotubes over non-metallic based catalysts: A review on the recent developments, Catal. Today (2012), http://dx.doi.org/10.1016/j.cattod.2012.10.023 CNT type, remarks Carbon yield Ref. Carbon source: EtOH Carrier gas: Ar/H2 Temperature: 850 ◦ C Reaction time: 10–30 min SWCNTs and DWCNTs, diameters [...]... room temperature CNT growth which is still considered a dream In addition, most of the current synthesis routes using metal-free catalysts focus only on the surface growth with a limited amount of SWCNTs obtained Also, the research is mostly carried out in small and low throughput reactors on small pieces of wafer, thus resulting in a minimum amount of SWCNTs produced The development of synthesis methods... of CNTs The simplistic view that the support only plays a catalytically passive role in the growth of CNTs is in need of further examination The two major drawbacks of the non-metallic catalyst growth of CNTs include the dramatically low synthesis yield and the high growth temperature required It is of utmost technological importance to synthesize CNTs at low temperatures without introducing any metal... Journal of Applied Physics 44 (2005) L803–L805 [74] T Maruyama, T Shiraiwa, N Fujita, Y Kawamura, S Naritsuka, M Kusunoki, Japanese Journal of Applied Physics 45 (2006) 7231–7233 [75] T Maruyama, H Bang, Y Kawamura, N Fujita, K Tanioku, T Shiraiwa, Y Hozumi, S Naritsuka, M Kusunoki, Chemical Physics Letters 423 (2006) 317–320 Please cite this article in press as: L.-L Tan, et al., Growth of carbon nanotubes. .. [121] A.J Page, K.R.S Chandrakumar, S Irle, K Morokuma, Journal of the American Chemical Society 133 (2010) 621–628 Please cite this article in press as: L.-L Tan, et al., Growth of carbon nanotubes over non-metallic based catalysts: A review on the recent developments, Catal Today (2012), http://dx.doi.org/10.1016/j.cattod.2012.10.023 ... SWCNTs produced The development of synthesis methods for the large-scale mass production of SWCNTs over non-metallic catalysts is highly desirable for applications in the field of nanoelectronics, photonics and biomedicine where metal particle impurities are harmful Acknowledgement This work was funded by the Ministry of Higher Education (MOHE) Malaysia under the Fundamental Research Grant Scheme (FRGS)... (2012) 4757–4764 [108] S.A Moshkalev, C Verissimo, Journal of Applied Physics 102 (2007) 044303–44306 [109] J.C Charlier, S Iijima, Journal of Applied Physics 80 (2001) 55–81 [110] P.T.A Reilly, W.B Whitten, Carbon 44 (2006) 1653–1660 [111] H Okamoto, T Massalski, Journal of Phase Equilibrium 5 (1984) 378–379 [112] M Hansen, K Anderko, Constitution of Binary Alloys, McGraw-Hill, New York, 1958 [113] L Oden,... New Carbon Materials 24 (2009) 307–313 R.T.K Baker, R.J Waite, Journal of Catalysis 37 (1975) 101–105 R.T.K Baker, M.A Barber, P.S Harris, F.S Feates, R.J Waite, Journal of Catalysis 26 (1972) 51–62 S.M Bachilo, L Balzano, J.E Herrera, F Pompeo, D.E Resasco, R.B Weisman, Journal of the American Chemical Society 125 (2003) 11186–11187 X Li, X Tu, S Zaric, K Welsher, W.S Seo, W Zhao, H Dai, Journal of. .. Chemical Physics Letters 423 (2006) 317–320 Please cite this article in press as: L.-L Tan, et al., Growth of carbon nanotubes over non-metallic based catalysts: A review on the recent developments, Catal Today (2012), http://dx.doi.org/10.1016/j.cattod.2012.10.023 G Model CATTOD-8288; No of Pages 12 12 ARTICLE IN PRESS L.-L Tan et al / Catalysis Today xxx (2012) xxx–xxx [76] T Uchino, K.N Bourdakos, C.H de... Ashburn, Journal of the Electrochemical Society 156 (2009) K144–K148 [41] Q Zhang, J.-Q Huang, M.-Q Zhao, W.-Z Qian, F Wei, ChemSusChem 4 (2011) 864–889 [42] J.-P Tessonnier, D.S Su, ChemSusChem 4 (2011) 824–847 [43] C.-M Seah, S.-P Chai, A.R Mohamed, Carbon 49 (2011) 4613–4635 [44] G Hong, Y Chen, P Li, J Zhang, Carbon 50 (2012) 2067–2082 [45] A Moisala, A.G Nasibulin, E.I Kauppinen, Journal of Physics:... Ajayan, Journal of Physical Chemistry B 107 (2003) 12161–12164 [115] S Hofmann, R Blume, C.T Wirth, M Cantoro, R Sharma, C Ducati, M ¨ S Zafeiratos, P Schnoerch, A Oestereich, D Teschner, M Albrecht, Havecker, ¨ J Robertson, Journal of Physical Chemistry C 113 A Knop-Gericke, R Schlogl, (2009) 1648–1656 [116] M Zhu, J Wang, B.C Holloway, R.A Outlaw, X Zhao, K Hou, V Shutthanandan, D.M Manos, Carbon 45 (2007) ... of carbon at high temperatures Generally, for supported metal catalysts, there are two proposed mechanisms for the growth of CNTs, namely tip -growth and base -growth model, both of which are based. .. sought for the proper understanding of CNT growth over non-metallic catalysts As oppose to metallic catalysts, the synthesis of CNTs over the nonmetallic catalysts still suffers from low yield,... developments in novel catalysts, interest in the use of non-metallic catalysts for CNT growth is gradually rising One of the most widely used catalysts in the synthesis of CNTs is SiC Growth of CNTs using