Effects of ferrite catalyst concentration and water vapor on growth of vertically aligned carbon nanotube

Iron acetylacetonate, oleic acid, oleylamine, 1,2-hexadecanediol, 1-octadecene (Sigma Aldrich) were of analytical grade and used without further purification. Silicon wafer with 100 nm thick SiO₂ was used as substrate for VA-CNTs growth.

In this work Fe₃O₄ nanoparticle catalyst was prepared by thermal decomposition of iron acetylacetonates (Fe(acac)₃). Briefly, Fe(acac)₃ (0.69 g, 0.63 mM), oleic acid (OA) (3.6 mL, 372 mM), oleylamine (OLA) (3.6 ml, 372 mM), 1,2- hexadecanediol (0.58 g, 75 mM) and 30 mL of 1-octadecene were added to 100 mL three-neck round-bottom flask. The mixture was magnetically stirred under nitrogen flow at room temperature for 30 min. Temperature was then adjusted to 100 °C and held for 30 min to remove the remaining water. In the next stage of the reaction, the mixture was maintained at 200 °C for 30 min and heated to 295 °C at a rate of 5–7 °C min⁻¹. After 60 min, the solution was cooled to room temperature. The product was washed with ethanol, centrifuged, and finally suspended in n-hexane at 0.01, 0.026, 0.033 g ml⁻¹ before use.

A piece of Si/SiO₂ (5 mm × 5 mm) was cut from a 4 inch wafer and sonically cleaned in acetone (20 min), ethanol (20 min) and water (20 min). The substrate was subsequently immersed in a mixture of H₂SO₄ and H₂O₂ (molar ratio: 3/1) for 15 min, rinsed by DI water and then dried under nitrogen flow. A volume of 0.002 mL (two drops) catalyst suspension was dropped on Si/SiO₂ surface. Preliminary experiments were conducted to compare drop coating and spin coating methods. It was found that catalyst particles were most uniformly distributed with spin coating at 2000–3000 rpm for 60 s.

Synthesis of the VA-CNTs was carried out at atmospheric pressure via catalytic decomposition acetylene (C₂H₂). The samples were placed in the furnace and calcined at 600 °C in the air for 30 min to remove the residual polymers. The furnace was heated to 750 °C at ramping rate of 20 °C min⁻¹ under an Ar flow of 800 sccm. The samples were incubated at 750 °C under flow of Ar/H₂(300/100 sccm) for 30 min to deoxidize the catalysts. Then C₂H₂ (30 sccm) was added for 30 min. After that, the furnace was cooled down to room temperature in Ar gas to prevent oxidation of the CNTs. In this work the effects of Fe₃O₄ catalyst concentrations (0.01, 0.026 and 0.033 g ml⁻¹) on the properties of VA-CNT were studied. The optimal concentration of catalyst was then used to evaluate the effect of water vapor on the growth of VA-CNTs. Water vapor was introduced during the CVD process by bubbling partly argon gas supply (60 sccm) through water prior to entering the furnace.

The morphology of Fe₃O₄ nanoparticle catalysts was observed by a transmission electron microscope (TEM) (JEM 1010 JEOL, Japan) at 80 kV and an atomic force microscope (AFM) (N9451A, USA). Sizes of Fe₃O₄ nanoparticle catalysts were analyzed by ImageJ sofware. The morphology of VA-CNTs was characterized by field emission-scanning electron microscopy (FE-SEM) (Hitachi S-4800) operating at an acceleration voltage of 10 kV. Diameter and length of the CNTs were estimated by analyzing SEM images using ImageJ software. Growth rate of the CNTs was calculated by the following equation:

$$\begin{eqnarray}&&v=\frac{l}{t},\end{eqnarray} \tag{ 1 }$$

where ν is growth rate, l is the length of the tubes and t is CVD time (30 min).

Raman spectra were recorded in the range of 500–2000 cm⁻¹ on a JobinYvon Lab RAM-1B (France), using He-Ne laser source of 632.8 nm as the incident light.

Figures 1(a) and (b) illustrate morphology and size distribution of Fe₃O₄ nanoparticles, respectively. Under given synthesis conditions, Fe₃O₄ nanoparticles are relatively well dispersed. The average diameter of the particles was estimated to be about 8 ± 1 nm. The distribution of as-synthesized Fe₃O₄ nanocatalyst on Si/SiO₂ substrate after spin-coating was examined by AFM and the smooth coating of catalyst particles was clearly observed (figure 1(c)).

Zoom In Zoom Out Reset image size

Figure 2 shows the SEM images of the VA-CNTs synthesized by CVD method using different concentrations (0.01, 0.026 and 0.033 g ml⁻¹) of Fe₃O₄ catalyst. It is clear that Fe₃O₄ concentrations strongly affect density, length and growth rates of CNTs (figure 2 and table 1). For the given catalyst concentrations, diameters of CNTs are virtually similar and range from 8 to 10 nm (figures 2(d), (e), and (f)).

Zoom In Zoom Out Reset image size

When Fe₃O₄ concentration is 0.026 g ml⁻¹, obtained VA-CNTs have the most uniform length and highest density (figure 2(b)). In addition, the length of CNTs produced at this catalyst concentration is highest among three concentrations studied. At 0.033 g ml⁻¹ Fe₃O₄, CNTs are shorter than those synthesized with 0.026 g ml⁻¹ Fe₃O₄ and amorphous carbon can be observed on the surface of the CNTs (figures 2(c) and (f)). This is possibly due to the formation of multilayers of catalyst particles on substrate surface at high Fe₃O₄ concentrations. The stacking of particles might lead to incomplete removal of residual polymer shell during initial heating stage. Eventually, amorphous carbon formed from the polymer shell would hinder the growth of VA-CNTs during CVD process. At 0.026 g ml⁻¹ Fe3O4, growth rate of CNTs is also highest among three concentrations used (table 1).

Significant changes in properties of VA-CNTs are observed with the addition of water vapor in the CVD process (figure 3). Water-assisted CVD produces CNTs with greater length (40 μm) (figure 3(b)) than the ones grown without water (6 μm) (figure 3(a), table 1). Some studies previously evidenced the effects of water on CNTs growth using various catalysts such as Fe₂O_3, cobalt [17, 18, 21]. The role of water in purifying amorphous carbon is well documented [17, 19, 20]. Water is considered as a weak oxidant which removes the deposited amorphous carbon on the active sites of the catalyst [21]. Another function of water is to inhibit Oswalt ripening by surface hydroxylation [17]. As a result, the introduction of water during CVD helps extend catalyst lifetime and promote growth rate, which are conducive to notably increased length of CNTs. However, exceeding a certain level, H₂O showed an overall negative effect on the growth of CNTs. In this case, it was difficult to keep the balance between the carbon supply and the solid carbon precipitation [22]. It was more and more difficult for the acetylene molecules to reach the catalysts. So, an amount of amorphous carbon was increased and the length of CNTs was reduced.

Zoom In Zoom Out Reset image size

Structures and properties of VA-CNTs grown with and without water were determined by TEM images and Raman spectroscopy. The TEM images can clearly distinguish amorphous carbon and structural defects of carbon nanotubes. As seen in figure 4(a), the rough surfaces of the CNTs are probably attributed to the formation of amorphous carbon and structural defects.

Zoom In Zoom Out Reset image size

During CNTs growth, under certain conditions, side walls can form cap-like structures and stack in the same direction to produce bamboo-like wall [23] morphology which are considered undesirable structural defects. The graphite layers consisting of one side of tubes are broken and are no longer continuous graphite sheets (figure 4(a). The addition of water not only reduces the formation of amorphous carbons but also promotes the growth of VA-CNTs into well-defined vertically standing and organized structures. Unlike the CNTs grown without H₂O, the CNTs have a hollow structure without a bamboo-like wall (figure 4(b)). It is known that the diameters of CNTs depend on the size of catalyst particle, amount of amorphous carbon formed on the CNTs walls and structural defects. Collected data point out that diameters and diameter distribution of CNTs grown with H₂O are smaller and more homogeneous than those synthesized without H₂O (figures 3(d), (e) and figures 4(a), (b)). Results in previous works show that the presence of reactive gases like H₂ or NH₃ can improve the homogeneity of nanoparticles which help narrow the size distribution of CNTs [21]. Hence, enhancement in uniformity of CNTs diameter may suggest a similar effect with respect to the presence of H₂O in the feed gas.

In this research Raman spectroscopy was used to detect the presence of amorphous and crystalline phases in CNTs samples. The Raman spectra (figure 5) of the CNTs samples grown with and without water vapor show the peaks at 1354 cm⁻¹ (D-line) and at 1594 cm⁻¹ (G-line), which are ascribed to amorphous carbon and to graphitized CNTs, respectively.

Zoom In Zoom Out Reset image size

The integrated intensity ratio of the G band over the D band can express the graphitization of the CNTs samples. Results from figure 5 show that values of I_G/I_D are 1.6 and 0.88 for the CNTs samples with and without water vapor. The analysis of Raman spectra evidences increased graphitization of CNTs samples with addition of water vapor during CVD process. These results are consistent with the SEM and TEM images which demonstrate the improvement in VA-CNTs structures in the presence of water vapor.