Study of surface defects and crystallinity of MWCNTs growth in FCCVD

1 Introduction

In the mid-20th century, the scientists discovered the unusual carbon structures known today as carbon nanotubes (CNTs) [1]. CNTs can be described as rolled-up sheets of graphite, with the carbon atoms arranged in honeycomb-like hexagonal structures. These hollow graphene cylinders can be open-ended or closed with a hemispherical structure, such as fullerene [2]. CNTs are classified either as single-walled or multi-walled depending on the number of graphene sheets that constitute them [2, 3]. Potential applications include their use in polymer reinforcements and composite materials for transistors, quantum wires, hydrogen storage media and power production [2–4].

Several methods for the production of CNTs have been reported in the past literature [5–8] including arc discharge method, laser ablation method, chemical vapor deposition (CVD), etc. Each method has its own merits and demerits. The major drawbacks of arc discharge method and laser ablation method are the consumption of large amount of energy, uncontrolled process parameters, and limited volume of sample, and the final product mostly consists of amorphous carbon. Therefore, comparing with the arc discharge and laser ablation methods, CVD is considered as simple, low cost with low deposition rates and easily scalable. CVD is a continuous process, and currently, it is the best known technique for high-yield and impurity-free production of CNTs at moderate temperatures. It is a flexible technique and can be employed in different ways, for example, floating catalytic chemical vapor deposition (FCCVD) is a kind of CVD that allows CNTs to grow in the reactive gaseous environment in the presence of a catalyst [8]. In addition, it has the capability of controlling the size, growth rate, shape, diameter, length and alignment of CNTs. It can produce relatively large amounts of CNTs under mild conditions and with reduced cost.

Andrews et al. [7] obtained aligned CNTs by catalytic decomposition of ferrocene-xylene mixture. Ren et al. [8] synthesized aligned CNTs by plasma enhanced hot filament CVD. Many researchers have reported that the catalyst also plays an important role in determining the properties of CNTs [9]. Bonard et al. [10] reported that the diameter of nanotubes depends on the catalyst particles, as the growth is initiated by the catalyst. Shin et al. [11] also indicated that the size of transition metal can control the diameter and length of CNTs. Basaev [12] proved that with the decrease of ferrocene concentration, the melt’s free energy increases, which leads to the growth of smaller diameter nanotubes with a lower wall number. It was also concluded that an increase in the ferrocene concentration in catalyst mixture can have an adverse effect on the quality of CNTs via degradation process.

Therefore, in this detailed note, growth of vertically aligned MWCNTs arrays was studied using a double zone FCCVD at 800°C reaction temperature and adding the ferrocene in process as a catalyst. The ethylene, argon and hydrogen were used as carbon precursor, carrier gas and supporting gas, respectively. The objective of the work was to study the effect of catalyst weight on yield, quality, type, diameter, alignment and crystallinity of MWCNTs.

2 Materials and methods

2.1 Preparation of substrate

The substrate used for CNTs growth is shown in Figure 1. A silicon wafer substrate with average thickness of 20 mm was prepared and cleaned using RCA method. The cleaning was carried out in two different steps. In first step, a mixture of de-ionized H₂O, H₂O₂ and NH₄OH (5:1:1; v/v) was used for Standard Clean 1 (SC-1). The silicon wafer was soaked in SC-1 for 20 min at 75°C and then rinsed twice with distilled water. After SC-1, the silicon wafer was soaked for 5 min in buffer solution of HF and distilled water (1:50; v/v). In second step, a mixture of H₂O, H₂O₂ and HCl (6:1:1; v/v) was used for SC-2. The wafer was cleaned for 20 min at 75°C and rinsed with distilled water. Finally, the wafer was dried in an oven and stored in moisture- and dust-free environment.

Figure 1:

Dimensions and structure of the substrate.

Cleaned silicon substrate was placed in a furnace for oxidation at 1100°C temperature with heating rate of 15°C per minute. After getting the desired temperature, the oxygen was purged into the furnace. A SiO₂ layer of 300-nm thickness was achieved after oxidizing the substrate for 4 h. The oxidized substrate was exposed to E beam followed by oxidation to form a buffer (Al₂O₃) layer. This time, the oxidation was carried out at 600°C temperature for 2 h. The formation of Al₂O₃ layer of 40-nm thickness helped in synthesizing the aligned and high-density nanotubes. The chemicals used in the given work were supplied by the Sigma-Aldrich Corporation (St. Louis, MI, USA).

2.2 Growth mechanism of MWCNTs

2.2 Schematic of a CVD reactor used for growth of MWCNTs is shown in Figure 2. This reactor was composed of two horizontal heating zones and a quartz tube of outer diameter of 80 mm, inner diameter of 72 mm and length of 1000 mm. Each heating zone was 150 mm in length. Initially, the quartz boat containing 0.1 g of ferrocene catalyst was placed in the low-temperature zone-1. The temperature of the zone-1 was set to 300°C. Ferrocene vaporized at 230°C in zone-1 and carried by argon flow to the high temperature zone-2, where the substrate (Si/SiO₂/Al₂O₃) was placed for CNTs formation. The temperature of the zone-2 was maintained at 800°C. The quartz tube was connected to a gas mixer, and the flow rates of argon, hydrogen and ethylene were controlled by mass flow meters.

Figure 2:

Schematic of a CVD reactor used in the current research work.

In CNTs’ growth process, first of all, only argon gas (100 sccm) was flushed into the system and waited unless the reaction temperature reached 800°C with a heating rate of 15°C per minute. Soon after this, argon flow rate was reduced to 40 sccm, and ethylene and hydrogen gases were introduced into the reaction chamber with flow rate of 60 and 100 sccm, respectively, as shown in Figure 3. Hydrogen gas in the process also ensures the excessive carbon vent and avoids carbon saturation in high-temperature zone-2 of the quartz tube. After 1 h of CVD process time, the ethylene and hydrogen supplies were blocked and the furnace was allowed to cool down to the room temperature using argon flow of 100 sccm. The substrate implanted with CNTs was taken out from the quartz tube and stored in dust- and moisture-free environment. The similar process was repeated again with 0.125, 0.15, 0.175 and 0.2 g of ferrocene catalyst. Morphology, diameter, thickness and crystallinity of the synthesized CNTs were studied through TEM, Raman spectroscopy and TGA techniques.

Figure 3:

TEM micrographs of MWCNTs grown with different weights of ferrocene. (A) 0.1 g ferrocene, (B) 0.125 g ferrocene, (C) 0.15 g ferrocene, (D) 0.175 g ferrocene, (E) 0.2 g ferrocene.

High-resolution transmission electron microscopy (HRTEM-Zeiss Libra, Helmholtz-Zentrum Berlin für Materialien und Energie, Berlin, Germany) was used to investigate the micro-structures and surface defects of the grown nanotubes. CNT samples were sonicated by mixing them with isopropanol. After 15 min of sonication, CNTs were completely dispersed in the isopropanol. The sonicated samples were collected and placed on a copper grid to dry overnight. The copper grid was then loaded in a platform equipped with a mechanical arm for holding the specimen and controlling its position. The TEM micrographs of CNTs were taken at different resolutions and further processed to elaborate the structures, surface defects and inner-outer diameters of the grown nanotubes.

Raman spectroscopy (Horiba Jobin Yvon HR800, HORIBA Scientific, Edison, NJ, USA) was used to determine the crystallinity of the grown CNTs. The type and chirality of CNTs can also be determined from their Raman spectra. The Raman spectroscopy provided the intensity plots based on the light scattering in the range of 200 cm^-1 to 4000 cm^-1. The laser light interacted with the molecular vibrations, phonons and or other excitations in the structures. These interactions produced up or down shifts in the energy of the laser photons, which were recorded and used to study CNTs structures. In the given Raman spectroscopy analysis, a 514.53-nm Ar laser was used as an excitation source under ambient air conditions.

For thermogravimetric analyses, Perkin Elmer TGA analyzer (model Pyris 1, Waltham, MA, USA) was used to generate the thermograms of CNT samples. TGA analyses were performed in the temperature range of 25°C to 900°C. In the weight reduction bend, onset and oxidation temperatures were measured as a function of ferrocene weight. The key findings of the work have been discussed in greater details in the following sections.

3 Results and discussion

In this study, formation of vertically aligned MWCNTs was confirmed from TEM micrographs of the samples. TEM technique is important in measuring the inner and outer diameters, inner-shell spacing and number of layers in a MWCNT. Figure 3 shows the bulk growth of MWCNTs at different magnifications. TEM images obtained at higher resolutions were considered for measurement of the inter-shell spacing. The results showed that the inner-shell spacing remains in the range of 0.350 nm to 0.355 nm. This range predicts that the inner-shell spacing was in good agreement with the graphite inter-planar spacing of 0.336 nm [13]. The lowest inner-shell spacing of 0.350 nm was obtained with 0.15 g of ferrocene, which is an indication of relatively pure CNTs growth. For all other ferrocene weights, the inner-shell spacing was somewhat higher than those obtained with 0.15 g of ferrocene. This increase in inner-shell spacing might be due to the graphene sheet curvature modification by the tube diameter and size. The surface defects were also seen in some of CNTs grown with 0.125 g and 0.175 g of ferrocene where outer and inner tube walls formed wave-like structures, as shown in Figures 3B and 3D, respectively.

The surfaces of nanotubes grown with 0.15 g of ferrocene were relatively smoother and symmetric; very few structural imperfections were seen in CNT structures. However, a further increase in ferrocene weight from 0.15 g to 0.2 g resulted in formation of defective CNT structures and few black spots were also seen in TEM micrographs. The presence of these dark spots in the tube structures was referred to Fe particles encapsulated by the tube, as shown in Figure 3E.

The structural defects can be attributed to the presence of impurities in the product or possible variations in the inter-layer spacing within a MWCNT [11, 14]. It reveals that as the ferrocene weight is increased from 0.15 g, the impurity level also increases. For larger amounts of ferrocene, the available catalyst particles do not take part in the chemical reaction and consequently the formation of nanotubes; therefore, the unused particles get trapped between the tube walls causing defects in their structures. For the optimized ferrocene weight, very few particles of amorphous carbon were seen in CNT structures, which was an evidence of growth of highly crystalline structures of MWCNTs. It has also been experienced that the average diameter of MWCNTs can be reduced and controlled by using optimized weight of the catalyst [5–9]. Furthermore, the number of tube walls correlates with the tube diameter; smaller is the tube diameter lower will be the walls number.

Hydrogen gas may also be added in the growth process to improve the structural quality and to minimize the amorphous carbon in the product. It has been noticed that once the growth process gets saturated with respect to the carbon, the large molecules of carbon appear on quartz tube. Therefore, crystallinity is one of the key parameters for the study of CNT quality in terms of degree of graphitization [13]. The Raman spectra of the grown MWCNTs are shown in Figure 4. These spectra were used to study D and G intensity bands of MWCNTs. D-band in the Raman spectrum appears due to the disorderliness of the structure, and G-band is referred to the graphitic nature of the nanotubes. In this study, D and G bands were found at around 1360–1369 cm^-1 and 1590–1610 cm^-1, respectively. The degree of crystallization of the grown CNTs was expressed in terms of band intensity ratio (I_D/I_G). The lower values of I_D/I_G ratio correspond to the higher degree of crystallization [14]. I_D/I_G rations for the pure and aligned CNTs should be below unity. These investigations predicted I_D/I_G ratios in the range of 0.700–0.893.

Figure 4:

Raman spectra of MWCNTs grown with different weights of ferrocene.

The lowest I_D/I_G ratios were calculated with 0.15 g of ferrocene, as expressed in Table 1. It reveals that increase in catalyst weight results in highly pure CNTs but up to a certain limit because further increase in ferrocene weight also produces high I_D/I_G rations and low degree of tube crystallinity [10–14]. These findings were in line with the results reported by Basaev [12], who revealed that the degradation of quality of the nanotubes increases with an increase in ferrocene concentration. Lim et al. [15] also reported similar results with highest degree of crystallinity (0.48) when 0.1 g of ferrocene was used rather than 0.2 g. In their study, they also varied the reaction temperature; however, in the presented research work, the reaction temperature was kept constant at 800°C.

Apart from TEM and Raman spectroscopy, TGA study was also conducted on the purity of the carbon product. It is a very simple and accurate analytical technique, which can quickly deliver the results on thermal degradation of the product. Purity of the nanotubes is often considered as a vital factor for their structural and chemical characterization. TGA may also be used to assess the thermal stability of the materials along with their purity. The imperative parameters measured in the weight reduction bend are the initial temperature and the oxidation temperature, as shown in the TGA thermogram in Figure 5. The initial temperature is characterized as the temperature at which the material begins to deteriorate. The oxidation temperature is often defined as the thermal stability of the material. Lima et al. [16] revealed that amorphous carbon decomposes at the lowest temperature range of 200°C to 400°C. Similarly, Dunens et al. [17] observed that amorphous carbon contaminants had lower oxidation temperatures in the range of 200°C to 300°C. Similar results were obtained in the given TGA study of MWCNT samples grown with different weights of the ferrocene, as shown in Table 2.

Figure 5:

TGA thermograms of MWCNTs grown with 0.1 g, 0.15 g and 0.2 g of ferrocene.

In case of 0.15 g of ferrocene, a negligible quantity of the amorphous carbon was observed in TGA thermogram of CNT sample, as shown in Figure 5. Nevertheless, the oxidation temperature for MWCNTs varies from material to material and typically ranges from 400°C to 650°C [17]. Ravindra and Bhat [18] measured the oxidation temperature at about 550°C for the growth of pure CNTs using ferrocene catalyst. In the present case, similar results were obtained with 0.15 g of ferrocene. The respective TGA thermogram revealed the presence of negligible amount of the amorphous carbon and iron content in the product. The amorphous carbon content disappeared at the temperature around 450°C, and oxidation temperature was much higher as compared to samples grown with the ferrocene weights other than the optimum. It was an indication of relatively pure MWCNTs production [18–20].

4 Conclusion

In this detailed note, MWCNTs were synthesized by using a CVD reactor at 800°C reaction temperature. The synthesized CNTs were characterized by using TEM, TGA and Raman spectroscopy techniques. The experimental results confirmed the formation of relatively pure MWCNTs having the graphitic structures especially for 0.15 g of ferrocene. The I_D/I_G ratio was decreased with an increase in catalyst weight. The lowest I_D/I_G ratio was noticed with 0.15 g of ferrocene. For ferrocene weights above and below the optimum, some of the grown nanotubes were found defective, where one side of the tube surface was relatively smooth while the wave-like structures were evident on the other side of the surface. This study also revealed that the average diameter of MWCNTs can be reduced and controlled using optimized weights of the catalyst (0.15 g). The number of walls also correlates with the tube diameter; smaller tube diameter corresponds to lower number of walls. The smallest inner-shell spacing was measured about 0.350 nm. TGA of the grown samples also confirmed the formation of relatively pure and aligned MWCNTs, especially for 0.15 g of ferrocene, with oxidation temperature of 625°C.