The morphology of the samples was observed by scanning electron microscopy (SEM) using a Carl Zeiss (ULTRA 55, Carl Zeiss, Oberkochen, Germany) with energy dispersive X-ray (EDX, INCA PentaFET × 3, Model: 7426, Oxford Instruments, Abingdon, Oxfordshire, UK) spectrometry mode. The Raman spectra were obtained using a Senterra R200-L Raman spectrometer CX-5461 purchase (Bruker, Germany) with a 514-nm line of laser source. Results and discussion To get the morphology, composition and the degree of graphitization of CNT arrays, the resultant SEM, TEM, EDX, and Raman spectra were used for characterization. As shown in Figure 1a,

the AAO template has flat surface with the regularly periodic pore structure. After completely removing AAO template framework, the resultant CNT arrays were obtained as shown in Figure 1b. The aligned CNTs have high density in consistent RAD001 mouse with that of the template. Figure

1 SEM images of the samples. (a) AAO template and (b) CNT arrays. Figure 2 is TEM image of CNT arrays after ultrasonic dispersion. It can be observed that CNTs with the assistance of the AAO template have good opening channels with the thickness of CNT walls of 8 to 10 nm, including about 25 layers. So CNTs prepared in our experiment are multi-walled ones. Compared with other reported research results [13], the obtained CNTs have clean and smooth surface with high degree of graphitization. Figure 2 TEM images of CNT. The inset is the low magnification image. Figure 3 presents the Raman spectra of CNT arrays with two kinds of diameters (80 to 100 and 110 to 150 nm). It is noted that there are two obvious peaks in the 1,350 and 1,580 cm−1, which are the D and G peak, respectively. By comparing the intensities of two peaks, the I G/I D of CNTs is about 2, which is better than those of other works using the same method [30]. Figure 3 Raman spectra of CNT arrays. In general, the diameter of CNTs is in consistent with pore size of AAO template. The roughness of CNTs has great relation with that of the hole wall of AAO template. In previously reported CVD experiments [12], the temperature of the system was increased quickly to reaction temperature

and then immediately started the CVD experiment. In this process, the temperature directly rose from room temperature to reaction temperature; in other words, the sample SPTLC1 has always been in a rapid heat treatment condition. Part of the internal thermal stress of the template was released through high-temperature deformation, but the majority of the thermal stress could not get released due to the rapid heating process. Thermal annealing is an effective method in thermal stress release [31]. In order to improve graphitization degree of CNTs, a heat preservation pretreatment for 1 h under 500°C was added during the fast heating process so that the template could be fully stretched and the deformation stress will be released completely.