Elongation of arrays of amorphous carbon tubes

Due to its special properties, amorphous carbon (a-C) materials have been reported to display novel functions for many potential optical [1,2], frictional [3] and electronic [2,4] applications. These include protective coating, ultra durable anti-reflection coatings [1,2], mechanical devices [2,3], and electron emitters [4]. The ability to fabricate a-C into desired physical shapes has attracted much attention. For example, anodic aluminum oxide (AAO) has been used as templates to shape a-C into tubular shapes by pyrolyzing propylene inside the AAO channels [5]. In our study herein we report on a process to grow a-C tubes (a-CT) by coupling two fabrication methods we recently developed [6,7]. First, a layer of short aligned a-CT was deposited by LPCVD from SiCl3CCl3 on an AAO substrate [6]. Then, by employing the Na-filled AAO (Na@AAO) reactive template method, we were able to significantly elongate the a-CT [7]. Using a hot-wall LPCVD reactor with a base pressure of 10 1 Pa, SiCl3CCl3 (Aldrich, 97%) was evaporated at 273 K without carrier gas. The precursor was decomposed at 1273 K and 10 Pa to deposit a layer of carbon on AAO substrates (Whatman Anodisc 13, observed pore diameter: 280 nm, thickness: 60 lm). After 3 h, the AAO substrates turned black and the product a-C film on AAO (a-C/AAO) was obtained. As shown by the scanning electron microscope (SEM, JEOL JSM-6330F at 15 kV) images in Fig. 1, the average diameter of the open-ended a-CT, 330–340 nm, is close to the channel diameter in the AAO. The average thickness of the a-CT wall was 30 nm. In Fig. 1(b), the average thickness of the array layer was estimated to be 800 nm, much thinner than the AAO thickness of 60 lm. We speculate that the activated species generated from SiCl3CCl3, :CCl2, cannot diffuse effectively into the AAO channels to form a thick carbon layer. Fig. 1(b) also shows that the a-CT array is connected as a layer of film on top of the AAO. Using energy dispersive spectroscopy (EDS), the composition of the isolated tube array layer was confirmed to be carbon. Thus, the data support the conclusion that we have successfully deposited a-C on AAO (a-C/AAO) in the form of short a-CT array. We also attempted to perform this process on a home-made AAO with a channel diameter of 30 nm. However, :CCl2 apparently could not diffuse into the narrow AAO channels to grow into an a-CT layer. After the short a-C/AAO layer was fabricated by LPCVD, we explored the possibility of extending the tube length by the sodium reactive template method that had been previously developed [6]. By decomposing NaH (Aldrich, 0.2 g, 8 mmol) on a-C/AAO inside a tube furnace at 623 K under 1 atm of Ar (flow rate: 30 sccm) for 0.5 h, Na filled the channels of AAO to generate Na@a-C/AAO. Further reaction of Na@a-C/AAO with Carbon 43 (2005) 2618–2641