Graphitization Mechanism during the Carbon-Nanotube Formation

New Energy and Industrial Technology Development Organization (NEDO), ... of the formation process and develop to form highly ordered graphene lay...
0 downloads 0 Views 293KB Size
J. Phys. Chem. B 2002, 106, 1849-1852

1849

Graphitization Mechanism during the Carbon-Nanotube Formation Based on the In-Situ HRTEM Observation Ayumu Yasuda,*,† Noboru Kawase,‡ Florian Banhart,§ Wataru Mizutani,| Tetsuo Shimizu,| and Hiroshi Tokumoto| Nanotechnology Research Institute (NRI), National Institute of AdVanced Industrial Science and Technology (AIST) Tsukuba Central 4, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8562, Japan, Joint Research Center for Atom Technology (JRCAT), c/o AIST Tsukuba Central 4, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8562, Japan, New Energy and Industrial Technology DeVelopment Organization (NEDO), c/o AIST Tsukuba Central 4, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8562, Japan, Nitto Analytical Techno-Center (NTC), 1-1-2 Shimohozumi, Ibaraki, Osaka 567-8680, Japan, Zentrale Einrichtung Elektronenmikroskopie, UniVersita¨ t Ulm, Albert-Einstein-Allee 11, 89069 Ulm, Germany ReceiVed: July 12, 2001; In Final Form: NoVember 6, 2001

The mechanism of the carbon nanotube (CNT) formation and its graphitization has not been well understood and is still under controversy. The authors have been studying a new process to form CNTs, where polyynecontaining carbons are heated and irradiated by an electron beam in a transmission electron microscope (TEM). The technique is compatible to an in-situ observation and was applied to a high-resolution TEM (HRTEM) to elucidate the graphitization mechanism. The observation shows that micro-graphenes exist from the beginning of the formation process and develop to form highly ordered graphene layers through fusion and rearrangement of the premature layers, accompanied by evaporation of micro-graphenes. The CNT formation mechanism focusing on its graphitization is presented and discussed.

Introduction After finding CNTs1-3 much attention has been paid to this material, while the mechanism of the CNT formation is still not well understood. The difficulty is in lack of experimental data, though the understanding of the mechanism is crucial for a material and process design. CNTs have been formed by an arc discharge, chemical vapor deposition, laser ablation, and so on. Great efforts have been paid to understanding the formation mechanism, where a lot of work is in progress at the moment to try and improve the situation.4-14 On the other hand, CNTs have been found to be formed by the thermal surface decomposition of SiC in a TEM, and the formation process was well analyzed.15-17 To cope with the experimental limitations, computer simulations have been carried out and presented.18-21 In contrast, the formation mechanism of fullerenes is better understood by in-situ observations. The mechanism has been studied by a modified TEM, and it has been found that the fullerenes are formed through micro-structural changes of carbon black primary particles.22-24 Parallel to these works, pioneering TEM studies on the formation of onion-like carbons have been presented.25-29 A beautiful series of photographs have revealed that those are formed by collapse of carbon nanoparticles and rearrangement of graphene layers.30 A new process for CNT formation, where polyyne-containing carbons are heated and irradiated by an electron beam in a TEM, has been developed.31-34 The polyyne-containing carbons are prepared in the following scheme, in which the reduction by * Corresponding author. E-mail: [email protected]. † NRI/AIST, JRCAT, NEDO. ‡ NTC. § Universita ¨ t Ulm. | NRI/AIST, JRCAT.

the magnesium cation radical is assumed,35 though it is not well understood as a reductant.

Scheme: (-CF2-CF2-)n + 4n Mg•+ f (-CtC-)n + 4n Mg2+ + 4n FSo far, polyyne-route formation of fullerenes has been assumed theoretically36 and been supported by experimental data.37-39 As for CNTs, the polyyne-route formation has been shown to work experimentally, as well.40,41 The process employed in this paper is associated with a special “thermally removable protection layer technique,”34 to handle the unstable polyyne-containing carbons, which are damaged by water, oxygen, and light in air. This process features two advantages: in-situ observation of the growth, and formation of CNTs directly on a substrate. The in-situ observation has been tried already by a TEM and the formation mechanism has been partially elucidated that rodlike carbons are formed first and graphtization follows. In this paper, the technique was applied to an HRTEM, focusing on the graphitization mechanism of the CNTs. Experimental Section The CNT formation is comprised of two processes: preparation of polyyne-containing carbons, and irradiation of an electron beam at 600-800 °C under low pressure. The details have been presented in other publications.34 For the preparation of the polyyne-containing carbons, poly(tetrafluoroethylene) (PTFE) films were reduced electrochemically by a two-electrode method (anode: magnesium; cathode: stainless steel) under argon at 0 °C, associated with a sacrificial anode technique. The PTFE films (10 mm × 10 mm × 60 µm) were charged with the solvent containing supporting salts

10.1021/jp012677u CCC: $22.00 © 2002 American Chemical Society Published on Web 02/02/2002

1850 J. Phys. Chem. B, Vol. 106, No. 8, 2002

Yasuda et al.

(tetrahydrofuran (THF): 30 mL; LiCl: 0.8 g; FeCl2: 0.48 g) in a flask. The DC voltage (40 V) was applied between the anode and the cathode for 10 h. After the reduction, the films were washed with THF and dried in a vacuum. The films were analyzed by an IR spectrograph (MagnaIR 760, Nicolet), Raman (HoloLab 5000, Kaiser), and XPS (ESCA 750, Shimazu). The prepared polyyne-containing carbon film was embedded in an epoxy resin (Araldite CY211, Ciba) and cross sections were cut by a microtome (UltraCut, Leica). The specimen was heated from a room temperature to 600 °C and irradiated by an electron beam (accelerating voltage: 1.25 MV, current intensity: 50 A/cm2) in a HRTEM (ARM1250, JEOL). The in-situ observation was carried out parallel to the irradiation. Results and Discussion The analysis of the prepared polyyne-containing carbons, i.e., the precursor for the CNT formation, is summarized first. The details have been published in other papers.34 The TEM photograph of the reduced PTFE film shows that only the surface of the film is reduced to the carbonized material. The reduced layer is ca. 1 µm thick and low in density, which is consistent with the reaction scheme, i.e., elimination of fluorines. The reduced layer contains sp and sp2 carbons, as concluded from the Raman, IR, XPS analyses, while sp3 carbon is negligible. The ratio of sp and sp2 carbons is not well analyzed in this work. The eliminated fluorines remain in the carbonized layer in the form of anion. The electron irradiation of the film after heating to 600 °C led to formation of CNTs. The yield cannot be measured because the size of the specimen is too small . So far, the lowaccelerating-voltage TEM (accelerating voltage: 100 kV, current intensity: ca. 1A/cm2) has been used by us,34 and is replaced by an HRTEM (accelerating voltage: 1.5 MV, current intensity: ca. 50 A/cm2) for this study. Though the accelerating voltage and the current intensity are much higher than before, the CNTs are formed almost in the same way. The difference is in the displacement of the CNTs during the formation. During the formation, the CNTs move sideways (like spring rods) by the highly accelerated electron beam (1.25 MV), in contrast to that the displacement is not observed at 100 kV. Therefore, the high-resolution observation of the whole CNTs is not carried out and a series of photographs are taken at the bottom, where the movement is negligible. In addition, the observation faces the difficulty of focusing. Quick focusing to follow the CNT formation is difficult and the photographs are sometimes not well focused. The temperature rise of the specimen during the observation is assumed negligible,42 though its direct measurement is experimentally difficult. The excitation of phonons, leading to a heating of the specimen, is mainly due to an inelastic scattering of the projectile by electrons. The mean free path of the projectile in the sample depends on its mass and energy, and is more than 100 nm for a highly accelerated electron, typical in a TEM. Therefore, in small objects such as carbon nanostructures, which are of some 10-100 nm in size, electron irradiation is expected not to heat the specimen by more than a few degrees. A series of photographs during the graphitization are taken at an interval of a few minutes (Figure 1). As already presented,34 the CNT formation is comprised of two steps: the fast formation of a rod, and the slow formation of a hollow inside the rod, accompanied by the graphitization. The first step is too quick (a few seconds at 800 °C and ca. 1 min at 600 °C) and failed to be taken into photographs. As of now, the photographs of the second slow step are successfully taken. Therefore, the rod-shape and the hollow inside the rod are

Figure 1. A series of HRTEM photographs showing the graphitization during the CNT formation. The arrows show the typical development of graphene layers through a defect reorganization. The region (a) and (b) are magnified in Figures 2 and 4. The photographs were taken every few minutes.

Figure 2. HRTEM photographs showing the rearrangement of the graphene layers.

already formed in Figure 1. The arrows show the graphitization process to proceed through rearrangement and fusion of the graphene layers, and the region indicated by the arrow (a) is magnified in Figure 2, (b) in Figure 4.

Graphitization during Carbon-Nanotube Formation

J. Phys. Chem. B, Vol. 106, No. 8, 2002 1851

Figure 5. Schematics showing the fusion of the round-shaped graphene (traced from Figure 4).

Figure 3. Schematics to show the rearrangement of the graphene layers (traced from Figure 2).

Figure 4. HRTEM photographs showing the fusion of the roundshaped graphene.

The rearrangement is observed in Figure 2, whose skeletons are shown in Figure 3 for an aid in understanding. The arrow indicates that the connection between the first layer from the top and the second (A in Figure 3) is broken (B). The connection is formed again between the second layer and the third (C). The connection is broken again, whereas the big strain is exerted to form the sharp edge (D). The connection is formed again between the second and the third (E), and moves further between the third and the fourth (F). The rearrangement includes sometimes many layers, which move dynamically. The fusion of the round-shaped graphene is observed in Figure 4, whose skeletons are shown in Figure 5. The roundshaped graphene appears (A in Figure 4) and fuses into a few layers (B). Through the rearrangement of a few layers (C), the highly ordered graphene layers are completed (D). The roundshaped graphene is covered by a graphene layer and does not come out to form fullerene-like carbons (A). The findings are summarized schematically in Figure 6. At the beginning of the graphitization process (A in Figure 6), the tubular shape is already formed by the partially aligned small

Figure 6. Schematics showing the graphitization during the CNT formation.

graphenes. The hollow is not formed and filled with the micrographenes. The rearrangement and the fusion of the small graphenes take place to form the aligned, larger-size graphenes (B). Parallel to that, the hollow inside the rod is formed by the diffusion of the micro-graphenes through the premature CNT wall. In the course of this stage, the round-shaped graphene, which is possibly a primitive form of fullerene, appears occasionally. The rearrangement and fusion of the graphenes proceed, parallel to the completion of the hollow formation (C). Finally, the graphitization of the CNT is completed (D). A primitive form of a CNT has been reported to be turned into onion-like carbon by electron-beam irradiation.25 This transformation is not observed in this experiment, and a selective formation of CNTs and carbon nanoparticles (CNPs) is observed instead. The precursor employed here is unstable and at a damaged place, where the precursor is oxidized and has higher surface tension, the CNPs are formed preferentially to reduce the surface energy.

1852 J. Phys. Chem. B, Vol. 106, No. 8, 2002 Conclusion The graphitization process during the CNT formation was observed in an in-situ way by an HRTEM. The observation shows that the micro-graphenes exist from the early stage of the formation process and develop to form highly ordered graphene layers, accompanied by evaporation of micrographenes. The graphitization is assumed to proceed through fusion and rearrangement of micro-graphenes and graphene layers. Acknowledgment. The authors acknowledge Prof. M. Ru¨hle and Dr. F. Philipp of the Max-Planck Institu¨te fu¨r Metallforschung, Stuttgart, Germany, for access to a HRTEM (ARM1250). References and Notes (1) Iijima, S. Nature 1991, 354, 56. (2) Iijima, S.; Ichihashi T. Nature 1993, 363, 603. (3) Bethune, D. S.; Kiag, C. H.; de Vries, M. S.; Gorman, G.; Savoy, R.; Vazquez, J.; Beyers, R. Nature 1993, 363, 605. (4) Saito, R.; Fujita, M.; Dresselhaus, G.; Dresselhaus, M. S. Mater. Sci. Eng. 1993, B19, 185. (5) Endo, M.; Kroto, H. W. J. Phys. Chem. 1992, 96, 6941. (6) Iijima, S. Mater. Sci. Eng. 1993, B19, 172. (7) Iijima, S.; Ichihashi, T.; Ando, Y. Nature 1992, 356, 776. (8) Smalley, R. E. Acc. Chem. Res. 1992, 25, 95. (9) Saito, Y.; Okuda, M.; Tomita, M.; Hayashi, T. Chem. Phys. Lett. 1995, 236, 419. (10) Ma, X.; Wang, E. G.; Tilley, R. D.; Jefferson, D. A.; Zhou, W. Appl. Phys. Lett. 2000, 77, 4136. (11) Diener, M. D.; Nichelson, N.; Alford, J. M. J. Phys. Chem. B 2000, 104, 9615. (12) Setlur, A. A.; Dai, J. Y.; Lauerhaas, J. M.; Chang, R. P. H. Carbon 1998, 36, 721. (13) Loiseau, A.; Willaime, F. Appl. Surf. Sci. 2000, 164, 227. (14) Amelinckx, S.; Zhang, X. B.; Bernaerts, D.; Zhang, X. F.; Ivanov, V.; Nagy, J. B. Science 1994, 265, 635. (15) Kusonoki, M.; Rokkaku, M.; Suzuki, T. Appl. Phys. Lett. 1997, 71, 2620.

Yasuda et al. (16) Kusunoki, M.; Suzuki, T.; Kaneko, K.; Ito, M. Philos. Mag. Lett. 1999, 79, 153. (17) Kusunoki, M.; Suzuki, T.; Hirayama, T.; Shibata, N. Appl. Phys. Lett. 2000, 77, 531. (18) Charlier, J.-C.; Vita, A. D.; Blase, X.; Car, R. Science 1997, 275, 646. (19) Ajayan, P. M.; Ravikumar, V.; Charlier, J.-C. Phys. ReV. Lett. 1998, 81, 1437. (20) Charlier, J.-C.; Blasse, X.; De Vita, A.; Car, R. Appl. Phys. A.; Mater. Sci. Process. 1999, A68, 267. (21) Roland, C.; Bernholc, J.; Nardelli, M. B.; Maiti, A. Mol. Simul. 2000, 25, 1. (22) Burden, A. P.; Hutchison, J. L. Carbon 1997, 35, 567. (23) Burden, A. P.; Hutchison, J. L. Carbon 1998, 36, 1167. (24) Burden, A. P.; Baigrie, S.; Hutchison, J. L. J. Microscopy 1998, 192, 7. (25) Ugarte, D. Nature 1992, 359, 707. (26) Ugarte, D. Chem. Phys. Lett. 1993, 209, 99. (27) Fuller, T.; Banhart. F. Chem. Phys. Lett. 1996, 254, 372. (28) Banhart, F.; Ajayan, P. M. Nature 1996, 382, 433. (29) Zwanger, M. S.; Banhart, F. Philos. Mag. 1995, B72, 149. (30) Ugarte, D. Chem. Phys. Lett. 1993, 207, 473. (31) Kawase, N.; Yasuda, A.; Matsui, T.; Yamaguchi, C.; Matsui, H. Carbon 1998, 36, 1234. (32) Kawase, N.; Yasuda, A.; Matsui, T. Carbon 1998, 36, 1864. (33) Kawase, N.; Yasuda, A.; Matsui, T.; Yamaguchi, C.; Matsui, H. Carbon 1999, 37, 522. (34) Yasuda, A.; Kawase, N.; Matsui, T.; Shimidzu, T.; Yamaguchi, C.; Matsui, H. React. Funct. Polym. 1999, 41, 13. (35) Tomilov, A. P. Russ. J. Electrochem. 1996, 32, 25. (36) Kroto, H. W.; Walton, D. R. M. Philos. Trans. R. Soc. London 1993, A343, 103. (37) Rubin, Y.; Kahr, M.; Knobler, C. B.; Diederich, F.; Wilkins, C. L. J. Am. Chem. Soc. 1991, 113, 495. (38) McElvany, S. W.; Ross, M. M.; Gorof, N. S.; Diederich, F. Science 1993, 259, 1594. (39) Hunter, J. M.; Fye, J. L.; Roskamp, E. J.; Jarrold, M. F. J. Phys. Chem. 1994, 98, 1810. (40) Kavan, L.; Hlavaty, J. Carbon 1999, 37, 1863. (41) Hlavaty, J.; Kavan, L.; Kasahara N.; Oya A. Chem. Commun. 2000, 737. (42) Banhart, F. Rep. Prog. Phys. 1999, 62, 1181.