Copied & Translated from Kikan Fullerene (April 2002) 5. Introduction of Laboratories (38)

Shigeo Maruyama Group
Department of Mechanical Engineering, The University of Tokyo

Updated: '02/5/3

Japanese Mode is Here


Contents

1. Overview
2. Research Topics
3. Introduction of Research Topics
3-1. Generation of SWNTs with Alcohol Catalytic CVD technique
3-2. FT-ICR mass spectroscopic studies of precursor clusters of SWNTs
3-3. Molecular dynamics simulation of formation process of SWNTs
3-4. Hydrogen Storage with Carbon Nanotubes
3-5. Molecular dynamics simulation of heat conduction of SWNTs
3-6. Other Molecular dynamics Simulations
References
Fig. 1. Group members in April 2002.
Fig. 2 Laser-oven SWNT generator.
Fig. 3 Alcohol catalytic CVD apparatus.
Fig. 4 TEM of SWNTs generated from alcohol CCVD method.
Fig. 5 Macro Raman spectroscopy apparatus.
Fig. 6 Raman spectra of SWNTs from laser-oven and alcohol CCVD.
Fig. 7 SPM (AFM/STM) apparatus with temperature controlled vacuum chamber.
Fig. 8 AFM image of purified SWNT bundle.
Fig. 9 FT-ICR mass spectrometer with direct-injection laser-vaporization cluster beam source.
Fig. 10 Negative ion clusters from Ni/Co loaded graphite sample.
Fig. 11 Chemical reaction of Ni- and Co-attached carbon clusters with NO.
Fig. 12 MD simulation of precursor cluster of SWNTs.
Fig. 13 An imperfect SWNT obtained from MD simulation
Fig. 14 MD simulation of hydrogen storage in a SWNT bundle.
Fig. 15 Dependence of thermal conductivity of SWNT on tube length.
Fig. 16 Phonon dispersion relations and phonon density of states directly calculated from MD.
Fig. 17 Formation of peapod.
Fig. 18 Water cluster inside a (10, 10) SWNT.

1. Overview

We are studying fullerene and nanotubes at Department of Mechanical Engineering in Hongo Campus of the University of Tokyo in Bunkyo-ku, Tokyo [1]. Because of the recent boom of nanotechnology and carbon nanotubes, we do not need to spend too much time answering the question "Why are you studying fullerene and nanotubes at Department of Mechanical Engineering?" I have described the detailed explanation why and how I have started this field of studies in more than 10 years ago, after my 2 year visit to Professor Richard Smalley at Rice University.

Here, we belong to Shoji-Maruyama Laboratory in the department. Maruyama group was in Engineering Research Institute of The University of Tokyo for 3 years from 1998 through 2001, and now it is back at Department of Mechanical Engineering. The members of our group are in Fig. 1. Three undergraduate students are joining in end of April. The current research topics of our group are all related with single-walled carbon nanotubes; a half of students is doing experiments and other half is using molecular dynamics simulations.


[Click for enlargement]


Fig. 1. Group members in April 2002. From left, Miwako Watanabe (secretary), Tatsuto Kimura (PD), Shuhei Inoue (D3), Shohei Chiashi (M2), Yasushi Shibuta (D2), Shigeo Maruyama (Associate Professor), Kazunori Teshima (M2), Satoru Ogawa (M1), Yuhei Miyauchi (M1), Yuki Taniguchi (M1). Mitsuru Inoue (Research Associate) is missing.


2. Research Topics

I have started research topics of fullerene and carbon nanotubes after my 2 year visit to Professor Rick E. Smalley at Rice University from 1989. Initially, I studied carbon, silicon and metal clusters with Fourier Transform Ion Cyclotron Resonance (FT-ICR) mass spectrometer [3,4]. At the time fullerene were called as "clusters". After the discovery of macroscopic generation technique of fullerene by Kratschmer and Huffman, all members at Smalley group started to generate fullerene in arc-discharge technique [5].

After coming back to the University of Tokyo, I have made the arc-discharge fullerene generator. At the same time, I have started molecular dynamics simulations of carbon cluster formation process and others, partially because I could not afford more experimental apparatus. When I visited Smalley group for a month in 1992, we have made a special carbon nanotube generator: we tried to grow carbon nanotubes from a tip of carbon fiber by focused Ar laser in a vacuum chamber filled with a few Torr hydrocarbon. At the time, the only detection technique we planned was the expected field emission from carbon nanotube. No SEM or TEM was there. It took a few more years before they have done interesting field emission experiments. Probably, with current knowledge of nanotubes, we can grow single-walled carbon nanotube by tuning metal catalyst and carbon source.

I have changed my research topics from clusters, fullerene, endohedral metallofullerene toward carbon nanotubes: by constantly increasing number of carbon atoms, in much slower pace compared with Smalley group. My main interest has been generation techniques and the self-assembly generation mechanism [6-10]. I have generated fullerene and SWNTs by using arc-discharge, laser-oven, and catalytic CVD generators. At the same time, precursor clusters were examined with FT-ICR mass spectrometer with direct-injection cluster beam source. Furthermore, molecular dynamics method was used for simulations of formation process of clusters, fullerene, and SWNTs. Recently, several applications of SWNTs were examined by molecular dynamics simulations such as hydrogen storage and heat conduction.

Research topics such as hydrogen storage and heat conduction are very natural even in Department of Mechanical Engineering and I am very happy that I can talk about nanotubes in conferences of my main field. Very recently, many research activities related with carbon nanotubes are actually being done in Mechanical Engineering field.


3. Introduction of Research Topics

3-1. Generation of SWNTs with Alcohol Catalytic CVD technique [11]

We have been working with the laser-oven apparatus in Fig. 2, since the purest SWNTs can be made with this technique. Dr. Masamichi Kohno (previous research associate at Engineering Research Institute, currently at AIST) designed the apparatus based on the apparatus at Achiba group at Tokyo Metropolitan University. He was former PhD student of Professor Achiba.
Since catalytic CCVD methods are recently employed for quite pure generation of SWNTs such as HiPco, we have converted the laser-oven apparatus to the CCVD apparatus as in Fig. 3.


[Click for Picture]


Fig. 2 Laser-oven SWNT generator


[Click for enlargement]


Fig. 3 Alcohol catalytic CVD apparatus


While we are trying the nanotube generation from C60 as carbon source, we accidentally found that very pure SWNTs were made from ethanol used for the solvent of catalyst metals [11]. Here, the Fe/Co metal catalysts were supported with Y-type zeolite according to the technique developed by Professor Shinohara at Nagoya University. ' method. Our technique is just very simple as follows.
(1) Catalyst on a quartz boat is set in the quartz tube in the electric oven.
(2) While heating up the oven, Ar flow is kept.
(3) At desired temperature, once evacuate the quartz tube with the mechanical vacuum pump.
(4) Introduce alcohol vapor flow from the room temperature reservoir.
The TEM image of 'as grown' SWNTs at 800 deg. C reaction for 10 min. is shown in Fig. 4. No amorphous carbon, MWNTs, carbon nanoparticles, metal particles are observed. Furthermore, this alcohol CCVD method favors much lower reaction temperature compared with normal CCVD using hydrocarbons. Considerable amount of SWNTs are made at 550 deg. C from methanol [11].


[Click for enlargement]
[Click here for SEM picture]

Fig. 4 Transmission electron micrograph of SWNTs generated from alcohol CCVD method.


Raman spectra is now unavoidable technique for SWNT characterization. The macro-Raman apparatus in Fig. 5 was constructed in a dark room just next to the CCVD apparatus. Dr. Masamichi Kohno had mainly designed the optics with kind advises from Professor Kataura at Tokyo Metropolitan University. Ar ion laser at 488 nm is used for the excitation. After plasma-line filter, laser light is mildly focus to the sample material, and back-scattered light is introduced through a notch-filter to 30 cm single monochrometer (Chromex 500is2-0419) with CCD detector (Andor DV401-FI).


[Click for enlargement]


Fig. 5 Macro Raman spectroscopy apparatus


Fig. 6 compares Raman spectra of SWNTs generated with Alcohol CCVD and with laser-oven method. G-band (graphitic band) at around 1590cm-1, broader D-band (disorder band) at around 1350cm-1, and RBM (radial breathing mode) at 150`300 cm-1 are clearly observed for both samples. Compared with the laser-oven sample, diameter of SWNTs from ACCVD are smaller and widely distributed. Because of the thinner metallic nanotubes at around 240`300cm-1 excited with 488nm light, Breit-Wigner-Fano (BWF) shape is observed for ACCVD sample. It is well known that the observed Raman of SWNTs is resonant Raman. The resonance behavior depending on chiralities of nanotubes are summarized as Kataura plot. Detailed explanations and useful pictures for the understanding and assignment of resonant Raman scattering is shown in our home page [12].


[Click here for temperature dependence of ethanol CCVD]


Fig. 6 Raman spectra of SWNTs from laser-oven technique and alcohol CCVD techniques (Excitation wavelength is 488nm).


We are using the TEM facility at Engineering Research Institute and SEM apparatus at Nakao Laboratory of our department. In addition to these observations, we have implemented the scanning probe microscopy (AFM/STM) as shown in Fig. 7. This SPM by SII can be operated with vacuum condition and special atmospheric gas condition. It also have the temperature control unit for -100 deg C through 800 deg. C. In addition to the simple observation of nanotubes, some handling experiments are possible. Fig. 7 is an example of AFM image of bundle of SWNTs which is prepared by laser-oven method and purified with H2O2 solution.


[Click for enlargement]


Fig. 7 SPM (AFM/STM) apparatus with temperature controlled vacuum chamber.




Fig. 8 An AFM image of purified SWNT bundle.


3-2. FT-ICR mass spectroscopic studies of precursor clusters of SWNTs [15-17]

In order to study the formation mechanism of SWNTs, generation experiments with laser-oven and alcohol CCVD methods are not enough. All we can observe are final results and intermediate reactions process are only speculated. Here, we use the FT-ICR apparatus with direct-injection supersonic-expansion cluster beam source for the detection of precursor clusters of SWNTs [15-17]. Experiments are the same as those for fullerene and endohedral metallofullerene [13,14]. The simple difference is the kind of metal atoms doped in graphite sample. The basic design of our FT-ICR apparatus in Fig. 9 is the same as Smalley group [4].

The experimental procedure is as follows.
(1) Solid sample is vaporized with focused laser beam.
(2) With the timed pulsed He gas, clusters are formed in the small nozzle and expands to vacuum.
(3) Supersonically expanded and cooled cluster beam is directly injected to the 6 T magnetic field of superconductive magnet.
(4) Cluster ions are trapped in the ICR cell for a minutes.
(5) Unwanted cluster ions can be excited away from ICR by SWIFT technique.
(6) Chemical reaction and/or laser photo-dissociation experiment can be done for trapped cluster ions.
(7) High resolution mass spectrum can be obtained from the ion-cyclotron resonance frequency.


[Click for enlargement]


Fig. 9 FT-ICR mass spectrometer with direct-injection laser-vaporization cluster beam source.


Fig. 10 is a mass spectrum of negative clusters generated by laser-vaporization of Ni/Co doped (0.6 at. %, each) graphite sample. The bottom panel in Fig. 10 is the simulated mass spectrum assuming the natural abundance of isotopes of each elements. It should be noticed that isotopes of H or C has almost integer amount of mass in amu units. However, with a metal atom such as Ni or Co, the deviation from the integer mass due to the principle of relativity are clearly observed. Some of carbon clusters with about 50 to 55 carbon atoms have a few Ni or Co atoms.

An example of chemical reaction experiment with FT-ICR is shown in Fig. 11.
(a) As injected clusters.
(b) After 2 s reaction with NO at 10-7 Torr.
(c) After 10 s reaction with NO.
Carbon cluster ions with Ni or Co are very reactive to NO. On the other hand, carbon clusters with La, Y, Sc metal atom were quite unreactive.


Fig. 10 Negative ion clusters from Ni/Co loaded graphite sample.




Fig. 11 Chemical reaction of Ni- and Co-attached carbon clusters with NO.


3-3. Molecular dynamics simulation of formation process of SWNTs

We have been studying the formation mechanism of C60, the beautiful soccer ball structure, with molecular dynamics method. Starting from random gas-phase carbon atoms, the formation process of C60 was reproduced. Base on those simulations, we have proposed a fullerene formation model [18,19]. Then, we have proposed the classical potential function between metal atoms (La, Sc and Ni) and carbon clusters. Using these potential functions, formation process of endohedral metallofullerene was simulated [20-22].

The experimental conditions of generation of SWNTs by arc-discharge and laser-oven method are the same as for endohedral metallofullerene except for the doped metals. Hence, the same molecular dynamics simulation as for endohedral metallofullerene is used for the SWNTs generation by simply changing metal atom to Ni [23,24].
As an initial condition all atoms are in random gas phase, and the growth process of cluster are simulated. Ni-carbon binary clusters with 50-100 atom range are shown in Fig. 12. Most of carbon clusters seem that they are trying to make spherical cage structure, just like the cases of fullerene simulations. However, one or two Ni atoms are disturbing the complete closure. A metal atom is always in the defect point (see the animation).Those structures can explain the results of FT-ICR experiments.


[Click for Annealing Animation of NiC60 Cluster]

Fig. 12 Molecular dynamics simulation of precursor cluster of SWNTs


A results of further collisions of those semi-spherical clusters is shown in Fig. 13. Since the very severe time-compression in the simulation, annealing of the structure is not enough. The simulation time is 4.5 ns after Fig. 12 but it corresponds to probably a few ms in the real experiment. Even though the structure shown in Fig. 13 is rather ugly, one can see that the tubular structure has grown longer by the collision and the coalescence. Ni atoms are slowly assembling to form Ni clusters, and they are diffusing around until finding the most stable position at the hemi-half-fullerene cap area.


[Click for Animation of Growth Process]


Fig. 13 An imperfect SWNT obtained from molecular dynamics simulation


3-4. Hydrogen Storage with Carbon Nanotubes

The high hydrogen storage capacity of carbon nanotubes has been reported for single-walled carbon nanotubes (SWNTs) and metal-doped graphite nanofibers. Because of its impact in applications in fuel cell vehicles, intensive experimental and theoretical works have been performed. Some of the most striking reports are questionable and theoretical models of physical adsorption cannot explain the hydrogen storage capacity of SWNTs more than 1 wt %. Current status of preparation of carbon nanotubes, hydrogen storage experiments, and theoretical works are being examined.

Here, we use molecular dynamics simulation for the hydrogen physisorption with SWNTs [25,26]DFig. 14 shows the snapshot (Fig. 14(a)), potential energy distribution (Fig. 14(b)(c)) of a bundle of 7 (10,10) SWNTs. Here, the temperature is 77K. Potential between hydrogen-hydrogen, hydrogen-carbon, carbon-carbon are expressed with Lennard-Jones (12-6) function. For the bundle of (10,10) nanotube with diameter 1.36 nm, relatively strong absorption site appears inside the tube (endohedral sites) and between tubes (interstitial sites).Unfortunately, the hydrogen storage capacity is lower than 1 wt % at room temperature [26]. However, the absorption in the extremely small pore of SWNTs are very exciting in general absorption applications.




[Click for Animation of Phase Transformation]


Fig. 14 Molecular dynamics simulation of hydrogen storage in a SWNT bundle [26].


3-5. Molecular dynamics simulation of heat conduction of SWNTs [27-29]

The thermal conductivity of carbon nanotubes is speculated to be higher than any other material along the cylindrical axis. Hence, CNTs can be used as the superior heat conduction devices. In order to examine the thermal conductivity, we are performing molecular dynamics simulations[27-29]. Heat conduction of finite length single walled carbon nanotubes (SWNTs) is simulated by the molecular dynamics method with Tersoff-Brenner bond order potential. Temperature at each end of a SWNT was controlled by the phantom technique, and the thermal conductivity was calculated with Fourier's law from the measured temperature gradient and the energy budgets in phantom molecules. As shown in Fig. 15, the measured thermal conductivity at 300K did not converge to a finite value with increase in tube length up to 404 nm, but an interesting power law relation was observed. Probably the thermal conductivity will converge for much longer nanotubes. However, some electrical devices using SWNTs may not be much longer than 100 nm. Hence, the lower apparent thermal conductivity at the length scale must be very important. The diverging feature of thermal conductivity is discussed in the one-dimensional model cases [see 28,29].

Another purpose of this study is the preliminary connection of molecular dynamics techniques to the solid-state heat conduction usually discussed as 'phonon transport' in solid physics. In principle, the molecular dynamics simulation can be used to obtain information for phonon transport dynamics such as phonon dispersion relation, group velocity, mean free path, boundary scattering rate and the rate of phonon-phonon scattering (Umklapp process). Especially, the phonon scatting at an interface is very important issue in recent thin film technology.

The phonon density of states and photon dispersion relations are directly extracted from simulation results for further analysis of heat conduction mechanism based on the phonon concept. Fig. 16 shows calculated phonon dispersion relations and phonon density of states for (5,5) nanotube with 101 nm length. Phonon dispersion relation is obtained as the 2D Fourier transforms of deviation of each atoms from the equilibrium position. And, the phonon DOS can be calculated as the Fourier transforms of velocity fluctuations. Here, Fig. 16 (d) is the theoretical solution of the dynamical matrix by R. Saito at University of Electro-Communications.


[Click for Animation of Vibration of (10,10) Nanotube]


Fig. 15 Dependence of thermal conductivity of SWNT on tube length.



Fig. 16 Phonon dispersion relations and phonon density of states directly calculated from molecular dynamics.


3-6. Other Molecular dynamics Simulations

Several classical molecular dynamics simulations related with carbon nanotubes are being performed. They are physorption of hydrogen on SWNTs, formation of peapod (Fig. 17), structure of fullerene inside a SWNT, generation of double walled-carbon nanotube from peapod, bundle of SWNTs. For example, in the generation process of DWNT from peapod, the fusion of five C60 inside a (10,10) SWNT was simulated. Depending on the simulation temperature, the formation of polymers, peanuts type structure, and DWNTs are simulated.

Fig. 18 shows the simulation of water cluster insertion to a (10,10) SWNT. A larger water cluster is bounced in the edge of SWNT but a small cluster enters into a SWNT (see the animation). We have been studying the condensation process of argon liquid on a metal surface [30] and structure of a water cluster on platinum surface by molecular dynamics simulations [31]. The wetting, phase-change, flow, heat transfer of liquid inside the narrow SWNT pore is extremely interesting from the theoretical stand point. Furthermore, the nanoscale pore of SWNTs is very fascinating in adsorption technology.


[Click for Animation]
[Click here for Another Animation]

Fig. 17 Formation of peapod.



[Click for Animation]


Fig. 18 Water Cluster inside a (10,10) SWNT.


References

(1) http://www.photon.t.u-tokyo.ac.jp/~maruyama/nanotube-j.html.
(2) S. Maruyama, Kikan Fullerene, (1996), vol. 4, no. 4, pp. 106-113 (in Japanese).
(3) S. Maruyama, M. Y. Lee, R. E. Haufler, Y. Chai and R. E. Smalley, Z. Phys. D, (1991), vol. 19, pp. 409-412.
(4) S. Maruyama, L. R. Anderson and R. E. Smalley, Rev. Sci. Instrum., (1990), vol. 61, no. 12, pp. 3686-3693.
(5) R. E. Haufler, Y. Chai, L. P. F. Chibante, J. Conceicao, C. Jin, L.-S.Wang, S. Maruyama and R. E. Smalley, Mat. Res. Soc. Symp. Proc., (1991), vol. 206, pp. 627-638.
(6) S. Maruyama, Kagaku-Kougaku, (1992), vol. 56, no. 6, pp. 422-423 (in Japanese).
(7) S. Maruyama, Kagaku, (1997), vol. 52, no. 5, pp. 20-22 (in Japanese).
(8) S. Maruyama, Kikan Kagaku Sousetsu, (1999), vol. 43, pp. 10-19( in Japanese).
(9) S. Maruyama, Radiation Chemistry, (2002), vol. 73, pp. 22-27 (in Japanese).
(10) S. Maruyama, Carbon Nanotube, Jyohou KikouC(2002), pp. 103- 119 (in Japanese).
(11) S. Maruyama, R. Kojima, Y. Miyauchi, S. Chiashi and M. Kohno, Chem. Phys. Lett., (2002), submitted.
(12) http://www.photon.t.u-tokyo.ac.jp/~maruyama/kataura/kataura.html.
(13) S. Maruyama, M. Kohno and S. Inoue, Fullerene 2000: Chemistry and Physics of Fullerenes and Carbon Nanomaterials, (2000), pp. 309-319.
(14) S. Maruyama, Y. Yamaguchi, M. Kohno and T. Yoshida, Fullerene Sci. Tech., (1999), vol. 7, no. 4, pp. 621-639.
(15) S. Maruyama, Perspectives of Fullerene Nanotechnology, (2002), pp. 131-142.
(16) M. Kohno, S. Inoue, R. Kojima, S. Chiashi and S. Maruyama, Physica B, (2002), in press.
(17) M. Kohno, S. Inoue, A. Yabe and S. Maruyama, Micro. Themophys. Eng., (2002), submitted.
(18) Y. Yamaguchi and S. Maruyama, Chem. Phys. Lett., (1998), vol. 286, pp. 336-342.
(19) S. Maruyama and Y. Yamaguchi, Chem. Phys. Lett., (1998), vol. 286, pp. 343-349.
(20) Y. Yamaguchi and S. Maruyama, Euro. Phys. J. D, (1999), vol. 9, pp. 385-388.
(21) Y. Yamaguchi and S. Maruyama, Fullerenes: Recent Advances in the Chemistry and Physics of Fullerenes and Related Materials, (1999), vol. 7, pp. 640-646.
(22) S. Maruyama, Endofullerenes: A New Family of Carbon Clusters, (2002), in press.
(23) S. Maruyama and Y. Shibuta, Molecular Crystals and Liquid Crystals, (2002), in press.
(24) Y. Shibuta and S. Maruyama, Physica B, (2002), in press.
(25) S. Maruyama and T. Kimura, Proc. ASME Heat Transfer Division 2000, Orlando, (2000), vol. 2, pp. 405-409.
(26) S. Maruyama, Ohyo-Butsuri, (2002), vol. 71, no. 3, pp. 323-326 (in Japanese).
(27) S. Maruyama, S.-H. Choi, Therm. Sci. Eng., (2001), vol. 9, no. 3, pp. 17-24.
(28) S. Maruyama, Physica B, (2002), in press.
(29) S. Maruyama, Micro. Themophys. Eng., (2002), submitted.
(30) T. Kimura and S. Maruyama, Micro. Themophys. Eng., (2002), vol. 6, no. 1, pp. 3-13.
(31) T. Kimura and S. Maruyama, Proc. 12th Int. Heat Transfer Conf., (2002), in press.



Contact: maruyama [at] photon.t.u-tokyo.ac.jp