Elsevier

Carbon

Volume 39, Issue 4, April 2001, Pages 547-558
Carbon

Synergism of Co and Mo in the catalytic production of single-wall carbon nanotubes by decomposition of CO

https://doi.org/10.1016/S0008-6223(00)00173-1Get rights and content

Abstract

The catalyst composition and operating conditions for the synthesis of single-wall carbon nanotubes (SWNT) from CO decomposition have been systematically varied in order to maximize the selectivity towards SWNT. A simple quantification method based on the standard Temperature Programmed Oxidation (TPO) technique has allowed us to determine the distribution of the different forms of carbonaceous deposits present on the catalysts after the CO decomposition reaction. A synergistic effect between Co and Mo has been observed. When both metals are simultaneously present, particularly when Mo is in excess, the catalyst is very effective. However, when they are separated they are either inactive (Mo alone) or unselective (Co alone). To understand this synergistic effect, X-ray absorption spectroscopy (EXAFS and XANES) has been used to characterize the state of Co and Mo on the catalysts before and after the production of SWNT.

Introduction

The fascinating properties of single-wall carbon nanotubes (SWNT) have opened a great number of potential applications for these unique materials. However, the high costs of the current production methods and the difficulty in making them available for large-scale manufacture have slowed down the process of bringing nanotube-based technologies to commercial practice. Following the original arc-discharge method [1], [2], other synthesis methods have been investigated, including laser ablation [3], [4] and plasma discharge. To develop a cost-effective operation for the manufacture of SWNT a drastic change in the production scale is necessary. The catalytic method for production of nanotubes has been known for a long time, but this method typically results in production of multi-wall carbon nanotubes (MWNT) or carbon nanofibers [5]. In recent years, several researchers have focused their attention to the production of SWNT by catalytic decomposition of carbon-containing molecules [6], [7], [8], [9], [10], [11], [12]. This method is sometimes referred to as chemical vapor deposition [12]. We have recently pointed out that the catalytic decomposition method is suitable for scaling up and for achieving a ‘controlled production’ of single-wall nanotubes (SWNT) [13]. By this term we imply the ability to control the selectivity towards SWNT by changing catalyst parameters and operating conditions, combined with the ability to obtain a reliable quantitative measurement of the amount of SWNT produced.

In our previous report [13], we described a simple quantification method of the SWNT produced by catalytic decomposition of CO, which allowed us to conduct a systematic screening of catalyst formulations. This quantification method was based on the standard Temperature Programmed Oxidation (TPO) technique, commonly employed in catalysis research to quantify and investigate the nature of carbonaceous deposits on spent catalysts. In this method, a continuous flow of an oxygen-containing gas is passed over the catalyst containing the carbon deposits while the temperature is linearly increased. Quantification of the CO2 evolved gives a direct measurement of the amount of carbon that gets oxidized at each temperature.

One has to be careful in using TPO to identify carbon species since the oxidation of carbon is a catalytic process, and as such, the peak positions may be affected by the metallic content as well as the amount of carbon on the surface. However, this method appears particularly suitable for the quantification of SWNT in a set of samples of similar composition and comparable carbon contents. In such set of samples, SWNT are oxidized in a relatively narrow temperature range, which lies above the temperature of oxidation of amorphous carbon and below the oxidation of MWNT and graphitic carbon. The position of the oxidation peak corresponding to the SWNT produced on the catalyst was exactly matched with that of pure SWNT obtained from Tubes@Rice (Rice University), mechanically mixed with the same catalyst powder used in the production of nanotubes. In a separate study [14] we have shown that the peak centered in the range 500–530°C corresponds to a first order process, while that at higher temperatures is a zeroth order process. This reaction orders support the assignment of those peaks to SWNT and MWNT, respectively. The oxidation of a single-shell carbon is expected to behave as a first order process, while when there exist a number of layers, and the burning is layer by layer, the process should be of zeroth order [15]. Also, it is well known that the position of oxidation peaks in TPO does not vary with carbon content for first order oxidations, but it shifts to higher temperatures for zeroth order oxidations. Therefore, we would expect that the carbon content will not affect the position of the peak corresponding to SWNT, but it would certainly affect that corresponding to MWNT. This has in fact been verified experimentally [14].

By using this methodology, we demonstrated that Co is not selective for the production of SWNT and Mo is inactive in the temperature range investigated, i.e., 600–800°C. However, bimetallic Co–Mo catalysts with low Co:Mo ratios exhibited a real synergism that resulted in high selectivity towards SWNT. Previous reports [16] have shown improved performances of bimetallic catalysts in comparison to monometallic, but none of the previously investigated systems showed the dramatic synergism exhibited by Co and Mo in the decomposition of CO.

In that case, the reaction temperature, CO concentration, and reaction time were kept constant at 700°C, 50% CO in He, and 30 min, respectively. In this contribution, we have employed the same methodology to investigate the effects of varying the operating parameters in order to optimize the production of SWNT. In addition, we have characterized the state of Co and Mo on the catalysts before and after the production of SWNT by X-ray absorption spectroscopy (EXAFS and XANES) in an effort to understand the synergism exhibited by these two metals.

Section snippets

Experimental

The Co, Mo, and Co–Mo catalysts were prepared by impregnating a silica gel support (SiO2, from Aldrich, 70–230 mesh, average pore diameter 6 nm, BET surface area 480 m2 g−1, pore volume 0.75 cm3 g−1), with aqueous solutions of Cobalt Nitrate and/or Ammonium Heptamolybdate to obtain the bi- or monometallic catalysts. The liquid/solid ratio was kept at incipient-wetness conditions, which for this support corresponds to 0.63 cm3 g−1. In all cases, the total metal content was kept at 6 wt%. The

Results

One of the objectives of this work was to investigate the influence of the various operating parameters on the carbon yield and selectivity to SWNT. Here we must point out the way that yield is defined in the gas phase catalytic method, which is different from that normally employed in the other methods of producing carbon nanotubes. We define yield as the ratio of the weight of SWNT deposited per total catalyst weight. Similarly, selectivity is defined as the fraction of carbon products

Discussion

In the first place, it is interesting to note that the growth of carbon nanotubes occurs for a limited period of time, after which the catalyst deactivates and the growth slows down. This was clearly observed when the growth was monitored by TPO as a function of time. During the first few min, the amount of carbon deposited on the catalyst increased very rapidly, but after about 10 min, the growth was much slower. This deactivation phenomenon may explain the inverse relationship between carbon

Conclusions

The present contribution has demonstrated a number of interesting characteristics of the formation of SWNT by decomposition of CO on Co–Mo catalysts. The main conclusions of this work are the following:

  • •

    A synergistic effect between Co and Mo has been shown. When both metals are simultaneously present, the catalyst is very selective. When they are separate they are either inactive (Mo alone) or unselective (Co alone).

  • •

    X-ray absorption spectroscopy (EXAFS and XANES) has shown that at the beginning

Acknowledgments

This research was conducted with financial support from the Oklahoma Center for Advancement of Science and Technology (OCAST), Conoco Inc., and the College of Engineering at the University of Oklahoma. The Raman spectra were obtained at the National Center of Catalysis (CENACA), Santa Fe, Argentina. One of us (A.B.) acknowledges support from Antorchas Foundation and the Fulbright Scholar Program.

References (29)

  • R.T.K. Baker et al.

    J Catal

    (1972)
  • J. Kong et al.

    Chem Phys Left

    (1998)
  • J.H. Hafner et al.

    Chem Phys Lett

    (1998)
  • H. Dai et al.

    Chem Phys Lett

    (1996)
  • H.M. Cheng et al.

    Chem Phys Lett

    (1998)
  • B. Kitiyanan et al.

    Chem Phys Lett

    (2000)
  • C.A. Querini et al.

    Appl Catalysis A

    (1994)
  • E. Flahaut et al.

    Chem Phys Lett

    (1999)
  • B.C. Satishkumar et al.

    Chem Phys Lett

    (1998)
  • J. Kastner et al.

    Chem Phys Lett

    (1994)
  • R.S. Weber

    J Catal

    (1995)
  • M. de Boer et al.

    Solid State Ionics

    (1993)
  • R.G. Leliveld et al.

    J Catal

    (1997)
  • J.J.C. Geerlings et al.

    Appl Catal

    (1999)
  • Cited by (0)

    View full text