Elsevier

Surface Science

Volume 590, Issues 2–3, 1 October 2005, Pages 127-137
Surface Science

Methane activation on Ni(1 1 1): Effects of poisons and step defects

https://doi.org/10.1016/j.susc.2005.05.057Get rights and content

Abstract

We have, theoretically and experimentally, investigated the dissociation of methane on the terraces and steps of a Ni(1 1 1) surface. Using Density Functional Theory (DFT) total energy calculations combined with Ultra High Vacuum (UHV) experiments, we find that the steps exhibit a higher activity than the terraces. We have, furthermore, investigated how carbon and sulfur present on the surface will deactivate the steps, leaving only the terraces active. We find the intrinsic sticking probabilities of methane on the steps and terraces at 500 K to be 2.8 × 10−7 for the steps and 2.1 × 10−9 for the terraces, in complete agreement with our calculated difference in activation energy of 17 kJ/mol.

Introduction

Methane is the dominating component of natural gas and used to synthesize CO and hydrogen, via the steam reforming process. This is a way to produce cheap hydrogen, and it is the first step in both ammonia and methanol synthesis. The catalyst used for the steam reforming process is nickel based, and hence, methane activation and C–H bond breaking on nickel has received much attention, both theoretically and experimentally [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11]. The majority of the work has been concentrated on surfaces with the lowest Miller indices, and minor attention has been focused on the possible effects of steps or defects on these surfaces. Experimentally methane activation is found to be structure sensitive, with Ni(1 1 1) being the least active and Ni(1 0 0) and Ni(1 1 0) increasingly more active [7], and it is possible that steps could exhibit an even higher activity.

According to Density Functional Theory (DFT) studies performed by Watwe et al., the dissociation of methane on Ni(1 1 1) is activated by 127 kJ/mol [4], but they did not take spin-polarization of the electrons into account, and therefore a slight energy discrepancy of the intermediates were expected. A subsequent DFT study by Bengaard et al. investigated the effect of larger lateral unit cells and magnetic ordering, and found that it had a significant influence on the overall binding and transition state energies. They reported an activation energy on Ni(1 1 1) below 91 kJ/mol [12]. Other theoretical studies by Yang and Whitten, using a configuration interaction approach, have found an activation energy as low as 71 kJ/mol. Bengaard et al. furthermore studied the effect of steps, using a Ni(2 1 1) surface, and found the activation energy to be lowered by 19 kJ/mol relative to Ni(1 1 1).

In the present paper we investigate the methane activation at steps further. By combining theory and experiment, we find that steps are more reactive than the closed-packed (1 1 1) terraces of a Ni surface. Furthermore, we show that both carbon and sulfur prefer the high coordinated step sites and will be present there during our experiments. We demonstrate that both carbon and sulfur present along the steps strongly affects the activation of the first C–H bond in methane, and that beyond a certain carbon or sulfur coverage the dissociation of methane will be dominated entirely by the terrace activity. The suggestion by Bengaard et al. [12], confirmed experimentally by Helveg et al. [13], that step edges on nickel surfaces act as growth centers for graphene is in perfect agreement with our findings.

Section snippets

Theoretical method

We have used the DACAPO code [14] for all total energy calculations. To investigate the influence of steps on Ni(1 1 1) we use a Ni(2 1 1) surface. This system is modeled by a nine layer slab periodically repeated in a super cell geometry. Each metal slab is separated by roughly 11 Å of vacuum. The thermodynamic total energy calculations are done in a 1 × 2 unit cell of Ni(2 1 1). This produces a surface with (1 1 1) terraces three atoms wide and steps with a (1 0 0) symmetry. A single methane molecule

Experimental

Contrary to what one might think, the difficult part, when measuring the importance of steps for a chemical reaction, is not measuring the activity of the steps, but measuring the activity of the flat terraces. It is not possible to create a single crystal surface without steps or defects. A standard single crystal, aligned to within 0.5°, will have about 1% steps or defects on the surface. To be able to measure the activity of the terraces the steps need to be deactivated. One approach to this

Theoretical results

The activation of the first C–H bond in methane, shown in Fig. 2, occurs over the top of a surface Ni atom on Ni(1 1 1). We obtain an energy barrier of 105 kJ/mol. This is in accordance with reported results by Bengaard et al. [12], in which a non zero-point corrected energy barrier of 101 kJ/mol was found. A similar calculation performed on Ni(2 1 1) resulted in a lowering of the activation barrier of 17 kJ/mol. The configuration in which CH3 is bonded to the bridge step edge site and H to the

Discussion

STM studies of sulfur adsorbed on Ni(1 1 1) by Maurice et al. [34] find that above 625 K sulfur will migrate to the steps, where sulfur islands will nucleate. This is in full agreement with our calculations, which predict sulfur to be 35 kJ/mol more stable on Ni(2 1 1) than on Ni(1 1 1). This provides a strong indication of a preferential decoration of the steps, given a temperature high enough for sulfur to be able to migrate to the steps. The strong effect, we observe upon the adsorption of up to 0.06

Conclusion

We have calculated the barriers for methane dissociation along a step on Ni(1 1 1) in the presence of carbon and sulfur, and found that both adsorbates will deactivate the steps. We have verified this experimentally, used this knowledge to measure the activity of a Ni(1 1 1) surface free from steps, and compared the activity of steps and terraces on Ni(1 1 1). The intrinsic sticking probabilities we find at 500 K are 2.1 × 10−9 for the terraces and 2.8 × 10−7 for the steps. This corresponds to an

Acknowledgements

We gratefully acknowledge financial support from the Danish Research Agency through the Centre of Excellence “Towards a Hydrogen-based Society” grant no. 2052-01-0054.

The Center for Atomic-scale Materials Physics (CAMP) is sponsored by the Danish National Research Foundation.

References (39)

  • H.S. Bengaard et al.

    J. Catal.

    (1999)
  • R. Watwe et al.

    J. Catal.

    (2000)
  • L. Hanley et al.

    Surf. Sci. Lett.

    (1991)
  • H.S. Bengaard et al.

    J. Catal.

    (2002)
  • G. Mills et al.

    Surf. Sci.

    (1995)
  • I. Chorkendorff et al.

    Surf. Sci.

    (1990)
  • M. Perdereau et al.

    Surf. Sci.

    (1970)
  • M. Valden et al.

    Chem. Phys. Lett.

    (1996)
  • J.B. Benziger et al.

    Surf. Sci.

    (1984)
  • G. Held et al.

    Surf. Sci.

    (1998)
  • H. Nakano et al.

    Surf. Sci.

    (2002)
  • C. Klink et al.

    Surf. Sci.

    (1995)
  • H. Robota et al.

    Surf. Sci.

    (1984)
  • B. Hammer et al.

    Surf. Sci.

    (1995)
  • J.J. Mortensen et al.

    Surf. Sci.

    (1998)
  • V. Pallassana et al.

    J. Catal.

    (2000)
  • R.C. Egeberg et al.

    Surf. Sci.

    (2002)
  • V. Maurice et al.

    Surf. Sci.

    (1997)
  • J.R. Rostrup-Nielsen

    J. Catal.

    (1984)
  • Cited by (225)

    • A review on bi/polymetallic catalysts for steam methane reforming

      2023, International Journal of Hydrogen Energy
    View all citing articles on Scopus
    View full text