546732 CO and H2 Adsorption on Co(11-20) - a Combined Experimental and Theoretical Investigation

Wednesday, June 5, 2019: 11:18 AM
Texas Ballroom EF (Grand Hyatt San Antonio)
Hilde Venvik1, Marie Døvre Strømsheim2, Mari Helene Farstad2,3, Ingeborg-Helene Svenum4, Mehdi Mahmoodinia2, Kees-Jan Weststrate5 and Anne Borg6, (1)Chemical Engineering, Norwegian University of Science and Technology, Trondheim, Norway, (2)Chemical Engineering, NTNU - Norwegian University of Science and Technology, Trondheim, Norway, (3)Jotun ASA, Sandefjord, Norway, (4)SINTEF Industry, Trondheim, Norway, (5)SynCat@DIFFER, Syngaschem BV, Eindhoven, Netherlands, (6)Physics, NTNU - Norwegian University of Science and Technology, Trondheim, Norway

Understanding the adsorption on and restructuring of cobalt (Co) surfaces is important towards controlling the phenomena occurring on Co-based Fischer-Tropsch catalysts. The surface of Co(11-20) is known to attain a (3x1) surface reconstruction under CO exposure [1], involving the anisotropic migration of Co atoms along [0001] to and from the step edges [2]. Co(11-20) was therefore chosen as a model system for the surface dynamics of a Fischer-Tropsch catalyst upon adsorption of CO and investigated with a combination of experimental and theoretical methods. The study is supplemented with investigations of hydrogen adsorption on different terminations of hcp cobalt to elucidate the nature of the Co(11-20) surface.

The CO-induced surface restructuring was studied with scanning tunnelling microscopy (STM), low energy electron diffraction (LEED), temperature programmed desorption (TPD) and density functional theory (DFT). The calculations were performed with the Vienna ab initio simulations package (VASP) [3]. Two theoretical model surfaces with either a missing row (MR) or an added row (AR) of Co atoms along [0001] were constructed for the CO/Co(11-20) system to represent alternatives for the (3x1) reconstruction. These were compared to (3x1) configurations of CO adsorbed on the unreconstructed surface (Figure 1). Transition state calculations with climbing image nudged elastic band (CI-NEB) [4] were performed to investigate the removal of a Co atom from the topmost layer and its mobility across the surface. In the case of (atomic) hydrogen, adsorption properties and coverage effects on Co(11-20) were studied by TPD and DFT and compared to results obtained for Co(0001) and Co(10-12).


Figure 1: the preferred adsorption geometries of (a) CO on unreconstructed (b) MR and (c) AR Co(11-20) surface.

It was inferred from LEED that, whereas the CO-induced reconstruction is facile at room temperature (RT), it may be kinetically hindered during CO exposure of clean Co(11-20) at low temperature. Previously, we have also reported that deposition of minor amounts of potassium (K) inhibits the rearrangement of CO atoms from the step edges, leading to a considerably slower and more disordered restructuring under CO [5]. Comparison of TPD data for the two cases shows significant differences in the range 200-350K, and somewhat higher temperature and coverage for the main desorption peak for the case where CO exposure was performed at RT before cooling in CO and subsequent TPD. The DFT calculations show that CO adsorption in coordination with the topmost layer of Co atoms is preferred, as this resulted in the most favourable adsorption energies irrespective of the Co(11-20) (3x1) model system configuration applied. The added row (AR) (3x1) reconstruction model yields the highest calculated adsorption energies for all coverages investigated so far, from 1 to 4 CO per (3x2) surface unit cell. Albeit small differences in adsorption energy, the unreconstructed surface remains the least stable. Moreover, surface energy values inferred increasing preference for the added row structure with increasing CO coverage. Jointly the computations and the experimental data suggest that if Co mobility is kinetically facile, the Co surface can rearrange to accommodate more and more strongly bound CO.

Co migration was investigated by comparing different initial and final configurations of a displaced single Co atom and an adsorbed CO molecule, and by obtaining the energy barrier for vacancy formation and displacement along the [0001] direction. The energy differences are not large, but imply that Co migration with CO attached is preferred over Co displacing alone, supporting the presumption that the migrating species is cobalt carbonyl. The relatively large barriers associated with the vacancy formation and migration further supports the notion of the restructuring being kinetically controlled.

From hydrogen TPD, a significantly lower desorption temperature was obtained for Co(11-20) than for Co(0001) and Co(10-12). In addition, a lower saturation coverage (H/nm2) was found. The DFT calculations support these findings, in terms of both total and differential adsorption energies. However, there is a discrepancy in the maximum anticipated H coverage in DFT compared to TPD experiments, since our DFT calculations predict almost doubled saturation coverages for all three surfaces. Previous work by van Helden et al. [6] already explored this for H atoms on Co(0001). By experimentally comparing flat and stepped Co(0001) they could explain the difference in saturation coverage by kinetic effects; i.e. that pre-adsorbed hydrogen kinetically inhibits the adsorption as higher coverages are reached. The results support the notion that, although H2 dissociation is facile on more open surfaces, atomic hydrogen binds stronger in (fcc) three-fold hollow sites. On Co(11-20), coordination to both top and second layer Co atoms is required for the preferred adsorption sites, and this weakens the binding. It still induces the same kinetic inhibition of H2 dissociation.

In context of Fischer-Tropsch synthesis under industrially relevant conditions, the results obtained may point to cobalt restructuring as important for the activation of CO, and coverage effects as important for the activation of H2. Defects such as step edges affect the adsorption of both CO and hydrogen.

References:

[1]   H. Papp, Surf. Sci. 149 (1985) 460–470.

[2]   H.J. Venvik, A. Borg, C. Berg, Surf. Sci. 397 (1998) 322–332.

[3]   G. Kresse, J. Hafner, Phys. Rev. B. 47 (1993) 558–561.

[4]   G. Henkelman, B.P. Uberuaga, H. Jónsson, J. Chem. Phys. 113 (2000) 9901–9904.

[5]   M. D. Strømsheim, I.-H. Svenum, M. H. Farstad, Z. Li, H. J. Venvik, Catal. Today 299 (2018) 37-46.

[6]   P. van Helden, J.-A. van den Berg, and C. J. Weststrate, ACS Catal. 2 (2012) 1097−1107


Extended Abstract: File Not Uploaded
See more of this Session: CO Activation I
See more of this Group/Topical: General Submissions