(457g) CO-Induced Layer Sequence Reconstruction By CO Adsorption on Cocu Model Catalysts

Straight-chain hydrocarbons with terminal oxygen functionalization (oxygenates) are very desirable chemical products, evident from the over 6 million tons/year produced in the mega-process of hydroformylation. In the 1970s, the Institut Français du Pétrole patented a heterogeneous CoCu catalyst for the production of short-chain alcohols (i.e. oxygenates), following CO hydrogenation [1]. Recent developments in our group suggest that both short- and long-chain oxygenates can be produced at both high selectivity and yield using CoCu-based  (either CoCuNb or, respectively, CoCuMn) catalysts in Fischer-Tropsch (FT) synthesis [2-4]. To make this technology truly robust and to direct future research into straight-chain oxygenate synthesis, the fundamental role each metal plays during catalysis needs to be elucidated.  However, little is known about the interaction of Co and Cu beyond the single-atom impurities in high and low index crystal facets investigated by Ruban et al. and later Nilekar et al., respectively [5, 6]. To this end, our present work uses density functional theory (DFT) calculations to investigate the segregation tendency of a Cu monolayer on Co(0001) (henceforth “Cu/Co(0001)”) both in the presence and absence of carbon monoxide (CO) - the principle FT reactant.

            According to the Ruban et al [5] and Nilekar et al [6] studies mentioned previously, single atom Cu impurities have a moderately high tendency to segregate out of a pure Co host. In accordance with this result, our calculations show that a 1 ML equivalent of Cu in Cu/Co(0001) does indeed tend to segregate completely, which is also in agreement with atom probe tomography evidence of a core@shell structure [2]. Our density of states (DOS) analysis shows little change in the electronic structure of Cu or Co due to this alloying regardless of segregation, which implies that a segregated Cu monolayer would be largely as FT-inactive in chain lengthening as pure Cu. This conclusion is bore out by the essentially unchanged CO adsorption energy on Cu in Cu/Co(0001) as compared to that on Cu(111). Conversely, CO’s adsorption on a surface Co in Cu/Co(0001) does in fact change – it is weakened. We then show that CO adsorption causes a reversal in segregation wherein Co is effectively “pumped” to the surface by the presence of the adsorbed CO. This phenomenon was tested as a function of CO coverage, and amongst the coverages and surface Co enrichments studied, a global energy minimum is reached when 0.25 ML CO is adsorbed on a surface enriched in 25% Co. Furthermore, high CO coverages (0.75 ML - 1.00 ML) can induce 100% surface Co enrichment: a complete inversion of the CoCu layer sequence. Furthermore, since Cu is FT inactive in chain lengthening, we can conclude that this surface Co is critical to the performance of CoCu-based catalysts.

            The next challenge facing the design of FT catalysts for the production of straight-chain oxygenates is the question of CO dissociation as evidenced by X-ray photoelectron spectroscopy in our group [3]. Regardless of the mechanism responsible for chain lengthening the C-O bond must be cleaved to produce CH_x and OH_y species (x=0,1,2,3; y=0,1), which we consider likely intermediates to chain lengthening. However, as our above DOS results have shown, the Co in the bimetallic catalyst is electronically similar to its pure metal counterpart, and its CO adsorption is actually weakened. Therefore, CO dissociation reaction energies on pure Co are likely a lower limit for those on CoCu catalysts. Still, CO dissociation on CoCu catalysts is not a straightforward process and needs specific surface atom arrangements and geometries to be taken into account. Work by Ge and Neurock showed that CO dissociation has prohibitively high reaction barriers on flat pure Co surfaces, and that the reaction is only truly feasibly on stepped and kinked pure Co surfaces [7]. In accordance with these results, our work has shown that CO has a reaction barrier of more than 2 eV on flat Cu/Co(0001), but we then show that that CO dissociation on stepped  is still very endothermic even when a modest concentration of surface Co is present. Cu essentially poisons the most reactive surface sites. Present efforts are devoted to design the surface configurations necessary to induce CO dissociation on such stepped surfaces. 

1.         Sugier, A. and E. Freund, Process for manufacturing alcohols, particularly linear saturated primary alcohols, from synthesis gas. 1978, US 4122,110.

2.         Xiang, Y., et al., Long-Chain Terminal Alcohols through Catalytic CO Hydrogenation. Journal of the American Chemical Society, 2013. 135(19): p. 7114-7117.

3.         Xiang, Y., R. Barbosa, and N. Kruse, Higher Alcohols through CO Hydrogenation over CoCu Catalysts: Influence of Precursor Activation. ACS Catalysis, 2014. 4(8): p. 2792-2800.

4.         Xiang, Y., et al., Ternary Cobalt–Copper–Niobium Catalysts for the Selective CO Hydrogenation to Higher Alcohols. ACS Catalysis, 2015. 5(5): p. 2929-2934.

5.         Ruban, A.V., H.L. Skriver, and J.K. Nørskov, Surface segregation energies in transition-metal alloys. Physical Review B, 1999. 59(24): p. 15990-16000.

6.         Nilekar, A.U., A.V. Ruban, and M. Mavrikakis, Surface segregation energies in low-index open surfaces of bimetallic transition metal alloys. Surface Science, 2009. 603(1): p. 91-96.

7.         Ge, Q. and M. Neurock, Adsorption and Activation of CO over Flat and Stepped Co Surfaces:  A First Principles Analysis. The Journal of Physical Chemistry B, 2006. 110(31): p. 15368-15380.