7KEYWORDS: Selective CO methanation, surface area, selectivity, particle size effects, metal-8 support interactions, Ru/TiO 2 9 10 ABSTRACT: Aiming at a detailed understanding of the role of metal-support interactions in the 11 selective methanation of CO in CO 2 -rich reformate gases we have investigated the catalytic 12 performance of a set of Ru/TiO 2 catalysts with comparable Ru loading, Ru particle size and TiO 2 13 phase composition, but very different surface areas (ranging from 20 to 235 m 2 g -1 ) in this 14 reaction. The activity for CO methanation, under steady-state conditions, was found to strongly 15 depend on the TiO 2 support surface area, increasing first with increasing surface area up to a 16 maximum activity for the Ru/TiO 2 catalyst with a surface area of 121 m 2 g -1 and then decreasing 17 for an even higher surface area, while the selectivity is mainly determined by the Ru particle 18 size, which slightly decreases with increasing support surface area. This goes along with an 1 increase in selectivity for CO methanation, in agreement with a model proposed previously for 2 non-reducible supports. In situ infrared measurements further revealed that also the adsorption 3properties for these catalysts, as evidenced by the CO adsorption strength, change significantly 4 with increasing catalyst surface area and that strong metal-support interactions cause a partial 5 overgrowth of the Ru nanoparticles for the highest surface area catalyst. The interplay between 6 catalyst surface area and reaction characteristics and the important role of metal-support 7 interactions in the reaction, in addition to particle size effects, will be elucidated and discussed. 8

Ru /TiO 2 catalysts are highly active and selective in the selective methanation of CO in the presence of large amounts of CO 2 , but suffer from a considerable deactivation and loss of selectivity during time on stream. Aiming at a fundamental understanding of these processes, we have systematically investigated the physical reasons responsible for these effects, using catalysts with different surface areas and combining time resolved kinetic and in situ / operando spectroscopy measurements as well as ex situ catalyst characterization. This allowed us to identify and disentangle contributions from different effects such as structural effects, adlayer effects such as site blocking effects and changes in the chemical (surface) composition of the catalysts. Operando XANES / EXAFS measurements revealed that an initial activation phase is largely due to the reduction of oxidized Ru species, together with a distinct change in the Ru particle shape, until reaching a state dominated by metallic Ru species (fraction RuO 2 <5%) with the highest Ru mass normalized activity. The loss of activity and also of selectivity during the subsequent deactivation phase are mainly due to slow Ru particle growth (EXAFS, TEM). Surface blocking by adsorbed species such as surface formate / carbonate or surface carbon species, which are formed during the reaction, contributes little, as concluded from in situ IR, TPO and XPS data. Consequences on the selectivity for CO methanation, which decreases with time on stream for catalysts with larger surface area and for the distinct loss of adsorbed CO and surface formate species, as well as the role of the catalyst surface area in the reaction are discussed.

No abstract

The carbon formation causing deactivation during CO methanation was studied for a Ni/Al2O3 catalyst. Sulfur-free methanation at low temperature (573 K) for various lengths of time was followed by temperature-programmed hydrogenation (TPH) providing information on carbon types involved in the deactivation of the catalyst. Three main carbon hydrogenation peaks were evident from TPHs following methanation: ∼460, ∼650, and ∼775 K. It is suggested that the ∼460 K TPH peak was composed of two peaks: a surface carbide peak at 445–460 K, and a peak due to carbon dissolved into the nickel at 485 K based on CO and CH4 adsorption measurements and XRD analysis. The 650 and 775 K temperature peaks are assigned to polymerized carbon structures and the ∼775 K peak was found to be the primary cause of deactivation as judged by a linear correlation between its amount and the degree of catalyst deactivation. The longer the duration of the methanation test, the more carbon was built up on the Ni surfaces and the highest observed amount was quantified to be as much as eight carbon atoms per Ni surface atom (8 C/Nisurf), which would roughly correspond to an average coverage of four monolayers of graphene. From H2 desorption measurements after reaction the 650 K TPH peak carbon structure is proposed to be partially hydrogenated, possibly resembling polycyclic aromatic-like carbon. The 775 K peak carbon species are likely more graphene-like. Results indicate that although carbon deposition nucleation may be initiated at the most active methanation sites, i.e., the Ni step sites, subsequent growth takes place over Ni terrace sites. A strongly inhomogeneous carbon growth distribution over the Ni nanoparticle surfaces could also account for our findings. Similar to suggestions regarding catalyst deactivation in Fischer–Tropsch synthesis, a surface CH* coupling mechanism is likely taking place, and our results suggest these polymeric hydrocarbon species become more ordered, aromatic, and eventually graphene-like over time.

No abstract