Titania Supported Ru Nanoclusters as Catalysts for Hydrodeoxygenation of Pyrolysis Oils

This study evaluates: (1) Ru-containing catalysts for low temperature hydrogenation of acetic acid in aqueous medium to simulate hydrogenating bio-derived oils, and (2) the development of the role of ruthenium–titania (Ru–TiO 2 ) catalytic interaction for this system. A series of 9 catalysts was screened, and a comparison of selectivity versus conversion indicated that selectivity was a function of conversion, with ethanol being the predominant product at low conversion, and light gases being favored at higher conversion by secondary reactions. Exponential trend curves for ethanol (decay) and methane (increase) with conversion were fitted. On a per gram catalyst basis, the most active catalyst under study in the temperature range of 120° through 220 °C with a hydrogen partial pressure of 1000 psi (7.0 MPa) was 5 % Ru on carbon; however, its selectivity for the conversion to ethanol was exceptionally low (5 % selectivity at ~70 % conversion and 180 °C) with the primary products being ethane and methane. This catalyst formulation displayed a negative deviation from the trend curve for ethanol and a higher methane selectivity deviation. On the other hand, a catalyst of Ru prepared by atomic layer deposition (i.e., Ru(ALD)/Ti(ALD)/Nb/Si) was also highly active (~55 % conversion at 180 °C) but displayed a significant positive deviation from the trend curve (~40 % selectivity). These results combined with those of EXAFS suggest that the interface between the deposited Ru and the titania support may be responsible for the increase in selectivity to ethanol. In general, the catalysts prepared by ALD were more active on a per gram catalyst basis than the catalysts prepared by standard aqueous impregnation. Samples of catalyst that were observed using transmission electron microscopy confirmed that the Ru was well dispersed in that no Ru nanoparticle morphology was observed within the resolving power of the JEOL JEM-3010 TEM instrument. Regarding the nanostructure of the support, TEM measurements revealed that the ALD method resulted in support domain sizes that were significantly smaller (10 nm), promoting defect formation. EXAFS characterization indicated that the best ALD catalyst (i.e., Ru(ALD)/Ti(ALD)/Nb/Si) had higher dispersion (i.e., smaller nanoparticles and thus greater metal-support interface) than the reference catalysts prepared by aqueous impregnation. A Ru-O support contribution was required in order to obt