(664b) First-Principles Kinetic Monte Carlo Simulations of Hydrogen Spillover across the Ru/TiO2 Interface

Catalytic
upgrade of bio-oil to fuels and chemicals requires the selective reduction of
its 28-40 wt% oxygen content while avoiding undesired
C-C bond cleavage reactions. Ruthenium supported on titania
(Ru/TiO₂) has recently evolved as one of the better performing
catalysts for such hydrodeoxygenation (HDO) processes.¹ Commonly proposed HDO
mechanisms on oxide containing materials invoke oxygen vacancy sites as
catalytically active sites, but the mechanism of their formation remains
disputed.² To
address this question, we use first-principles kinetic Monte Carlo (kMC) simulations to
investigate hydrogen spillover, reduction and
hydroxylation of rutile TiO₂(110) in the presence and absence of Ru
clusters, with particular focus on the Ru/TiO₂ interface. First-principles
kMC is a multiscale modeling approach that extends
atomic-scale energetic information obtained from density functional theory
(DFT) to realistic time and length scales. A key advantage of kMC over more traditional microkinetic modeling is that it
explicitly considers correlations, fluctuations, and spatial distributions of
the chemicals on catalyst surface.³

To eliminate the strong effect of Lewis acid-base pairs
on adsorbate-adsorbate interactions as illustrated by Metiu
et al.⁴, we extract
pairwise interaction parameters from a series of co-adsorbed configurations
that do not exhibit any electronic artifacts on the periodically repeated rutile
TiO₂(110) unit cell. Our kMC results show
that the defect-free, stoichiometric surface of rutile TiO₂(110) can
hardly be reduced by hydrogen exposure. The presence of Ru nanoparticles, however,
dramatically increases the reduction rate of the TiO₂(110) surface attributed
to H spillover and a low energy, heterolytic H₂ cleavage pathway across
the Ru/TiO₂ interface as depicted
in Figure 1. We also qualitatively reproduce experimental results for water
adsorption on TiO₂(110).⁵ Notably, we show that the final
surface termination of TiO₂(110) upon exposure to H₂O/H₂
depends on the initial vacancy concentration, while the same equilibrated
surface termination is found for any initial condition when Ru nanoparticles
are present. Our findings provide new mechanistic insight regarding hydrogen
spillover and the significant effect of Ru nanoparticles on reduction and
hydroxylation behaviors. These important results build a foundation for further
investigations of the complex mechanistic interplay of oxide supported metal
catalysts, not only for hydrodeoxygenation, but also for (selective) oxidation
reactions.

Figure 1. The overall potential energy diagram for the
elementary steps that comprise different ways to reduce TiO₂. Values
shown in the figure are activation energies at the zero-coverage limit. Subscripts
represent four different active sites: ruthenium (Ru), ruthenium at the interface
(RuIF), oxygen bridge site at the interface (ObrIF), oxygen bridge site (Obr).

(1) P. M. Mortensen,
J. D. Grunwaldt, P. A. Jensen, K. G.
Knudsen, A. D. Jensen, Applied Catalysis A: General,
407 (2011) 1-19

(2) T. O. Omotoso, B. Baek, L. C. Grabow, S. P.
Crossley,
ChemCatChem, 9 (2017) 2642-2651

M. Scheffler, K. Reuter, Phys. Rev. B, 73 (2006) 045433

(3)
Metiu, H.; Chretien, S.; Hu, Z.; Li, B.; Sun, X. J. Phys. Chem. C, 116 (2012) 10439-10450

(4) Ketteler,
Guido; Yamamoto, S.; Bluhm, H.; Andersson,
Klas; Starr, D. E.; Ogletree, D. F.; Ogasawara, H.; Nilisson, A.; Sameron, M. J. Phys. Chem. C, 111 (2007) 8278-8282