(223q) Development of Reaxff Simulation Techniques for Assessing in Situ Phase Behavior of Pd-Based Catalysts

A detailed understanding of the structure of a catalyst’s surface as a function of varying reaction conditions is essential for optimizing active and selective catalysts. Surface phase reconstruction under operating conditions is a particular concern for Pd-based catalysts, which are widely employed in industrial-scale oxidation and hydrogenation applications. Pd readily forms surface and bulk oxide phases when employed under oxidizing conditions, and can form complex hydride/carbide phases under a hydrocarbon-rich reactant phase. Density functional theory (DFT) can assess phase stability through the formalism of ab initio thermodynamics; however, the computational expense of DFT limits such investigations to idealized surfaced models. This has motivated the use of computationally inexpensive classical-force field methods that can reach larger size scales. This allows for more complex surface models that can better represent the catalytic behavior of the system. We have developed a Pd/O/C/H interaction potential for the ReaxFF reactive force field by fitting the interaction parameters of the potential to an extensive set of DFT data encompassing bulk and surface properties. Additionally, we developed a ReaxFF-based hybrid grand canonical Monte Carlo/molecular dynamics (GC-MC/MD) approach to assess the thermodynamic stability of oxide, hydride, and carbide phases in Pd clusters. Using this simulation method, we derive theoretical phase diagrams for Pd clusters under reactive gas phase environments, which demonstrate agreement with experimental results in the literature. This poster will present the detailed force-field development and the algorithmic approach using in the hybrid MC/MD method. The GC-MC/MD method is novel to ReaxFF, and is well suited to systems where phase restructuring influences catalytic activity, selectivity, and stability.