A heterogeneous catalytic process can be divided into three steps: (i) adsorption and/or diffusion of the reactants from the surrounding bulk phase to the active site, (ii) catalytic conversion at the active site, and (iii) desorption and/or diffusion of products from the active side to the surrounding bulk phase. Depending on the catalytic process, active sites may be located at the surface or inside the pores of the catalyst and the surrounding bulk phase may be either a gas, liquid, or supercritical fluid. The overall rate and selectivity of a particular reaction is the result of a delicate balance of these three steps.
· Mass transport processes: Large-scale extended dynamical simulations are crucial to understand and evaluate the mass transport processes that precede and follow catalytic conversion. Particle-based simulations will be used to investigate diffusion of reactants and products, free energy barriers that govern the rate of adsorption and desorption processes of reactants and products onto the catalytic particle, diffusion of reactants and products on the surface or inside pores of the catalyst, and excess free energies that control the adsorption equilibria. The main challenge with classical particle-based simulation methods lies in the description of phenomena that involve significant changes in chemical structure.
· Accurate and efficient potentials: Catalytic conversion at the active site requires an accurate and efficient modeling for the bond breaking and forming process. Electronic structure methods with high performance-per-price are highly desired. Moreover, the presence of complex catalytic environments is likely to perturb the reaction process at the active site, and a realistic description for such a perturbation demands accurate and efficient modeling of long-ranged electrostatic forces.
· Reaction pathways: A great challenge for characterizing chemical reactivity in the condensed phase lies in the determination of reaction pathways. Using transition state theory, reaction rate constants can be expressed in terms of the free energy of activation or alternatively the potential of mean force along a reaction coordinate (PMF). When tunneling is important, as for acid-base chemistry in zeolites, condensed-phase tunneling approximations may also be based on the PMF. The maximum of the PMF along the reaction coordinate enters the TST rate expression as the location of the dynamical bottleneck. Previous studies of multidimensional PMF surfaces have generally involved small numbers of coordinates that actively participate in the reactions.
We will develop an array of methods for multi-time-scale simulation. For potential energy functions, we will examine a new adaptive scheme with effective pair potentials, many-body tight-binding theory, and new kinds of QM/MM methods employing the second moment approximation to tight binding as the way to join the quantal and classical regions. For dynamics calculations, we will develop a molecular dynamics code with the following convenient features: nonlocalized active site (the subsystem that needs a reactive potential function need not be localized; it evolves with the transport process), efficient algorithms for sampling and long-range electrostatics, reaction pathway search in the condensed-phase, and quantum mechanical tunneling. The programs will be general enough to apply to a variety of problems, for example, catalysis in zeolites, catalysis on surfaces of metals and metal oxides, catalysis by and on nanoparticles, and partitioning and reactivity of aqueous solutions containing electrolytes at metal/water interfaces.
Updated: Feb. 18, 2008