Development of a Reactive Force Field for Simulations on the Catalytic Conversion of C/H/O Molecules on Cu-Metal and Cu-Oxide Surfaces and Application to Cu/CuO-Based Chemical Looping

Abstract

A fully interactive Cu/C/H/O reactive force field (ReaxFF) was developed for the Cu-metal surface catalysis system following three steps: (1) re-optimization of the Cu force field by an extended training set including additional Cu cluster properties, (2) combination of this re-optimized Cu force field and an existing C/H/O force field, and (3) fitting of interactions between Cu and C/H/O with extensive DFT data involving the various binding energies and elementary reaction steps. In addition, we developed an algorithm to search the transition state (TS) of elementary steps, which is the first TS searching program based on the ReaxFF theory framework, and a new algorithm is proposed to create reaction paths and coordinate scans with highdegrees of freedom. The comparison of results of DFT and ReaxFF indicate that the developed force field is capable of describing the properties related to reactive interactions between the Cu metal and C/H/O molecules. To test this developed Cu/C/H/O force field, a series of molecular dynamics simulations were performed. In Cu/C/H/O surface simulation, elementary C/H/O reactions involving H transfer and H 2 /CHO dissociations were observed supporting the complex C/H/O interactions on a Cu surface.In addition, two Cu/CHO example cases relevant to the chemical looping combustion process were also simulated: metal oxide (CuO) generation from reactions of metallic Cu with glucose and hydrocarbon fuel oxidation using a copper oxide as the oxidizer. Our simulation results indicate that the current Cu/C/H/O ReaxFF is able to capture the reaction details and distinguish the redox performances of different fuels.

Plasma Oxidation of Copper: Molecular Dynamics Study with Neural Network Potentials

“…This procedure requires a starting energy engine to generate the very first data set. For the present study, a ReaxFF parametrization reported in the literature 57 was chosen. A slight reparametrization of the inner-wall repulsion terms was necessary to stabilize the training set generation MD simulations, as the original potential was used for thermal surface catalysis conditions.…”

On the Origin of Nonclassical Ripples in Draped Graphene Nanosheets: Implications for Straintronics

“…On the other hand, none of the widely used classical force fields�which are generally suitable for simulating large systems�such as AMBER, 39 CHARMM, 40 GROMOS, 41 or OPLS, 42 can be used to correctly reproduce the mechanical properties of graphene while simultaneously satisfying chemically justified carbon−copper interactions. Fortunately, a recently developed Cu−C ReaxFF parameter set 43 can be used to model a graphene sheet at a copper step edge, as it not only includes the parameters for interactions between carbon atoms that correctly describe the mechanical properties of graphene but has also been shown to accurately model experimentally observed draping of a suspended graphene ribbon at a copper step. 44 Thus, we used the ReaxFF method with a recently updated version of this parameter set to further stabilize the carbon−copper interactions to perform atomistic simulations and understand the mechanical properties of our system.…”

Atomistic-Scale Simulations on Graphene Bending Near a Copper Surface

“…The ReaxFF carbon parameters presented in the CHO-2016 force field [21] were developed to obtain better mechanical properties for graphene. The CHON-2019 force field [26] proposed for carbonization simulations incorporates the carbon parameters proposed by Srinivasan [21], and these Srinivasan parameters were also included in the recent extension of ReaxFF to copper interfaces with C/H/O species [27]. Unfortunately, the older C/H/O parameter sets-like the Chenoweth 2008 C/H/O combustion force field [25]-are still used in ReaxFF simulations on graphitic materials, although they were not trained for this.…”

“…To test the possible effect of a copper surface on the graphene conformation, we use the Zhu 2020 ReaxFF parameters set [27], which includes the Srinivasan 2015 carbon parameters [21]. We placed the graphene sheet initially simulated using an NPT ensemble without a copper surface (Figure 1a) at the top of the copper surface (Figure 3a).…”