Theoretical studies on structural properties of Iron Oxides

Abstract

Iron oxides are one of the most vulnerable and stable materials utilized in many kinds of applications such as pigments, reducing agents, catalysts, and so on. Beyond such an importance of iron oxides, their complicated phase relations often make it difficult to identify their structural characteristics. Particularly iron oxides formed at the nano-sized interfacial regime or under extremely high temperature and pressure conditions near the core of the earth can hardly identified using current experimental methods. Thus development of correct computational simulation methods for studying iron oxides has its own importance. Here, we carried out reactive force field (reaxFF) simulations and quantum mechanical density functional theory (DFT) simulations for iron oxide systems. We note that reaxFF can describe the bond dissociations and formations with inexpensive computational cost but with reliable accuracies using the well-developed force field parameters based on DFT energetics; however, DFT methods are known to have critical limitations in describing oxide systems due to the strong electron correlation that is usually incorrectly handled within DFT framework due to the self-interaction error (SIE). First, we carried out reaxFF simulations to model the oxidation process of bare iron under ambient conditions, and investigated the oxidation mechanism and final oxide structures. By comparing reaxFF results with DFT energetics, we found that previous reaxFF parameters by M. Aryanpour, et al., [Journal of Physical Chemistry A, 2010. 114(21), 6298-6307] cannot describe the correct energetics among B1, B2, B3, and B4 phases of FeO. More interestingly, we further found that even the most widely using DFT method based on generalized gradient approximation (GGA), PBE, and its Hubbard correction method for oxide systems (PBE+U) have problems in investigating polymorphic phases of FeO. Using hybrid functional method where Hartree-Fock Hamiltonian is mixed into DFT formalis...