Mechanical, Optical, Transport, and Catalytic Properties of Iron Oxides from First Principles

Abstract

Iron oxides have been known to human beings for a long time, in the form of hematite ores, lodestones, or rusts. They are still subjects of current research interest. For example, wüstite (FeO), magnetite (Fe3O4), and hematite (α-Fe2O3, "α-" is omitted henceforth) are products of steel corrosion. The weak mechanical strengths of these oxides lead to failure of steel-based devices. But the mechanical properties of these oxides are not straightforward to measure experimentally. In addition, one of the iron oxides, hematite, is a promising candidate for photoanodes in photoelectrochemical cells. A full understanding of its optical, transport, and catalytic properties is crucial to developing practical photoanodes with high efficiency. To address the above issues, this dissertation employs first principles calculations to study various properties of iron oxides.

Theoretically, it is challenging to provide a physical description of these oxides. A modified density functional theory (DFT), called DFT+U, has been proposed as one way of treating correctly these iron oxides that contain strongly correlated Fe 3d electrons. In this thesis, we propose an ab initio scheme to evaluate the on-site Coulomb (U) and exchange (J) energies that are required as inputs for DFT+U theory. Converged U-J values for Fe(II) in FeO and Fe(III) in Fe2O3 are obtained from electrostatically embedded cluster calculations at the level of unrestricted Hartree-Fock theory (UHF). Then the tensile and shear strengths of the three iron oxides are investigated with the ab initio DFT+U theory. The direction dependence of predicted mechanical strengths can be understood by analyzing local bonding strains. We predict that both tensile strengths and shear strengths follow the order of FeO < Fe3O4 < Fe2O3, and that the iron oxides will shear before they fracture.

The rest of the dissertation focuses on examining or improving properties of hematite for use in photoanodes. The key properties to evaluate are electronic excitations, band gaps, electron and hole transport, band edge positions, and finally surface chemistry. To understand the optical properties of hematite, we use embedded correlated wavefunction methods to characterize the optical excited states of hematite. We find that the lowest-lying excitations of hematite are Fe d-d transitions. The ligand to metal charge transfer excitations appear at much higher energy. Various versions of the GW approximation are adopted to investigate the photoemission and inverse photoemission gap of hematite. We find that the G0W0 approximation using ab initio DFT+U wavefunctions and energies as input gives the most accurate predictions for pure hematite.

Both electron and hole transport in pure or doped hematite are studied within the small polaron model. Electrostatically embedded iron oxide clusters are modeled at the UHF level. The electron transport in hematite is predicted to be between Fe centers. We suggest that Si-, Ge-, or Zr- doping is favorable because they are capable of increasing the concentration of electron carriers without causing trapping. We predict that hole transport in hematite is primarily between O anions, with a slightly higher activation energy than for electron transport in hematite. We suggest that Mg-, Ni-, and Cu-doping perform similarly, and the conductivity of Mn-doped hematite might be significantly improved in the high doping concentration limit.

We proposed a first principles scheme to evaluate band edge positions for transition metal oxides. The band gap center is determined within the framework of DFT+U theory. The valence band maximum (conduction band minimum) is found by subtracting (adding) half of the quasiparticle gap obtained from a non-self-consistent GW calculation. Fe2O3 is shown to be a promising parent material for water splitting and CO2 reduction since both reduction and oxidation reaction potentials appear within its band gap.

Water oxidation catalyzed by a hematite surface is studied with DFT+U theory to elucidate detailed reaction mechanisms and energetics. We consider cation doping via substitution of Fe by Ti, Mn, Co, Ni, or Si and F anion doping by replacing O within a fully hydroxylated surface. The reaction energetics on pure or doped hematite surfaces were analyzed using the typical volcano plot employed in catalysis. Our results suggest that Ni- or Co-doping can be an effective means to reduce the overpotential of hematite photoanodes.