Abstract. Magnetism of transition metal (TM) oxides is usually described in terms of the Heisenberg model, with orientation-independent interactions between the spins. However, the applicability of such a model is not fully justified for TM oxides because spin polarization of oxygen is usually ignored. In the conventional model based on the Anderson principle, oxygen effects are considered as a property of the TM ion and only TM interactions are relevant. Here, we perform a systematic comparison between two approaches for spin polarization on oxygen in typical TM oxides. To this end, we calculate the exchange interactions in NiO, MnO, and hematite (Fe 2 O 3 ) for different magnetic configurations using the magnetic force theorem. We consider the full spin Hamiltonian including oxygen sites, and also derive an effective model where the spin polarization on oxygen renormalizes the exchange interactions between TM sites. Surprisingly, the exchange interactions in NiO depend on the magnetic state if spin polarization on oxygen is neglected, resulting in non-Heisenberg behavior. In contrast, the inclusion of spin polarization in NiO makes the Heisenberg model more applicable. Just the opposite, MnO behaves as a Heisenberg magnet when oxygen spin polarization is neglected, but shows strong non-Heisenberg effects when spin polarization on oxygen is included. In hematite, both models result in non-Heisenberg behavior. General applicability of the magnetic force theorem as well as the Heisenberg model to TM oxides is discussed.

Please be advised that this information was generated on 2018-05-09 and may be subject to change. PHYSICAL REVIEW B 91, 045420 (2015) Modeling Klein tunneling and caustics of electron waves in graphene We employ the tight-binding propagation method to study Klein tunneling and quantum interference in large graphene systems. With this efficient numerical scheme, we model the propagation of a wave packet through a potential barrier and determine the tunneling probability for different incidence angles. We consider both sharp and smooth potential barriers in n-p-n and n-n junctions and find good agreement with analytical and semiclassical predictions. When we go outside the Dirac regime, we observe that sharp n-p junctions no longer show Klein tunneling because of intervalley scattering. However, this effect can be suppressed by considering a smooth potential. Klein tunneling holds for potentials changing on the scale much larger than the interatomic distance. When the energies of both the electrons and holes are above the Van Hove singularity, we observe total reflection for both sharp and smooth potential barriers. Furthermore, we consider caustic formation by a two-dimensional Gaussian potential. For sufficiently broad potentials we find a good agreement between the simulated wave density and the classical electron trajectories.

Correlation between geometry, electronic structure and magnetism of solids is both intriguing and elusive. This is particularly strongly manifested in small clusters, where a vast number of unusual structures appear. Here, we employ density functional theory in combination with a genetic search algorithm, GGA+U and a hybrid functional to determine the structure of gas phase Fe In nano technology there is an ever increasing demand for increasing the density of electronic and magnetic devices. This continuous downscaling trend drives the interest to electronic and magnetic structures at the atomic scale. In essence, two things are required: first, novel materials and building blocks with exotic physical properties. Second, a fundamental knowledge of the physical mechanism of magnetism at the sub-nanometer scale.Atomic clusters, having highly non-monotonous behavior as a function of size, are a promising model system to study the fundamentals of magnetism at the nanoscale and below. Such clusters consist of only tens of atoms. Quantum mechanics starts to play an essential role at this small scale, adding extra degrees of freedom. Since these clusters are studied in high vacuum, they are completely isolated from their environment.To use these clusters as a model system, as a starting point, a detailed understanding of the relation between their geometry and electronic structure is required.Even in the bulk, iron oxide has a wide variety of chemical compositions and phases with many interesting phenomena, such as the Verwey transition in magnetite. 1,2Experiments performed on small gas phase Fe x O y clusters beyond the two-atom case are scarce. The structure of one and two Fe atoms with oxygen has been studied in an argon matrix using infrared spectra.3,4 The corresponding vibration frequencies have been identified using density functional theory (DFT).Iron-oxide nanoparticles have been investigated for their potential use as catalyst in chemical reactions.5 Furthermore, since the iron-oxygen interaction has a fundamental role in many chemical and biological processes, there have been quite some studies, both experimental and theoretical, of the chemical properties of Fe x O y clusters. 6-12The possible coexistence of two structural isomers for stoichiometric iron-oxide clusters in the size range n ≥ 5 was experimentally measured using isomer separation by ion mobility mass spectroscopy for Fe n O n and Fe n O n+1 (n = 2-9).13 Furthermore, the formation of Fe x O y clusters has been studied in the size range (x = 1-52). 14The number of theoretical studies is, however, manifold. The magic cluster Fe 13 O 8 was extensively studied and identified as a cluster with C 1 but close to D 4h point group symmetry.15-19 However, also the geometry and electronic structure of other cluster sizes have been studied theoretically.15,20-25 The prediction of geometric structures requires a systematic search of the potential energy surface to find the global minimum.The majority of theoretical studies were performed using DFT. 4,6,9,1...

From magnetic deflection experiments on isolated Co doped Nb clusters we made the interesting observation of some clusters being magnetic, while others appear to be non-magnetic. There are in principle two explanations for this behavior. Either the local moment at the Co site is completely quenched or it is screened by the delocalized electrons of the cluster, i.e. the Kondo effect. In order to reveal the physical origin, we conducted a combined theoretical and experimental investigation. First, we established the ground state geometry of the clusters by comparing the experimental vibrational spectra with those obtained from a density functional theory study. Then, we performed an analyses based on the Anderson impurity model. It appears that the non-magnetic clusters are due to a complete quenching of the local Co moment and not due to the Kondo effect. In addition, the magnetic behavior of the clusters can be understood from an inspection of their electronic structure.Here magnetism is favored when the effective hybridization around the chemical potential is small, while the absence of magnetism is signalled by a large effective hybridization around the chemical potential.

We study the infrared (IR) resonant heating of neutral niobium carbide clusters probed through ultraviolet (UV) photoionization spectroscopy. The IR excitation not only changes the photoionization spectra for the photon energies above the ionization threshold, but also modulates ion yield for energies significantly below it. An attempt to describe the experimental spectra using either Fowler's theory or thermally populated vibrational states was not successful. However, the data can be fully modeled by vibrationally and rotationally broadened discrete electronic levels obtained from Density Functional Theory (DFT) calculations. The application of this method to spectra with different IR pulse energies not only yields information about the excited electronic states in the vicinity of the HOMO level, populated by manipulation of the vibrational coordinates of a cluster, but also can serve as an extra indicator for the cluster isomeric structure and corresponding DFT-calculated electronic levels.

The electronic structure of metal clusters is notoriously difficult to detect spectroscopically, due to rapid relaxation into the ground state following excitation. We have used IR multiple photon excitation to identify a low-lying electronic state in a tantalum carbide cluster. The electronic excitation is found at 458 cm , and is confirmed by experiments on isotopically labeled clusters. Time-dependent density functional theory (TD-DFT) calculations confirm the current assignment, but a second predicted electronic state was not observed.

We report experimental vibrational spectra of small neutral niobium carbide clusters in the 350-850 cm spectral range. Clusters were first irradiated by IR light and subsequently probed using UV light with photon energies just below the ionization threshold. Upon resonance with an IR vibrational mode, the number of cluster ions increases, allowing to record a vibrational spectrum. In complementary density functional theory calculations, we have simulated the IR spectra for several low-energy isomers. We were able to assign the spectra experimentally obtained for each cluster size to a specific geometric structure based on the match with the computed spectra. The number of the cluster sizes investigated here allows to follow the evolution of the geometric structure of the niobium and carbon components of clusters separately. For NbC (m = 4, 5, 6), we observe the emergence of the cubic crystal structures similar to the bulk.

The sensitivity of material properties to the atomic and nanoscale morphology is most clearly demonstrated in small gas-phase clusters. In particular, magnetism serves as an extremely sensitive probe of the smallest modifications of atomic environment. This Rapid Communication demonstrates the drastic changes in both the exchange mechanism and the atomic moments in iron oxide clusters as compared to the bulk. In particular, the exchange is essentially non-Heisenberg and the exchange interactions are increased by an order of magnitude compared to bulk hematite. In addition, very large atomic magnetic moments are observed on oxygen sites due to strong spin polarization. As Fe 3 O + 4 and Fe 5 O + 7 are pure trivalent, double exchange is excluded and super exchange is the dominant exchange mechanism in these iron oxide clusters. Therefore, the non-Heisenberg behavior is attributed to covalent magnetism as the hybridization between Fe 3d and O 2p orbitals for clusters is strongly enhanced compared to bulk hematite and becomes magnetic state dependent.