The recently proposed theoretical concept of a Hund's metal is regarded as a key to explain the exotic magnetic and electronic behavior occuring in the strongly correlated electron systems of multiorbital metallic materials. However, a tuning of the abundance of parameters, that determine these systems, is experimentally challenging. Here, we investigate the smallest possible realization of a Hund's metal, a Hund's impurity, realized by a single magnetic impurity strongly hybridized to a metallic substrate. We experimentally control all relevant parameters including magnetic anisotropy and hybridization by hydrogenation with the tip of a scanning tunneling microscope and thereby tune it through a regime from emergent magnetic moments into a multi-orbital Kondo state. Our comparison of the measured temperature and magnetic field dependent spectral functions to advanced many-body theories will give relevant input for their application to non-Fermi liquid transport, complex magnetic order, or unconventional superconductivity.1 arXiv:1604.03854v1 [cond-mat.str-el] Apr 2016Recent examples of exotic phases of matter, including unconventional superconductivity in iron pnictides and chalcogenides [1][2][3] as well as non-Fermi liquid behavior in ruthenates [4][5][6], depend subtly on the complex interplay of magnetic moments and delocalized electron states taking place in transition metal d-shells. All these materials combine sizable Coulomb interactions and hybridization, which are comparable in their strength. In such cases, it is generally unclear, to which extent local magnetic moments exist, how they can be described using quantum impurity models [7], and how far electronic correlation effects such as Kondo screening [8,9] modify material properties, particularly magnetism, as a function of temperature and magnetic field. The recent concept of a Hund's metal [2,10,11] has been introduced in order to describe exactly this regime, where charge fluctuations in the orbitals are not negligible due to the presence of strong hybridization, but where local magnetic moments can still survive.The fundamental constituent of such a Hund's metal is a magnetic impurity strongly coupled to the electron states of a metallic host, which we coin Hund's impurity. This concept is described in the following for the particular case of a 3d transition metal atom that gets adsorbed (adatom) onto a metallic substrate (Fig. 1). If the atom is still in the gas phase an integer number of electrons is filled into the five 3d orbitals according to Hund's first rule: [12,13] The orbitals are first filled up by electrons having the same spin, before being filled with the remaining electrons of opposite spin. This is driven by the intraatomic exchange energy, or so-called Hund's rule exchange J Hund , which has to be paid if one of the electron spins is flipped. If the 3d transition metal atom is adsorbed onto the metallic substrate, electrons can hop on or off of these orbitals into the bath of substrate conduction electrons, which has ...

The aim of this work is to unravel a basic microscopic picture behind complex magnetic properties of hexagonal manganites. For these purposes, we consider two characteristic compounds:YMnO 3 and LuMnO 3 , which form different magnetic structures in the ground state (P 6 3 cm and P 6 3 cm, respectively). First, we establish an electronic low-energy model, which describes the behavior of the Mn 3d bands of YMnO 3 and LuMnO 3 , and derive parameters of this model from the first-principles calculations. From the solution of this model, we conclude that, despite strong frustration effects in the hexagonal lattice, the relativistic spin-orbit interactions lift the degeneracy of the magnetic ground state so that the experimentally observed magnetic structures are successfully reproduced by the low-energy model. Then, we analyze this result in terms of interatomic magnetic interactions, which were computed using different approximations (starting from the model Hamiltonian as well as directly from the first-principles electronic structure calculations in the local-spin-density approximation). We argue that the main reason why YMnO 3 and LuMnO 3 tend to form different magnetic structures is related to the behavior of the single-ion anisotropy, which reflects the directional dependence of the lattice distortion: namely, the expansion and contraction of the Mn-trimers, which take place in YMnO 3 and LuMnO 3 , respectively. On the other hand, the magnetic coupling between the planes is controlled by the next-nearest-neighbor interactions, which are less sensitive to the direction of the trimerization. In the P 6 3 cm structure of YMnO 3 , the Dzyaloshinskii-Moriya interactions lead to the spin canting out of the hexagonal plane -in the same direction as the single-ion anisotropy. Finally, using the Berry-phase formalism, we evaluate the magnetic-state dependence of the ferroelectric polarization, and discuss potential applications of the latter in magnetoelectric switching phenomena.

We have examined a complete phase diagram of Y 1−x Lu x MnO 3 with 0 Յ x Յ 1 by using bulk measurements and neutron-diffraction studies. With increasing Lu concentration, Curie-Weiss temperature and Neel temperature are found to increase continuously while the two-dimensional nature of short-range magnetic correlation persists even in the paramagnetic phase throughout the entire doping range. At the same time, the lattice constants and the unit-cell volume get contracted with Lu doping, i.e., chemical pressure effect. This decrease in the lattice constants and the unit-cell volume then leads naturally to an increased magnetic exchange interaction as found in our local spin-density approximation band calculations. We also discover that there is strong correlation in the temperature dependence of a volume anomaly at T N and the magnetic moments.

Han purple (BaCuSi2O6) is not only an ancient pigment, but also a valuable model material for studying Bose-Einstein condensation of magnons in high magnetic fields. Using precise low-temperature structural data and extensive density-functional calculations, we elucidate magnetic couplings in this compound. The resulting magnetic model comprises two types of nonequivalent spin dimers, in excellent agreement with the Cu63,65 nuclear magnetic resonance data. We further argue that leading interdimer couplings connect the upper site of one dimer to the bottom site of the contiguous dimer, and not the upper-to-upper and bottom-to-bottom sites, as assumed previously. This finding is verified by inelastic neutron scattering data and implies the lack of frustration between the layers of spin dimers in BaCuSi2O6, thus challenging existing theories of the two-dimensional-like Bose-Einstein condensation of magnons in this compound.