Neutron scattering studies on layered ruthenates

Abstract

Transition metal oxides (TMOs) exhibit a large variety of magnetic, electronic, and structural phases and have received much attention from the community. The tight competition between different interactions and ordering phenomena typical for such systems result in phase diagrams which are characterized by a multitude of transitions. These often depend on external variables, including temperature, magnetic or electric fields, pressure, and chemical doping. Early research focused on oxides of light transition metals exhibiting flat electronic bands and strongly correlated systems. Prominent examples include the families of copper oxides (cuprates) that exhibit high-temperature superconductivity, and manganites that show colossal magnetoresistance. For a long time, oxides of heavier transition metals were not expected to exhibit particular exciting phenomena: with increasing atomic mass and ionic radii, the Coulomb repulsion decreases while the extension of the d orbitals enlarges, consequently increasing the orbital overlap and the electronic bandwidth W. Such heavy-metal based systems were therefore expected to be metallic, without the intricate competition between different ordering phenomena seen in their lighter analogues. Recently, however, it was recognized that the spin-orbit coupling (SOC) can profoundly change the phase behavior of 4d- and 5d-electron materials. The strength of SOC scales with the atomic number Z as ∝Z^4 , which—in contrast to systems including 3d TMOs—renders SOC a driving force in oxides of heavy transition metals. As the interplay between SOC and electronic correlations brings about novel quantum ground states, these systems have received increasing interest during the last decade. One such systems is Sr2IrO4, where this interplay generates a Mott-insulating state with total angular momentum J_{eff} = 1/2 . 4d-electron compounds, which are characterized by moderate SOC, have until recently been modeled akin to oxides of 3d-electron systems, treating the SOC as a minor perturbation only. However, even moderate SOC proved to be enough to realize exotic phenomena that are not captured by such approaches, and can lead to a variety of competing structural and magnetic phases. Consequently, the role of SOC in 4d TMOs has been underestimated, calling for re-evaluation of the underlying physics. In this work we focus on the antiferromagnetic Mott insulator Ca2RuO4, in which the interaction is limited to the two-dimensional layers of RuO6 octahedra. The low-spin 4d^4 configuration of Ru^{4+} leads to a S = 1 spin, while the lattice symmetry results in an effective orbital momentum of L_{eff} = 1. Previous studies have shown that Ca2RuO4 undergoes an insulator-metal transition upon heating and exhibits a series of phase transitions upon isovalent substitution with Sr. The wide variety of phases makes Ca2RuO4 a prime material platform to investigate the role of moderate SOC in magnetism. We concentrate on the magnetic excitation spectrum, which reflects the combined influence of the exchange interactions between the Ru ions and the inter-ionic SOC. The first part of this PhD project is dedicated to the growth of high-quality crystals of Ca2RuO4 and related ruthenium oxides. To this end, we used the optical floating zone technique. The several hundred crystal shards were then co-aligned to be used in inelastic neutron scattering experiments. With a map of the magnetic scattering intensity in the full magnetic Brillouin zone, we observe and distinguish all transverse magnon (Goldstone) modes as well as a longitudinal amplitude (Higgs) mode. The results can be consistently interpreted in an excitonic magnetism model with a dominant influence of the SOC. We then used inelastic neutron scattering to investigate the magnetic excitations of the Sr-substituted Ca2RuO4 crystals and found a modified set of exchange interactions. We also investigated the closely related Ca3Ru2O7 system; here, a double-layers of RuO6 octahedra are interleaved by a CaO barrier layer. We find that this bilayer system exhibits a metallic phase where the impact of the SOC is less pronounced. Surprisingly, a chemical substitution of the 4d^4 Ru^{4+} ions with magnetically inactive 3d^0 Ti^{4+} ions renders the system insulating even for Ti concentrations less than 1 %. In this phase that the system’s magnetic excitations are similar to Ca2RuO4 suggesting the same excitonic magnetism. Our studies demonstrate the crucial role of SOC for the magnetic properties of ruthenium oxides, and call for a general reevaluation of the impact of SOC on the ground state and excitations of 4d-electron systems.