Tetrahedral Magnetic Order and the Metal-Insulator Transition in the Pyrochlore Lattice ofCd2Os2O7

Abstract

Cd 2 Os 2 O 7 shows a peculiar metal-insulator transition at 227 K with magnetic ordering in a frustrated pyrochlore lattice, but its magnetic structure in the ordered state and the transition origin are yet uncovered. We observed a commensurate magnetic peak by resonant x-ray scattering in a high-quality single crystal. X-ray diffraction and Raman scattering experiments confirmed that the transition is not accompanied with any spatial symmetry breaking. We propose a noncollinear all-in/all-out spin arrangement on the tetrahedral network made of Os atoms. Based on this we suggest that the transition is not caused by Slater mechanism as believed earlier but by an alternative mechanism related to the formation of the specific tetrahedral magnetic order on the pyrochlore lattice in the presence of strong spin-orbit interactions. The metal-insulator (MI) transition is one of the most dramatic phenomena observed for electrons in solids. Generally, a strong electron correlation or Fermi surface instability causes a solid to transition from showing conductive behavior at high temperature to showing localized insulator behavior at low temperature. The transition caused by the strong electron-electron interaction is known as the Mott transition, which is of the first order and in many cases is accompanied by a simple collinear-type antiferromagnetic ordered state [1]. A prime example of the MI transition caused by Fermi surface instability is the charge-or spin-density wave (CDW or SDW) transition; in this case, the magnetic structures exist in a nonmagnetic singlet state or an incommensurate state. In addition, in weak coupling systems, another type of MI transition-the Slater transition-is theoretically expected to occur via folding of the Brillouin zone owing to the antiferromagnetic order [2]. Given that frustration typically inhibits the antiferromagnetic order, previously documented MI transitions with magnetic ordering are expected to be less probable on frustrated lattices. Consequently, an MI transition due to new mechanism leading to a peculiar ground state that eliminates the frustration is expected.One of the most extensively researched materials is the pyrochlore oxide [3]. Two classes of pyrochlore oxides with MI transitions have been identified: one class exhibits transitions with a large structural distortion and the other class superficially shows no distortion.

Towards properties on demand in quantum materials

“…A family of pyrochlore R 2 Ir 2 O 7 materials (where R stands for rare earth elements) and Cd 2 Os 2 O 7 compounds exhibit magnetic ordering of the Ir (Os) 5d moments with an all-in all-out (AIAO)-type structure 106 . The AIAO magnetic structure implies that all magnetic moments on one tetrahedron point either inwards (all-in) or outwards (all-out) from its centre (Fig.…”

Lifshitz metal–insulator transition induced by the all-in/all-out magnetic order in the pyrochlore oxide Cd2Os2O7

“…What has been suggested for the mechanism based on this in previous study is a kind of Lifshitz transition 27 in the characteristic semimetallic band structure. 14,25 We would like to address this issue in more detail in the present paper.…”

“…25 Contrary to this, the 17 O NMR experiments find a power-law temperature dependence close to T 3 in the NMR relaxation rate at a wide temperature range of 25-150 K, indicating the lack of a large Ising anisotropy gap in the AIAO state. 16 On the other hand, Bogdanov et al performed many-body quantum-chemical calculations and find that even a small trigonal distortion of the OsO 6 octahedron, as actually observed for Cd 2 Os 2 O 7 , causes an easy-axis anisotropy of 6.8 meV, which stabilizes the AIAO order in conjunction with an antiferromagnetic exchange interaction of 14 This imposes a stringent constraint on the mechanism of the MIT since any of the mechanisms associated with the Mott, charge/spin-density wave and Slater transitions would require a structural change more or less. What has been suggested for the mechanism based on this in previous study is a kind of Lifshitz transition 27 in the characteristic semimetallic band structure.…”

“…The following infrared spectroscopy study supports this mechanism, which observes a clear opening of a BCS-type gap with 2∆ = 5.2k B T MI ∼ 1200 K. 13 More recently, Yamaura and coworkers investigated the spin arrangement below T MI by means of resonant x-ray magnetic scattering on a small single crystal and found a q = 0 order that keeps the original cubic primitive cell. 14 They suggest a noncollinear AIAO spin arrangement for the q = 0 order, which is a "tetrahedral order" 14,15 with all the magnetic moments on the four vertices of each Os tetrahedron pointing either inward or outward, as schematically depicted in Fig. 1.…”

“…The process involved two steps. First, a polycrystalline pellet was prepared from a mixture including 5%-10% excess CdO and Os in a sealed quartz tube under the supply of an appropriate amount of AgO as the oxygen source at 1073 K; 14 addition of too much AgO should be avoided, as highly toxic OsO 4 might be produced. In the second step, the pellet was placed at one end of a sealed quartz tube of 30 cm long, which was heated for a week in a furnace having a temperature gradient of 1040-1200 K with the pellet at the high-temperature side.…”