Oxygen-hole pairs drive bond order and superconductivity in bismuth oxides 1

Superconductivity in hole-doped BaBiO3 (BBO), namely, Ba1-xKxBiO3 (BKBO) with Tc as high as ~30 K, was discovered immediately after the discovery of high-Tc cuprates [1]. (Prior to them, Pb-doped BaPb1xBixO3 BPBO with a lower Tc of ~12 K had been known for more than ten years [2].) BKBO differs from the cuprates in that they have three-dimensional (3D) crystal and electronic structures against the twodimensional (2D) electronic structures of the cuprates. Moreover, BKBO does not contain transition elements and hence no magnetism should play a role in Cooper-pairing mechanism nor in competing order. Indeed, “charge-density-wave (CDW)” states have been known as phases competing with superconductivity in BKBO. In the “CDW” states, the BiO6 octahedra are expanded and contracted alternatingly. The “CDW” formation is closely related to the instability of the Bi4+ (6s1) valence state toward the charge-disproportionated Bi3+(6s2) + Bi5+ (6s0), referred to as valence skipping [3]. This would naturally lead to bipolaron formation of doped holes and to a Bose-Einstein condensation (BEC) of bipolarons [4]. Even if two holes are not localized on one Bi atom but are extended over many Bi atoms, the on-site attractive interaction between two holes may survive and help the s-wave superconductivity. In fact, BKBO is found to be a strong-coupling BCS superconductor [5]. The above scenario appeared consistent with experiment, but some fundamental questions remain, too, and the problem turned out to be much more complicated and deeper than expected: