An explanation for the need of a reduction process in electron-doped superconductors offers new insight into the conductivity mechanism of these lesser-known superconductors.
Immediately after the discovery of high-temperature superconductivity (HTSC) in layered cuprates two decades ago, it was thought that introducing holes (positive charge carriers) in two-dimensional CuO2 planes is essential to realize a high transition temperature (TC) to the superconducting phase. For example, in La2–xSrxCuO4 (LSCO; TC = 40 K) the partial replacement La3+ with Sr2+ in the antiferromagnetic insulator La2CuO4 removes electrons from the CuO2 plane and creates current-carrying holes1. A big surprise came in 1988 when electron-doped superconductivity was discovered in R2–xCexCuO4 (RCCO, where R = La, Nd, Pr, Sm or Eu) with TC = 25 K, where R3+ is substituted with Ce4+ and extra electrons are doped into the CuO2 planes2,3. In addition, a peculiarity of all RCCO systems has left scientists puzzled since their discovery, namely, the necessity of annealing them in a reducing atmosphere. Writing on page 224 of this issue, Hye Jung Kang and co-workers4 propose an intriguing picture for the reduction process based on a careful structural analysis of oxidized and reduced samples.
Samples of RCCO doped with electrons can be easily synthesized in air but show the behaviour of a poor metal or weakly localized insulator. To achieve bulk superconductivity, the as-grown samples must be annealed in a reducing atmosphere close to their decomposition point and then quenched. Although the creation of oxygen vacancies is a well-known source of extra electrons, experimental evidence suggests that the role of reduction is much more substantial than simply controlling the carrier density through a change in oxygen content. For example, superconductivity always appears in the same range of cerium concentration, and TC as a function of cerium content x always peaks at around x ∼ 0.15, independent of the reducing condition. The real mystery is what happens to the RCCO lattice during the reduction process. In the scenario proposed by Kang and co-workers, copper vacancies, which are always present in the as-grown material, act as defects to destroy electron superconductivity in the CuO2 plane. This plane is self-repaired through a structural rearrangement that occurs during annealing. The emergent picture for the reduction process is not only interesting in terms of defect chemistry of complex transition metal oxides, but also impacts on the physics of high-TC cuprates.
The structural refinement of Kang and co-workers on RCCO (R = Pr and La) samples show a 1–2% level of Cu vacancies in oxidized non-superconducting samples. The authors argue that these Cu vacancies suppress superconductivity through strong disorder effects but they can be repaired through cooperative structural changes that result in enhanced superconductivity. In X-ray-diffraction patterns, the signature of these structural changes is the detection of 1% of R2O3 (epitaxial to the RCCO lattice), which only appears in reduced superconducting samples and disappears in oxidized, non-superconducting samples. The appearance of this R2O3 impurity phase, which had been observed in previous diffraction studies5,6,7, is related to the fact that copper atoms are removed from some parts of the copper oxide plane and are collected in other planes where the vacancies become filled. This becomes possible because of the structural relation between the fluorite R2O2 layer unit in the RCCO structure and the structure of R2O3, which can be can be viewed as an oxygen deficient (by 25%) fluorite structure.
There is a global symmetry of the phase diagram (see Fig. 1) between hole-doped and electron-doped superconductors in that, on doping, an antiferromagnetic insulator first turns into a superconductor and then eventually into a metal. However, in many aspects electron-doped HTSCs are distinct from hole-doped HTSCs. For example, it can be seen that in the phase diagram, an antiferromagnetically ordered phase seems to be much more robust against doping in an electron-doped HTSC. To capture the essential ingredients for HTSC and sort out the material-dependent details, an exploration of electron-doped HTSC in comparison with hole-doped HTSC is extremely important. There are, however, only two electron-doped HTSCs other than RCCO: Sr1–xRxCuO3 (ref. 8) and LixSr2CuO2Br2 (ref. 9), but they are both quite difficult to synthesize. Thus, among many cuprate HTSCs, RCCO occupies a special place in the quest for the mechanism of superconductivity. The complex chemistry originating from the reduction process had forced scientists to solve many difficult materials issues, such as the artificial magnetic signatures produced by the secondary magnetic phase R2O3 in neutron studies5,6,7. This new proposal for the chemistry behind the reduction process in RCCO will help to eliminate materials-specific issues of the electron-doped HTSC, and also improve the quality of RCCO crystals.
Kang and co-workers show that, after the reduction, a 1% level of oxygen vacancies persists in the CuO2 plane as a by-product of the self-repairing mechanism. This implies that superconductivity might be quite robust against the creation of oxygen vacancies but not Cu vacancies. It could be inferred that doped electrons in cuprates reside mainly on the Cu sites, in contrast to the holes residing mainly on the oxygen sites. The authors argue that oxygen vacancies might even help superconductivity in contrast to Cu vacancies. But the opposite argument could be made: if oxygen vacancies were indeed harmful, filling them would be a way to further enhance superconductivity. Indeed, early work suggested that annealing samples in oxygen at a low temperature of 500 °C, seemingly too low to create Cu vacancies but high enough to fill in oxygen vacancies, improved superconducting properties3.
Many challenges remain to be tackled, and they are entwined with chemical phase and composition issues that can be much more complex and fascinating than physicists might imagine.
Bednorz, J. G. & Müller, K. A. Z. Phys. B 64, 189–193 (1986).
Tokura, Y., Takagi, H. & Uchida, S. Nature 337, 345–347 (1989).
Takagi, H., Uchida, S. & Tokura, Y. Phys. Rev. Lett. 62, 1199–1102 (1989).
Kang, H. J. et al. Nature Mater. 6, 224–229 (2007).
Matsuura, M. Phys. Rev. B 68, 144503 (2003).
Mang, P. K. et al. Phys. Rev. B 70, 094507 (2004).
Kimura, H. et al. J. Phys. Soc. Jpn 74, 2282–2286 (2005).
Smith, M. G. Nature 351, 549–551 (1991).
Kajita, T. et al. Jpn J. Appl. Phys. 43, L1480–L1481 (2004).
About this article
Physical Review B (2016)
Applied Physics Express (2015)
Journal of the Physical Society of Japan (2014)
Physical Review B (2014)
Physical Review B (2011)