New insight into superconductivity

NOVEMBER 2008 | VOLUME 11 | NUMBER 11 13 The physics of high-temperature cuprate superconductors in the overdoped regime, where superconductivity is suppressed, is something of a mystery. But first-principle calculations by researchers in the US and Switzerland appear to confirm a recent suggestion that ferromagnetism can destroy superconductivity at high doping levels [B. Barbiellini and T. Jarlborg, Phys. Rev. Lett. (2008) 101, 157002]. Superconductivity in the low or undoped regime has been well studied, but standard theories cannot readily explain the sudden destruction of superconductivity at high doping concentrations. “What was missing was a reliable study of high doping,” says Bernardo Barbiellini of Northeastern University. Together with Thomas Jarlborg of the University of Geneva in Switzerland, Barbiellini undertook calculations using the Linear Muffin-Tin Orbital (LMTO) method for nano-sized supercells of La2CuO4 doped with Ba. The calculations show that weak ferromagnetism appears around clusters of high Ba concentration. The results confirm the suggestion that ferromagnetism and superconductivity compete in overdoped samples. “Rather like a Swiss cheese, when the ferromagnetic domains become bigger and bigger, superconductivity vanishes,” Barbiellini explains. But it is not just the amount of doping that is crucial, the arrangement of the Ba atoms is also important. Ferromagnetism can be induced when the concentration of Ba is high, but also at lower doping concentrations if the Ba atoms are clustered. “Spontaneous magnetization is not found in configurations with well-separated single Ba atoms,” says Barbiellini. “The Ba clustering is a crucial ingredient to stabilize ferromagnetism.” The work represents a significant advance, believes Stephen B. Dugdale of the University of Bristol in the UK. “It addresses a gaping hole in our understanding of the physics of the cuprates in the overdoped regime,” he says. The conclusions of the researchers, however, could be seen as controversial in some circles. Theorist Sudip Chakravarty of the University of California, Los Angeles, nevertheless believes the study is important. “Ferromagnetic fluctuations, such as discussed in this paper, should be taken very seriously.” Barbiellini and Jarlborg’s approach is also a good example of the kind of insight that firstprinciple calculations provide, believes Dugdale. By comparison, commonly used approaches such as rigid band approximation are not as effective for studying the effects of doping in La2CuO4 on ferromagnetic properties. “[The] interplay between superconductivity and ferromagnetism hints that further studies of over-doped Cu oxides will reveal important secrets about the mechanism of hightemperature superconductivity,” adds Barbiellini. Cordelia Sealy