BERKELEY, CA — A team led by investigators from Stanford University and the Department of Energy's Lawrence Berkeley National Laboratory has gathered surprising information about the electronic structure of the "stripe phase," a new electronic state of solids. Their report, in the October 8, 1999 issue of the journal Science, may help resolve an apparent paradox between different theories of superconductivity and may explain how copper-oxide ceramics can become superconducting at high temperatures.
Xingjiang Zhou, working with Zhi-Xun Shen of Stanford, Zahid Hussain of the Advanced Light Source (ALS), and other colleagues, used the High Energy Resolution Spectrometer at the ALS to uncover new clues to the mysterious behavior of the high Tc superconductors.
"High-temperature superconductivity was discovered in 1986, but after a dozen years we are still struggling to understand it," says Zhou, a physicist who has a joint position at Stanford and the ALS. "Of the 50 or so high-Tc materials discovered so far, all of them are copper oxides."
The researchers used angle-resolved photoemission spectroscopy (ARPES) to study the stripe phase, or charge- and spin-ordered state, in the compound Nd-LSCO (neodymium-substituted lanthanum strontium copper oxide). The technique employs beams of synchrotron light to knock electrons out of a sample, probing its electronic structure by measuring both the energy and the direction of the emitted photoelectrons.
Plots of the resulting "spectral weights" at high energies were consistent with charges moving through the Nd-LSCO sample along one-dimensional lines, so-called stripes, but at lower energies the pattern was more easily explained if the charges were moving in two dimensions -- behavior that appears to require two different theoretical explanations.
The parent compounds of cuprates are insulators; their complex structure, similar to that of the mineral perovskite, alternates two-dimensional layers of oxygen and copper atoms with layers of other atoms. Cuprates are made more metal-like, and in some cases superconducting, by doping -- adding elements which contribute extra electrons or create holes to carry negative or positive charges.
"With Nd-LSCO we found that at about one-eighth doping level, the picture that best fit the data was the stripe phase -- charge carriers segregating themselves into one-dimensional lines," says Zhou. "In the regions between these lines, electronic spins are arranged anti-ferromagnetically," that is, the spins are arranged so that each spin points opposite to those around it, producing insulating regions.
Nd-LSCO has static stripes and is not superconducting, but in those cuprates that are superconductors, dynamic stripes may come into play and become associated with superconductivity. Whether the stripe phase is actually responsible for high-temperature superconductivity, however, remains the subject of vigorous debate.
The stripe phase was predicted by Jan Zaanen and Olle Gunnarsson in 1987 and discovered experimentally in cuprates by John Tranquada and others in 1995, using neutron scattering. However, the underlying theory which gives rise to the stripe phase, mean field theory, paradoxically suggests that the stripe phase should always be insulating!
The traditional electronic theory of metals describes how quasiparticles -- collective entities with particle-like properties such as energy and momentum -- experience the field of a solid's crystal lattice. At low energies, the charge carriers in Nd-LSCO indeed appear to interact with variations in the field due to the crystal lattice. The carriers move back and forth in a two-dimensional manner and exhibit low-energy states comparable to those observed in good superconductors. But other expected two-dimensional effects are missing from the ARPES data.
Established theories also describe a characteristic Fermi surface that marks where (in momentum space) a given material's uppermost energy level is filled with electrons. At high energies, the Fermi surfaces in Nd-LSCO plot as straight lines set at right angles, indicating that the charge carriers move along one-dimensional bands -- features which could not arise from the quasiparticles of traditional electronic theories.
Writing in a Science Perspective, theorist Jan Zaanen points up the conundrum represented by these results. "How can it be that the electrons which are 'one-dimensionalized' at high energies can rediscover the two-dimensional world at low energies? Within established electronic structure theory this appears as a paradox, and new physics is here at work."
"The solution to this paradox may be a new basic starting point," Zhou suggests. "Instead of the quasiparticles that are responsible for superconductivity in ordinary metals, in the cuprates one may have to start with the stripes themselves, along which charge flows freely. The stripes appear stable at high energies, but at lower energies they may exhibit quantum fluctuations that give rise to two-dimensional effects."
Zhou adds that "the compound we investigated is not itself a superconductor, but through understanding its electronic structure we can make a major advance in understanding its high-Tc relatives."
The High Energy Resolution Spectroscopy (HERS) endstation used in studies of highly correlated materials at the ALS was supported by the Facilities Initiative of the Department of Energy's Office of Science, and "is unique in the world for this kind of science," says the ALS's Zahid Hussain.
Besides X. J. Zhou, Z.-X. Shen, and Z. Hussain, other collaborators included P. Bogdanov and S. A. Kellar of Stanford, and T. Noda, H. Eisaki, and S. Uchida of the University of Tokyo. The report by Zhou et al and Zaanen's Perspective appear in the same issue of Science, October 8, 1999.
The Berkeley Lab is a U.S. Department of Energy national laboratory located in Berkeley, California. It conducts unclassified scientific research and is managed by the University of California.