ORNL Review Vol. 34, No. 2, 2001

Neutrons, "Stripes," and Superconductivity

How electrons behave in high-temperature copper oxide superconductors is a mystery, but progress is being made in resolving that mystery on both theoretical and experimental fronts. One of the leading theories postulates that electronic matter organizes itself into fluctuating regions where the charge (holes) and the magnetism (spins) are separated in space in one-dimensional regions called stripes. These striped phases were predicted in theories developed by physicists at the Leiden Institute in the Netherlands, DOE's Brookhaven National Laboratory, and the University of California at Los Angeles.

Herbert A. Mook and Pencheng Dai, both of ORNL's Solid State Division, and colleagues from the University of Washington in Seattle have found evidence for this stripe theory in experiments on high-temperature superconducting materials such as yttrium-barium-copper oxide (YBCO). The results were obtained in neutron-scattering experiments on superconducting samples prepared by Rodney Hunt of ORNL's Chemical Technology Division.

Superconductivity in standard low-temperature materials is understood in terms of the theory developed by John Bardeen, Leon Cooper, and Robert Schrieffer (known as the BCS theory), who found a way to make superconductivity work by pairing up electrons of opposite spins. Normally, electrons bump into each other, impeding conductivity. However, the pairs glide through the superconductor like couples waltzing across a ballroom dance floor, completely unhindered by the other electron couples on the floor. The distance between the electrons in each pair is called the superconducting coherence length. Both electrons in each pair have a negative charge so that they repel each other, but the coherence length is long in the standard materials, so the electron partners are quite far apart and, in fact, have many other electron pairs dancing between them.

In the copper oxide superconductors, however, the coherence length is very short, so the paired electrons are dancing close together. This makes it very hard to find a pairing interaction strong enough to provide the glue to keep the pairs together.

Aerial view of the High Flux Isotope Reactor, where some of the neutron evidence was obtained to support a leading theory for explaining high-temperature superconductivity. (Photo by Curtis Boles.)

"This is where stripes come in," Mook says. "Rather amazingly in one dimension, electrons can split in two parts with one of the parts carrying the charge and the other the spin. In this case, the spins can form superconducting pairs without their charges trying to keep them apart. The stripes provide the one-dimensional regions that make it possible for the electrons’ spins to separate themselves from their charges."

The neutron scattering results on stripes have been published in three papers in the prestigious journal Nature. The experiments led by Mook and Dai were performed between 1998 and 2000 at the Rutherford Appleton Laboratory's spallation neutron source at the ISIS Facility in the United Kingdom and at ORNL's High Flux Isotope Reactor (HFIR). Their first experiment at the ISIS spallation neutron source showed that the electron spins separated into distinct regions in the high-temperature, superconducting material.

"In our experiment at ISIS," Mook says, "we determined the spatial distribution of the electrons' spins, based on the neutron scattering pattern." These measurements provided the first direct evidence that striped phases occur in the YBCO superconductors.

Arrangement of the spins and charges in a striped phase. The circles represent the copper atoms in the copper-oxygen planes of the superconductor. The spin stripes are represented by the arrows while the open circles are the hole stripes. The shaded circle shows that the holes are not uniformly distributed on the hole stripe and may move along the stripe.

A second set of experiments at the HFIR showed that the charge distribution was also consistent with stripes. "At HFIR we saw how the charges were distributed in crystals by measuring their effect on the lattice vibrations, which are called phonons," Mook says. "The changes in the phonons allowed us to see a periodicity in the pattern of these vibrations, which demonstrated that the charge part matched up with the spin part of a striped phase."

A third experiment, also conducted at HFIR, demonstrated that the spin distribution observed earlier was really one dimensional in nature. "This was a key issue in the stripes argument," Mook says. "We initially submitted the paper to Nature as a letter, but the head physical sciences editor wanted it expanded into a full article."

It will probably be some time before a consensus is reached on the mechanism behind high-temperature superconductivity. However, because of the neutron scattering experiments, the "stripe" theory is regarded as one of the leading theories for explaining high-temperature superconductivity.

Related Web sites