Dan Dessau of the University of Colorado has worked with students and colleagues at the University of Tokyo and Berkeley Lab's Advanced Light Source to attack a recalcitrant puzzle in contemporary electronic theory: what can explain colossal magnetoresistance, or CMR?
In the May 25 issue of Science, Dessau and his student Yi-De Chuang, with their coauthors, describe using high-resolution angle-resolved photoemission spectroscopy (ARPES) to determine the key electronic properties of a manganese-oxide compound noted for CMR. Their observations suggest that several electronic phases of the material exist side by side, sometimes cooperating with and sometimes competing against one another.
Dessau says, "The general concept of multiple electronic phases coexisting and competing with each other at nanoscale dimensions appears to be a general phenomenon for novel electronic materials -- including these CMR materials and the high-temperature superconductors as well."
Putting up resistance
Consider an electric current running in a material like iron. Placed in a strong magnetic field, its resistance drops or increases by several percent, depending on orientation -- hence magnetoresistance, or MR.
In 1988, thinly layered materials were found that increased or decreased their resistivity by 20 percent or more in relatively weak magnetic fields -- hence "giant" magnetoresistance, or GMR. While not completely understood, the basic effect depends on the alignment of electron spins at the interface of different kinds of magnetic materials.
Read heads using MR outperformed conventional ones using magnetic inductance and soon found favor in computer hard drives and other storage media. More sensitive GMR materials allow data to be stored even more densely and read more quickly, and GMR technology is now used in most modern hard drives.
Then in 1993, materials were found that could increase or decrease resistance not by a few percent but by orders of magnitude. Hence "colossal" magnetoresistance -- an effect no existing theory can explain.
Not the usual suspects
CMR is found in materials with the crystalline structure called perovskite, whose atoms are arranged in discrete layers of differing composition. Many perovskites have unusual electronic properties; all the known high-temperature superconductors are perovskites incorporating copper-oxide layers.
CMR is pronounced in the perovskite studied by Dessau, Chuang, and their colleagues, one that has double layers of manganese oxide plus specific proportions of lanthanum and strontium, some of which, with oxygen, form additional layers. "The bilayers are where the current flows, and they are responsible for the CMR effect," says Chuang.
The most successful traditional theories cannot explain CMR. Double-exchange theory, for example, describes the relation between magnetic transitions and metal-insulator transitions in terms of the orientation of electron spins at adjacent sites. For lanthanum strontium manganese oxide there should be only about a 30 percent change in conductivity.
"In fact," says Chuang, "the observed changes are much bigger, by orders of magnitude. Thus the 'colossal' effect."
The remarkably successful quasiparticle concept (see sidebar), which uses fictitious properties to simplify the modeling of real particle behaviors, also fails. In lanthanum strontium manganese oxide, the calculated lifetime of quasielectrons is anomalously long, even though the material is about as poor a metal as can be imagined.
Chuang sums up the paradox thus: "The compound has a Fermi surface" -- a parameter indispensable for understanding a material's electronic properties -- "so it's a metal. But if only 10 percent of the electrons are contributing to conductivity, it can't be a metal! Why is the conductivity 10 times worse than it should be?"
Dessau and Chuang and their colleagues found a clue in their map of the Fermi surface: evidence of a "pseudogap," a phenomenon previously observed in some of the copper-oxide perovskites and suspected of playing a role in high-temperature superconductivity.
Unlike the well-defined energy gaps of most semiconductors, which specify the energy difference that must be overcome if the insulating material is to become conductive, a pseudogap is a more porous boundary between insulating and metallic states. Many electrons that might otherwise be available for conduction are effectively "swallowed" in the pseudogap.
Quasiparticle theory cannot account for pseudogaps, even if fictitious quasielectrons are assigned quantum numbers different from real electrons. The appearance of a pseudogap indicates that additional factors must be at work.
A straight-edged surface
Instead of smooth curves, typical for ordinary solids, a graph of the Fermi surface of lanthanum strontium manganese oxide (and some other perovskites) has remarkably straight segments.
This suggests that, even though the material is in many ways two-dimensional, with conduction occurring in the thin manganese-oxide layers, some conducting particles are confined to a single dimension -- a kind of organization which has been called the "stripe phase."
Such organization might arise if an applied magnetic field physically distorted the lattice structure of lanthanum strontium manganese oxide in a periodic way. By bringing molecular orbitals closer together and farther apart, insulating regions might alternate with one-dimensional "rivers" along which electrons could flow, hopping readily from one conveniently oriented molecular orbital to the next.
Thus several electronic states could exist simultaneously, and odd features like suprising conductivity and the long lifetimes of quasielectrons might be explained by fluctuations among different electronic states.
"Only a small proportion of the electrons in the material would be involved in conduction at any given time," says Chuang, who compares the state of affairs to a short in an electric circuit. "No matter how small the short, current will flow."
By invoking a number of factors operating together, a role for traditional quasielectrons can be preserved. Whether quasiparticle theory continues to survive, as more sensitive investigations of these extraordinary materials progress, has yet to be seen.
"Studying these materials is intellectually stimulating and great fun," says Dessau, adding that "it is also nice that the understanding we gain can better equip us to engineer new materials or devices, with even more exciting or technologically desirable properties."
"Fermi surface nesting and nanoscale fluctuating charge/orbital ordering in colossal magnetoresistive oxides," by Y.-D. Chuang, A.D. Gromko, D.S. Dessau. T. Kimura, and Y. Tokura, appeared in the 25 May 2001 issue of Science magazine.