PORTLAND, Ore. Magnetoresistive materials being developed by researchers at the University of Chicago and Argonne National Laboratory may offer a promising alternative to the semiconductor industry's current chip-making processes. Instead of relying on atomic accuracy, theorists say that the researchers, who have harnessed nanoscale imperfections in a material, could use it as a thin film for chip fabrication.
"The semiconductor industry is based on the fact that you control things beautifully-exquisite control-but there are other approaches to developing devices. Depending upon your underlying physics, you may be able to allow nature to do some of the work for you," said physics professor Thomas Rosenbaum at the University of Chicago's Materials Center, working with researcher Marie-Louise Saboungi at Argonne National Laboratory.
In 1997, Rosenbaum's collaborative research group discovered a "disorderly" material that behaved like a colossal, linear magnetoresistor with an immeasurably wide dynamic range. By randomly lacing nanoscale filaments of silver in sulfur, selenium or tellerium, the researchers were able to change its resistance linearly, as a function of the strength of the magnetic field it detected.
"If you have a perfect material, where you've forced it to be exactly two parts to one, [for example], then you wouldn't get this effect. We are harnessing the lack of homogeneity of the material" said Rosenbaum. "I think, intellectually, that is the important part."
The magnetoresistive material with ultrawide dynamic range could theoretically be fashioned as a thin film suitable for chip fabrication. Magnetic sensors using it would be able to measure from a small fraction to over 1 million gauss.
"If you take a compound that is two parts silver and one part tellerium, then it won't do anything interesting when you apply a magnetic field. But if you put in just a trace amount, say one part in 10,000, then all of a sudden it responds hugely to a magnetic field, it changes its resistance, its electrical properties-the ability for electrons to flow-and it does it in a way that is unusual.
"You would usually expect a quadratic behavior with magnetic fields that would just stop changing at a certain point. And what we discovered was that not only was it linear, but that it didn't stop changing up to the highest fields we measure, almost a million gauss," said Rosenbaum.
The material family, dubbed silver chalcogenides (Ag2S, Ag2Se, Ag2Te), only responds well to a magnetic field when trace amounts of silver are partnered with sulfur, selenium or tellerium. What's most unusual is its wide dynamic range and linear response-making it unlike almost any other material.
"If you look on the scale of, say, hundreds of atoms, there could be a lot of silver there, so it looks very conducting. But if you go to another region, there's virtually no silver-so it's highly resistive. And the really beautiful thing about this is that the system self-organizes in this way," said Rosenbaum.
His group at the University of Chicago pioneered the material by characterizing the magnetoresistive effect. Then the group got the 2003 Nobel prize winner for the physics of type-II superconductors, Alexei Abrikosov, to advance a complicated and refined theory, based on quantum-mechanical effects, to explain the silver-chalcogenide magnetoresistive effect.
"Abrikosov's theory is essentially three-dimensional. It's not just a two-dimensional simulation, and it matches some of the data well," said Rosenbaum. "It's a terribly complicated theory, but Abrikosov speaks of the quantum energy levels as being very important."
A simulation of a two-dimensional thin-film version of the material was demonstrated more recently by University of Cambridge researcher Meera Parish with professor Peter Littlewood. Their advantage was that their model could potentially be used to begin fabricating magnetoresistive sensor chips.
"The Parish-Littlewood simulation is beautiful, but since it's a 2-D model, it can't address all the issues of our 3-D material," said Rosenbaum. "Theorists have fertile imaginations, and can describe the same results in different ways. It's not a question of whether you subscribe to one or another-the question is whether they yield clear predictions that can be tested experimentally."
The benefit of the Parish-Littlewood model, regardless of its merit as a theory, is that it proposes a strategy by which semiconductor makers could begin to design a new class of magnetoresistive materials to be fabricated as thin films.
Parish and Littlewood simulate a resistor network in their model of silver chalcogenides, which has been done before. But instead of using normal two-lead resistors, these resistors are shaped like flat disks with four leads coming out in the four directions of the compass. This permits two voltage drops to exist between different sets of opposing leads from the four-lead resistors. Here the key is that the voltage drop induced by the externally sensed magnetic field is perpendicular to the direction that normal current is flowing through the device.
"If you think of a resistor as a linear element, then it is very hard to understand how you can get a voltage drop perpendicular to the direction of current. But if you think about it like a disk-a two-dimensional object as opposed to just a line-then you can see why [in the presence of a magnetic field] you get a voltage drop in a different direction from the current flow," said Rosenbaum.
In the presence of a magnetic field, something called Hall's coefficient measures the off-diagonal term of the resistivity tensor, according to Rosenbaum. "If you put a magnetic field perpendicular to a current, then the current wants to run in a circle, so the notion is that the material builds up a charge to oppose the magnetic field, which is trying to make the charge carrier flow in a circle."
This electrical field, perpendicular to normal current flow, is a Hall effect called an induced "Hall voltage" drop across the "extra" perpendicular leads of the model's four-terminal disk resistors. If the applied magnetic field is sufficiently large, current can actually be induced to run around in a circle in the x-y plane (called a cyclotron orbit).
"The key fact here is that the Hall voltage is linear in the magnetic field, so if you have enough of these funny current patterns-because each resistor is different, since each little filament is different on the nanoscale-then you can be dominated by the Hall effect.
"That is what the new model shows-that if you have enough disorder, that is, enough differences among the resistors, then you wind up with a linear magnetoresistor," said Rosenbaum.
For the future, the Parish-Littlewood research group vows to get into the experimental side of research by enlisting the help of some physics researchers at the University of Cambridge. Together, they hope to prove that their two-dimensional model can be implemented in thin films.
"We are planning to begin working with University of Cambridge experimentalists, instead of just doing more simulations. We want to see our theory realized," said Parish.
For the future, Rosenbaum hopes to conduct new experiments too-ones that will make different predictions for the Parish-Littlewood model. "One thing we would like to do is look at the longitudinal resistivity," said Rosenbaum. "Now we apply the magnetic field perpendicular to the current, but the question is, if you put the magnetic field parallel to the current, what is the response of the material? Here the two theories have different predictions, so in terms of testing the theories, that is the next step I would like to take."
Some thin-film results are already here, Rosenbaum claimed. It seems the most promising areas of application for the material may be as high-field-strength sensors, he said.
"Next, we want to look at reduced dimensionality-what would nanowires of this material look like?" said Rosenbaum. "Length scales in the problem are very important. One way to control length is just to force it, to confine it, which might emphasize the quantum effects if they are the important ones. . . . I'd like to see how the effect changes as a function of size, and if we can enhance the effect by going to lower dimensions."
Rosenbaum also plans to more fully characterize the material. For instance, some preliminary measurements of the material indicated that it might also have giant thermopower responses.
"There were some measurements made showing a giant thermopower effect with magnetic fields, probably because the filaments of silver are there, because it's known that [filament-like] carbon nanotubes also have a large thermopower effect," said Rosenbaum.
The Materials Center, working with Argonne National Laboratory, is conducting neutron and X-ray diffraction studies to determine whether modulation is responsible for the magnetoresistance.
In this simulation of a silver chalcogenide in a high magnetic field, color represents voltage levels and the black arrows represent currents, which loop around to generate large magnetoresistance.