BUFFALO, N.Y. By applying atomic-dimension "nanocontacts" to magnetic media, an experiment at the State University of New York here has revealed the potential of an effect known as "ballistic magnetoresistance." The tiny metal contacts showed a 3,000 percent change in magnetoresistance at low switching fields of a few hundred oersted.
The effect, observed by SUNY professor Harsh Chopra working with doctoral candidate Susan Hua, has the potential to generate huge increases in hard-disk drive density. Whereas "giant magnetoresistance" (GMR) read heads have enabled hard drives to approach the theoretical limit of magnetic media, about 20 Gbits per square inch, ballistic-magnetoresistance heads could redefine those limits, enabling atomically small domains to pack terabits per square inch.
"What we do is take a few atoms called a nanoconductor or nanocontact and pass some current through it," said Chopra. "But because it is such a small conductor and has special properties regarding how the electrons pass through it, what we found was a 3,000 percent change in resistance."
Researchers are trying to boost the magnetoresistance effect, which has created a revolution in disk-drive data densities, while also working on ways to reduce the size of the sensing probe. Ballistic magnetoresistance (BMR) would solve both problems at the same time.
Rapid gains in the magnitude of the magnetoresistive effect have continued since its discovery in the late 1980s. At first a 10 or 20 percent change in the electrical resistance of metallic sandwiches placed in magnetic fields was observed. The discovery of this GMR effect kicked off the development of a new kind of disk-drive read head that could sense very small magnetic-field changes, allowing smaller magnetic domains on rotating disks and hence much higher storage capacities.
About 10 years later a new family of perovskite-based materials (lanthanum-magnesium-oxide) was found to change its percent resistance in orders of magnitude. This was called "colossal magnetoresistance."
Other groups, Chopra said, have been shooting for BMR effects for several years, reporting changes of around 300 percent, but Chopra's device is the first to jump an order of magnitude, as has been theorized to be possible for BMR.
Hard-disk drive read heads originally used magnetic induction to sense magnetic domains. But soon after Peter Gruenberg, of the KFA research institute (Julich, Germany), and Albert Fert of the University of Paris-Sud discovered in 1988 that magnetized metallic layered systems would change their electrical resistance by as much as 50 percent in the presence of small magnetic fields, read heads based on the effect began to appear. The storage capacity of disk drives grew by 60 percent a year during the 1990s as a result, and that rate is now at 100 percent a year.
Nanometer domain walls
Instead of relying on layers of magnetic insulators, BMR read heads would depend on the odd behavior of electrons traveling through a conduit that is atomically small. The nanocontacts created by Chopra and his team reduce the size of the magnetic-domain walls within it to just 1 nanometer.
"When current passes through such a small group of paramagnetic atoms nickel in our case they behave quite differently than when traveling through a normal conductor," said Chopra.
Because the nanocontacts are so small, he said, they have very different interior magnetic-domain walls the regions between oppositely oriented magnetic domains than those of a normal conductor. The width of a domain wall in a normal bulk conductor is about 100 to 200 nm, but the domain wall of a nanocontact must fit inside the nanoconductor itself. Therefore, the length of the nanoconductor is essentially the width of the domain wall only about 1 nm.
Electrons forced through such a small conduit literally do not have space to bounce around inside the conductor, producing heat, as in a normal conductor. However, with a nanocontact's 1-nm-wide domain walls, electrons stream through, creating a draftlike effect that prompted the label "ballistic" in BMR.
The electrons stream though at ballistic speeds only when the magnetic field is not switching (and heads are at their lowest resistance). However, when the field switches and the heads are at their highest resistance, the electrons must switch magnetic orientation as they pass through the domain wall. In GMR heads, electrons assume many transitional, in-between magnetic orientations as bits take their time, gradually flipping magnetic orientation while traveling over the 100-nm width of the domain walls. However, with a nanocontact's nanometer-wide domain wall, electrons must switch in a very short space, accounting for the theoretically high resistance of BMR heads.
In future BMR read heads, the extremely high resistance changes would result from the media domain forcing electrons to switch magnetic orientation in just 1 nm an almost impossible feat, according to Chopra.
"In principle, you should be able to get infinite resistance so that no electrons are passing through [when the magnetic orientation is switching], because a spin down would have to switch to a spin up in only 1 nm," he said. "But of course, nonideal conditions make it just a very high resistance instead of infinite."
Chopra constructed the nanocontacts using the same method that serves to make scanning tunneling microscopy tips. The technique, electrodeposition, lays down a dot of nickel atoms in a 1-micron gap left between two 100-micron-thick nickel wires, one of which has been "sharpened" to an atomically sharp point with a 40-nm diameter. By covering the wires with epoxy, except where the sharp point meets the second wire, and monitoring the resistance between the two wires during electrodeposition, Chopra was able to control the size of the nanocontact. He claims to be able to grow any size nanocontact needed for an application, from 1 to 10 nm.
"We have a self-terminating procedure now. If you tell me that you want a nanocontact of a certain radius, then we can give it to you each and every time," he said. "Let us say you want a nanocontact whose radius is 5 nm or 2 nm; you can actually preset the equipment in such a way that the deposition automatically stops at the right time, because it is controlled electronically."
Chopra foresees other applications of his nanocontacts beyond disk drives. As sensors, they could be fabricated in arrays on chips to create a kind of nanoscale magnetic probe for sensing nearly any size magnetic domain in the environment right down to molecular sizes.
"These nanocontacts are so small that you could fabricate thousands of them on a chip. You could also make a device that measures the magnetism of individual molecules, because they are of comparable size," said Chopra.
Chopra's next step is to use e-beam lithography to incorporate his BMR nanocontacts onto test chips so that his research group can more fully characterize the behavior of nanocontacts, in preparation for tackling real-world applications.
"It takes about a decade for basic scientific discoveries to be realized in applications," he said. "With BMR, there is still so much that is unknown such as what is the nature of these very small domain walls? What are their dynamics? How fast can they move? And how can you control the morphology of such small nanocontacts as they are used over time?"
Morphology the study of the form and structure of the heads without regard to their electrical functions should yield answers to handling read heads that consist of only a few atoms, for instance, or how to attach wires or prevent wear. The researchers are preparing three more papers that will describe the various details of their discovery, as well as offer some preliminary characterization of nanocontacts within a working circuit.
An audio recording of reporter R. Colin Johnson's full interview with Harsh Chopra can be found online at AmpCast.com/RColinJohnson.