Portland, Ore. -- Harnessing electron spin for optomagneto-electronic devices will depend on materials that, like silicon, can separately adjust their densities and dopant levels to designer specifications. Indium oxide doped with chromium may fill the bill, according to a research team from the Massachusetts Institute of Technology and Boise State University.

Lacking the limitations of gallium arsenide-based ferromagnetic ma- terials, chromium-doped indium oxide could enable durable, transparent thin-film spintronic devices, said researchers at MIT's Francis Bitter Magnet Lab.

"Our system is easy to fabricate on a large scale," said team leader Jagadeesh Moodera, a senior research scientist at the magnet lab. "We can control the carrier density independently of magnetic dopants."

Today, ferroelectric random-access memories use ferroelectric oxides that can permanently switch electric dipole moment without an external electric field and hence are nonvolatile. Bulk ferroelectric materials spontaneously form into nanoscale dipoles that can be electrically switched in the lab from within quantum dots.

FRAMs have been researched by Ramtron International Corp., Texas Instruments Inc., Hynix Semiconductor Inc, Macronix, Samsung Electronics Co. Ltd., Sanyo Electric Co. Ltd., Toshiba Corp., Infineon AG and Celis Semiconductor. TI licensed FRAM technology from Ramtron to reduce the cost of nonvolatile embedded memory compared with flash.

Ferromagnetic memory is attractive because it permits electronic devices to work from a single-layer memory architecture. Recent announcements, such as for Ramtron's new 8051 microcontroller with integrated FRAM (search www.eetimes.com for article ID: 188700327), will permit info to remain stored after power-down and enable single-byte erasures, rather than the block erasures required by flash.

According to the researchers, ferromagnetic memories will need to harness spintronic approaches within a decade as they scale down into molecular-sized magnetic domains to store information. Chromium-doped indium oxide and similar formulations could enable the magnetic spin of even individual molecules to be flipped from "up" to "down," potentially packing a bit of data into every atom.

By the time FRAMs scale down to molecular-sized domains, MIT researchers hope to have chromium-doped indium oxides fully characterized and ready to build spintronic devices. For now, they are merely reporting that they have overcome the basic limitations of other ferromagnetic formulations, especially those using GaAs, by adjusting the location of molecules in the crystalline lattice of indium oxide and by setting the dopant levels separately.

"Ferromagnetic semiconductors based on gallium arsenide cannot separately control carrier density and magnetism," Moodera said, "but we can control carrier density by leaving vacancies [missing oxygen atoms] in the indium oxide lattice, whereas its ferromagnetic properties are mainly controlled by doping with chromium atoms."

The team has found that just 2 percent doping with chromium enables the material to become ferromagnetic. The researchers were able to adjust indium oxide's parameters to create a ferromagnetic metal, a semiconductor and a paramagnetic insulator. They plan to characterize the various formulations before attempting to build spintronic devices.

After characterizing their material, the researchers intend to demonstrate that chromium indium oxides can be used to inject spin-polarized electrons into traditional semiconductor circuitry. By adjusting the carrier density and ferromagnetic moment separately, the team hopes to tune chromium-doped indium oxide for optical and magnetic behaviors as may be required by future spintronic optomagneto devices.

The team also predicts that chromium-doped indium oxide circuits will be ultralow-power because energy will need to be expended only when switching the spin state of a magnetic bit. The optically transparent ferroelectric material could find applications in everything from consumer devices to high-efficiency solar cells.

Moodera performed the work with professor Patrick LeClair at the University of Alabama and professors Alex Punnoose and Byung Kim at Boise State, along with Boise State postdoctoral research associate Kongara Madhusudan Reddy and undergraduate Joseph Holmes; MIT postdoctoral associate John Philip, grad student Tiffany Santos and undergrad Scott Layne (now at Princeton); and Biswarup Satpati, a postdoctoral associate in the Paul Drude Institute in Berlin. MIT, Boise State, the Korea Institute of Science and Technology, the National Science Foundation and the Office of Naval Research provided funding.