The method promises to be important not only in the quest to understand the operation of future computer circuit and data-storage elements as they shrink toward atomic dimensions, but also to lay the foundation for new materials and computing devices that leverage atom-scale magnetic phenomena.

"We have developed a window into the atomic heart of magnetism," said Andreas Heinrich, research staff member at IBM's Almaden Research Center in San Jose, in a statement. "We can now position atoms and then measure and control their magnetic interactions within precisely designed structures."

IBM researchers expect to use the technique to explore the limits of magnetic data storage, by engineering the energy required to flip the collective orientation of a small number of magnetically coupled atoms.

In addition, the researchers hope to determine the feasibility of spin-based wires and a spin version of the molecular-motion cascade, which IBM scientists successfully applied in 2002 to form a working computer circuit some 260,000 times smaller than its conventional design in silicon.

Also slated for study is how engineered spin interactions could be applied to quantum information systems, such as quantum computers.

Spin-excitation spectroscopy uses IBM's low-temperature scanning tunneling microscope designed for use with magnetic fields up to 140,000 times stronger than the earth's. The researchers first move atoms into position and then measure the interactions between their atomic spins, which are the fundamental sources of magnetism.

In their experiments, the IBM researchers created chains of up to ten manganese atoms atop an extremely thin electrically insulating surface and measured how the magnetic properties changed as each new atom was added. They found that chains with an even number of atoms had no net magnetism, while chains with an odd number of atoms showed net magnetism.

"This kind of exploratory research is essential for the long-term future of the computer industry," said Gian-Luca Bona, manager of science and technology at IBM Almaden. "Sometime in the next couple of decades, it will be impossibly difficult to continue improving transistors and other traditional microelectronic circuit elements by simply shrinking them. We will then need alternative structures and, perhaps, altogether different ways of computing. Techniques like this can help us gain the knowledge needed to create those alternatives."

The new technique builds upon the IBM team’s development in 2004 of spin-flip spectroscopy-- a method to measure magnetic properties of single atoms. Those experiments were done using an aluminum oxide insulating surface, which was not conducive to atom movement.

The experimental results are published in today's issue of Science Express. The report's authors are Cyrus Hirjibehedin, Christopher Lutz and Heinrich.