By color-coding atoms, new Cornell electron microscope promises big advance in materials analysis

25-02-2008: A new electron microscope recently installed in Cornell's Duffield Hall is enabling scientists for the first time to form images that uniquely identify individual atoms in a crystal and see how those atoms bond to one another. And in living color.

"The current generation of electron microscopes can be thought of as expensive black and white cameras where different atoms appear as different shades of gray," explained David Muller, Cornell associate professor of applied and engineering physics. "This microscope takes color pictures - where each colored atom represents a uniquely identified chemical species."

The instrument is a new type of scanning transmission electron microscope (STEM), built by the NION Company of Kirkland, Wash., under an instrument-development award to Cornell from the National Science Foundation (NSF). John Silcox, the David E. Burr Professor of Engineering at Cornell, and Ondrej Krivanek of NION are co-principal investigators on the project.

The microscope incorporates new aberration-correction technology designed by Krivanek that focuses a beam of electrons on a spot smaller than a single atom - more sharply and with greater intensity than previously possible. This allows information previously hidden in the background, or "noise," to be seen. It also provides up to a hundredfold increase in imaging speed.

The capabilities of the new instrument in analyzing a test sample are described in an article in the journal Science by Muller, Silcox, Krivanek and colleagues at Cornell and in Korea and Japan.

It allows scientists to peer inside a material or a device and see how it is put together at the atomic scale where quantum effects dominate and everyday intuition fails. One of the most important applications of the new instrument will be to conduct what Silcox calls "materials pathology" to aid researchers in their development of new materials to use in electronic circuits, computer memories and other nanoscale devices. "We can look at structures people have built and tell them if they've built what they thought they did," Silcox explained.