June 1999

Incoherent clarity

ORNL’s Z-contrast microscope gives world’s finest atom images

TEM today
These images of a strontium titanate grain boundary show what’s available now (top) with Z-contrast TEM and what Steve Pennycook expects in the near future. Brighter blobs are strontium atoms, weaker are titanium. Just visible in the lower simulation are oxygen atoms.
TEM tommorrow
A “see what we get” transmission electron microscopy experiment has produced the world’s highest-resolution image of a crystal structure, enabling researchers to see atomic details less than an angstrom in size. It’s a doubly improved clarity over previous TEM images.

The atom images of silicon crystals were captured on ORNL’s Z-contrast scanning transmission electron microscope by the Solid State Division’s Steve Pennycook and a former ORNL postdoc, Peter Nellist of the University of Birmingham (U.K.). Pennycook and Nellist are both native Englanders who have found East Tennessee to be very fruitful ground for science.

In Pennycook and Nellist’s images, atomic details only 0.78 angstrom apart can be discerned, which is an improvement over the conventional limit of 1.6 to 1.8 angstroms by a factor of two, says Pennycook.

“In the materials we image, we know where the atoms are; it’s just a matter of resolving,” Pennycook says. “Our Z-contract TEM is the only machine of its type and the only one that could do this.

“Pushing the resolution wasn’t our major aim. We didn’t realize we could achieve less than one-angstrom resolution until we tried it.” Pennycook explains that the TEM’s electron probe can work either “coherently” or “incoherently.” With coherent TEM, the electrons are focused much like a laser; a magnetic field is used as a lens. However, blurring, caused by effects called spherical and chromatic aberration, limits resolution because spurious electrons distort the image.

“It’s like trying to do light microscopy through the bottom of a wine bottle,” Nellist told the Physical Review Letters.

Nellist experimented with incoherent TEM, in which electrons are more broadly applied, more like a light bulb than a laser. As counterintuitive as it seems, the incoherent beam produced finer details in the image.

“That triggered our thinking,” Pennycook says. “The coherent TEM was too sensitive to instabilities, and the image just fades away at high resolution. However, with the incoherent method the spurious electron waves were canceling each other out, providing much better resolution.”

Pennycook and Nellist produced images of silicon-444 (the numbers refer to the planes of atoms in the crystal) in which rows of previously indiscernible atoms can just be made out. Seeing atoms is a very cool thing (they are actually images inferred from the TEM’s data), but why is it a scientifically important thing to do?

“We spend much of our time looking at materials and seeing how they work, particularly grain boundaries in semiconductors and ceramics,” Pennycook says. “Impurities in these structures can change totally the way these materials behave.”

These grain boundaries are interfaces between similar materials. Other interfaces occur where materials have been deposited on other materials to obtain certain properties. Much of superconducting and semiconductor materials research hinges on how successfully these boundaries are structured.

“Grain boundaries are a mystery. Some are strong; some are weak,” Pennycook says. “It’s usually because of impurities in the materials—they can be anything really—that affect the strength of an alloy or the electrical barrier in a ceramic. They can completely change the property of the material.

“You need tremendous sensitivity so see these. Knowing where the impurities occur can show how a boundary is constructed, which can show why it behaves the way it does. We often ‘dope’ a material with impurities to study these effects.”

Pennycook refers to a diagram of a grain boundary: first, a desired structure that is shingle-shaped; second, with an impurity introduced. Because the impurity has varying electrical affinities for the different elements in the material, it bonds to one and not the other, drastically changing the molecular structure.

“Our incoherent Z-contrast imaging is going to be great for a new Basic Energy Sciences Condensed Matter Physics program where we’ll push our resolution limits further,” Pennycook says.

He explained that spherical aberration has dogged the art of microscopy for the past 60 years. Researchers have tried to compensate by resorting to multipole lenses—four to eight magnetic lenses instead of one. However, the data are so immense and ever changing that the lenses are almost impossible to tune. “You just couldn’t set it up,” he says.

Pennycook believes that advances in computers, which can now automatically tune the multipole lenses, combined with the new technique for increased resolution will allow him to hone the TEM probe’s beam of electrons even finer.

“Today we have the smallest beam in the world. Tomorrow it will seems blunt compared with the much sharper and more concentrated, brighter beam of electrons we expect,” he says.

“The new program will test these techniques on our ORNL microscopes, using the older Z-contrast microscope first and then the newer one. We hope to get a half-angstrom resolution from it.”

Although Nellis returned to Cambridge after his postdoctoral term at ORNL a few years ago, he has returned in the summer to experiment, including this summers to work again with ORNL’s Z-contrast TEM lab.

“We have the highest sensitivity in the world for detecting impurities,” Pennycook says. “We predict that by shrinking the probe we’ll be able to see oxygen, in fact all atoms, wherever they are—every atom of every element. We’ll have single-atom sensitivity, even the ones we don’t expect to see.”—B.C.

Powerful Z-contrast TEM also has staying power
ORNL obtained its Z-contrast transmission electron microscope in 1988, and a bigger and better model in 1993. The Solid State Division’s Steve Pennycook provided specifications for that specially built, 300-kilovolt direct-imaging machine. At the time, it was the most powerful microscope in the world.

“Z-contrast” refers to protons: Atoms with higher atomic numbers, and thus more protons, show up brighter in the TEM’s images. Z is the designation for protons.

Years have passed—Pennycook became a Lockheed Martin corporate fellow in the meantime—but ORNL’s Z-contrast TEM can still claim to be the world’s finest. That’s pretty amazing in a world where state of the art technologies reach obsolescence with dismaying rapidity. Pennycook has a pretty straightforward reason for the scope’s staying power: “The company that built it, VG Microscopes in England, was bought out and the new owner stopped making them. It seems that the microscope was too specialized to generate a profitable number of orders.

“It makes keeping the microscope running and furnished with spare parts a challenge,” Pennycook says. “But other companies are catching up.”

With interest in materials research increasing, Pennycook should expect no shortage of things to look at with this unique instrument.