The two companies have created an electron microscope that can accurately depict structures measuring less than an angstrom--that is, one ten-billionth of a metre, or less than the width of a hydrogen atom, according to Philip Batson, the researcher who managed the project for IBM's research arm. Their efforts will be documented later this week in the journal Nature.
The high degree of resolution will be important to the semiconductor industry as it continues to push the limits of Moore's Law, which dictates that chips get smaller as time goes on. And as they get smaller, their wires and insulating layers are thinning out to nearly the atomic level.
In turn, this requires that designers know exactly where each atom is located, so that the chips can be set to function properly.
"When we had a layer of insulation that was 100 atoms thick, it didn't really matter if one or two atoms were out of place, but now we are at a place where the insulation layer is 20 atoms thick," Batson said. With the new microscope, researchers will be able to see how materials stick together."
Biologists will also likely be able to use the microscope to more precisely examine the 3D structure of complex molecules, he added. Current electron microscopes have a maximum resolution of 2 angstroms, or 0.2 nanometers. The distance between atoms of silicon--the central ingredient in microprocessors--is 1.3 angstroms. So in current circumstances, designers can't definitively tell if there are two atoms or one in a given space, Batson added.
The new microscope has a resolution of 0.75 angstroms, and its vision will improve over the next few years, he said.
The higher level of resolution largely comes from a complex optical system developed by Nion. In other electron microscopes, four to five magnetic lenses focus a fine beam of electrons. In the IBM-Nion microscope, there are 38 lenses.
Nion developed the optical elements, while IBM worked on controlling the magnetic fields and managing the microscope. The two companies began working together in 2000, and "it took me about a year to prove the subangstrom resolution," said Batson.
Microscopes have played an important part of semiconductor development since the beginning.
In 1971, researchers at Intel were struck with a sense of dread when the first-ever batch of microprocessors failed to work. Federico Faggin, one of the three inventors, quickly slid a sample into a microscope and breathed a sigh of relief when he saw that some circuits failed to connect during the manufacturing process and that the problem was fixable.
