North Carolina State University researchers have developed a technique that sheds new light on how silicon bonds with various other materials at the atomic level. 

Dr. Kenan Gundogdu, co-author of the study and assistant professor of physics at North Carolina State University along with Dr. David Aspnes and Bilal Gokce have created a new method for observing how silicon bonds with different materials at the atomic level in an effort to understand and control bond formation at this level. This could lead to the invention of more efficient microchips and other new devices.

Silicon-based devices are built using layers of various materials, and these materials obtain their distinctive characteristics from the bonds, which are the chemical interactions between adjacent atoms. Bonds act as a "glue" that holds atoms together, and has the ability to determine material characteristics like transparency and hardness.

"Bonds are formed as materials come together," said Gundogdu. "We have influenced the assembly process of silicon crystals by applying strain during bond formation. Manufacturers know that strain makes a difference in how bonds form, but up to now there hasn't been much understanding of how this works on the atomic level."

Now, through the use of optical spectroscopy, researchers are able to observe what happens when strain is applied to a silicon crystal at the atomic level. 

"Application of even a small amount of strain in one direction increases the chemical reactivity of bonds in a certain direction, which in turn causes structural changes," said Gundogdu. "Up to now, strain has been applied when devices are made. But by looking at the effect on the individual atomic bonds, we now know that we can influence chemical reactions in a particular direction, which in principle allows us to be more selective in the manufacturing process."

While this new development sheds light on these atomic bonds, Aspnes notes that further research is needed to "identify the relevant hidden variables" in order to make silicon-based devices more efficient. 

This study was published in Proceedings of the National Academy of Sciences in September 2010.