Crystal of the Week: Silicon!
It’s the fifth annual National Robotics Week celebration, so those of us celebrating International Year of Crystallography realized we could not miss out! After all, at the National Science Foundation, robots are a big deal! In conjunction with the U.S. Navy and Woods Hole Oceanographic Institution, our Division of Ocean Sciences helped fund recent upgrades and certifications for the retrofitted and rebooted Alvin underwater remotely operated vehicle. Research funded through NSF’s Directorate for Engineering embraces robots big and small: Bin He at University of Minnesota has developed a flying, mind-controlled craft, while Sarah Bergbreiter at the University of Maryland College Park designs insect-like robots smaller than a penny! And through the Advanced Technological Education program, NSF funds continuous improvements to associate degree-granting tech programs throughout the country where thousands of students at these schools learn how to build their own robots!
So how do crystals add up to robots? It’s all in the chips — the microchips, that is. For this week’s Crystal of the Week, we focused on the one material that almost all robots have in common – silicon-based microchips (a.k.a., the integrated circuit). Microchip fabrication technology has a long history of NSF support, and many major advances in microchip engineering were developed in American laboratories.
Pure silicon is an electronically neutral insulator (electrically non-conducting). Silicon has a similar electronic structure to carbon, which means the crystal structure of silicon takes on the same form as diamond, called “diamond cubic.” Large, cylindrical, high-purity single crystals of silicon are grown from molten vats of silicon (1500oC!) in a method called “pulling,” also known as the Czochralski process. Various improvements in this process have increased the diameter of the cylindrical silicon crystals that can be pulled from the melt from a once-standard 6 inches to a behemoth 12 inches! Once the large cylindrical crystals have been formed, they are sectioned into very thin slices (less than 1 mm), called silicon wafers. The resulting silicon wafers must be as much as 99.9999999% pure for many applications, including solar technology. However, microchip technology requires silicon wafers to have electrically conducting regions, in addition to the insulating pure silicon.
Those conducting regions are incorporated into silicon chips during subsequent processing steps. Two types of conducting silicon can be fabricated: “n-type” and “p-type,” where electrons and holes are the respective charge carriers. Doping with phosphorus creates “n-type” silicon, whereas doping with boron creates “p-type.” Other dopants can also be used. When silicon is doped, each dopant atom replaces a silicon atom in the lattice. Because dopants can be used to induce conductivity in silicon, the material is known as an extrinsic semi-conductor, i.e., silicon conducts under certain conditions.
Looking at robots from this crystallographic perspective is a bit different. But then again, like so many of our greatest technologies, robotics rely on the most minute details to work successfully and find new ways to move technology forward in trailblazing fashion—clearly a case where the whole robot is far greater than the sum of its very smallest parts.