As horror fiction author Stephen King writes,
everything's eventual, especially death.
In living organisms, the Gompertz law is a single tool that predicts death as a function of age.
For computers, Purdue University electrical and computer engineering professor Ashraf Alam says his team has developed "the first single tool for accurately predicting how new designs for both types of CMOS transistors will degrade" and die over time.
This "unified model" can predict the reliability of new designs for silicon transistors, "a potential tool that industry could use to save tens of millions of dollars annually in testing costs," explained Purdue University spokesperson Emil Venere.
As the Bond Breaks
Electronic devices equipped with CMOS -- complementary metal oxide semiconductors -- use two types of transistors -- PMOS (positive polarity) and NMOS (negative polarity).
Only one circuit type is on at a given time, so CMOS chips require less power than chips using just one type of transistor.
Bonds between silicon and hydrogen are critical to the proper performance of both transistor types.
"Even for the tiniest transistor today, there are perhaps thousands of these silicon-hydrogen bonds," said Purdue engineering doctoral student Haldun Kufluoglu. "But these bonds gradually break."
As the bonds break, the two types of transistors degrade over time, but differently -- like having two hearts, with one aging at a different rate than the other.
"The important point is that the mechanisms by which the silicon-hydrogen bonds break are different for these two types of transistors," Alam said.
If one heart dies, the entire organism dies. Similarly, if one CMOS transistor dies, a computer could crash; an airliner could fall from the sky; or a national security system could fail.
"Over a period of time, when lots of the bonds begin to break, the different transistors within an integrated circuit start getting out of synch," Kufluoglu said.
Separate, But Not Equal
Predicting exactly when one transistor will die relative to the other could be a lifesaving maneuver.
But tests that help make the predictions cost companies millions of dollars annually, Alam said. "If you could explain both within the same framework, then you could cut down significantly on the number of measurements required to characterize the performance of the transistors."
The Purdue model is able to relate these two different classes of degradation simultaneously, Alam added.
The new model is a real breakthrough, because "for the past 30 years, we have treated these processes with separate models," Alam explained. "We didn't know how to put them in a common framework or a common language, mathematically," (continued...)
|