Intel Airs New Type of Transistor

It’s almost like going back to the drawing board, but not quite; it’s more akin to rewriting what’s on it. Intel Corp.
continued its push for faster, cooler and altogether more efficient processors Monday with the release of a chip
that banishes performance limits brought on by heat and power consumption.

The leading chip manufacturer posed its new steps as part of an “effort to maintain the pace of Moore’s Law and remove the technical
barriers that Intel and the semiconductor industry have only recently begun to identify.” Moore’s law states that that the number of
transistors per square inch on integrated circuits doubles every year.

Santa Clara, Calif.’s Intel said it would go into more depth about the new processor structure, called the TeraHertz transistor because it will be able to switch on and off more than one
trillion times per second, at the International Electron Device Meeting (IEDM) in Washington D.C. on Dec. 3.

“Our research has shown that we can continue to make smaller and faster transistors, but there are fundamental problems we need to
address around power consumption, heat generation, and current leakage,” said Gerald Marcyk, director of components research, Intel
Labs. “Our goal is to overcome these barriers and produce chips that have 25 times the number of transistors of today’s
microprocessors at ten times the speed with no increase in power consumption.”

Specifically, Intel will reveal a new type of transistor called a “depleted substrate transistor” and a new material called a “high
k gate dielectric.” Combined, these solutions will offer the power needed for the future chips of mobile devices without letting
transistors overheat and short out circuitry.

Intel has created the “depleted substrate transistor,” — a construct in which the transistor is built in an ultra-thin layer of
silicon on top of an embedded layer of insulation. This layer is designed to create maximum drive current when the transistor is
turned on, allowing it to switch on and off faster. Conversely, when the device is shut off power leakage is reduced in this
architecture because the layer is so thin.

In terms of groundbreaking material, Intel is casting aside traditional silicon dioxide for “gate-dielectric,” a material that
separates a transistor’s “gate” from its active region. This, Intel said, is because silicon dioxide, though effective for today’s
products, consumes too much power. The new material reduces gate leakage by more than 10,000 times compared to silicon dioxide.

The high k gate material is grown by technology called “atomic layer deposition” in which the new material can be grown in layers
only one molecule thick at a time.

Intel expects to include elements of this new structure in its chips by 2005.