[MLUG - DISCUSSION] Motorola Makes Chip Breakthrough

[Mike Miller <EMAIL:PROTECTED>]

full text:
http://www.nytimes.com/2001/09/04/business/04CHIP.html

excerpts:

N.Y. Times
September 4, 2001

Motorola Makes Chip Breakthrough

By BARNABY J. FEDER

Motorola plans to announce today a breakthrough in the manufacture of
semiconductors that allows high- performance materials like gallium
arsenide to reside on top of standard silicon, a decades-old dream for
microchip designers.

Motorola said that it hopes to license its discovery to other
manufacturers widely and quickly. The technology, which takes advantage of
equipment already in use, will make it possible to make specialty chips
like power amplifiers in cell phones more cheaply as early as next year,
the company said.

"It's a monumental change in the constraints on the construction of
semiconductor systems," said Dennis Roberson, chief technology officer of
Motorola. "We've opened the door on a whole new world."

Over the long term, Motorola expects the breakthrough to lead to speedier
alternatives for a range of chips that now rely on silicon alloys. In
addition, chip designers should be able to invent new integrated circuits
that combine the low-cost robustness of silicon with characteristics of
higher-performance semiconductors, like the ability to emit and absorb
useful quantities of light in wavelengths widely used in communications
equipment.

Such integrated circuits could sharply reduce the price of connecting
high-speed optical networks for Internet and data communications to the
home. They might also spur the development of new wireless devices, like
radar systems that would help automobiles avoid collisions, and new
semiconductor-based lighting systems, according to Motorola.

[snip]

The discovery resulted from an experimental failure. The company had been
trying to create extremely thin transistors in essence, gates for
electrons on top of crystalized silicon out of a material called strontium
titanate, or STO. The process involved bathing the crystal of the
strontium compound in oxygen as it grew on top of the silicon.  As the STO
crystal thickened, however, oxygen slipped through to the silicon below
and combined with it to form a glass- like layer between the two crystals.

The irregular arrangement of atoms in the middle layer impaired the
performance of the STO transistors. Motorola, like others who had been
doing similar work, initially tried to get rid of it. Eventually though,
the researchers noticed that the spacing between the atoms in the STO,
which had been squeezed to align the STO with the silicon crystal, relaxed
as the middle layer formed. The spacing became wide enough for crystals of
gallium arsenide and related materials to be grown on the STO.

The excited researchers soon realized that they had stumbled on to a
possible solution to a constraint that had stumped Government, academic
and industry semiconductor experts for some 30 years.

Silicon, which is basically purified sand, had become the workhorse of the
semiconductor industry because it was plentiful, cheap and relatively easy
to work with. The main method for improving silicon chip performance had
been to shrink components to boost speeds and cut power consumption. In
the 1990's, chip designers also improved performance by growing related
alloys like silicon germanium on top of pure silicon.

What no one had been able to do, though, was deposit materials like
gallium arsenide and indium phosphide, much more efficient semiconductors,
on silicon. As a result, when the performance characteristics of such
materials and their chemical relatives have been needed, users like the
builders of radar equipment, power amplifiers and optical components like
lasers have been unable to take advantage of the economies of scale of the
vast silicon-based semiconductor business. Nor can such chips be easily
integrated with silicon devices.

The barrier, which Motorola says it has solved, was the difficulty of
bridging the difference between the spacing of the silicon atoms and the
spacing in the other materials.

[snip]

--
To unsubscribe, go to http://mlug.missouri.edu/members/edit.php

Archives are available at http://mlug.missouri.edu/list-archives/