Scientists find more room for chip-making at atomic level
Posted at 2:11 a.m. PDT
Thursday, June 24, 1999
BALTIMORE (AP) -- Bell Labs scientists say they have found what they
believe is the limit for making ever-smaller computer chips using current materials.
Five atoms is the minimum thickness possible for the silicon dioxide
film at the heart of computers -- a part of the transistors that turn on and off at
lightning speed, the scientists report in Thursday's issue of the journal Nature.
For practical purposes, the layer will have to be about twice that
thickness.
Currently, the so-called gate oxide films used are about 25 atoms thick,
said Greg Timp, one of the study's authors.
While the researchers say they believe limit is absolute, the good news
is the limit is lower than had been thought. So, manufacturers will be able to keep
shrinking chips at the current rate for 10 to 12 years, instead of six years using current
technologies, as some had estimated, the authors said.
``This is an extremely important issue to the semiconductor industry,''
Timp said. ``Everyone wants to know how far you can go with the current technology because
no one wants to spend money to rebuild their plants.''
Top-of-the line computers currently sport chips with 600 megahertz of
power. Timp said a chip with the smallest features possible would allow for computer
processing of at least 10,000 MHz.
Such small devices could also find other applications. Such devices
could include a combination handheld computer, organizer, cellular phone, the researchers
said.
Chip makers, who are constantly seeking to make faster, smaller, less
expensive and more energy-efficient devices, have been told before that they would not be
able to shrink beneath certain limits.
This time it may be the truth, Max Schulz of the Institute of Applied
Physics at the University of Erlangen-Nurnberg in Erlangen, Germany, wrote in an
accompanying commentary.
``The new limit to thickness is fundamental, however, and cannot be
simply overcome by technological improvements,'' Schulz wrote. ``The science community and
the semiconductor industry will have to come up with new ideas to avoid a bottleneck in
growth.''
One analyst, meanwhile, said the increasing cost of building smaller and
smaller chips will stop their development before science runs out of ways to shrink them
further.
Drew Peck, a Boston-based technology analyst for SG Cowen Securities,
said increasing costs may force manufacturers to move away from current technologies
before ``we plumb the limits of the physics.''
``The challenge is to continue to push the limits of the technology to
make each transistor smaller without watching costs spiral out of control,'' Peck said.
``Is it possible to build smaller chips in the laboratory? Yes. Is it
possible to build millions of them commercially? Possibly not.''
The cost of building a semiconductor plant is now well above $1 billion,
and has been going up by a factor of 10 every five years, Peck said.
Dave Johnson, head of the Material Sciences Institute at the University
of Oregon, said he agreed with the findings.
``I think they showed that once again silicon is going to be around for
a long time,'' Johnson said.
Others, however, said new roadblocks could pop up for chip makers before
they reach the limit described by the Bell Labs researchers.
Richard Doherty, director of research for The Envisioneering Group,
which advises companies on trends in the semiconductor industry, said the research may
prompt other companies to come forward with developments they have been keeping secret.
``One the one hand, equipment makers will be pleased to hear we can go
another 10 to 12 years, but there are other start-ups chomping at the bit to show there
are other ways to go,'' Doherty said.
