During the past 30 years, microchip companies have developed manufacturing
processes by combining some understanding of their underlying physics and
chemistry with a tremendous amount of iterating.
The result has been a wildly successful set of manufacturing techniques that
have given us VCRs, cell phones, minicomputers in everything from
automobiles to toaster ovens and desktop computers with far more computing
muscle than NASA had when it sent men to the moon.
Engineers call this kind of educated guessing the "Edisonian Approach." It_s
named for Thomas Edison, who tested more than 2,000 substances before
finding a filament for light bulbs.
The Edisonian approach has worked so well that when microchip manufacturers
started to develop gas phase technology _ which uses volatile gases to clean
wafers between production steps _ they saw no reason for things to change.
In fact, ten years ago industry pundits confidently predicted that gas phase
cleaning technology would be a mainstream manufacturing process by 1999.
It didn_t happen _ largely because it is just too complex for Edisonian
So gas phase research efforts have shifted to university laboratories, where
engineers are methodically working through the atomic-level details to
develop a deeper understanding of the basic science involved.
"Industry is no longer set up to do this kind of basic science," says
Anthony Muscat, who conducts his research in the NSF/SRC Engineering
Research Center for Environmentally Benign Semiconductor Manufacturing. The
center, which is administered from the University of Arizona in Tucson, is
funded by the National Science Foundation and the Semiconductor Research
Corp. It includes researchers from Arizona State University, Stanford, MIT,
Cornell, UC Berkeley and MIT_s Lincoln Lab.
"This kind of basic research is perfect for an academic institution because
it involves deeply and methodically analyzing the chemical and physical
properties of surfaces to gain a complete understanding_ and, therefore,
control _ of the processes," says Muscat, an assistant professor of chemical
and environmental engineering.
"We are trying to engineer the microchip wafer surface on a molecular
scale," he explains. "And the strategy is to develop gas phase chemistries
that will remove contaminants and leave the surface with just the right
chemical, material, and electronic properties as liquid phase chemistries
based on water do today." The contaminants include organic molecules,
metals, polymers, and various oxides that are left over from processing
steps that either etch the microchip wafer surface or deposit materials that
make up its microscopic circuitry.
Manufacturers are focusing on gas phase cleaning technology because it
promises to be cheaper and makes possible smaller, faster devices.
Muscat and his students conduct the research inside vacuum chambers that
mimic the emptiness of deep space. This allows them to know what is in the
chamber and on the wafer surface down to the last molecule.
Then the trick is to find a gas mixture that will combine with the surface
contaminants to produce volatile byproducts. These byproducts evaporate off
the surface, and when the gas is pumped out of the chamber, the contaminant
molecules go with it.
To study the chemistry and physics underlying these reactions, Muscat
bombards the surface with x-rays, electrons, or infrared light. "We know
that an electron energy or absorbance of certain wavelengths of light
characterizes certain atoms or molecules that are bound in a certain way on
the surface," he says. "So the electron and light energy serve as
fingerprints telling us what atom is bound to the surface, how it is bound
and how much of it is there."
When Muscat and his students evaluate these processes, they consider not
only the technical merits, but also the potential environmental problems.
Developing environmentally friendly manufacturing processes is part of the
mission for the Environmentally Benign Semiconductor Manufacturing Center,
which is directed from UA by Farhang Shadman, a professor of chemical and
"We want to advance technologies in parallel with minimizing their
environmental consequences," Muscat says. "This is not an either/or
proposition. We_re designing completely new chemistries from the ground up
and we take the environment as another design constraint. We haven_t found a
chemistry yet that is so superior that it is worth environmental
degradation. In terms of cleaning, we want to develop promising chemistries
that use fewer resources _ water, energy and hazardous chemicals. Moreover,
we want to train students in this way of thinking
about new technology."
Muscat notes that the semiconductor industry has made advances in
environmental protection without being forced to do so by regulatory
"Electronics manufacturers fund this center because they believe we can push
the technology and make the processes environmentally acceptable at the same
time," Muscat says.
While Muscat and his students pursue the basic science of cleaning
technologies, it will be up to tool manufacturers to integrate that science
into marketable processes and equipment.
"After we_re finished with the basic science, an enormous amount of
engineering work still will be needed to scale these processes to production
levels," Muscat says. "The tool manufactures are set up to do this and
eventually will sell the tools and processes to the microchip