http://www.eetimes.com/story/OEG20000731S0019
The researchers recently discovered that errant bacteria survive in the
cleanest of clean rooms by inducing the top semiconductor layer on chips to
grow over them, thereby embedding themselves inside the chip.
When SUNY researcher Robert Baier started to investigate the role bacteria
play in the yield problem, he knew they were avoiding even the most stringent
attempts to eradicate them — the bugs just wouldn't die. How could bacteria
survive where no other living thing can? Baier, funded by the National
Science Foundation, found the answer, but it wasn't what he expected.
"When we started this study, we were just trying to find the source of
bacteria in the fab, and how they could remain alive after all the heroic
measures to eradicate them with ultraviolet light, ozone and everything else
including a dollar a gallon to purify the water," said Baier, who is director
of the Center for Biosurfaces at SUNY.
Other scientists and engineers participated in the research at the Center for
Microcontamination Control at the University of Arizona, the Rensselaer
Polytechnic Institute in New York and the Center for Environmentally Benign
Semiconductor Manufacturing at the University of Arizona.
The problem wasn't people in the fab getting sick — those kinds of bacteria
were easy to kill. Rather, it concerned some clever bugs that just wouldn't
die, no matter what — bacteria that can survive in the vacuum of space, or
inside a volcanic vent at the bottom of the sea. They can hibernate
indefinitely and only need the slightest bit of light to wake up and thrive
anew.
"We found that these extremely hard-to-kill bacteria were coming in with the
ultrapure water, and the way they survived our calculated assault was to
capture a tiny bit of semiconductor that had dissolved in the ultrapure water
and start growing," said Baier.
Once the bacterium sticks a molecule of semiconductor to itself, other
dissolved crystals spontaneously attach themselves to the formation, growing
islands atop silicon wafers during a subsequent vapor deposition step.
In short order, the bacteria have encased themselves inside armored shells of
semiconductor, making them impervious to all the attempts by clean-room
personnel to kill them. "These bacteria can cause a lot of problems in the
clean room, like shorting out adjacent lines on chips, and inside these
armored shells they are almost impossible to kill," said Baier. "Now we are
turning a problem into a feature. A plant is basically a single-electron
photonic device converting light into electricity. If we embed a
photosensitive bacteria inside a chip, we have the beginnings of a
biotransistor."
Baier's goal of harnessing bacteria as the active element in a transistor may
not be as far-fetched as it sounds — at least his theory sounds convincing.
He points out that copper is a conductor because it has one free electron per
atom to contribute to current flow. Semiconductors are called "semi" because
they have only about one free electron per thousand atoms, depending on
doping levels. Current flow in those semiconductors, unlike copper, can be
precisely controlled by parameters that match the parameters of bacteria,
according to Baier, enabling regular transistors to switch from an insulator
into a conductor by changing state.
'Many uses'
These small charge transfers, Baier contends, are just what happens in common
biological processes like respiration and photosynthesis. In fact, he
believes that the current flowing in a semiconductor can be controlled by the
chlorophyll in a single cell. For instance, when light shines on a
photosensitive bacterium, it yields up an electron that could be used to
switch a primitive biotransistor. "This is a new class of biochips, which I
believe can be adapted to many uses, but at present it's at a primitive
stage, like the crude crystal detectors that preceded today's radios," said
Baier.
The theory is that doping semiconductors is always done to disrupt the
perfect lattice, making free electrons or "holes" available in "n-" and "p-"
type semiconductors, respectively. Likewise, if a biological atom, say
phosphorus with five electrons from a bacterium's cell, is doped into a
silicon crystal that only needs four, it then makes the fifth electron
available, enabling biological cells to serve as metabolic "sources" and
"sinks" for electrons and "holes" in biotransistors.
"Biological membranes have been known to develop potentials of a million
volts per centimeter. They are perfect for semiconductors, and biochips using
them could be made as small as five microns on a side," said Baier.
According to Baier, most of the steps used to build chips today will continue
to be used in manufacturing his biochips. For instance, masks will be used on
the semiconductor doped with bacteria, as will diffusion, sputtering and
other common deposition techniques. All the other common devices like
resistors and capacitors will be built into circuits connected to the
biotransistors.
Optical amps
Biotransistors will also simplify optical communications by amplifying
optical beams the way a normal transistor amplifies electrical current. And
Baier envisions the construction of light-sensitive heterojunctions where
lattices of different energy gaps are "biodoped" so that their crystalline
lattices mesh imperfectly, creating atomic-scale defects and strains with
useful photonic and electrical side effects to drive circuitry. Such
heterojunctions would harness complicated biophysical reactions to provide a
tunable variable for design problems.
For now, Baier will be satisfied if he can build a "crystal" radio with his
biochip architecture. His approach will be to follow the recipe of Bardeen
and Brattain when they invented the world's first transistor. As with the
original, Baier plans to attach two fine wires only micrometers apart, but
this time in a biodoped germanium crystal that has been bonded to a metal
disk. The assembly will be housed inside a metal cylinder electrically
grounded to the disk. Bardeen and Brattain's "cat's whisker" will be
connected to a meter to register success or failure.
This archive was generated by hypermail 2b29 : Mon Oct 02 2000 - 17:35:47 MDT