WHEN NASA鈥檚 Cassini probe blasted off in October 1997, hundreds of protesters
gathered outside Cape Canaveral to demonstrate against the launch. The problem
was 150 kilograms of plutonium packed into the craft鈥檚 nuclear batteries. During
Cassini鈥檚 long journey into the icy depths of space, the plutonium would be
needed to produce several hundred watts of power. But should the launch fail, or
should NASA miscalculate the trajectory of the craft as it returned to fly past
Earth in August 1999, the spacecraft could be destroyed and the plutonium
released into the atmosphere.
In the event, everything went without a hitch. But now the anti-nuclear
protesters who gathered to send Cassini on its way have something else to worry
about. Engineers making the tiny silicon-based devices known as
MEMS鈥攎icroelectromechanical systems鈥攁re also looking for a power
source for their creations. And some think nuclear batteries could be just the
thing.
MEMS will be created by the billion. And even if only a small proportion of
them are nuclear powered, this could mean hundreds of thousands or possibly
millions of nuclear batteries being used in all kinds of places. Some MEMS
devices are even being designed to be spread on the wind like dust. Most
worrying of all, there is little consensus on how and where these batteries
should be used. And if they are distributed widely, what are the potential
consequences for the environment and our own health?
Advertisement
The idea of carving tiny machines from silicon to work as sensors and
actuators was dreamed up in the late 1950s by the Nobel prizewinning physicist
Richard Feynman. Today MEMS are appearing everywhere from motion sensors to
computer joysticks and accelerometers. With the potential global market
estimated at up to $14 billion a year, even the most pessimistic believe
that MEMS devices will soon have a major impact on our lives.
But these tiny machines are greedy for power. That shouldn鈥檛 be a problem in
applications where they can be linked to the electricity mains, or to power
supplies such as a car engine. But many of their potential uses are in places
where power is hard to come by. MEMS-based pressure sensors in smart tyres, for
example, will have to work for years, broadcasting data at regular intervals.
For this, conventional power sources such as bulky batteries simply won鈥檛
do.
鈥淭he physics is working against you,鈥� says Paul McWhorter, a MEMS expert at
MEMX, a New Mexico company that designs optical switches based on MEMS. The
power that a battery can provide depends on its volume, and it drops
dramatically as devices get smaller. The major drain for many MEMS sensors is
transmitting the data they produce. 鈥淚t takes a finite amount of power to
broadcast data over a certain distance and that doesn鈥檛 change with the size of
your device,鈥� says McWhorter.
One solution is to siphon energy from the environment, a trick perfected over
200 years ago by the Swiss clock maker Abraham-Louis Perrelet, who invented the
self-winding watch. In July this year, Stephen Beeby at the Department of
Electronics and Computer Science at the University of Southampton unveiled the
21st-century version of Perrelet鈥檚 idea: a device a little larger than a sugar
cube that picks up the vibrations from its surroundings, absorbs their energy,
and turns it into electrical power. The MEMS device is essentially a magnet on a
minute springboard inside an inductive coil. Twang the springboard and the
movement of the magnet induces current in the coil. With a number of
springboards of different resonant frequencies, the device can siphon power from
a wide range of ambient vibrations. It can produce up to 1.5 milliwatts, which
is more than enough for many applications, says Beeby.
His ambition is to harness vibrational energy to power sensors buried deep
inside engines or embedded in the fabric of bridges or roads. Most of the time
these devices would be dormant, simply absorbing energy from their surroundings.
Then, when enough power had been stored, they would wake up, make a measurement,
and broadcast it to a receiver.
Beeby鈥檚 devices could even harness the energy of human movement, like the
self-winding watches that inspired them. Gilbert Schiltges, a mechanical
engineer with Disetronic Medical Systems in Burgdorf, Switzerland, suggests that
within five years his company鈥檚 insulin-delivery devices, which are themselves
no bigger than a credit card, could be powered in this way. This would remove
the need for expensive batteries that now have to be replaced every three or
four weeks. Pacemakers are another potential application.
A similar idea is being pursued by Wen J. Li and his colleagues at the
Advanced Microsystems Laboratory at The Chinese University of Hong Kong, using a
traditional helical spring. Their device is currently under wraps while the team
applies for patents. An array of them could one day recharge a mobile phone, Li
suggests.
Another promising non-nuclear power source for MEMS is sunlight. At the
University of California, Berkeley, Kris Pister is creating 鈥渟mart dust鈥�. Each
mote is a silicon sliver designed to take a reading from its surroundings,
process it, and beam the data out for collection. It can also pass the data on
to its neighbours so the cloud can act like a distributed network of processors.
He envisions releasing them like dust into the environment to keep watch for
chemical and biological weapons, or monitor growing conditions on farmland. He
has already tested a 鈥渃loud鈥� of 800 motes, each just 5 millimetres across, that
measure temperature and relay the data back to him.
The trouble with solar power is that the energy yield is limited by the
surface area receiving sunlight. So as MEMS get smaller鈥攅ventually Pister
hopes each mote will be less than 2 millimetres across鈥攐utput drops
dramatically. At the moment, Pister鈥檚 motes are powered by tiny solar panels
only a few millimetres square. 鈥淥utdoors I can generate 100 microwatts per
square millimetre but indoors it鈥檚 only 1 microwatt,鈥� he says. And 1 microwatt
is just 5 per cent of the power he needs, which means each device must sleep for
95 per cent of the time, storing energy until it has enough to take a
measurement and transmit the results.
These devices鈥� thirst for power is pushing Pister towards the nuclear option.
鈥淲e are about to start a project looking at radioisotope thermal generators,鈥� he
says. Just a cubic millimetre of polonium-210 produces 1 watt of heat and has a
half-life of 138 days. Other isotopes produce less energy but last far longer.
Nickel-63, for example, has a half-life of a century, but generates just 1 per
cent of the power of polonium-210. But even if more than 90 per cent of this
energy is lost in the conversion to electricity, such a power source could
change the landscape for MEMS designers. 鈥淭he amount of power available is
awesome,鈥� says Pister.
At the University of Wisconsin-Madison, Jake Blanchard and his colleagues
have built a number of nuclear microbatteries, funded by almost half a million
dollars from the US Department of Energy. Forty years before Cassini blasted
off, the US Navy tested large nuclear batteries as power sources for remotely
operated buoys. Now Blanchard is investigating a more refined鈥攁nd much
smaller鈥擬EMS version.
He is currently testing a battery that generates current by bombarding a tiny
semiconductor diode with beta radiation鈥攈igh-energy electrons, in other
words. The diode is made up of two layers of silicon: one, called n-type
material, is doped with an element that gives it an excess of conducting
electrons, while the other, called p-type material, has a deficit of electrons.
When beta radiation strikes the junction, it knocks electrons out of the n-type
material which then flow across the junction. This is similar to the way photons
generate a current when they hit photovoltaic materials.
To squeeze as much current as possible from the device, Blanchard needs to
maximise the area of the diode in contact with the radioactive material. This is
tricky because solid radioactive materials are difficult and dangerous to shape
on this scale. So he uses a liquid made with the beta-emitting isotope
nickel-63, which he pours into a series of fine channels in the top of the chip.
The resulting microbattery (see Diagram)
produces only a few nanowatts, but it proves the principle works, Blanchard says.
The next step should be to raise the power output, but this is difficult.
Silicon begins to break down when the energy of the electrons hitting the
lattice rises above 250 kiloelectronvolts (keV). While nickel-63 produces
electrons with a maximum energy of 66.9 keV, there are few other isotopes that
won鈥檛 damage the silicon.
So Blanchard intends to go back to the more conventional approach, and simply
transform the heat created by radioactive decay into electricity. To do this he
uses thermocouples鈥攄evices in which two junctions, each made from a pair
of dissimilar conductors, are held at different temperatures. The output is
determined solely by the temperature difference between the junctions, so
thermocouples can run with any radioactive material regardless of the energy of
the particles it emits. 鈥淵ou can get fantastic amounts of energy out of these
devices,鈥� says Pister.
So far, Blanchard鈥檚 prototype of this simple device produces a few tens of
nanowatts. This is by no means enough for Pister鈥檚 purposes, but it should be
possible to scale it up by increasing the number of junctions in each device,
says Blanchard.
The few environmentalists aware of the work are unconvinced that nuclear
generators will ever be a practical power source for MEMS. 鈥淭his has more to do
with [scientists鈥橾 desperation to promote nuclear technology than any genuine
breakthrough,鈥� says Karl Grossman, an anti-nuclear campaigner and author of
The Wrong Stuff: The space program鈥檚 nuclear threat to our planet. 鈥淚
sympathise with their problem, but this is dangerous stuff. Do we really need to
take these risks?鈥� he asks.
Blanchard insists that the risks are negligible. His minuscule batteries are
designed so that the nuclear material cannot escape and he is adamant that
anyone unfortunate enough to breathe in or swallow one would receive a total
radiation dose well within recommended safety limits. 鈥淲e鈥檝e paid special
attention to environmental concerns, which is why we鈥檙e not trying to boost the
power at this stage,鈥� says Blanchard.
But will the public accept them as safe? Perhaps it will be possible to sell
the idea of MEMS-based nuclear batteries as power sources for small hand-held
devices such as computers or PDAs. But Blanchard also talks about using nuclear
microbatteries for powering MEMS devices similar to Pister鈥檚, and sprinkling
them 鈥渓ike breadcrumbs鈥� onto battlefields to detect chemical weapons, for
example. Tiny nuclear-powered sensors could also be mixed with oil or grease and
added to the lubrication system in heavy machinery to detect when maintenance is
needed. 鈥淥ur batteries would work well for devices that must have a really long
life. We could build them now if there was the demand,鈥� Blanchard says. So what
about the protesters? Are they already up in arms at the prospect of nuclear
batteries being scattered far and wide? 鈥淚 expected to get opposition to the
idea,鈥� says Blanchard. 鈥淏ut I haven鈥檛 heard a single complaint.鈥�
Nevertheless, Pister intends to be cautious. 鈥淚 wouldn鈥檛 want these things
spread around my planet,鈥� he says. 鈥淏ut I wouldn鈥檛 mind them on other planets.鈥�
One promising application for nuclear microbatteries is to power swarms of
nanospacecraft that could be released into the atmosphere of Mars or Venus, or
put into orbit around the Earth. Their power source would allow scientists to
build up a picture of environmental conditions over huge areas of a planet or in
a huge volume of space for years to come. And if a few chips were to fail, who
would miss them?
鈥淚f none of this works out, the MEMS revolution will continue quite happily,鈥�
says Pister. 鈥淏ut if we do find ways of powering these devices remotely, the
applications could be bigger than all the others put together.鈥� For the moment,
the public and many environmentalists remain strangely quiet about the prospect
of tiny nuclear batteries being sprinkled around their planet. But they may not
stay silent much longer.