New Scientist 13 December 1997 p32
Forget the idea of endless energy. People could soon be using a fusion machine the size of a football to detect bombs at airports and impurities in mineral ores.
Bennett Daviss reports.
A hollow stainless steel sphere about the size of a football sits on a laboratory bench. From inside the globe, a purple glow radiates through a small glass window. George Miley peers in and glimpses a tiny luminescent ball hanging in the centre and spires of light that seems to radiate from it. “It’s a beautiful sight,”Miley sighs.
Miley is a professor of nuclear, electrical and computer engineering at the University of Illinois in Urbana-Champaign and his sphere is a fusion machine. Unlike other fusion machines, this one is small enough to sit on a desktop, it can be switched on and off at will and it produces no radioactive waste.
The sphere is different in other ways too. Its primary purpose is
not to make energy but to generate neutrons. Billions of them, every second.
Neutrons are subatomic particles with no electric charge that are
extraordinarily useful. Scientists use them for materials analysis-they
can help to identify the most common elements in seconds. By contrast,
chemical analysis can take hours. neutrons can also help to work out the
structure of new molecules and crystals. Beams of these particles can
even be used for cancer treatment.
The trouble is that neutrons are notoriously difficult to make. Nothing less than a nuclear reactor or a high powered particle accelerator will do the job. This means that neutron analysis can only take place with the help of a handful of specialised laboratories. Until now.
Miley is about to begin selling his spheres. The University of Illinois has licensed the technology to Daimler-Benz Aerospace, which in return is helping to finance his research. Text year, the spheres will go on sale for as little as $60.000-a tiny fraction of the cost of the nuclear reactors or particle accelerators that are now needed to produce neutron beams. Likely users include mining companies, who will be able to spot impurities in an ore as it is being mined, specialist metal smelters who will be able to monitor the composition and quality of their alloys in real-time, and airport security staff who will use the neutron beams to spot bombs in suitcases as they pass by.
Miley wants to go even further. The holy grail of fusion research is to create a source of cheap, clean energy. Although his spheres currently use up much more energy than they produce, he says they have the
potential to generate more in future. And unlike any other type of fusion research, he hopes to fund his work with the profits from the desktop neutron generators. “By marketing inexpensive neutron generators, we hope we can finance the work that might transform these devices into cost-effective power generators,” he says.
Fusion spheres were first conceived more than 40 years ago by Philo Farnsworth, an inventor who developed much of the technology behind early televisions. But Farnsworth was never able to reap the financial benefits of his inventions, and in the late 1950s he started work for the defence and electronics company, ITT, based in Fort Wayne, Indiana. There he set out to create the ultimate energy generator by fusing nuclei together. What makes Farnsworth’s idea different is the way he chose to initiate and control the fusion reaction.
The conventional approach is to take deuterium – a harmless, stable
isotope of hydrogen that has a neutron as well as a proton in its
nucleus, and heat it to many millions of degrees. This strips away the
electrons from individual atoms, leaving a soup or “plasma” of positively
charged nuclei and negatively charged electrons. At these temperatures, a
small proportion of the nuclei have enough energy to fuse with each other
when they collide, forming a highly energetic neutron and nuclei of
hydrogen, tritium and helium-3 in the process.
Sometimes researchers use a mixture of deuterium and another isotope
of hydrogen, tritium, which has two neutrons and a proton in its nucleus.
This reaction results in a far higher number of neutrons but tritium is
highly radioactive and difficult to handle. In either case, the big
problem is how to keep the plasma contained. It is much too hot for any
ordinary container to withstand, so it is held in place by a magnetic
Farnsworth chose a different route. Instead of heating deuterium
gas, he used an electric field to accelerate individual ions to the
energy at which they would fuse, in the same way that electrons are
accelerated towards the screen in a television tube.He argued that this
was far more efficient than heating an entire volume of deuterium,
particularly when only a small por- tion of the gas would reach the
energies required for fusion. He then aimed several beams of deuterium
nuclei towards the centre of a sphere where he hoped they would collide
and fuse. Of course,the beams have to be precisely aligned for any chance
of fusion to occur, and even then only some of the nuclei actually
But this is not the only chance they have to fuse. Farnsworth
calculated that the build-up of positive ions near the centre of the
sphere would attract negative electrons. This shell of negative charge
would then trap the positively charged ions at the centre, multiplying
the chances that they would collide and fuse. To distinguish it from the
traditional magne- tic confinement technique, Farnsworth’s method of
fusion is known as inertial electrostatic confinement (IEC).
When Farnsworth retired in 1967, four years before his death,he and
a newly minted physics PhD fron1 the University of Illinois named Robert
Hirsch seemed to have proved the notion workable. Measuring the energy
released in the conventional way – as the rate at which a fusion reaction
liberates neutrons -Hirsch’s final machine, which was fuelled by a
deuteriuni-tritium mixture, delivered more than 10 billion neutrons per
second, a generous number even by today’s standards. Unable to raise
enough noney to continue the work, Hirsch joined the US Atomic Energy
Commission in 1968 and event- ually became the director of its fusion
program. He now works as a power technology consultant in Washington DC.
But before Hirsch gave up IEC fusion research, he had developed an
entirely new way to accelerate his ions – this time using pure deuterium
in place of the deuterium-tritium mixture. In his original experinments
Farnsworth used accelerators around his sphere to fire in beams of
nuclei. Hirsch replaced this arrangement with a spherical grid roughly
the size of a tennis ball, made of wire 1 millimetre thick.
To begin the fusion process, Hirsch admitted a small amount of
deuterium gas into an evacuated sphere. Next, he set up an electric
potential of 60.000 volts between the grid and the outer sphere, setting
up an electric field that is strong enough to ionise the deuterium gas.
The field then draws the positive ions towards the grid. Some of the ions
collide with the wires of the grid and play no further part in the
process. But others pass between the wires and enter the region within
it. There, a significant proportion collide in the centre of the sphere
and fuse, producing neutrons and energy.
When Hirsch left fusion research,he donated one of his experimental
spheres to his alma mater, the University of Illinois. Two decades later,
Miley returned to the concept and developed the rudimentary desktop
fusion device that now sits on his bench. The neutrons stream out of the
sphere in all directions, because they carry no electrical charge, and so
cannot be directed by electric or magnetic fields. “Neutrons go anywhere
they want to. That’s why increasing the yield and efficiency is so
important,” says Miley. His initial yield was 10^6 but has now reached
10^9 neutrons per second.
The design of the spherical grid is crucial to the success of
Miley’s fusion machine, and he has spent several years perfecting the
design. The wires combine to create an electric potential shaped like the
surface of a sphere. “Originally, we thought we had to make this sphere
as smooth as possible,” says Miley. This meant a tightly wound grid with
small gaps between the wires. “But eventually we found a better
design”.The grid Miley now uses is like the lines of latitude and
longitude On a globe and has large holes.This produces dimples in the
basically spherical electric field. “It’s more like the surface of a golf
ball,” says Miley. Only those deuteriurn ions that are heading for the
exact centre of each dimple are accelerated towards the middle of the
sphere. This leads to the loss of many of the deuterium ions, but turns
out to improve the machine’s perform- ance over-all. The reason for this
is that the grid promotes the formation of beams of ions that avoid
hitting the wires, preventing corrosion. This is a key advantage for
“Now we have all sorts of calculations and modelling to explain why
this particular arrangement of wires aligns the ions into near-perfect
beams,” Miley says, “but at the time we just stumbled onto it.” The form
of the grid is the key break-through that makes Miley’s sphere work. “It
is this that makes the design so forgiving. You no longer need the
precise align- ment that Farnsworth struggled with,” he says.
His spheres already produce neutrons more cheaply and safely than
existing methods. One of the most common of these is nuclear fission in
which atoms of heavy metals are split apart, liberating neutrons inside a
nuclear reactor. But nuclear reactors are expensive, complex and
potentially dangerous machines. What’s more, they produce highly
radioactive waste, which must be disposed of safely. By contrast, Miley’s
machine produces less radioactive waste in a year than is contained in a
single luminous cinema “Exit” sign.
Another approach uses a particle accelerator to shoot ions of
deuterium into a metal target impregnated with tritium. But this is an
expensive option too. Because tritium is a radioactive gas, it requires
special handling facilities and licences. Particle accelerators are
complex machines, and the target must be rotated and cooled to prevent
overheating and damage. Yet another technique uses high-energy protons to
chip neutrons out of heavy metal atoms. Known as spallation,this
technique is perhaps the most expensive of all: its advantage is that it
can produce intense neutron beams of up to 10^16 particles per second.
Reactors and accelerators are big machines, so samples to be
analysed usually have to be taken to the neutron source rather than the
other way round. One way to provide a more portable neutron source is to
use a reactor to manufacture californium, a radioactive element that
produces neutrons when it decays. But this solution has its
limitations:californium produces no more than 10^7 neutrons per second
and requires special hand- ling facilities for the gas.
“Then there’s the big problem,” notes physicist Richard Nebel of the
Los Alamos National Laboratory in New Mexico. “When you’re done using a
radioactive material, you can’t turn it off.” Californium continues to
decay whether the neutrons it produces are needed or not. Within two
years, it must be replaced with a fresh batch. Nebel is part of a team at
Los Alamos that is developing spherical neutron generators. In addition
to spherical grid machines like Miley’s, they are considering another
method that works on a different principle. In place of a grid, they use
a pulsating field tuned to the resonant frequency of the plasma to give
deuterium nuclei the energy they need to fuse.
Being safe and easy to use is what gives machines like Miley’s an
edge over today’s neutron sources, says John Sved,an engineer with
Daimler-Benz Aero- space. “Many of our potential customers are concerned
about liability – the ‘Chernobyl effect’, if you will,” Sved says. These
companies do not want to own nuclear reactors or handle radioactive
gases, because of the risk of contamination. “If there were a fire or an
accident in their plant and these isotopes were set free, they would face
a contamination problem.”
With Miley’s desktop neutron generators they avoid these risks. They
are fuelled by harmless deuterium, and the only waste is helium-3 gas, a
whiff of hydrogen and negligible traces of radio-active tritium. ” A
small IEC neutron generator could run for decades without creating enough
radioactive waste to exceed minimum regulated levels,” Sved says. “The
machine could be completely consumed in a fire and there would be
virtually no concern about escaping radiation.” And to allay fears about
even these small amounts of radiation, Daimler-Benz plans to remove the
tritium from the spheres safely each time they are recharged with fresh
And there should be no shortage of buyers, Sved predicts. When
neutrons collide with atomic nuclei, they generate gamma rays with an
energy signature unique to that material. “If you wanted to check airport
luggage for bombs, you could bombard the area in question with neutrons.
If a 10-megaelectronvolt gamma ray comes back, that’s a signature of
nitrogen, a primary constituent in virtually every form of high
explosive,” explains Nebel. The technique – known as neutron activation
analysis – is often faster, neater and more thorough than messy chemical
assays and other conventional forms of analysis. The problem until now
has been finding a safe, portable source of neutrons.
Miley believes that the first customers for his neutron spheres will
be manufacturers of high-quality alloys in which traces of impurities
have a huge effect on the properties of the metal. “With a neutron source
they could measure the composition of the metal as it was made,” he says.
Mining companies would be another potential buyer. A neutron source would
allow them to measure the impurities in an ore as it is being mined, or
the proportion of minerals remaining in mine waste and whether it is
worth reprocessing. To protect workers who might be around the machine
for extended periods of time from the neutron bombardment, the machines
would need to be screened with a material such as boron that captures
neutrons. “But this is straight- forward, says Miley. He and his team use
a simple shield around the sphere to protect them.
The practical applications envisaged for these machines require a
yield of between 10^7 and 10^10 neutrons per second. Miley believes his
spheres will fall well within this range, but he hopes to produce higher
rates in future. “There’s a ladder of applications,” Sved notes. “The
larger the volune of neutrons you produce, the faster the analyses you
can perform; the faster the analysis, the more uses neutron scanning can
be put to – as long as the devices are cost-effective and of manageable
At very high yields of 10^16 neutrons per second, medical
applications become possible. Researchers in Japan and America have
developed a cancer treatment in which patients are first given a boron
isotope that lodges in their tumour. Bombarding the boron isotope with a
beam of neutrons produces energetic ions of helium, which in turn destroy
the tumour. While this treatment is only possible today at the small
number of clinics where these high yields of neutrons are available,
Miley’s machines could bring it to many more hospitals.
But boosting the neutron yield to these levels will not be easy.The
spheres will have to be larger, and the wire grid will have to cope with
thousands of amps rather than the small currents it handles today. Since
fusion releases energy, the grid and the sphere will have to be cooled,
probably by pumping water through cooling tubes. The vessels themselves
could be protected by lining them with heat-resistant cladding developed
by Daimler- Benz Aerospace to protect spacecraft re-entering the Earth’s
Then there is the “confinement problem”. In Farnsworth’s device,
nuclei that failed to fuse were quickly lost. Ideally, the energetic
deuterium nuclei should continue to circulate, retaining their energy,
until they collide and fuse. This is the problem that magnetic
confinement fusion researchers have struggled to solve. Miley argues
that his method is more efficient than magnetic confinement, which has to
heat a large mass of plasma to give just a small proportion of the nuclei
sufficient energy to fuse. “A temperature of about 22 C is really a
distribution of velo- cities with a mean energy of about O.02
electronvolts,” says Miley. To give a significant number of nuclei an
energy of 10 kiloelectronvolts needed for fusion requires a temperature
of millions of degrees.
Miley’s design follows the more efficient strategy of accelerating each deuterium ion individually to this energy, and gives each ion at least 10 chances to fuse before it strikes the grid or picks up an electron and drifts away as a neutral atom. But this is still not enough even to reach higher yie1ds, says Miley. “To economically deliver 10^10 neutrons per second or more we need to confine the ions long enough to give each perhaps a thousand chances to bounce back inside the grid and fuse,” he says.
That would also be major step towards the goal of making a fusion energy generator. “To make economical fusion energy, we’d have to give each ion not thousands, but tens of thousands of chances to fuse if we’re going to reach the energy breakeven point. The question is, how many times can we bounce an ion back and forth before it gets lost?”
Only future experiments with large spheres will tell. For the moment, the physicists can do little more than speculate. But one thing is for sure: with fusion power as with television, someone else will reap the benefits of Farnsworth’s pioneering work. “But wouldn’t it be the perfect vindication if his ideas solved the world’s energy problems?” Nebel muses. “If that could happen, I’d like to think that somewhere he’d have a smile on his face.”
Bennett Daviss is a freelance science writer