Non-Statistical Fusion Reactions
In Atomic Scale Accelerators
Brian T. Donovan*
Brian T. Donovan
This is an abstract
for a poster to be presented at the
Foresight Conference on Molecular Nanotechnology.
There will be a link from here to the full article when it is
available on the web.
A sub-nuclear resolution microscopic accelerator just a few
centimeters long and only a few nano-meter wide might
theoretically initiate individually controlled fusion reactions
without requiring thermonuclear temperatures or confinement.
In the last few years, researchers have created precise atomic
structures using atomic force microscope equipment (piezo
electric and magnostrictive crystals used to position sensing
probes and atomic position probes). Recently researchers at IBM
have created an atomically perfect ring of copper atoms to create
an electron standing wave pattern. Other researchers have created
various other atomic scale structures such as
nano-meter-sized-tubes and Buckminster Fullerenes.
Variations of these structures might be assembled to create an
atomic scale particle accelerator capable of accelerating a
tritium nucleus and a deuterium nucleus and accomplishing their
collision with sub atomic, sub-nuclear, or even sub-proton,
accuracy. Conventional fusion reactors are all based on confining
the reaction products at a high enough temperature and pressure
to achieve a statistical probability of enough high energy random
collisions to induce a sustainable fusion reaction. In contrast,
individual atoms collided with sub-nuclear accuracy can be
induced to fuse with only 51 KeV of energy.
Figure 1 shows a somewhat schematic cross sectional view
of the basic atomic scale accelerator. The underlying crystal
substrate and numerous technical details have been left out.
I understand that this is a highly speculative idea. Apart
from the many possible theoretical design problems, formidable
obstacles must be overcome before even one device could be built
and even more formidable obstacles must be overcome before the
millions of parallel units necessary for any significant power
output can be built. Nonetheless, the great potential for
microscopic fusion reactors deserves attention.
Nuclear fusion with nuclear accuracy atomic scale accelerators
does not require extremely high temperatures and confinement
required by thermonuclear reactions:
Thermonuclear reactions take advantage of E=(3/2)kT. Thus,
using a standard formula for deriving temperature from
average molecular collision energy, assuming 10 KeV
collisions taking place in a 1015 density plasma
the temperature must be:
T=2/3(104 eV)(1.6x10-19 J/eV)/1.38x10-23J/K
T=77M degrees K
In practice, an ignition temperature of 400M K is needed to
compensate for lost energy. For plasmas of 1015
density, this incredibly high temperature equates to particle
collisions of a relatively modest 51 KeV. This energy can be
imparted using standard electrostatic acceleration or one can add
laser or microwave assisted acceleration. Existing 51 KeV
accelerators can impart this much energy to colliding ions, but
cannot guarantee that many of the ions will actually collide
because the beam is wide and atoms are small. If the atoms can
individually be accelerated and aimed with sub-nuclear precision,
then a high percentage of the ions can collide and fuse. If even
only 10% of the ions accelerated actually collided and fused a
large net energy gain would be realized.
The potential electrostatic repulsive energy of deuteron
centers 3 nuclear radii from each other is 2.72x105 eV
from a standard handbook.
E=kQ1Q2/R. Q's are in coulombs and R in
meters. k = 9x109
Thus 51 KeV will force deuteron nuclei to overlap closely enough
to fuse. By comparison, Tokamaks have reached the equivalent of
20 KeV and the NOVA laser system has reached only 3 KeV particle
Vacuum Breakdown voltage is 1.25MV/cm in vacuum or 125 KV/mm.
Thus a 50 KeV accelerator might be as small as 406 Ám long. At
these small sizes, other factors may strongly influence the
practical length required.
Various recent technological achievements indicate that the
accuracy for constructing and aligning the accelerator can be
high enough. As mentioned earlier, various atomic structures have
been constructed in the laboratory. Several sensing methods also
have resolution in the nuclear range, that might facilitate their
use for alignment. Capacitance micrometry, for instance, is a
very sensitive method for detecting small displacements. This
method works by measuring the change in the impedance of a
parallel plate capacitor as the spacing or area of the parallel
plates changes. Displacements as small as 10 femtometers (10-14m,
about the diameter of an electron, or 10,000 times smaller than
an atom) can be measured using this technique .
The Deuteron ion trajectory can be electrostaticaly deflected.
The atoms in the accelerator can be accurately placed on the
substrate crystal lattice. The thermally induced vibrations of
the accelerator atoms will by averaged out over some 109
atoms. The electrostatic potential on the deflection plates can
be adjusted with very high precision including down to the adding
and subtracting of individual electrons. The actual frequency of
occurrence of fusion events can be used as feedback for the
accurate deflection of the accelerated particles.
We can start by looking into direct assembly using piezo or
magnostrictive elements and structures similar to those used in
Scanning tunneling microscopes and used by other researchers to
assemble nanny structures (IBM electron corrals). Much more
advanced nano-assemblers will be required and some other method
will be needed for any kind of economic mass production.
Alternatively substructures can be built using various techniques
and then assembled. Subassemblies might include microtubules 
and thin film coatings (such as MBE). Micro structures could be
mass fabricated for most of the accelerator with some misplaced
atoms. Then the output end of the accelerator might be assembled
atom by atom. This level of accuracy may not be needed since the
electric potentials will be used to aim the beams and thus
compensate for any small misalignments.
Choice of atomic reaction:
Compared to Deuterium-Deuterium the Deuterium-Tritium reaction
ignition temperature is only 4.4 KeV for a reaction energy of 14
MeV vs. an ignition temperature of 48 KeV. for reaction energy
around 3.7 MeV. However the Deuterium-Tritium reaction puts out
high energy neutrons that destroy the reactor chamber and make it
radioactive. The possibility of accelerator based reactions
reliably achieving high energy nuclear reactions suggests the
possibility of looking at other nuclear reactions that are less
radioactive then the deuterium-tritium reaction.
The Deuterium-Deuterium reaction is a good reaction to start
with since it is well understood and creates less radioactivity
then tritium deuterium reactions. Another disadvantage of tritium
is that it must be made in a reactor using lithium and thus is
not as common as deuterium.
It would be, of course, be good to eliminate any
radioactivity. The following reactions are reported to be
1H1 + 11B5 => 4He2
+ 8.68 MeV. b
Boron 11 and 1H1 are the dominant
isotopes and both are common.
Another possible radiation free reaction is:
2H1 + 3He2 => 4He2
(3.6 MeV) + 1H1 (14.7 MeV)
Both reaction products are charged so that direct electrical
energy recovery is possible. He3 is rare on Earth: 1.4 atoms per
M atoms air. Moon rocks and Jupiter and Saturn have more of it.
Generating useful amounts of power
The energy from each reaction for the Deuterium-Deuterium
reaction products is 4 and 3.3 MeV. Taking 3.7 MeV as the energy:
3.7 MeV =2.884x10-18Joules. Thus 3.468 x1017
reactions per second must take place to generate 1 W of energy.
Given the speed of the particle particles = [(2*Volts*1.6x10-19)/mass].5,
1 , generating 1 watt with a single accelerator requires a line
of 10 meters of hydrogen ions packed 2 radii apart per second. 1
W is also an extremely large amount of energy to dissipate for
one nano-meter sized reaction chamber. Thus multi-watt devices
will require many multiple accelerators running in parallel.
The energy theoretically available from fusion is of course
vast. A 3 GW fusion plant would require .455 grams/day of
deuterium, from 13 k Kg/day of sea water. Thus 100KW, a 100 HP
car or a couple of American homes for instance, would require 15
mgm/day from just .433 kg of sea water per day.
References And Notes
 (FEBRUARY 1994 LASERFOCUSWORLD)
 J. J. McClelland et al., Science 877, (931105)
 (Science News '94)
 Nature 93 IBM Switzerland:researches make electron coral
 Laser Focus world 93. (40 nm chromium rows.)
Brian T. Donovan, postal address with Zip, phone, fax, email: email@example.com