Over the Summer of 2007, physics major Andrew Grzankowski worked with Longwood faculty member Keith Rider (Chemistry) to construct a Farnsworth Fusion Reactor. Here’s a breakdown of the project.
Introduction to Fusion
Nuclear fusion reactions result when two light nuclei collide with very high kinetic energy and fuse to produce a heavier nucleus. When two deuterium nuclei fuse, the reaction products are helium and a fast moving neutron. The kinetic energy of the neutron is much higher than the initial kinetic energy of the deuterium, so fusion could, in principle, be used to produce cheap, clean energy.
Philo T. Farnsworth is credited with conceiving the first Inertial Electrostatic Confinement (IEC) fusion device. His is best known for inventing the first fully electronic television and demonstrating in publicly and the first electron microscope. In the early 1960’s, Farnsworth’s IEC device was comprised of spherically symmetric (other symmetries are possible) ion guns focusing towards a center and the ions were further accelerated by single grid. Robert L. Hirsch worked with Farnsworth in the late 1960’s to construct more practical devices that did not utilize ion guns but two concentric spherical electrodes introducing the concept of recirculation; increasing the number of high-energy ions that pass through the center. This type of IEC device is more common among amateur scientists and academic research institutions. Having the chamber act as one of the two concentric electrodes presents some electrical safety concerns but is still practical. This also decreases the amount of grid damage that becomes a limitation in the Hirsch IEC device.
There are many fuel choices for an IEC fusion device, and the best is not often the most abundant, inexpensive, or safe. Examples include the Deuterium-Tritium, Proton-Boron-11, and Tritium-Helium-3 mixtures. We chose to use the simplest fuel source, the Deuterium-Deuterium (D-D) mixture, which has its own disadvantages and advantages. It is a neutronic fusion fuel, meaning that one of the byproducts of this reaction is a high-energy neutron capable of doing severe damage to animal tissue. However, with the right precautions in place D-D can be used by any amateur experimenter. Deuterium is also widely abundant on earth and can be easily extracted by heavy water electrolysis; and therefore, is very cheap.
The Project and Its Construction
Before beginning this project, electrical safety procedures were addressed. Voltage and current in the order of that required for fusion is very dangerous; it is possible to be severely injured and can prove to be fatal. Radiological precautions were also discussed. Fast neutrons, like the byproducts of D-D fusion, can be damaging to biological molecules such as DNA. It is also possible at such high voltages to encounter x-rays if the device is not properly shielded.
A crucial link in the vacuum system that connected two different size flanges together was intentionally not purchased new for this project allowing ample education in machine shop and scrounging techniques. This component was full nipple reducer between a common flange standard and an archaic flange standard, and included a tube. The common flange was a CF450 and the uncommon flange was an ANSI/ASA Lightweight 2. The uncommon flange was made from a 304 stainless steel plate. This flange was first cut from the plate with a plasma torch and then was finished in the machine shop. The tube was machined from a piece of 304 stainless steel DOM tubing. The lathe was used to remove excess material around the side of the flange, on both faces, and to bore out a center for the tube to be inserted. It was used to remove material from the tube and bore out the center to be equivalent in diameter with both flanges. The vertical mill was used to bore out all bolt holes in the stainless steel components and used in creating the mounting apparatus that holds the vacuum chamber.
The cart was acquired and is modified to hold the device with rough pump, diffusion pump and cooling system, gauge controllers, power supply, neutron detector, and video camera (this does not include the baffle cooling system, gas tank and lecture bottle). Aluminum plates were chosen to support the device mounting setup and gas handling valves. Clamping bars were machined on the vertical mill to have the same curvature of the tube under the chamber and clamped by stainless steel bolts.
The rough pump is a BOC Edwards RV12 and provides sufficient vacuum at the diffusion pump outlet that is required for its proper function. The rough pump was drained of oil and refilled with Varian brand direct drive pump oil. The diffusion pump is a CVC PMC-2 and capable of reaching pressure far beyond that of a rough pump. This link provides a description of how a diffusion pump works. The diffusion pump was drained of oil, fully disassembled, subject to numerous chemical washes to remove all old oil and other impurities, reassembled, and filled with new Santovac 5 pump oil that has a much lower vapor pressure than the originally issued oil. The diffusion pump is continuously cooled by tap water in order to ensure ideal operation and prevent overheating (which will eventually ruin the oil). A peristaltic pump provides low pressure cooling water with a very precisely controlled flow rate, which helps to maintain the diffusion pump temperature in the correct range.
Three separate gauges are used in our system to measure the pressure inside the chamber and the backing pressure between the diffusion pump and the rough pump. The Convectron gauge can accurately report pressures from 5000 Torr down to 10-3 Torr. The gauge tube contains a temperature compensated heat loss sensor which utilizes conduction cooling to sense pressure at lower pressures. At higher pressures, it utilizes convection cooling in which gas molecules are circulated through the gauge tube by gravitational force (SISWEB.com). The Ion gauge can accurately report pressures from 10-3 Torr down to well beyond 10-12 Torr. The hot cathode ionization gauge is composed of three electrodes contained in a glass enclose; the collector plate, filament, and grid. Electrons are emitted from the filament and make their way towards the grid. They will eventually reach the collector plate after several collisions that result in electron ionization of the molecules not yet evacuated from the chamber and the pressure is estimated by measuring the ion current (Wikipedia.org). The Thermocouple gauge can accurately report pressures from 1000 Torr down to 10-3 Torr. The pressure is indicated by measuring the small voltage of a thermocouple spot welded directly onto the hot wire. The wire is fed with a constant current and its temperature depends on the thermal conductivity and pressure of the gases present (ThinkSRS.com).
The vacuum chamber is a 316 stainless steel sphere with an outer diameter of 4.880” and an inner diameter of 4.800”. There are six 4.5” and eight 1.33” Conflat ports radially symmetric about the center. One port is occupied by a 2.75” quartz viewport for observation purposes. The Conflat fitting is a standard vacuum component system that is available in diameters from 1.33” up to and beyond 18”. The topmost port has an electrical feedthrough which when set at a negative potential will accelerate positive ions toward the center. The copper feedthrough is rate for a maximum of 30 kV @ 55 A. The stock that holds the grid together is made out of 304 stainless steel and machined on the lathe. Six precision 0.033” holes were drilled to hold the six ends of the grid in place on the cutting mill. Nickel-Chromium, 0.032” in diameter, is used for the three rings that make up the grid. The wire was wound into a spring using the lathe and the three rings with short ends for connection to stock were hand cut to ensure minimal differences in ring diameter.
A power supply with a minimal 400 Watt output is recommended. The power supply must be of negative polarity DC with respect to ground. We chose to use a Spellman power supply rated at a maximum of 20 kV @ 18 mA (only 350 Watts), which will be cooled by convection fins on the rear panel. The power supply is capable of reaching potentially fatal levels and therefore the connector and high voltage rated cable must be well-insulated by large diameter PVC pipes. The power supply comes with a flying lead adapter that will be connected to the electrical feedthrough connector. The voltage and current will be monitored by a data acquisition module for a pc and controlled by a pair of potentiometers.
From the gas storage tanks, copper Swagelok piping was connected to a variable rough valve for on/off flow control. Piping then traveled to a much finer needle valve for fine-tuning the gas flow in to the chamber. We acquired gas regulators for both argon and deuterium tanks and piped to a tee-valve for interchanging both gas supplies. A gas feedthrough had to be constructed to connect the 1/8” Swagelok piping to the chamber. This was achieved by drilling a threading a hole into one of the eight 1.33” port blanks on the vertical mill. Once drilled, the Swagelok connection was inserted and installed with Teflon tape and the blank was remounted to the chamber.
Plasma and the Eventual Fusion
A plasma is a mixture of ionized atoms or molecule and the free electrons not bound to them. A plasma, due to the free electric charges, is heavily affected by electromagnetic fields. Plasmas may take on an extremely wide range of temperature and in some cases there is no defined temperature due to differences in ion and electron energies and it is not in thermal equilibrium. The energy and temperature of plasma are interchangeable as temperature is just a measure of the thermal kinetic energy of a particle. Some measurements have shown that the temperature of a D-D plasma such as that of an IEC device to reach at least 4 keV. The first plasma created in this device was with argon using 6 kV @ 1.3 mA. The plasma emitted a purplish glow as seen through the viewport. The second plasma was with deuterium at approximately the same levels as the first plasma, and emitted a lighter purple glow.
Given enough available power with minimal losses, D-D fusion is possible in this apparatus. As the pressure, voltage, and current are adjusted accordingly, a deuterium plasma will first appear around the exposed grid wires. Increasing the voltage and current, a concentration of plasma will amass in the center of the grid reaching brighter intensities and increasingly energetic particles. There are several interactions that the positive deuterium ions may encounter. The acceleration towards the center can result in ion bombardment of the inner grid causing current losses and damaged grids. Ions may travel untouched through the center or scatter off of each other due to Coulomb repulsion forces. Ions may also collide resulting in fusion, which is verified by measuring the neutron flux about the vacuum chamber. The neutron count will be measured by the Eberline-123 counter and REM ball, a 12” polyethylene sphere that increases the accuracy of the readings by slowing down highly energetic neutrons.
The goal of this project is to attain a significant neutron flux while simultaneously decreasing the input power required in maintaining controlled nuclear fusion; this is referred to as the fusion gain. Here are some ways the fusion gain can be increased in this IEC device:
- Grid alterations to increase transparency and ion focusing or decrease damage from ion bombardment
- Varying ranges of applied voltage and current
- . . .
Original patents (H-M Fusor)
Thesis – Carl Dietrich 2007 (MIT)
Tom Ligon – The world’s simplest fusion reactor, and how to make it work
Todd Rider – Is there a better route to fusion? (MIT)
Prof. Kim Molvig – Fusion without neutrons using p-B11 (MIT Fusion Seminar)
EMC2 Fusion Development Corp. – Inertial Electrodynamic Fusion: The answer to interplanetary space travel?