At approximately 10:00PM, Wednesday, August 23rd, two years of independent research, development, construction, and experimentation came to fruition when, for the first time, I measured neutron emissions from my homemade Farnsworth–Hirsch inertial electrostatic confinement nuclear fusion reactor, or fusor. I can now conclusively say that I have achieved nuclear fusion.
Just over two years ago, I began this project following a simplified, and in some places entirely incorrect guide published in Make: Magazine.¹ ² This guide lead me to construct Fusor 0, a (relatively) low cost, very simple demo fusor with the intent to prove that I could embark on a fully fledged fusion project.
Over my junior and senior year of high school, I continued to learn more about the underlying physics and technical theory behind nuclear fusion and commensurately upgraded my own device until nearly every component had been replaced or rebuilt.
With this fusor, I finally created and measured neutron radiation, one of a small number of irrefutable fusion indicators.
As the above images clearly demonstrate, a great deal of progress was made between my first and latest fusor prototypes. I learned an even greater amount about physics and engineering between these attempts. Although I endeavored to maintain an up-to-date record of these events on this blog, I very clearly fell short. Additionally, I am now a full time student at Columbia University studying applied physics, so time needed to bring this site up to speed will be in even shorter supply than it already has been. For that reason, this post will be dedicated solely to describing Fusor 1.7 and the run in which I measured nuclear fusion. Subsequent posts will discuss the series of incremental upgrades and information learned between my earlier update in April and now.
I hope to maintain this blog both as a record of my independent high school work and as a place to continually post my latest experimentation as I plan to continue work on my (now collegiate) fusor when I am able.
(Very) Simplified Theory of Fusor Operation
Abstractly, the fusor operates as a voltage gradient based particle accelerator. In technical terms, it can be considered to be an “electrostatically focused and accelerated deuteron collider … relying on inertial electrostatic confinement to allow fusion to take place in velocity space.”³ Simply, a fusor accelerates hydrogen isotopes (actually, a specific isotope called deuterium (D2)) to a sufficient speed such that collisions between ions can result in fusion events. Approximately half of these events result in the emission of a high energy (2.45MeV) neutron particle, the measurement of which will be discussed later.
(Please refer to the third citation below for a complete discussion of the theory surrounding fusion itself.)
In a Farnsworth–Hirsch type fusor, like my Fusor 1.7, deuterium ions accelerate down a focused voltage differential between a conductive, symmetrical chamber wall and the charged geodesic “grid” contained therein. In a device of this type, fusion is created and controlled through the careful balancing of three variables: The potential difference between the center and wall of the fusion chamber, the pressure of deuterium gas within the chamber, and the current flowing through the fusion chamber itself.
The first and most important of these variables is the potential difference, or voltage, within the chamber. Vitally, a potential difference is required first to strip the deuterium atoms within the chamber of their outer shell of electrons, creating fuse-able deuterium ions (duterons). Once ions are created, a potential gradient of at least several kilo-electron volts is required to accelerate the positive deuterons to a sufficient kinetic energy that allows fusion.
Although deuterium fusion events peak at approximately 20keV, more than 20kV potential difference between the chamber grid and walls is ideal. This is because duterons are created all over the fusion chamber, not just at its center. This distribution is roughly Maxwellian, meaning that a fusor contains ions of very different energy levels colliding together. Because of this, almost all of the duterons within a 20kV potential difference would have energies below 20keV. For this reason, the potential difference in a fusor should be higher than 20kV for measurable fusion to occur.
The second variable controlling fusion is deuterium pressure. At greater pressures, more deuterium atoms are available to be ionized, which increases the amount of collisions between duterons and thusly the amount of fusion events. Notably, as described by Paschen’s law, the voltage required to ignite plasma from a gas is directly related to the gas pressure.
Closely related to pressure is the current flowing between the cathode and anode of a fusor system. Similarly to pressure, the system current simply determines how many available deuterium atoms are stripped of their electrons and ionized into fuse-able duterons.
As you can see above, there is no theoretical maximum value for any one of these variables (at very high voltages, fusion cross section drops off dramatically, but this is not relevant to amateur fusors) and no theoretical minimum value for pressure or current. The “required” values for fusion become apparent instead through the interplay between these three variables in a practical device.
Most influentially, the vast majority of amateurs cannot attain or safely operate a power supply capable of over 30kV. A number of factors including x-ray radiation emissions of high energy plasma, electrical arcing and corona discharge, as well as the economic difficulty of acquiring and driving a very high voltage transformer all limit most fusioneers from running fusors above this limit. (Fusor current is limited for this reason too, as high voltage and current power supplies are exceedingly rare and incredibly expensive to buy OR to build.)
As stated above, higher deuterium pressure, of course, means that there are physically more deuterium atoms within the chamber and therefore more chances for fusion events to occur. Likewise, greater current means that more deuterium atoms are ionized and converted into fuse-able duterons. However, the gas pressure within the chamber is inversely related to the electrical resistance between the cathode and anode of the fusor, meaning that when chamber pressure increases, amperage also increases.
Heating of the grid plays a vital role in determining the current and pressure within a fusor. Firstly, metals heat up proportionally to the square of the power flowing through them. Because the voltage through the grid should be maximized in an amateur fusor, this provides a practical limit to the allowable current before the grid begins to melt.⁹ Additional heating is caused by high energy ions from the chamber impacting the grid which is a function of both current and pressure.
These two heating effects, combined with the interplay between pressure and current, provide a functional limit to both.
In a common fusor, these variables work out to approximately:
- 30kV potential difference attained by a -30kV grid and a grounded metal chamber ⁴
- 5–20 mTorr of deuterium pressure within the chamber ⁵
- 10mA electrical current between the cathode and anode ⁶
Now that the characteristics of fusor fusion have been established, how does an amateur fusor such as my Fusor 1.7 create this extreme environment?
First, a suitable vacuum must be created within the fusion chamber. Atmospheric pressure is approximately 760 torr. This means that not only must the chamber pressure be about 0.002% of one atmosphere, but that 0.002% must contain almost no air molecules (nitrogen, oxygen, carbon, etc.). Instead, as much of the air within the chamber (well over 99.998% of it) must be evacuated and replaced with deuterium gas.
This is always achieved with a sophisticated array of vacuum pumps attached to an almost perfectly sealed fusion chamber. After pressure has reached ideally below 1 mTorr (to prevent contaminating deuterium with other molecules), deuterium gas is injected into the chamber at a very slow rate (specific leak rates are dependent on many different factors). Because of contaminants such as leaks of outside air into the chamber, heat and/or pressure related outgassing of trace residues within the chamber, and any air remaining in the system after the initial pumpdown, the vacuum pumps in use within then system are left running but throttled back significantly to balance out with the injection rate which keeps deuterium pressure level and improves deuterium purity.
Once a very low pressure deuterium atmosphere has been created (~1mTorr), the grid can be energized. This is frequently achieved with either an off the shelf power supply capable of the above voltage and amperage or with various homemade systems of varying complexity. For ease of operation, linear supply solutions are almost universally used.
Electricity is carried from the power supply to the grid by some sort of high vacuum rated feedthrough. These feedthroughs allow electricity to pass into the chamber but are sealed from air leaking in. Within the chamber, a (sometimes insulated) conductive stalk carries power to a geodesic grid positioned in the geometric center of the chamber (in most cases). The grid itself must be designed to be electromagnetically transparent such that the average mean free path of duterons does not intersect with it. This limits heating of the grid as well as loss of high energy duterons before they can collide with other duterons. Nevertheless, grids heat significantly, so a material with a high melting point such as tantalum must be chosen.
Once fusion conditions have been reached, the ability to quantify fusion events is, of course, crucial. A variety of methods exist for this but are beyond the scope of this summary. Suffice it to say that the majority of these methods rely on measuring the neutron radiation emitted by ~50% of D-D fusion reactions. These high energy neutrons are not found commonly other than as output from a neutron source, so verifiable measurement of significant neutron counts is a clear indication that fusion is taking place.
Beginning as a proto-fusor, a model device that emulates true fusion by generating plasma within a metal vacuum chamber without necessarily using deuterium gas or very high voltage, Fusor 1 is the second complete prototype I have built and currently by far the most advanced. In its current state, Fusor 1.7 has peaked at 142.6 neutron counts per minute as measured with my neutron meter (this is significant as discussed below).
My vacuum system is built around a Precision D-25 roughing pump backing a Pfeiffer TPU 170 turbopump. I use a right angle bellows sealed choke valve to manually control the pump rate out of the chamber. This setup is capable of pumping my 6" CF cross chamber down to slightly below 1 mTorr of air pressure. This is notably higher than I would like, and I am aware of one leak in particular that hampers my ability to get to a lower pressure, however this leak is located between the turbopump and the choke valve and as such does not cause air to enter the chamber. Still, it’s something to fix in the future. As of now, the chamber leak rate when the choke valve is sealed is about 4 mTorr per minute (which is still higher than I would like, but not too bad). Chamber pressure is measured with an MKS 901P transducer serially outputting to a laptop and foreline pressure is measured with a Varian 531 thermocouple connected to a Varian 804-a controller.
I get high voltage from a Hipotronics 30C high voltage transformer rated for 30kV at 10mA fed by a 5A variable AC transformer (variac) through a 3A fuse. This is fed through a 30kV high voltage feedthrough (with a naked steel feedthrough stalk) to a tantalum geodisic grid that is friction fit onto the end of the feedthrough. Transformer output voltage (grid voltage) is metered with a Fluke 80k-40 high voltage probe connected to a multimeter. Transformer output current (grid current) is measured with a multimeter as the voltage drop over a 10Ω resistor placed between the Hipotronics 30C’s positive output lead and ground
My deuterium comes from Hydrosticks purchased from an online mentor who happens to be a researcher at MIT. The canisters screw into a hydrostik branded flow regulator that is attached via silicone hose to a 10 SLPM mass flow controller that I run at the lowest flow setting. This gives me about 90 mTorr/minute of deuterium inflow when the chamber is sealed. This is quite high, however I have not yet been able to attain a more suitable 1 SLPM flow controller.
Gamma radiation is measured in uSv/hr with an MKC-1 Russian geiger counter connected to a Бета 5 sensor containing 4 SBM-20 tubes. Neutron radiation is measured in CPM with an Aberline ASP2 neutron detector connected to a BF3 tube in a homemade parafin moderator (this system was loaned to me by the incredible James Luck who has helped me immeasurably. According to Mr. Luck, the BF3 tube is slightly under-moderated and is thus less sensitive than it should be). This tube measures only the neutrons whose path takes them through the small tube itself. This means that in reality, the number of neutrons emitted in total from all fusion reactions is significantly higher, although it is impossible to say how much higher without detailed information about the sensitivity of the tube that I was using which is unavailable to me.
Video recording is taken by a cheap usb camera looking into an unleaded borosilicate viewport (that is pointed away from any inhabited areas).
I achieved the first successful fusion run of Fusor 1.7 on August 23rd at 10:00PM. The successful run was preceded by two failed attempts, both failures having been caused by a defective neutron detector. The run lasted 36 minutes total. Plasma was created and stable for 29 minutes.
Fusor Standby Conditions
As follows were the conditions of Fusor 1.7 preceding the successful fusion run:
The throttle valve between the fusion chamber and the Pfieffer TPU-170 turbopump was completely sealed following the preceding fusion attempt. This left chamber pressure at 73 mTorr. The foreline turbopump isolation valve was closed which left the turbopump at an indeterminate internal pressure below 760 Torr. The foreline was at atmospheric pressure, having been vented to atmosphere immediately preceding the run. The mass flow controller was turned off and sealed shut; No deuterium hydrostik was attached to the flow regulator.
Fusion Run Operating Procedure
- All metering devices activated and confirmed to be at expected levels (all ambient besides the chamber pressure which read 73mTorr)
- Precision D-25 roughing pump was activated and allowed to run until foreline pressure reached 100 mTorr as measured by the Varian 531 thermocouple
- Foreline turbopump isolation valve was opened fully to allow the roughing pump to depressurize the turbopump internal mechanism. This lead to a slight represurization of the foreline
- Upon again reaching 100 mTorr as measured by the thermocouple, the throttle valve was fully opened and the run timer started
- After 1:30 minutes, the chamber pressure as measured by the MKS 901-P transducer gauge calibrated for atmospheric gas 87 mTorr. The TPU-170 was turned on and brought into standby mode
- A hydrostik deuterium canister was firmly screwed into the flow regulator
- At t+4:00, chamber pressure measured 2mTorr and the mass flow controller was turned on and given 1V signal power (lowest flow setting)
- Immediately following that, the throttle valve was completely closed, beginning the deuterium backfill phase of operation
- A serial command was sent to the transducer vacuum gauge to recalibrate it to measure hydrogen gas pressure
- At t+5:47, chamber pressure measured approx. 202 mTorr and the throttle valve was opened slightly, allowing chamber pressure to again decrease
- At t+7:00, chamber pressure stabilized at 15.5 mTorr. Variac was turned to supply 39Vac to the Hipotronics transformer, which energized the grid to 9kV at 5mA, creating plasma ignition.
- Over the following 29 minutes, grid voltage and amperage were gradually increased and pressure gradually decreased
- Neutron radiation counts definitively above background levels were measured starting at t+19:00
- At t+36:00 the last data point was taken and the variac was gradually turned to zero before being turned off
- The mass flow controller was quickly deactivated and closed
- The turbopump power supply was turned off and the throttle valve was completely closed
- After five minutes, the foreline turbopump isolation valve was closed
- The Precision D-25 roughing pump was deactivated and the foreline was vented to atmosphere
- All metering equipment deactivated
Fusor Run Data
I intend to make a separate, complete data report with all of my collected information. However, as digitization and analysis of this data will be rather time consuming, I have elected to post preliminary data highlights here for now instead.
The run lasted 36 minutes total. Plasma was created and stable for 29 minutes and the data below was collected over the final 10 minutes of the run.
Voltage averaged about 27kV over the 10 minute period. Voltage was brought to above 30kV for the final two minutes.
Current was between 9 and 15mA over the period but was generally held at 10mA.
Pressure was between 10.5 and 13.1 microns and gradually decreased over the 10 minute period.
The Бета 5 sensor averaged 34.78uSv/hr over the 10 minute period and peaked at 106.5uSv/hr
The BF3 detector averaged 75.6 CPM over the 10 minute period and peaked at 142.6 CPM (12.3 microns, 25.3kV at 10mA)
Two other smaller peaks were recorded at 128CPM (11.4 microns, 27.7kV at 10mA) and 116CPM (10.5 microns, 30.1kV at 10mA)
Background before the run was ~2.8 CPM averaged over a 30 minute period (peaks of ~12CPM were measured, but only single pulses followed immediately by the gauge going back down)
This data, when presented to the community on fusor.net in this post was sufficient to confirm that I had achieved fusion and awarded me membership into the neutron club.
The below photo was taken during the run at a moment of peak neutron output:
I am honored beyond words to have been accepted into fusor.net’s neutron club. This award certifies that not only that I have achieved nuclear fusion, but that I achieved nuclear fusion as an individual enthusiast on a limited budget from my family’s attic before attending college, and that I did so with enough rigor to certify this claim to many highly knowledgeable researchers.
This achievement would not have been possible in the slightest without the dedicated mentorship of the contributors to fusor.net nor without the mentorship of other enthusiasts whom I have learned from over the past two years (my Spicy Neutron Thing friends spring to mind here as they have helped me more than almost anyone else). I intend to remain active in and give back to this community in whatever ways that I can while away at school.
When I am home, I plan on continuing my project down many new avenues of endeavor. There are, of course, improvements that I am excited to make to Fusor 1.7 (many of which center around increasing D2 gas purity). Beyond that, though, metal foil activation, neutron spectroscopy, beam target experimentation, and even neutron bombardment of fissionable materials are all on the table as things I would like to try!
For the near future, however, I am now a full time student at Columbia University! I intend to major in applied physics and am already looking for ways to get involved with Columbia’s own fusion research.
As I indicated at the beginning of this post, I hope to make more updates illustrating the many months of progress between my last post and this one (that should prove to be interesting as some things did blow up in that interim). I will likely also update this post several times to include new information and clarifications should they be needed.
Beyond all else, I am so excited to have finally achieved my goal of nuclear fusion and I truly cannot wait to see where my project goes from here!
Notes & Citations
² https://fusor.net/board/viewtopic.php?t=13463 fusor.net post rebutting the makezone article
³ https://fusor.net/board/viewtopic.php?t=10776 Much of my theory of operation section simply restates what is said here
⁸ Pressure is bounded on the low end because the voltage required to ignite plasma at very low pressures is simply too high to be attained in amateur systems, but lower pressure means less fusion so the lower bound should be avoided. Note as well that at even lower pressures (near the single digit mTorr range), the voltage required to ignite plasma can be so high that the average energy of duterons goes beyond the 20keV point and fusion events drop off anyways.
⁹ Obviously, materials with low resistivity and high melting points make good grids because they allow more current and pressure before melting.