Cylindrical Fusor design

[Daniel Itkis]

I'm not quite sure what category to post this in, but it seems like it sort of fits in here.
Over the last year two friends of mine (16, 18) and me (just turned 19) built a fusor. Here are finally some of the details about our baby

The idea of this Fusor was NOT to maximize neutron yields, as so many before us have already set this goal for themselves. What we had in mind was to explore a different geometry of the chamber and of the cathode. Instead of the usual setup of two concentric spheres we chose to use to cylinders. This does of course dramatically decrease fusion efficiency as the central focus is lost, but it's interesting to know how the plasma behaves nonetheless.
The chamber itself is regular cheap 4”-steel double nipple used in plumbing with two fitting caps for the top and the bottom. In the top cap a 7,0cm hole was drilled for the window.
The height/ diameter ratio of both cylinder-electrodes is 2 : 1.

The bottom cap has a hole fitting for 1”-steel hydraulic tubing which is welded into that hole. The central part is equipped to 8mm steel tubes, one at exactly half the height of the piece used as the high voltage feed-through, the other one a bit higher 90 degrees relative to the first one used as an inlet for various gases. Those two are welded in place, too.
The hydraulic tubing on the bottom cap additionally has a self-made adapter to the CF-30-flange of the Pirani-probe.
The bottom cap is sealed by a weld all the way around (this one gave us some trouble at first as the was a smell hair fracture in the weld, almost invisible, which was a constant leak; it was later found and sealed with epoxy two-component glue).
The top cap is held in place by its original thread and some epoxy and high vacuum grease. The window is polycarbonate, mostly because of a lesser risk of fracturing compared to glass. Although being an organic polymer the PC doesn't melt, swell up or gas out until much lower pressures or high temperatures. The plasma did not seem to degrade it in any visible way for the time of it being used, except for a thin layer of conductive copper which probably eroded off the cathode as a consequence of constant bombardment with high energy ions.

The high voltage feed-through was a bit tricky to build, however we found that for voltages up to 20kV (the most we could produce) a combination of a ceramic tube, soft PVC-isolation and heat-shrinking tubes worked flawlessly. The only part that is exposed to the vacuum is the ceramic tube, proper sealing was insured by a special boric oxide glass seal based on a regular Housekeeper-seal we developed specifically for this purpose. It consists of a molten down and dehydrated glass based on a mixture of boric acid and glass powder with more that 95% B2O3 in the final composition, and has quite pleasant properties. It's quite cheap and easy to apply, it's compatible with the vacuum and shows no deterioration from plasma exposure over time, it is quite hard and resistant to the pull from the vacuum. Most importantly however unlike regular glass, which because of the high application temperatures cracks when trying to connect substances with different heat expansion coefficients like ceramic and some metal, our mixture can be used to connect ceramic and steel and copper. Glasses with this low of an operating temperature are often expensive and based on lead oxide, which is both toxic and eager to oxidize more reactive metals like iron. The cathode material is copper wire.
The gas inlet is simply a glass stopcock often encountered in chemical laboratories.
The apparatus seems to be more or less tight, vacuum measurements show that after a night (i. e. approx. 16 hours) with turned off pumps the pressure in the chamber changed from 10^-2 mbar to a bit less than 10 mbar (which is still enough to generate some plasma).

For the vacuum production a combination of a Vacuubrand RZ-5 and a water-cooled Edwards E-04 diffusion pump is used.
The high voltage is achieved with a 2kV microwave transformer connected to a Cockcroft-Walton-Cascade. In theory up to 24kV should be possible, however with a 1000MΩ-High Voltage Probe only up to 20kV have been measured.
The whole setup is processed with a computer via a modular Arduino Mega system, which controls the pumps, HV-generation and measuring equipment. A self-coded Java program in a Processing 3.0 surrounding graphically represents the voltage, pressure (from output voltage of the Pirani-probe), neutron measurements and the output of multiple temperature sensors in real time.
The neutron measurement is performed in two separate ways; firstly, a piece of silver is activated by the neutron radiation to yield the short-lived radioactive isotopes Ag-110 and Ag-108. The decay of these isotopes is measured with a regular Geiger-Counter. Taken together the results from multiple experiments the decay-curves indicate we have had an about doubled count number (relative to background avg.) for the first and second minute. This alone would mean, taken the radioactive background as a normal distribution, that we have indeed activated the silver with a probability of more than ±3σ (99,73%).

The second way to measure the neutrons is using a soviet SNM-13 neutron detector tube filled with BF3. This more quantitative method implies a neutron emission rate of about 2*10^2 neutrons/second at about 10kV, assuming the emissions are isotropic. False counts arising from induced currents in the counter because of electrical breakdown and sudden changes in current in the Fusor are eliminated from the measurements by comparison to an adapted circuit for detection of atmospheric discharges, which can now precisely measure even small discharges and malfunctions in the cascade and the Fusor chamber.
This very low emission rate is in our eyes a byproduct of the decreased symmetry of the Fusor. Plasma emission spectroscopy in UV-Vis shows a mostly clean D2-Plasma, with a small impurity which should be Oxygen leftover from the production of the D2-Gas from D2O.

Our production setup involves a controlled reaction with small amounts of D2O with Sodium metal in the vacuum, which has the advantage of using less than 0,1mL per fusion run, which makes the method quite cost effective. The emissions spectrum shows strong Balmer-emission-lines of atomic deuterium with a small Fulcher-α-band of molecular Deuterium emissions around 560-640 nm.
The interesting part however is the plasma distribution in the cylinder:
We have, before building the device, run a few simulations of what to look for. As a cylinder is not a sphere, it's internal electric fields and electric potentials are different from a sphere's. Some of the images below show these electric fields and potentials visualized by a program by Prof. Girwitz (LMU, Munich), which is easy and free to use for purposes of education and research. We have in our thoughts assumed a fairly simple viewpoint in based on electrostatic lensing. The interesting part are herein the funnel shaped zones on top an bottom. Assuming a uniformly distributed ionization probability of a deuterium molecule those funnels imply a strongly increased particle density in the center of the top and bottom cathode ring. Additionally, the leftover cylinder with two cones cut out would concentrate the particles to a more diffuse zone around the central point. This same prediction is obtained using the physics based graphic program “Blender”, which predicts a plasma distribution with three separate zones of high plasma density not unlike a 3d-z^2 orbital of the hydrogen atom.

Electric field in a cylinder

Equipotential lines in a cylinder

In reality the plasma distribution is much like the spherical counterpart until a pressure of about 1,3 mbar is reached (depending on gas filling). At this point the plasma visibly organizes into three separate zones, one diffuse one around the center and two rather sharp secondary focal plasma balls (actual balls, not jets) in the center of the top and bottom cathode rings. It would be interesting to know if these secoundary focal balls already have a name, in our team we have been calling them "Hot Zones" throughout the project. We've not been able to find any similar experiments except for Doug Coulter's recent work with cylinders.
These Hot Zones have, depending on the gas, only a limited life range. In H2-Plasma they only last for 50-100ms at the most, while we have been able to produce stable Hot Zones with N2-Plasma in a pressure range of 0,35-1,4 mbar for more than 30 minutes (before turning the HV-generator off). For pressures any lower than these 0,35mbar there doesn't seem to be enough material in the chamber to form the three separate focal zones and only a sharper central poissor remains.

Overall this project has given all of us an more in depth understanding on plasma physics, handling of high voltage and vacuum chambers, applied programming and many other interesting issues we had to solve all on our own (except for the UV-Vis-spectrometer, we asked a Professor at the LMU in Munich to use one of theirs). With a total cost of about 1000€ our Fusor experience didn't even cost us that much, the main costs being the used vacuum pumps off of ebay.

We're now working on a improved HV-generator that could provide up to 50kV, upgrading the HV-feedthrough and once that is done getting some nice voltage-neutron emission-correlation as well testing for possible isotropic effects in neutron emission.

So that's pretty much what I have to say, now I'm curious what you guys have to say...

[Rich Feldman]

Nice work, with creative and novel solutions in many places. Thank you for coming to present it.

Can you tell us more about your sodium-and-D2O reactor, and how it's regulated? For a moment, I foolishly expected the evolved gas to be 33% O2. Am reminded of miners' carbide lamps, with their tiny and bright acetylene flames.

carbide.JPG (14.19 KiB) Viewed 3972 times

I'm also interested in your detector for electrostatic discharge events, and how it correlates with pulses in your BF3 tube circuit.

[Daniel Itkis]

The setup for the deuterium production is a bit archaic, but is does work quite reliably and efficiently and is very cheap to set up.
The idea is that you produce the deuterium directly from D2O in a vacuum; for this we do not use a electric current, but in fact
a piece of sodium metal 2 D2O + 2 Na --> 2 NaOD + D2. It does not necessarily have to be sodium, Li should work for this as well, potassium
and the higher analogues would probably be a bit to reactive. The analogy to carbide lamps is not all that far off, the idea came as
a variation of a döbereiner lamp. But just like in a carbide lamp the oxygen from the water is bound as Ca(OH)2, in our case most of it is bound as
NaOD.
For the injection we connect the glass tube of the inlet-stopcock to another piece of chemical glassware, a very small glass filter crucible
with a layer of porous sintered glass. Maybe crucible is the wrong word here, it actually looks more like a funnel.

The end of the crucible is plugged with a rubber stopper. Through the rubber stopper a 0,5mL syringe's needle is pierced. For a deuterium run
the syringe is filled with D2O, as mentioned an amount of even less than 0,1mL can be used and accurately measured out this way.
A small piece of Na is placed between the stopper and the barrier of sintered glass.
As the crucible is directly connected to the vacuum through the stopcock, so the chamber can quickly be evacuated. As the
syringe has an extremely narrow outlet the vacuum can't exert enough force (Force = pressure * area) on the syringe to release any of the D2O
prematurely. Evaporation rates through the opening are also neglegible. To see what the actual needle of the syringe looks like one can
google "diabetes syringe 0,5 ml". You'll get the feeling. Once the desired overall pressure is reached the stopcock can be closed again and a tiny amount of D2O can
be squeezed onto the Na. The reaction starts upon contact releasing D2. The sintered glass barrier keeps both the D2O and the Na from being sucked into the system.
Now the cock can be reopened to let the D2-Gas flow into the chamber. By carefully alternating inlet of D2 and D2O pretty much stable pressure can be achieved (as long as there's D2O and Na in the gas generator).
Anyhow, a small amount of D2O (water has a vapor pressure of about 30mbar at RT, D2O is a tiny bit below that) would still evaporate before reaching the Na. From a fusion efficiency standpoint even that small impurity is
fatal as O with its Z=8 has a 64 times higher bremsstrahlung loss compared to D. According to the plasma emission spectrum the
dominant emission seems to be D. From measuring the spectra of both H2 and O2 (pure) I can say that the emission of Oxygen at around 780nm is very strong in intensity compared to the strongest Deuterium-signal
at around 656nm at the same pressure (6 times stronger at 0,1mbar), so I would from this assume that there is only a very small amount of Oxygen present in the plasma. However I wouldn't want to give any
figures of exact compositions as many factors play a role in the intensity of emissions of certain quantum transitions.

[Richard Hull]

Your work appears impressive. Some additional questions must be asked. How did you activate the silver? Details please. Voltage during activation? current? How long did you expose it? Images of you activation setup? A real win dpends on these critical issues as the silver activation looks good, but needs much more clarification.

Richard Hull

[Daniel Itkis]

Thanks. As to the actication:
We used a silver coin, moderated with paraffin wax poured into a two stacked blocks to moderate the neutrons.
Both taken together about 6,5cm in height. The moderators and the Ag were placed above the chamber's window.

The diameter of the silver piece is 40mm, the total distance from the cathode was about 13cm, i.e. about 2% of the
total radiation are absorbed by the silver.

The Ag has been at that position for the entire experiments, but the time of actual neutron exposure was about 15min per run.
The time between the deactivation and the start of the measurement was in each case 30s seconds. The voltage was changing
with pressure as a consequence of the inlet system and the steepness of the Paschen-curve: Throughout
a the voltage was rising and falling from 10kV to full 20kV, although we tried keeping it around 18-19kV for most of the time.
After finished exposure the Voltage for the Fusor was turned off, the neutron counter tube quickly exchanged for a regular one, the
counter restarted and the silver coin placed directly upon the new tube. CPM-values have been measured and noted for the first 20min,
the counter was then lest to run for another 40min as to get a good average background.

[Richard Hull]

From your report and follow up on activation, I have placed you into the Neutron club. You have done a good job of it.

Richard Hull