Peltier Cooled Cloud Chamber

A cloud chamber is one of the easiest way to build your own particle detector. With a cloud chamber you can visualize the tracks left by cosmic radiation, environmental radioactivity or that of radioactive samples like minerals containing uranium or thorium. Historically cloud chambers belong to the oldest type of particle detectors while today they are mainly used for demonstration purposes. In a diffusion-type cloud chamber a supersaturated vapor (usually alcohol) condenses along the trails left by ionizing particles thereby making the tracks visible. For a more detailed description of the working principle you can refer to the Wikipedia article.

The cloud chamber described in this instructable has the following features

• 60 x 60 mm2 active area
• peltier cooling (no dry ice needed)
• high voltage source to increase visibility of tracks
• white LEDs for chamber illumination
• all electronics mounted in laser cut enclosure with fancy illuminated signs

There are a couple of other cloud chamber projects which helped me with this build. The design was basically adapted from this cloud chamber by David Aceituno. Also the following instructable by nothinglabs was very helpful.

I used the following materials for this project

• TEC2-25408 (ebay.de)
• ATX power supply (amazon.de)
• Cooler Master Hyper 412S (amazon.de) + second fan (amazon.de)
• 11 x 16 cm glass bell (amazon.de)
• white LED strip (amazon.de)
• electric fly swatter (amazon.de)
• 60 x 60 mm copper plate, 2 mm thick
• 400 x 700 mm black acrylic plate, 3 mm thick
• flip switch
• 3 mm LED + bezel holder
• sponge
• self-adhesive black vinyl foil
• PVC pipe, ID 5 mm (amazon.de)
• 28x M3 screws, 10 mm long + nuts + washers
• 8x M5 plastic screws, 20 mm long + nuts + washers
• 4x M4 screws, 40 mm long + nuts + washers
• thermal grease (see recommendation in next step)
• 99.9% isopropanol

In addition, I used some 3D printed components, lots of "Dupont" wires, ATX Molex cables, some packaging foam and, hot glue (because there is no way to build a project without hot glue) and white baking paper for diffusing the LEDs. The used tools include a 3D printer, laser cutter, soldering iron, drill and an IR thermometer. As radioactive samples I used thoriated welding rods (amazon.de) and a small piece of pitchblende (uraninite) which you can sometimes find on ebay.

For the cloud chamber to work the temperature should be below about -25°C. In many other instructables this is achieved by cascading two peltier elements and running the bottom one at a higher voltage. I tried many different configurations including two or even three TECs running at different voltages but always ended up with about the same temperature. In the end, I used a two-stage TEC element which is maybe a bit more expensive than two single-stage elements with the same cooling power but it requires only a single voltage. The TEC was attached to a CPU cooler with heat pipes. While these CPU coolers are designed to work with the heat pipes facing upwards they also seem to provide sufficient cooling power when used upside down. The TEC was attached to the CPU cooler with thermal grease. At first, I was using the popular Arctic MX-4 but I do not recommend it because it has a very high viscosity and is difficult to spread evenly on larger surfaces. Also it is difficult to remove because it is so sticky. Apparently, I even managed to bond two TECs permantely to each other with it. Finally, I switched to a different thermal grease (which was also cheaper) as shown on the picture, this decreased the temperature by almost 10°C (!). The TEC was connected to the 12 V output (yellow wires) of the ATX power supply. As shown on the picture a jumper wire was connected to the power supply (see also step 11) so that it turns on immediately after activating the switch at the back. The fans of the CPU cooler were also run by the power supply. Take care to place the TEC with the correct side facing upwards, i.e. the red wire should be facing to the right side as in the picture. I was able to reach a temperature of about -42°C after a few minutes. If you do not reach a temperature lower than -25°C check your thermal coupling.

WARNING: Be careful when handling high voltages!

While an electric fly swatter is probably not able to kill a human person (unless maybe your DNA got messed up in a failed teleportation attempt ;-)), you will be fiddling with the electronics so try to keep it safe.

I removed the high voltage generating circuit from the electric fly swatter. The circuit was running on two 1.5 V batteries so it could later be attached to the 3.3 V wires (orange) of the ATX power supply. The PCB contained a momentary push button which charges a capacitor, once the button is released the HV slowly discharges again. To keep the voltage constant, I replaced the push button with a flip switch. Also the status LED was replaced by a panel mountable LED and I attached Dupont connectors at the input and output. I am not sure how high is the generated voltage, I can only say it is somewhere above the measurement range of my multimeter (1 kV) but below the point where it gets destroyed.

I have attached stl files for all 3D printed parts. The top plate is used as base for the glass bell and to mount the LED strips. The top bracket is needed twice to connect the PVC pipes which are used to mount the sponge and the high voltage grid. The other bracket is used to mount the copper plate to the CPU cooler.

Drill four holes (4.5 mm diameter) in the copper plate with a distance of 50 mm (see attached pdf). Also attach the black vinyl foil to the plate. I left the edges free because there needs to be an electrical connection between the plate and the screws fixing the plate for the high voltage. The M4 screws go into the 3D printed bracket which is later used to fix the copper plate.

The grid for the high voltage was cut from the electric fly swatter and conveniently already had a wire attached. I mounted the grid about 4 cm above the copper plate, the smaller the distance the higher the electric field will be. The plastic screws were glued to the PVC pipes and the grid was attached with the corresponding nuts. The sponge will go above the grid and I adjusted the length of the PVC pipes so that it still fits inside the glass bell. I figured it is convenient to have the sponge at a long distance from the cold plate so that more alcohol will evaporate.

I designed the box using the Makercase tool and then later modified it in Fusion 360. As already mentioned the enclosure was cut from 3 mm thick black acrylic and I have attached the dxf files. You may ask yourself why the housing looks like swiss cheese. This is because I first had problems with the temperature not going low enough once the CPU cooler was mounted in the enclosure, so I started adding more and more holes for ventilation. Finally, as already mentioned using a different kind of thermal compound got me a large improvement in temperature so probably the chamber would also work with fewer ventilation holes.

I thought it would be nice to add some illuminated signs to the enclosure because LEDs make just about everything look cooler. So I 3D printed some elements using transparent HD Glass filament (stl files attached). The parts were printed with 0.4 mm layer height and 100% infill. I also used the "enable ironing" feature in Cura to get a smoother top layer. The quality of the prints did not turn out great on my 3D printer because the filament is very sticky. After printing I attached some pieces of white LED strip to the back with hot glue. Since the HV sign only required a single LED but the LED strip contains three LEDs I just blocked the other LEDs with black tape. I also put some hot glue on the LEDs before to help diffuse the light (a trick I learned in this instructable). Finally, the parts were attached to the housing also using hot glue.

Prepare the top plate by gluing the LED strip on the inside and feed the wires through the opening. Here, I attached baking paper to the LEDs to diffuse the light. The 3D printed plate was then attached to the top plate of the laser cut enclosure using M3 screws and nuts.

Spread thermal paste on the top of the peltier element then place the top plate (enclosure + 3D printed ring) on top. After that the copper plate goes on top of the TEC element. I also cut a square pieces from some packaging foam and put it in between. The copper plate is fixed with the 3D printed bracket using M4 screws. I attached a wire to one of the screws which is later connected to the high voltage. After assembly it is a good idea to connect everything to the power supply and measure again the temperature. As you can see I was able to reach about -36°C. In my experience reaching a low temperature is most important for getting the chamber to work. If the temperature is not below -25°C you will probably not be able to see any tracks.

Assemble the housing as in the attached pictures. The walls are held together by 10 mm long M3 screws and nuts which go in the t-slots. I also reinforced the housing with hot glue where possible and glued small feet made of foam to the bottom. The ATX power supply was attached to the back plate using M3 screws. The enclosure was more or less intentionally designed to be a little bit higher than needed. To get the CPU cooler to the correct height I put small pieces of foam below. The high voltage PCB was mounted on the bottom plate using a PCB spacer. The LED and flip switch were mounted to the front panel. The fans, TEC and all LED strips were connected to the 12 V output of the power supply (yellow wires). The HV PCB was connected to the 3.3 V output (orange wire). The HV output gets connected to the copper plate via the wire on one of the screws. The other output can be fed through the hole in the 3D printed top plate and is later connected to the metal grid. From my experience the polarity of the high voltage does not matter. This seems to refuse that the high voltage will increase the sensitivity by drifting ionization tracks in the active region, as suggested in the Wikipedia article. I rather believe that it helps to clear excess charges which build up in the chamber as other people have suggested. Since there is only little space inside the housing the cables have to be cramped in a little. I tried to hold all cables in place using cable binders.

After the housing is assembled the high voltage grid and sponge can be placed on top of the cold plate. Connect the wire of the high voltage grid and seal the chamber with the glass bell after that the build is finished.

To get the chamber running, thoroughly soak the the sponge with isopropanol and place a sample in the middle of the copper plate. When switching on the chamber you should see a mist of alcohol forming at the bottom after a few minutes and tracks should start to appear. If you switch on the high voltage the visibility of the tracks should increase greatly. In particular, the high voltage makes the tracks more focused and less spread out.

So far I tried two samples, the first one was a few pieces of thoriated welding rod containing 2% thorium oxide. These samples are easily available and rather save to handle because they do not produce any dust or high amounts of radioactive gas (Th-232 produces Rn-220 but the half-life of the latter is very short). Th-232 is the first isotope in one of the natural decay chains, i.e. the sample will contain a number of radioactive isotopes and emitts alpha, beta and gamma radiation. The second sample I got was a piece of stone containing the mineral pitchblende (uranium oxide). You can already see from the pictures and the video that this sample has a much higher activity. U-238 is the first isotope in another natural decay chain and therefore, the sample contains all isotopes of this series among which are several alpha, beta and gamma emitters. I would not recommend to keep larger amounts of this mineral at home because the activity is rather high. In addition, it does produce some dust and also gaseous Rn-222. That means even if you keep it in a container made of lead radioactive gas will diffuse out.

What could be the issue if you do not see tracks? In the beginning, it happened very often that I did see a mist of alcohol forming but no tracks. Usually the chamber did work without the enclosure but not after I closed the housing. This led me to the suspicion that the temperature does not get low enough so I tried different combinations of various TEC elements. I also had only a single fan for the CPU cooler at the beginning and fewer slits in the housing for ventilation. I also tried a liquid CPU cooler but I never really reached much lower temperatures than in the current configuration. In the end what got me the largest improvement was changing the thermal compound. I am not completely sure that this was solely a temperature problem it could also be related to air turbulences inside the chamber caused by the fans which prevent a stable vapor cloud from forming. In addition, the ambient temperature and humidity might also play a role.

I would say one of the biggest drawbacks is the small active area of this chamber. This is not a problem if you are measuring small samples but to visualize ambient radioactivity or cosmic muons it would be nicer to have a larger area. In principle one could just increase the size of the copper plate but to get to reach the same temperature it would be necessary to also increase the size of the TEC elements or use multiple elements. In this case you would probably also need a different cooler for the peltier's hot side. To reach really large areas there is also the possibility to use a compressor for cooling as shown in this impressive cloud chamber by CloudyLabs.

As it is the chamber can run continously for at least 15 minutes. To achieve real continous operation one would have to add a pump which collects the condensed alcohol from the bottom and moves it back to the top. Finally, one could also implement a heating wire at the top to increase the temperature gradient and thereby the thickness of the active layer.