First attempt at building an alpha beta detector

[Sesselmann]

Hi Guys,

Inspired by the project Lukas did, I attempted this myself with mixed success.

My design is similar to Lucas, but with a few differences. I machined the base from 130 mm HDPE and fitted a 1/8" Hose barb with a channel through to the chamber, I also drilled an outlet for the gas on the opposite side. This configuration made it easier to operate.

The dome as you can see is fitted with and SHV connector, and since I had Argon welding gas in my workshop I connected Argon to the chamber with a silicone hose.

The 0.001" Tungsten wire was kindly sent to me by Christopher Lloyd, but I believe it is came from Ebay.

So far I have only tested two configurations of the wire, a simple loop and a 3d printed wire grid support. The simple loop worked better than the 3d printed grid, so maybe the PLA filament is slightly conductive, don't know, but it didn't work as well.

Working with the 0.001" wire is challenging to say the least, I can barely see it even with a 10X watchmaker loupe, and soldering with one eye through a loupe is tricker than you might think.

Experimental results so far are very vague, but this is what I did with the single wire loop.

Hooked up a standard GS-USB-PRO set it to 1500 Volts, put a small Na22 source into the chamber and let some Argon flow through the chamber.

As the chamber filled up with argon pulses began to appear on the scope, the pulses were about 200 µs long and about 20 mV tall, from what I could see there wasn't much pulse height variance, neither would I expect this from the Na22 source.

This was a very rough experiment because the GS-USB-PRO is designed for PMT's consequently it is not ideal for high impedance proportional counters. Ideally it should have a 10-50M Ohm load resistor and a pF coupling cap, this should result in a taller narrower pulse. I also need to try different gas mixtures.

For a first attempt I think it was a success.

Steven

[GigaBecquerel]

Hello Steven,

It's nice to see people here building other detectors than scintillation counters ;-)
But IMO there are a few issues with your design. The chamber floor has to be conductive, both to shape the field, and to act as an eMI shield.
It looks like your wire loop is almost touching the floor, that's way too big! In my experience 0.5 to 1 cm loop diameter works great.

I am confused as to why you'd expect a constant pulse height from Na22?
The (annihilation) gammas will almost not interact with the gas, but the main energy will be from the betas directly, and that is a continuum.
I recommend this detector for alpha, or beta-minus. Also, how did you get 200 µs pulses out of it?
That's very slow even for stuff like GM, usually prop counters give <<µs signals. Also, the long cable length (and resulting capacitance) will strongly attenuate your signal.
I recommend something like 50 pF maximum between detector and amplifier.

Despite the points I mentioned I really like your design, it looks way cleaner than mine.
With some changes this will be a great detector!

[Geoff]

That’s really interesting! I can’t imagine soldering something that thin haha. I’m interested to see how this project evolves.

[Sesselmann]

But IMO there are a few issues with your design. The chamber floor has to be conductive, both to shape the field, and to act as an eMI shield.

Thanks for reply. Yes I agree, that makes perfect sense, I must have been blinded by the fact that plastic machining was so much easier :)
I can make a round metal sheet and attach it to the top of the platform with a tab that touches the dome, that should work.

It looks like your wire loop is almost touching the floor, that's way too big! In my experience 0.5 to 1 cm loop diameter works great.

The picture is from my first attempt, then I made the 3D printed holder and finally I actually made a slightly smaller loop than the first one, so yes I figured that one out already.

I am confused as to why you'd expect a constant pulse height from Na22?
The (annihilation) gammas will almost not interact with the gas, but the main energy will be from the betas directly, and that is a continuum.
I recommend this detector for alpha, or beta-minus.

I only had a short play with this and I don't fully understand what is going on yet. I was just looking at the pulses on my scope and they seemed to be about the same height.
It is possible that the voltage was too high and that it was operating in geiger mode rather than proportional mode.

Also, how did you get 200 µs pulses out of it?
That's very slow even for stuff like GM, usually prop counters give <<µs signals. Also, the long cable length (and resulting capacitance) will strongly attenuate your signal.
I recommend something like 50 pF maximum between detector and amplifier.

I simply connected an unmodified GS-USB-PRO and plugged the scope into the audio output, this is obviously not optimal as it is designed to be used with PMT's.
For my next experiment I have ordered some 10M load resisters and some 50pF/3kV caps and I intend to modify a GS and try to adapt it better for proportional counters.

Despite the points I mentioned I really like your design, it looks way cleaner than mine.
With some changes this will be a great detector!

Thanks, there are some more improvements I think can be done..

1)
Less gas waste, maybe a better seal around the groove and a one way valve on the outlet, this would give me more time to experiment without having to flow the gas continuously.

2)
I read several articles that suggest a gas mix with 90% Argon and 10% Methane, this mix can be ordered from BOC gases and costs around $270.00 for a D size cylinder, which is not too expensive.

also I have some beryllium windows and it might be interesting to build a proper beryllium windowed proportional counter. The LND web site literally gives you 90% of the information you need to build one.

https://www.lndinc.com



Steven

[GigaBecquerel]

Hello Steven,

At regular pressures it is close to impossible to get enough gain for GM, you will have discharges before it reaches any plateau.
And if you did your pulses would be way bigger, GMs usually drop to 1/3 their applied voltage during a discharge. But, since your were using straight argon, you might have had some weird not proportional, but also not quite gm discharge. Without additives argon becomes gain-instable above a gain of around 100, because the UV photons will ionize more gas somewhere else in the chamber. That might explain your pulse height spectrum.
Additives like CO2, or most commonly methane, absorb those photons without being ionized, effectively quenching the UV avalanches. I have started to use Ar/CO2 82/18 Welding mix for my counters, and that gives me great results. P-10 (10% Methane) would be even better, but that's only needed for spectroscopic applications. If you're just after counting Ar/CO2 works great and has the advantage of not containing a flammable component, as well as costing about 1/3 as much.

Now, on the topic of sealed detectors. I am currently experimenting with them, and they come with all sorts of issues.
You really need to remove *all* the air inside, because oxygen really quenches your avalanches, even in the ppm range. Because I was too lazy to set up my turbomolecular pump I tried pumping the counters with a rotary vane pump, and only after around 10 cycles of pumping the detector, backfilling it with Ar/CO2, and pumping again you start to see counts.
A beryllium (or similar) window adds the difficulty that they are usually so thin that they don't hold atmospheric pressure, and so what LND does is they put their counters in a vacuum chamber that is pumped / filled at the same rate as the detectors, to never have any pressure differential across the window.

[Sesselmann]

That all makes a lot of sense, and I already figured one would need good vacuum, I have all the vacuum gear but it's not worth setting up for one experiment.

Not sure how to get a good ceramic to metal seal for the feedthrough, might have to buy these parts and braze, or do you think JB Weld glue would suffice?

Steven

[GigaBecquerel]

Heheh, yes, the feedthroughs are an issue, but they can be bodged.
Since the prop counters operate close, or at atmospheric, and organics don't bother that much, you can use epoxy to seal everything.
I don't have direct experience with JB Weld, but I tried many Epoxies and they all seem to work well for vacuum stuff, with proper baking and turbopumping I got atmospheres of well below 10^-7 mBar.
But there's a "lifehack" that makes things really cheap and easy: Use PCBs as feedthroughs!
FR4 is fiberglass and epoxy, they're really stable and hold vacuum well, if they're metal plated they barely outgass, and a soldered THT connection or via is vacuum tight.
And from places like JLC they are *really* cheap.

Here's a video and some images from my first sealed detector:
https://www.youtube.com/watch?v=JpIAbbr2Cxw

The wire is held between the two PCBs, and soldered under tension.
The tube is 28 mm Cu piping from the hardware store and the fill tube is an old capillary from a fridge, brazed onto the tube. Everything else if soldered with regular Sn40Pb60.

[Sesselmann]

Cool, I really like that hack using PCB to close the ends of the tube. I'm not sure if I want to put that much effort into making a GM tube, but if we could make a proportional tube the same way, it would be worth the effort.

But as you said above, I would need quite an elaborate vacuum system to evacuate the whole tube and a system of valves to exchange the gas and seal the tube.

Step #1 is to improve my tiny dome detector.

Steven

[GigaBecquerel]

Technically this was a proportional counter, but I could not resist the "pun" if saying I built a geiger counter ;-)
The detector is not very usable as is, but it was only a proof of concept.
You could introduce a window (would you believe it, household Al foil holds a vacuum if the window is below ~20 mm in diameter, and can be soldered to the copper), or line it with boron on the inside, or try something else completely!

As far as your detector goes, I think the biggest changes can be achived by replacing the bottom with a metal one, and using Argon with some (organic) additive.
You could try adding 5-10% propane and see how that changes the detector response!

[Sesselmann]

Hi guys,

Just a quick update on this proportional counter experiment.

I modified a new GS-USB-PRO with a higher value 10M load resistor and a 50pF coupling cap, this works much better for high impedance proportional counters including Neutron counters.

Based on Lukas advise I placed some aluminium foil on the base of the dome detector to give it a ground potential plane, and poked a few holes for the gas.

DSC_0045.JPG

I then placed a few different sources inside the dome and recorded a few spectra to compare.

I found that the spectra are indeed different even though they look almost the same, because the detector efficiency drops off so rapidly above 100 keV that there are virtually no counts.

The higher energy counts can be seen in the log scale image of Polycrase, probably some alpha in this source as well.

I can see that there would be quite a bit of R&D to build a good proportional counter, simple in principle but complex to fine tune.

Steven

cs-137.png

Ba123.png

polycrase-log.png