Silvered GM: High/Low End Performance

RobertTubbs · Post by **RobertTubbs** » Fri Sep 18, 2009 1:11 am

This is a two parter on using silvered geiger muller tubes to detect neutrons at medium to low neutron fluxes. I will go into the construction of the device itself, and then the testing of the device. It's advised that anyone who isn't already familiar with neutron activation read Richards activation FAQ. (See link: viewtopic.php?f=13&t=5899#p34343)

Part one: Construction

Having a surplus of CDV-700s I decided to undertake the challenge of building a silvered GM type neutron detector. The first step was to go visit the local electronics store (NOT Radioshack) and pick up some MG Chemicals Silver Print. After applying a nice smooth coating to one of my victoreen geiger muller tubes (which I lacked the foresight write the number down for) I picked out a nice looking coffee can with which to house the tube in. Threw together a double boiler and melted some paraffin wax and cast the wax into the coffee can. There would be roughly three inches of fine paraffin wax on all sides of where the GM tube was to be fitted. After the wax cooled I got out the drill and drilled one hole of varying size so that both the tube and connector would fit snugly inside. Threw a cap on top of the can and I was finished. I was now the proud owner of a neutron detector.

See next page for testing.

Construction pictures below:

Picture One: The Silvered GM
Picture Two: Can O' Wax
Picture Three: Detector Setup

RobertTubbs · Post by **RobertTubbs** » Fri Sep 18, 2009 1:12 am

Part two: Testing

Having a big shiny neutron detector doesn't seem to do much good unless it's detecting neutrons, and I had not the faintest idea as to how sensitive it was. So out came the Am/Be neutron source from days of old! I had originally estimated this source at around 50uCi. Pitifully low neutron output compared to a fusor but enough to spot a slight increase when comparing the irradiated silver to a background count.

The method was simple, take a ten minute background count in one minute increments while making tick marks. For this I used the head phones to listen to the counts. Once I had a background count to work with I set the neutron source by the side of the can for it to "cook" for about ten minutes. I backed the neutron source with a stack of wax bricks to help reflect more neuts into the can. After ten minutes I removed the neutron source and quickly tossed it in a lead box while counting the clicks in one minute increments for another ten minutes. Overall there was about a 15-20 count difference between the background and the irradiated silver. Background in my shop is a little high (call the NRC) and stands at roughly 130 counts over 10 mins or 13 CPM. The irradiated silver CONSISTENTLY gave me between 145-155 counts over 10 mins. Naturally this test was repeated many times and the background was refreshed after a 10 minute cool-down on each run. It was now time for the next test.

And out came the nuclear density gauge with a 40mCi Am/Be neutron source spewing out over 70,000 neutrons per second!!! A pittance to many among us I'm sure, but hey, we can't all have CW multiplier stacks can we. Tipped the NDG on it's side and set the coffee can against the butt while backing it with paraffin bricks.

The results from the NDG are as follows:

Background was 130 counts over ten minutes or 13CPM on average. Upon removal of the nuclear density gauge after ten minutes of cooking the results were 180-220 counts over 10 minutes, they were then averaged over the four ten minute long tests. See below for minute by minute counts.

First minute: 37 counts
Second minute: 22 counts
Third minute: 20.75 counts
Fourth minute: 14.5 counts
Fifth minute: 16.25 counts
Sixth minute: 17.5 counts
Seventh minute: 11.5 counts
Eighth minute: 17 counts
Ninth minute: 18.5 counts
Tenth minute: 17.5 counts

You can easily see where the decay of Ag110 drops off and leaves the Ag108 to troop it out alone. One would expect more counts out of a NDG perhaps, however, this being a point source contained within a big steel box and surrounded by obscene sums of cadmium, tungsten and boron I think what I got was quite good for being parked on the side of the box.

So In the end I'm pretty satisfied with what I'm seeing and believe I can make use of this over caffeinated behemoth. Good luck making your own!

RT

-------------

See pictures below for testing:

Picture One: Weak Am/Be Neutron source
Picture Two: Nuclear Density Gauge Alone
Picture Three: Nuclear Density Gauge Against Can

Post by **Richard Hull** » Fri Sep 18, 2009 9:48 am

Nice work Robert!

The use of a steady source was a plus in this effort.

There is a possible better scenario than just wadding a moderator around the silvered tube. One would seek to, ideally, search out the ideal moderator thickness for the given neutron energy sought to be detected so that maximum advantage might be had from the 10,000 + barn cross section of silver through the resonance range just prior to full thermalization. Full thermalization is great of course, but you are back at the 1/v limitation seen in most materials. While even in silver this is high, that sweet range of immense resonances would, measurably, boost the count.

Jon had posted on this effect related to a water bath he used for Indium, I believe, some years back. Carl and I noted this in the first part of this posting. I would imagine or at least hope that an ideal moderator thickness might boost the effect by an order of magnitude or more.

The work would be extremely tedious requiring hundreds of runs as you slowly built up a moderator around the detector. The key is that your work proceed with neturons from the exact reaction you wish to detect. In our case this is D-D fusion. I might image the moderator thickness might only be less than an inch, but might even turn out to be an inch or more. Certainly, a 4 inch moderator would put you at or near 100% thermalization.

Some sort of long counter or neutron telescope might even work better. A lot of "playing about" with this would be required.

I would seek Carl's counsel on this as he be da' man.

Again, superb effort

Richard Hull

RobertTubbs · Post by **RobertTubbs** » Fri Sep 18, 2009 11:45 am

Thanks Richard,

Yes, you caught me red handed! Unsure of the ideal thickness for paraffin I placed 3" on all sides in accordance with your FAQ. I'll defiantly hit up Carl and ask for his thoughts on the ideal thickness for a paraffin moderator. He did actually mention a few days ago that one would find slightly better performance with a thinner moderator than would be used for BF3 or He-3 tubes.

I certainly think this is worth exploring and I've still got a few coffee cans left.

RT

Post by **Richard Hull** » Fri Sep 18, 2009 3:04 pm

I am not rebuking you or your work, for it stands as a shining example of "the doing" and "hands-on-imperative". You just might be able to be boosted to greater sensitivity, that's all.

I sat back and queried myself as to how I would do this. I would definitely use a water moderator.

1, I would wrap a 1B85 in silver foil and secure it.
2. I would place this inside a piece of thin walled pvc pipe that would just contain it. (1 inch I.D.?)
3. Glue a PVC end cap on both ends and seal the shielded cable end coming out of one end with hot glue, RTV, whatever, to make it water tight.
4. Obtain several closed cylinders of varying diameter 1.5", 1 3/4", 2", 2.25" 2.5", etc.
5 Drop the probe created in #s 1,2 and 3, above, into these cylinders filled with water so that it was concentrically located and contained water above and below the probe equal to the thickness of water in the concentric cylinder region.
6. Run a series of tests with a fixed source almost in contact with the outer cylinder that produces similar neutron energy to the D-D reaction. Each test would use a different size "water tank" cylinder as in #4 above.

The best result would be the near ideal moderator thickness. Once you made a finalized moderated assembly with probe, you could place it next to the fusor and then back it with massive PE blocks or whatever to try and scatter/reflect some of the failed resonance activation neutrons, which are now thermalized, back for at least a 1/v activation response.

Again, the limitation here is the time to maximum sensitivity for a given flux, i.e., a few minute lag on even a steady fixed flux and the absolute refusal of this system to reflect what is happening at any givien instant in time.

The advantage...Very cheap neturon detection verification provided you make a decent amount of neutrons and provide a good baseline background count before, counts during, and decay count numbers over intervals following a run.

Perfect for a novice who can't grab the dough for an He3, BF3 or Bubble system provided they understand the operation and required data handling to make this a good qualitative indicator for the presence of real continuous fusion success.

Richard Hull

Jon Rosenstiel · Post by **Jon Rosenstiel** » Fri Sep 18, 2009 4:02 pm

Do keep the PVC as thin as possible, or use some other material, because Cl-35 (76% of elemental chlorine) has a fairly large cross section and will capture some few neutrons before they’ve had a chance to interact with your silver wrapped GM tube.

Jon R

RobertTubbs · Post by **RobertTubbs** » Fri Sep 18, 2009 5:36 pm

Richard,

I really like the idea and would love to see more experimentation done with silvered-GMs. Like you said there really is something to be said for a low cost and fairly sensitive neutron detector available to the amateur. I frankly believe these things are sensitive enough for even the first time fusioneer as I've detected neutron fluxes as low as a few tens of neutrons per second. Generally the first timer is either fusion go or no go so the trick would just be keeping the neutron flux stable.

I'm in agreement with Jon, I think it's best to avoid PVC. Also be aware of the chlorine in the tap. Here in Seattle there's so much chlorine in the water you might as well be drinking out of the pool. So try using pure distilled water. A good substitute for the PVC might be aluminum.

Edit: Is there anything stopping us from just using glass or pyrex? Using the "good china" to detect neutrons might have some consequences later on in the evening but if you're looking for varied container sizes it might be worth a shot.

RT

Dan Tibbets · Post by **Dan Tibbets** » Fri Sep 18, 2009 5:54 pm

A plastic bag and rubber band should work as a water barrier for temperary testing so long as the top (rubberband on the cable) is out of the water, and after you test the bag to make sure there is not a pinhole in it.

How narrow is the resonance band, would you need to vary moderator thickness tests by a mm to catch it, or would cm steps be adiquate? Is there a linier relationship between the neutron energy and the thickness- or some formula to calculate for D-D fusion neutrons, once you know the thickness for some radioactive sample?

And, how does the neutron absorption of chlorine compare to that of boron10?

Dan Tibbets

Richard Hester · Post by **Richard Hester** » Fri Sep 18, 2009 6:53 pm

I wouldn't use aluminum because of corrosion issues. Polyethylene or polypropylene would be better, and are both decent moderators to boot. Polypropylene can be welded with a plastic filler and hot air pencil.

I was thinking over the weekend, and a piece of ABS pipe would be perfect for protecting a geiger or other tube in a water bath. There's no pesky chlorine to eat neutrons, and the pipe and fittings and glue are readily available at just about any decent hardware store. True, fresh ABS pipe does stink a bit, but this is likely not an issue for something that will be immersed in water.

Carl Willis · Post by **Carl Willis** » Sat Sep 19, 2009 1:15 am

Hi Robert,

Sorry I'm slow to get to your post. Very nice work on developing the silver-lined detector. I think it does have a lot of potential as an accessible and potentially-viable concept for neutron detection. I'm curious if your line of work or your personal connections give you access to the Troxler gauge, or if you...um...well...just kinda picked it up somewhere. No need to answer that unless you want to.

A few advantages of the activation + Geiger tube approach come to mind immediately. Since Geiger tubes generate huge pulses and capacitance of the detector doesn't matter much, and since they present a well-defined counting plateau essentially independent of tube particulars and operating voltage, they lend themselves to being hooked up in parallel. One tube may have limited surface area with which to detect activation beta particles, but you could easily make a bundle or "wall" of tubes, all in parallel, that could effectively increase the active surface area. A hundred tiny tubes, all silver-coated, would do a lot to bring the active material into homogeneity with the detection medium (the tube gas). Geiger tubes are also much more resistant to electrical noise than proportional tubes and scintillators. Their downsides are the limited sensitivity to neutrons (even with a coating) and high sensitivity to gamma, x-rays, and cosmic rays. (Hence the Albatross idea--using a tin foil on a second co-located tube to subtract a gamma-only background.)

To address the questions and build intuition in regards to detector effectiveness versus moderator depth, I have attached a number of files from some MCNPX computer simulations. The top image shows neutron capture rate in Ag-109 distributed evenly throughout a paraffin block (density 0.898 g / cc), in units of captures per Ag-109 atom per source neutron. The silver is not actually present in the problem materials; only the product of its cross-section and the neutron flux is tallied. Thus there are no self-shielding or dilution effects to consider. The data graphically answers the question of where to best put your silver foil in the wax. Also, I should point out that the problem intentionally considers a POINT source of radiation on the surface of the wax block. So there is not only material moderation and attenuation, but geometric attenuation as well--just like in the real-world situations you are measuring. Shown are results for both DD and AmBe neutrons. The AmBe spectrum is the ISO 8529-1 standard AmBe spectrum. The DD spectrum is En = 2.4 MeV.

There are two graphs showing capture rate versus depth in wax. One pertains to Ag-109, the other to B-10 for comparison. Ag-109 has high resonance capture cross-sections and much lower cross-sections below the resonance energy band, while boron is a strong (1/v) absorber of low-energy neutrons. To sum the graphs up, it doesn't much matter if you are trying to detect AmBe or fusor neutrons, despite their different spectral character; you want a wax moderator of about 3 cm thickness. With boron, a thicker moderator would be called for, but not by much (~4 cm). You can see that moderator thickness is not a paramount concern. You can quantitatively estimate detector sensitivity with the aid of these graphs. If you do that and want a check on the calculations I can help out. One thing to remember is that the tallies shown are for neutron capture, NOT subsequent beta decay of the silver. The production of activated Ag from (n,g) capture is split between the ground state (Ag-110) and an excited state (Ag-110m).

Nice work, and I might try my own version of your experiment shortly. Thanks for pursuing this idea.

-Carl

RobertTubbs · Post by **RobertTubbs** » Sat Sep 19, 2009 4:07 am

Carl,

Credit must be given where it's due, this wealth of information you've provided will not go unused. I love the idea of putting a number of silvered GMs in parallel. I think I'll trim the fat off my moderator and give it another go. After that I'll cannibalize my stash of GMs and attempt a GM bundle. Now you've really got the gears turning. Thanks.

RT

P.S. In regards to the Troxler, lets call it a mix of work and connections. I've had many opportunities to get my own, however, the regulatory baggage that comes with one of these things (assuming you come into possession of it legally) is absurd. Also, the prospect of having someone randomly come out to my shop once a month to wipe the whole place down in search of "contamination" is a bit disturbing. Sort of like a urinalysis but for misfits like us.

Starfire · Post by **Starfire** » Sat Sep 19, 2009 9:29 am

Robert - the Chlorine in water is less than one part per million [ recomended 0.5 ppm max ] as per the regulation - I don't think it is a factor.

RobertTubbs · Post by **RobertTubbs** » Sat Sep 19, 2009 10:23 am

Truly, if it's that low then it probably isn't a factor as far as the garage scientist is concerned, I keep having to remind myself I'm using a coffee can and technology from the 1960's. A good knock on the head brings me back, thanks.

Upon retrospect I never really did the tap justice when I was testing different moderators, I think I'll go back and give that another go as well.

Thanks John

RT

Jon Rosenstiel · Post by **Jon Rosenstiel** » Sat Sep 19, 2009 2:14 pm

Very, very cool, Carl.

I don’t recall if I posted the results here or not, but some time back I ran a bunch of moderator thickness tests. Neutron source was my fusor, moderator material was UHMW, and the activation medium was a 2” disc of silver foil.

After each activation period with a different thickness moderator I would wait one-minute and then start a 10-minute count on a GM detector. During the one-minute wait 24.6 s Ag-110 would have, to a great extent, decayed away, so basically what I was counting was 2.4 m Ag-108.

After many, many runs and counting periods the moderator thickness that produced the highest GM count was around 3.2 cm thick.

Jon R

Steven Sesselmann · Post by **Steven Sesselmann** » Sat Sep 19, 2009 9:32 pm

Carl,

How would the optimum moderator thickness need to change for a He3 tube?

The HDPE moderator on my He3 tube is quite a bit thicker then 3.5 cm..

Would it be wise of me to machine it down to 3.5 cm?

Steven

Jon Rosenstiel · Post by **Jon Rosenstiel** » Sat Sep 19, 2009 9:41 pm

Hi George,

Yes I did / do use a reflector, an 8 cm thick piece of UHMW-PE.

And you’re definitely right-on about its effectiveness; it does make a huge difference.

I remember your post on moderators, but for some reason I can’t seem to find it… did you delete it for some reason?

Jon R

Carl Willis · Post by **Carl Willis** » Sun Sep 20, 2009 3:32 am

Hi Steven,

He-3 has a (1/v) cross-section dependence that is no different from B-10, so what's good for a boron-based detector will be good for He-3.

While thinning out the "front" side of your moderator is probably a good idea, keep in mind that my modeling assumes a semi-infinite reflective moderator behind the detector. So if your moderator is made out of a cylinder with the tube running down the axis, you could advantageously mill a flat into the side of the cylinder, but I'd leave the bulk of the material.

-Carl

Jon Rosenstiel · Post by **Jon Rosenstiel** » Sun Sep 20, 2009 1:49 pm

George,

I think you should scrap your vehicle mounted N detector and instead sell that 2" diameter He3 tube to me.

BTW, what kind of background count does that monster He3 tube generate?

Jon R

Post by **Richard Hull** » Mon Sep 21, 2009 11:39 am

I have to be at work to be on line and missed the weekend chatter. Many thanks to Carl on the data. He always comes through for us. It is mightily appreciated and his efforts on our behalf are not wasted, I am sure.

I look forward to communing with a number of you in a couple of weeks at my HEAS conference. I had hoped to have a setup like this Silver foiled-GM system to show and tell, but I'll be lucky to orgainze the lab by the time all show up.

I just sealed the leaky fusor IV and ran it over the last two weeks at over 200,000 n/s in casual runs. I dare not push it and melt something prior to the gathering. It must fuction well, naturally, for the big event and now it does. The leaks were traced and the worst forced me to blow $130.00 for a brand new viewport from Duniway.

Of course, my He3 works fabulously, though I have a bad 10 turn pot in my HV supply and, on occasion, the counter just instantly goes from a roar of 1900 cpm to zero until I tap on the pot or jiggle the knob. (maddening).

Richard Hull

Doug Coulter · Post by **Doug Coulter** » Mon Sep 21, 2009 8:08 pm

Nice work indeed, now I'm going to have to scurry off and try this myself.
Too bad it's slow, but some applications that don't need immediate feedback to the operator can benefit from the integration this provides, so it's a useful part of a suite of instrumentation indeed. And it should be as reliable as it gets besides, always a nice thing indeed. Self testing -- as long as you're seeing background counts you know it's working.

We're having enough troubles with cosmic backgrounds here on sensitive scintillators that we're looking at making a big "geiger" tube and putting it above the stuff, making it so numb it only sees cosmics, and doing anti-coincidence kinds of things. Waiting for a long count, even with automatic background subtraction can be excruciating, and if most of your background is cosmics, inaccurate -- the flux is far from constant at least here on the mountain.

Fusor Forums

Fusor Forums