Activation measurement technique

This area is for discussions involving any fusion related radiation metrology issues. Neutrons are the key signature of fusion, but other radiations are of interest to the amateur fusioneer as well.
User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 7:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Activation measurement technique

Post by Doug Coulter » Sat May 15, 2010 2:49 pm

Let us suppose I want to make a more-sensitive neutron flux integration measurement via activation.
For the moment, I am not worried about absolute calibration, but would simply like to make the current process more efficient than it now is, so lower fluxes can be measured with decent repeatability.

The current method I'm using is to put silver or indium into my neutron oven, which is an affair made of 4" HDPE rod split up so a foil can be placed between the pieces. With about 1.5" of HDPE between the foil and the fast neutron source, and several more inches behind it, the process works quite well with my existing setup, particularly when the fusor output is high -- I have a lashup to count the resulting activation as the sample decays, uniform sample sizes and all the rest -- that's handled fine. But it takes a pretty good flux for a pretty long time to get the samples as hot as I'd like them to be compared to background on the sensitive 2" diameter pancake detector -- I want to extend the useful range downward from where it is, and be able to follow the decay curve longer out without getting into too much noise from background cosmic rays and so on.

Now, looking at the absorption cross sections of silver and indium I see that they have some really big capture resonances at energy levels way above the usual .025 eV thermal neutrons. In the case of silver, for ideal sensitivity it seems you want about 5 eV neutrons, and for indium you'd want about 1.9 eV assuming I'm reading these plots correctly. For silver this means I want energies around 200 times thermal. This shouldn't be a huge problem as we are starting with DD neutrons that are plenty hot by comparison to either -- it's just that I don't want to slow them all the way to thermal for this.

What I am wondering is has anyone worked out what would be the best moderation scheme to get the original neutrons down to these levels more efficiently, rather than all the way to thermal, where the cross sections of my samples are much lower than they are at resonance? Seems in general that many of the other things we'd use would be in a similar range for a large capture cross section. As you can see in the plots, this is a pretty big deal compared to thermal cross sections, so in theory the sensitivity for activation could be raised immensely.

Would one use a non hydrogen containing moderator for this, to lower the possibility that neutrons would be lowered all the way down to thermal in only a couple of collisions? Would carbon be better, or perhaps some oxide or carbide or sulphide of something that has a negligible neutron cross section itself (meaning the other element in there would be more or less inert for this process)? If so, how thick would it want to be between the source and the sample? Since thermal neutrons are still useful (just not as much so) of course you'd back this with some more moderator, and getting some thermals is still better than nothing if the goal is ultimate sensitivity.

You can see where this could be really useful particularly for cases where then neutron flux is either small, or in short bursts that the detector tube types cannot time-resolve, and also good for people just starting out and getting low fluxes at first, since we hope they have a geiger counter before they start out anyway, for their own safety. It would of course be good if whatever moderator substance didn't itself activate of course, so carbon looks good -- less neutron energy loss per scattering event, and perhaps some other medium-light elements would be good too. Most of those have the nice property of being available cheap, and most are fairly easy to form to some shape, but if I had to go to a container full of some powder that would do if nothing better can be devised.

Does someone know how to calculate/predict this so I am not reduced to "trying everything"?
I realize there is no way to get them all down to just this or that particular energy, but it would be nice to get a larger fraction to this range than a normal H-containing moderator would.

Since even a moderator designed to take neutrons to thermal isn't perfect, and has some fraction that escape that -- but an unpredictable fraction it seems, or has to be so big to get them all to thermal that too many are scattered out of the moderator in the process, this seems something worth looking into.
Even a factor of 10 increase in sensitivity would be a boon, especially if it were accompanied with better repeatability, and from these plots, it looks to me like threre's one heck of a lot more than that possible. Since a few here like to activate other, more difficult elements, a plan to get the bulk of neutrons into their resonance ranges (if any) would have good usefulness too.

I believe Carl has some of this info and some software that may help, but anyone is of course invited to comment on this.
Attachments
Cross_Ag.gif
Cross_Ag.gif (29.95 KiB) Viewed 4631 times
Cross_In.gif
Cross_In.gif (34.49 KiB) Viewed 4631 times
Why guess when you can know? Measure!

User avatar
Steven Sesselmann
Posts: 2111
Joined: Thu Aug 11, 2005 1:50 am
Real name: Steven Sesselmann
Location: Sydney - Australia
Contact:

Re: Activation measurement technique

Post by Steven Sesselmann » Sat May 15, 2010 11:32 pm

Doug,

When you moderate neutrons with HDPE at room temperature, you get room temperature neutrons right?

How about using hot oil as a moderator, you can heat the oil to 200 C˚ then activate your silver sample in the oil before removing it and counting.

This might almost double the temperature right?

There might be other materials that could work as a moderator if they were hot enough.

Steven
http://www.gammaspectacular.com - Gamma Spectrometry Systems
https://www.researchgate.net/profile/Steven_Sesselmann - Various papers and patents on RG

User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 7:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: Activation measurement technique

Post by Doug Coulter » Sun May 16, 2010 12:33 am

Yes, or even below room temp in some cases (polycrystalline graphite will scatter all the but the coldest neutrons out of the column via a Bragg type phenomenon and make super cold ones.

A single collision with a proton (hydrogen) and reduce the neutron energy to essentially zero, so anything containing hydrogen is out. Works like the pool shot where the cue ball just stops, imparting all its energy to the struck ball.

So, you certainly want something that won't stop them cold (or to room temp more or less) in just one collision. D would of course be better than H -- but heavy water isn't cheap or that easy to deal with for making a a neutron oven out of.

I note that here I am lookng for neutrons roughly 200 times room temperature, lessee....call that 291k * 200 or 58,200 degrees K. Nope, don't have any medium weight elements that can do that one.

IE thermal neutrons are roughly .025 ev energy and I want 5 volts more or less.

Things like carbon, oxygen, sulfer, and other stuff in that weight range only take about 20 something percent off a neutron per collision. Of course, some of those things might also absorb the neutrons I want to use, so some are right out of consideration. Or they do complex things with resonance energy neutrons (like Al).

I am right now thinking carbon or some chemical with carbon and other stuff might do, but I once heard a certain very smart guy on the forum (Carl) knew how to calculate things like this feedforward, and I was kind of hoping he'd chime in here so we all can benefit -- else I'd have just mailed him direct.

The idea would be to get a gaussian spread (the best you can do anyway) around the resonance region for all these cool things we want to activate, instead of thermal. If you look at those plots there's rather a large advantage if you can do that without losing too many -- the cross sections are nearly 1000x higher at resonance than for thermal....now that's an advantage worth trying for I think.

I like activating things for certain classes of measurements simply because you can't fool it with electrical noise, gamma rays, or anything else, and it can handle huge short bursts and still catch them all, unlike any other detection method yet mentioned on this board. And since I'm getting real high Q in pulse mode these days, all the 3he tubes etc do is inform me that one pulse happened, as all the neutrons in that pulse come out inside a single resolution time of any gas tube device.

Even my very short pulse scintillator that sees knock-on neutrons in a scint plastic also sees some stragglers....the tube can see 5 ns or better, but the neutrons may farkle around in the scint plastic and time-smear things much worse than that -- so this is the single candidate method that can work for short, intense bursts. Ideally, I'd like to be able to measure the output of just a few pulses so I can tune without making my lab too radioactive...and though I realize I won't get any 1000x (theoretical best possible) 100x would sure be nice....and probably possible.

See, going thin with something like HDPE -- many neutrons don't hit anything so they are still at full energy -- and the ones that do mostly get to thermal in 1 or a few collisions, so just making the HDPE thinner isn't going to work. Or so I think -- I was asking a question I actually don't know the answer to, hoping someone did.
Why guess when you can know? Measure!

User avatar
Steven Sesselmann
Posts: 2111
Joined: Thu Aug 11, 2005 1:50 am
Real name: Steven Sesselmann
Location: Sydney - Australia
Contact:

Re: Activation measurement technique

Post by Steven Sesselmann » Sun May 16, 2010 9:23 am

Doug,

Yes, I agree it would be out of the question to heat the moderator to 52,000 K˚.

Would it be an option to cool the specimen sample instead?

Not sure if cooling changes the resonance...

Steven
http://www.gammaspectacular.com - Gamma Spectrometry Systems
https://www.researchgate.net/profile/Steven_Sesselmann - Various papers and patents on RG

tligon
Posts: 587
Joined: Tue Jan 02, 2007 6:58 pm
Real name:
Contact:

Re: Activation measurement technique

Post by tligon » Sun May 16, 2010 3:46 pm

Doug,

We played around with indium at EMC2 a little, while using a fusor running at 10 kV. Our neutron flux was too low to produce good results, so we stuck with a He-3 detector.

I had no idea what I was doing when I put ours together. I bought some indium wire and draped short lengths of it over rods and suspended it in water in a Tupperware container. After activation, I gathered the lengths together and dumped them on a GM counter, but the geometry would have been useless for calibration.

Dr. Bussard crunched some numbers and said the wire was the wrong form. What was better was foil about 0.002 inches thick. I have not checked his calculations, but apparently at that prominent resonant capture cross section a neutron is unlikely to make it more than about 0.002" before being soaked up. His conclusion was that a sandwich of indium foil and a moderator such as polyethelene would be ideal.

Just thinking about the math needed to accurately predict the sensitivity of such a detector from basic physics makes my head hurt. But if you can get a readibly detectable count, tracking the decay rate over an hour or so should be convincing proof that neutron activation occurred. Thus, the technique would prove that fast neutrons had been produced.

Overall, bubble dosimeters probably make more sense, but I've always wanted to get a second chance at indium neutron activation. There is something cool about making new elements at will.

User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 7:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: Activation measurement technique

Post by Doug Coulter » Sun May 16, 2010 5:56 pm

Yes, it seems foils are the best form for activation. For one thing, with a thick piece activated all through, only the radiation (assuming beta decay is the result) from the outermost layer/area can ever get out to be seen on your detector, so the rest is wasted. So that is what we use here. We got silver as thin sheets, and roll our own indium in a metal rolling tool -- it's real easy to do even with just a small hammer. We use fixed size foils that just cover the sensitive area of a pancake detector for post-activation counting. I even bought a piece of very pure gold sheet to fool with, but compared to the others, it's relatively "numb" (in agreement with theory). And in this case, changing gold to mercury is kind of going the wrong direction if one is primarily interested in transmutation!

What I am attempting to do here is make what amounts to a better BTI -- more sensitive and more reliable and accurate. We get such disparate results with those compared to other measurement gear that we got kind of disgusted with them. Not to mention short lifetimes and how much they cost.
I think we're about at the $1000 level with those -- just the ones that have hit the trashcan at end of life. Not a great deal.

Example: Cylinder fusor, inner grid about 1" diameter, and about 4" long in a 6" ID pipe with BTI's side by side standing on end about 4" from fusor grid center, centered on grid center axis (eg in the middle of the 4" length). One BTI 8.6 bub/mrem, the other 26.6 bub/mrem. Take two runs, where all of the other gear (in this case, silver activation, 3He tube, BF3 tube) all read within 10% of the same for the duration of the two runs. In one orientation, the two BTIs have *the same* number of bubbles as each other, switch them for the other run, and they *almost* show the correct sensitivity ratio between them as the labels -- but not quite, and they vary 50% comparing the BTI's to the other gear between the runs. According to all the literature and math I can find, the output of DD fusion should result in uniform neutron output across all the steradians -- perfect istotropy, not small beams....which is about the only way to explain that that doesn't indicate serious problems with the BTI's.

That is not acceptable at all -- I don't even have consistency run to run with BTI's much less absolute accuracy, and it's far easier to believe that the stuff that doesn't even agree with *itself* is the problem, not the other gear. I know of things that can fool the tubes -- fast bursts and such; sometimes you can see this with a scope on the raw signal from the tubes, which beats a counter in some cases -- a counter will count one neutron or a million that happen in a few uS the same -- one pulse, but at least the scope shows a larger pulse for the latter case of a burst.

I know of *nothing* that can fool the activation oven which is fixed in place and consistent from run to run, and tracks the tubes quite accurately when we're not bursting. Both of those, of course, collect neutrons from a fairly large angular area -- so "beams" would explain that -- but the theory doesn't seem to indicate the possibility of beams at these energy levels, so the game's afoot. In fact, we just spent the last day starting to build a directional fast-neutron "camera" to test for the existence of beams, just in case, and will report if/when we have anything interesting to report on that. I will personally be pretty surprised if that much of the literature is wrong, though.

So, it seems activation is the "gold standard" if there is one, for repeatability, though absolute calibration is going to be fairly suspect, also true of the tubes if you want absolute accuracy better than some double digit percent (I don't care about that for now, myself -- "Am I doing better this time than last time" is enough for what we are doing now.). As the curves I show indicate, tiny differences in the neutron energy spectrum from the moderator can have enormous effects on net activation for a certain number of neutrons entering -- nearly 1000::1 in some cases.

So what I *am* looking for here is a moderator design that results in mainly neutrons centered around a few eV, rather than thermal, starting of course with neutrons of DD reaction energies. This should make the measurement both much more sensitive and much more repeatable. If in a thermal moderator, the "tails" are doing all the "work" then that's also where the variability is going to happen according to basic statistics - fewer samples generally means an inferior signal to noise ratio in results.

It occurs to the engineer in me (what I started out as) that moderators that can take 2/3 of the neutron energy on average per collision would not be ideal for meeting this goal. EG hydrogen based, and that one ought to be able to do better with something (carbon?) that takes more collisions to get to thermal. In that case, it seems more likely there ought to be a spot along the moderator column that satisfies this condition, and I am hoping to find that someone knows how to calculate that. It would be far nicer to have a reasonable starting point than having to simply "try everything".

I am sure that I could get fairly close with some sort of sandwich affair that may (but as you point out, due to absorption by other sandwich layers, may not) give me a clue in one test where the right thickness is of whatever the moderator part of the sandwich should be for the final design.

But I'm only well off, not insanely wealthy - and that much gold/silver/indium and the issues of getting many samples counted simultaneously right after a run are well, daunting, so for me it's not practical.
(very pure gold foils we obtained cost the equivalent of about $4k/oz, silver sheet was more like twice the bullion price, and indium is $99/.25lb at rotometals). But getting say 20 pancake detectors hooked to counters and cross calibrated isn't trivial either -- not to mention getting them all loaded and started together quickly after a run.

So I'm after something very particular here -- a moderator design that optimizes several eV neutrons as well as can be done, rather than just going all the way to thermal + tails. Or worse -- some thermals plus the rest full or nearly full energy, which is what I'd predict from a thin, mostly H based moderator (but only based on my gut). So I'm looking for someone who has the math to get me closer in a feedforward fashion so I can just tweak from there.
Why guess when you can know? Measure!

User avatar
Rich Feldman
Posts: 1351
Joined: Mon Dec 21, 2009 11:59 pm
Real name: Rich Feldman
Location: Santa Clara County, CA, USA

Re: Activation measurement technique

Post by Rich Feldman » Mon May 17, 2010 2:25 am

Remember the fatal criticality accident in a Japanese uranium-processing plant, about 15 years ago? Every time the froth subsided in their acid tank, it would go critical again.

According to one technical report, the best estimate of total fission yield was from measuring activation of some coins found in the office.
All models are wrong; some models are useful. -- George Box

User avatar
Richard Hull
Moderator
Posts: 12851
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

Re: Activation measurement technique

Post by Richard Hull » Mon May 17, 2010 2:58 pm

The ideal is about 1.5"- 2" of moderator to place your neutrons in the resonance range for silver. Jon Rosenstiel did a complete exposition on this in the radiation forum some years back. This is what I use for silver to gain the max activation. It puts you a bit closer to the source, too.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
Retired now...Doing only what I want and not what I should...every day is a saturday.

User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 11:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: Activation measurement technique

Post by Carl Willis » Mon May 17, 2010 4:36 pm

Hi Doug

I last considered the silver activation problem in this reply to Robert Tubbs last year:

viewtopic.php?f=13&t=6021#p39812

The best thickness of moderator depends quite a bit on geometry. Localized sources have a lot of geometric attenuation of the flux close to the source, so ideal moderators are thinner if you're close to localized sources.

Sometimes other ideas for moderation arise in the context of improving neutron economy. Light moderators with low cross-sections, like heavy water or beryllium or graphite, where the moderation is due to elastic scattering, can sometimes replace hydrogenous media for better results. Inelastic scattering can sometimes be exploited to preferentially cause the loss of energy from high-speed neutrons; such moderators may contain fluorine, magnesium, or aluminum. But for neutron activation with a fusor, the most effective moderators will be made out of dense hydrogenous materials because of the geometry of the source and its neutron field.

Heating the activation target can cause Doppler broadening of the narrow capture resonances in the epithermal band, leading to a higher effective cross-section. I doubt this would have a noticeable effect for indium or silver activation with a fusor, but the phenomenon is reliably modeled in computer codes (see link) and can be checked easily. Similarly, the geometry in the problem shown can be changed to whatever you have in order to get a highly-accurate result. But in general, I concur with what Richard said earlier about the optimum moderator.

-Carl
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277

User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 7:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: Activation measurement technique

Post by Doug Coulter » Mon May 17, 2010 5:16 pm

Ah, finally a response to the actual question asked!

Yes, I agree that for many things about 1-1.5" of good old HDPE is great, and it's what I'm using now.

It works, but my best guess is that it's mainly the tail off the mainly thermal neutrons (and of course nearly all from the reflector side of the oven are going to be thermal) that are doing all the "work" on silver or indium, as their thermal cross sections are pretty small compared to their resonances, nearly 1k::1 ratio -- big enough to look seriously at getting some good out of.

But lets do a little reductio ad absurdium here to get at what I'm really asking/trying to find out.

Let us suppose we have a moderator so thin that we either see neutrons with one scattering event, or zero. For H type things, the math I have means we either have full energy, or ~2/3 energy neutrons at that point -- some completely unaffected, some greatly (but of course still nowhere near taking a 2.xxx MeV neutron down to 5v). That's the highest moderation per scattering you can get, and after not many, you have primarily thermal neutrons -- with a tail going upwards in energy, still some hotter ones.

So the tradeoff I'm interested in exploring here is more or less, would carbon (pure) vs hydrogen be better at getting a narrower net neutron distribution around 2-5eV than something lighter.

The lighter (H or D containing) moderator would need fewer scattering events to reduce the neutron energy, but since each removes so much energy, it might be harder to hit the target energy without going past it to thermal, or getting scattering at large angles (out of the moderator and lost entirely).

The heavier (carbon or similar that does moderate but not absorb) would take more scattering events so there "ought" to be a smoother drop to energy level desired, and a smaller distribution of energies at any point in the column length, so perhaps you could pick a spot in the colum where you got a more decent distribution of energies (for silver and its ilk) someplace -- but due to more scattering events needed to get to 5 eV or so, may also scatter more neutrons out of the moderator entirely.

This could "in theory" be improved by some kind of bragg scattering if the moderator column walls were made of the "right stuff in the right crystal orientation" for neutrons in the 1 to about 20ev range, you can get pretty decent reflections back into the moderator, again "in theory" in the literature.
I have seen neutron energy spectrometers working in this range using a crystal of something just like a prism would work for light (the internal mechanism is more like a diffraction grating, of course, but the results are the same).

So the real question is:
Is is possible, by any means described above, or any I've left out, to do a much better job of getting a fixed amount of 2-something MeV neutrons down to roughly 2-5ev than the HDPE at 1.5" or so?

My gut says "yes". My experimentalist side says "that's one heck of a job of testing to find out", therefore the request for prior data or good math.

Make this a little thicker so we have several scatterings per -- and we still have a "wide" distribution, as the loss per event is so high. But for some purposes maybe we don't want such a wide distribution,and of course the other side of the tradeoff is that the more scatterings there are (and the more energy lost in each) the more there are scattered right out of a finite volume moderator.

FWIW for those who haven't seen it, my moderator/oven looks like this:
viewtopic.php?f=13&t=6069&hilit=neutron+oven#p34513

One end is a curved cut in an HDPE rod that fits over the cylinder outside the grid, with a reflector over the top, between which I put samples. The whole thing is coated in lead to help cut down the X rays that would otherwise leak out through the plastic. Due to close proximity, it does a pretty good job of capturing neutrons, as the entrance subtends a fairly wide angle of the fusor output.
I have cut various HDPE spacers to put the samples either closer or farther from the entrance point, with the range mentioned by Richard being about the best -- in HDPE. The real question is, is HDPE really best for above-thermal results, or would something else (and maybe a different geometry) be better?

It seems well worth asking due to that 1000::1 ratio shown in the cross section plots -- and of course once you have a bunch of single digit eV neutrons, going the rest of the way to thermal for other tests isn't hard at all..

Carl's link above does not address this -- all the moderators used are hydrocarbons, no pure carbon or anything else tried. Are we really sure that those are the only, or best, usable substances for things like this?

I want to point out again, that when working with lower total neutron inputs (lab safety and all that) I can't just count silver or indium for longer to get a better estimate in the wildly varying cosmic background I have here, unless I first make it very hot. So, the point is getting more hot for less neutrons.
Which speeds up other testing as each run could be say one minute vs 5 or more (assuming you have a batch of fresh cold samples to use each run, which I do).
Why guess when you can know? Measure!

Post Reply