Activation measurement technique

[tligon]

There was a criticality accident during a lab demo back in the early nuclear days, where a researcher brought two pieces of fissible material too close together and then the apparatus would not allow them to be pulled apart. He reached in with his hands and pulled the two masses apart, saving the day but absorbing a fatal dose of radiation.

They were able to tell the exact dose by measuring the activation of a gold belt buckle worn by a researcher across the lab. Dr. Bussard knew the wearer ... said he showed no ill effects at that distance, although the activation was unambiguous.

[Carl Willis]

Geometry matters greatly to the questions of optimum thickness and composition, and if you want case-specific guidance on what to use and how thick to make it in order to get the highest specific activity in a particular activation sample, then a detailed problem has to be specified. Otherwise, one can merely only suggest in the most qualitative terms what the relevant tradeoffs are.

I have sometimes looked at other materials and other geometries in MCNP, e.g.:

viewtopic.php?f=13&t=5384#p33828
viewtopic.php?f=6&t=2998#p18567

With some specific input from you, the same code package will likely answer most detailed questions about moderator design for the purposes of activation. Without more detail, I just concur with Richard that a few cm of a dense hydrogenous moderator and a thick reflector are likely "where it's at," that graphite is too low-lethargy for the situation, beryllium and D2O too expensive in the necessary quantities and likely not as good as HDPE, and so on as I wrote previously. All this just reflects on my knowledge of apparatus that I presume is similar to yours in layout, but may be different in some critical respects.

Most people would probably say these problems are beyond the domain of paper-and-pencil mathematics. In lieu of direct experiment, a multigroup transport equation has to be solved deterministically, or a Monte-Carlo simulation performed, in order to calculate the energy-dependent flux and the activation rate, and realistically that's done on a computer.

-Carl

[Chris Bradley]

Being the inventive soul that I am; I am picturing in my head a series of HDPE discs spinning on a common axis, each with a series of radial slits (sectors) cuts out of them and displaced by some given distance, according to the speed of the rotation and that of the neutrons you want. If a neutron comes by and hits the HDPE part of the first disc, so it is slowed down a little. Then again at the next, etc., etc., until eventually it passes through the slit in one of the discs and if all is set correctly then it will then pass through the next slit in the next disc and so on, unimpeded provided its energy has dropped to the desired energy. If it is still a little too high, so it will get to the next disc just ahead of the cutout slit and so be slowed down a little further [each time], just enough that it then matches the slit in the next disc.

Each disc would need to have slits fine enough that the slits in the next disc are offset so as to obscure fast neutrons, but close enough (azimuthally) that an achievable rotation speed would bring the slit into alignment timed for neutrons of the desired velocity to get to the next disc as a slit comes into alignment.

To do some maths, say we have discs containing 120 slits of 1 degree coverage each, with 2 degrees of HDPE between each slit and each disc is 5cm apart from the previous. So if we want to discriminate 5eV neutrons at 30,000m/s, they cover the 5cm in 2us, so the rotation would have to be once every 240us or 4,000Hz = 250,000rpm.

Seems like it might be a bit quick for a rotor made out of HDPE, but I'll leave that one for some engineers to solve, to optimise for number of slits and rotation speed! (e.g. 1000 laser-cut slits would mean only 25,000 rpm is required)...Doug, I am sure you could knock a working prototype up!!!

[Doug Coulter]

Carl,
Thanks for those, we are now getting somewhere with this! Still all mostly hydrogen based, no pure carbon (or other somewhat heavier element than H that could be good -- you'd know better than I what that might be).

Of course, if it was an easy paper/pencil question I'd be posting the results instead of asking the question! I appreciate that this is hard, and I'm for sure asking a big favor here, to explore a thing that no one else seems to have bothered with so far (or at least published about where I can find it).


For purposes of the assumptions you have to make with your model, let's suppose these conditions:

4" round moderator column (length TBD), starting at 3" from source (grid) center, in line with the main neutron output, which I assume for now is straight out from the grid (in my case a line source about 4" long, which might be a little better than a point source in a spherical arrangement, but shouldn't be a real big deal).

We can forget for now that I'm actually using a piece of HDPE there that is cut to match the curve of the 6" pipe the grid is inside of -- any "wings" that provides are just bonus after all -- the chances of neuts caught by that scattering up into the column are probably not real great - mechanically it helps the thing stay put is all, and the reason I made it that way. If I get a little more from the thin part that wraps the pipe sides some, fine. Some materials you wouldn't be willing to waste that cut out of anyway.

So let us assume (because other theory says we should, and it's probably correct) that we have a line source ~4" long, ~3" away from the start of the moderator column, which is 4" round (more or less, I could make it a little bigger if needed but there's not a lot of room close in). We assume the source is isotropic, that is, there's no preferred direction that the neutrons come out, no beams or anything, also what theory says. (I am about to test that assumption with a directional neutron detector, but for now, we assume it's correct) I am using 2" diameter foils placed inside the column, flat to the direction the neuts are presumably coming from (before scattering, by the time they reach my sample I assume they're going in all directions pretty much). The big picture question is of course, what to make the column out of to get the best net neutron flux at a couple eV and how far down the column to put the sample.

This means we're catching neutrons that are coming both straight at us, and also some that are at a slight angle, perhaps +/- 20 degrees max that our column base subtends from the POV of the source, and try with some things NOT H containing and see what the energy distributions along the length of the column would be. Or rather, what distributions have the bulk of the neutrons in the desired range for resonance activation.

My gut is really telling me that more gradual moderation than you get with H will win out for single digit eV neutron flux at some distance, even though it also should increase the number of neutrons scattered all the way out of the column (because more scattering would be needed to get to the energy). I just can't find any data in the literature I have access to.

In this case, the things that made carbon such a pain for non enriched fission reactors don't apply -- if a percent or so of neutrons are captured in this moderator column by some impurity, it would be in the noise here compared to scattering right out of it anyway (which means all the calx done so far for fission reactors are meaningless for this problem). So I don't need the fancy nuclear grade of carbon (or whatever else) if it just maximizes the net neutrons at 2-5 ev in a moderator I can actually make.

For example, I'm unlikely to be able to make one of that scale from Be....hard to machine and kinda expensive. But it *would* be cool to know if that would be best!

Again, making the bottom of the thing form-fit to my cylinder, or someone's sphere would just be bonus-extra from this POV, so need not be modeled. Just use a flat plate/end touching the chamber should be fine to get a feel for what material would be best at whatever net thickness gets to the required speeds/energies, here. I'd assume that any moderator behind the sample would mostly reflect neutrons that by then would be thermal, but I'd love to stand corrected on that one too -- could it be that a really high Z moderator behind the sample that would tend to reflect without much further energy loss improve things further (kind of a metamaterial)? A tantalizing thought, at least to me.

[Frank Sanns]

It is not a one dimensional problem but a three dimensional problem. Nearly as many neutrons arrive at the detector from past the detector. To capture the greatest number of slow neutrons, the moderator surrounds the detector and is not just between the dector and the source.

Frank Sanns

[Chris Bradley]

Frank S. wrote:
> It is not a one dimensional problem but a three dimensional problem. Nearly as many neutrons arrive at the detector from past the detector.
Ordinarily, because of the distrubution of energies. But if you can slow down neutrons to a non-normal and more mono-energetic 5eV, hitting the peak cross-section of the activation material, then why would the same assumptions apply?

[Carl Willis]

OK Doug, I have the following noted:

Line source, 2.5 MeV isotropic, length 4"

Right cylindrical moderator, 4" diameter, effectively infinite length, face normal to a source radius and displaced 3", each of these materials:
-HDPE @ 0.945 g/cc, composition CH2
-Graphite @ 1.88 g/cc, composition C (< 1ppm B-10 eq.)*
-Beryllium metal @ 1.85 g/cc, composition Be
-Heavy water @ 1.11 g/cc, composition D2O
-Anything else?

Activation targets are 2" dia. thin (non-self-shielding) foils of Ag or In at their typical densities and isotopic compositions, disposed co-axially in the moderator.

Deliverable results: specific radiative-capture rate in the foils per source particle, as a function of axial position in the moderator. Capture in silver produces Ag-108 in two states and Ag-110 in two states; capture in indium produces In-116 in three states. (I can't provide detail on the breakdown from this calculation, just total specific radiative capture rate, which equals the saturation activity. You can estimate the product ratios to within 5-10% by hand.)

If that all looks interesting to you I will write and run this problem in the next couple days, time permitting.

-Carl

[Doug Coulter]

Yes, Carl, beautiful, and more than I had any right to ask for -- but that is precisely what I want, and you have the specs I think I want precisely.

Those plots you provided in that first link you gave above were quite beautiful themselves, and helped me get some insight on those materials as well -- good stuff. It's now in the library here, thanks.

As to other moderator materials? Well, it's pretty unlikely I'll ever get a mass of Be that big and then dare to try and machine it (I can do without lung failure), but what else would be "interesting" as regards the most of us?
(goes and looks at periodic table) I guess there's not much there going down from carbon, and until we see carbon, no idea if there's any point in going up farther I guess. Li is available, cheap and moderately easy to deal with (and a pound is a large lot of Li) so maybe that one. Boron is obviously out as it would react with anything getting slow and make gammas -- Li may have the same troubles I suppose if there's significant 6Li in it, dunno about 7Li if that has the same problems. Sulfur? I don't know what I'm asking here as I don't know how hard it is to plug in a new element into your model, so don't take that all that seriously if it's a pain to do or too many computer cycles. And I don't have reaction cross sections for a lot of the nearby elements, or whether they do nasty things with neutrons (like capture them and make hot gammas) in my literature collection in any detail worth using.

I think just a comparison between H and C will tell what's needed about which direction I should be looking in. We have all this nice data on hydrocarbons due to you already....so we also know what the mixture looks like and only need to tease apart the contributions of each part.

Covering Ag and In pretty much spans the range of interest, as most of the other elements we'd activate also have resonances in there someplace close to or in between those -- and correct me if I'm wrong, but aren't those about the "easiest" to activate and get a nice loud resulting count from?
I'd heard that manganese might also be a candidate (and the price is right for MnO2), but I'm thinking that I want something with pretty short half life so I "get all the radiation back I can" during a fairly short measurement after each run. Not to mention not having to wait weeks for a sample to cool back down for re-use.

This isn't for the more exotic stuff I (and I hope others) someday might activate, it's just going to be part of a measurement system I am hoping to increase the sensitivity of, over what I have now.


As I mentioned, the thing here that's killing me in getting reasonably accurate measurements here is my wildly varying cosmic counts, which I'm also setting up an anti-coincidence rig for, but as we know, those things are kind of a band-aid and never perfect. Here, in ten second counts, normalized to CPM, I will see anything from 6 to 90 cpm from cosmics alone -- sometimes both extremes in two consecutive counts, which kinda puts a rather big error bar on things that only count a few hundred CPM in the first interval and decay fast after that. So that's the motivation here.

So yes, please do as you describe, I'll be waiting gratefully for the eventual results.

[Doug Coulter]

Sounds like what you are describing here is kind of like the chopper used in time of flight neutron energy spectrometers, but with built in variable-moderation ability. Quite clever, and I might indeed figure a way to make one at some point. Yeah, I don't think I'd try spinning HDPE that fast, and as Sam Barrows once found out, even CD's can't quite take 20k rpm and explode in a most satisfying way (if that's what you wanted) when you try. I'll do some skull sweat on that one and see what I come up with -- that actually might be something not just I am interested in, you know -- could have value to others with (much) more money than we have...Surprised no one else has mentioned that one. Perhaps a carbide of something that would take the forces would do....have to look into that.

Thanks for the idea -- you get the credit, I'll do the fab when the time comes around for doing it.

[Doug Coulter]

Yes, of course Frank, you're right.

I knew this and it's why I tossed in the idea that what goes behind the sample in the moderator column might ideally be something different than what goes in the front -- something in the back that tends to just reflect the neutrons without slowing them much would seem ideal in a perfect world, eh? In other words, something pretty high atomic weight for that part.

It may not be so 3d though -- if we're hitting those fantastic-big resonances, the sample will capture virtually all the neutrons on one pass at reasonable sample thicknesses, and of course side-on it will get them all at 2" thick in that dimension, as I am assuming 2" diameter foil samples here.

So the only stuff that comes back from the "reflector" would be either stuff that went around the sides of the sample, or was too fast or too slow to get captured on the first go around. If too slow, well, nothing will fix that up, if too fast we might get another crack at that one, but my gut says that would be relatively minor -- although my gut has been known to be wrong now and again ;~)

[Frank Sanns]

Chris,

I was applying it to your chopper approach. One slit would be open but just behind it would be a solid that would destroy the linearity of the path of a neutron if it would happen to interact with it. The path then of a neutron through a series of spinning slits would not be a simple linear problem because no slit would be in "clear air" if you will and any collisions.

Frank

[Richard Hull]

It is always nice to have a kind heart like Carl run the numbers at his level of knowledge and experience. However, even the big boys tend to tune for minimum smoke on these things. The best calcs only ball park and you will never ever know to 1% the number of neutrons you have there.

Since 1999 I came to the realization and instruction from neutron metrologists in conversations, as well as readings, that neutron counting and acquistion is a crap shoot where 10% precision is quite good in the sense of absolute numbers.

As we are looking at improvements in and modifications of the fusor environment, relative measurements based on fixed measurement scenarios can, indeed afford +/-2% or better RELATIVE neutron readings if the counts are high enough and instrumentation is bolted to the floor.

To tune for minimum smoke here, (maximum response), from a given moderator and detector geometry, we need a rigidly fixed neutron source so that relative improvements in moderator-detector design can be witnessed. Unfotunately, the fusor is not a good candidate here as no two runs, no matter how well regulated, warrant any reasonable degree of precision in the source.

Here is where even a crude but satisfactory neutron source would be of value. While not potent enough to rival a fusor's "numbers", you do have all the time in the world to collect data from each iteration and alteration of the activation-moderator-detector setup. Again, you are source limited and statisitcally bound, mathematically.

Probably playing with the moderator as Jon did way back when would the be the answer. He used a water tank moderator and could slide his silver and indium back and forth in the tank to increase or decrease his activation. (older posting)

Activation in fusor operation is proof positive beyond all doubt as to fusion taking place, but it is never going to be a quick and ready, non-statistical method for gauging operation the way an 3He tube will. While I might imagine a doubling of activation numbers between resonance-thremal activation and a delicately adjusted, fixed, resonance activator system, the effort would not compare to the numbers and relative indication of system improvements announced by a good He3 system.

Still, intellectually and construction wise, it would be an interesting effort, but Jon has already trod some of the ground long ago.

I attach an image of a simple test rig that might serve well for experimental location of ideal resonance activation for silver. The drawing is not to scale. I doubt if backing this 6 X 6 cube or even surrounding it with more moderator would increase the activation to any genuinely significant degree. Naturally, placing this almost in contact with the fusor would be the ideal. Quick removal and measurement is key to this operation.

Finally, if you did not mind counting some proton recoils with the activation, embedment, from the rear of a 2" pancake GM tube facing the foil in situ might also be an idea provided you do not have "shine through" x-rays from the fusor shell. If you did, the thinest piece of lead that would stop them might be employed between the fusor and the moderator assembly.

Richard Hull

[attachment=0]7419.resonance moderator.JPG[/attachment]

[Doug Coulter]

Yes, indeed, I am quite grateful for Carl's willingness to help here.

As you say absolute measurements to a few percent are tough indeed, and this is acknowledged all over the literature.

As I stated above, either theory of isotropy in low energy DD collisions is dead wrong or the BTI's we've compared to one another are off by far more than that -- factor of 3 or more -- compared to each other on the *same* run, sitting touching side by side at the same distance from the action. That's not acceptable for that money and in my eyes, makes any "absolute" calibration of them or via them a complete waste of time. This was done with two brand new certified BTI's. Either they forgot how to make them right, or something else interesting is going on. I intend to find out which, and prove it beyond any question.

However, run-to-run measurements should be a lot better (with other techniques than BTI) if, as you say the "equipment is bolted to the floor", and relative measurement is very much more important to me, personally, than absolute, at this point. When things stop getting better as I make mods, and it seems I've reached some kind of plateau or am banging my head against some wall, I'll get more interested in absolute numbers to perhaps get a better clue as to why I am stuck and where I'm stuck.
But that hasn't happened yet -- each new thing I learn and mod I make based on that has improved the results so far.

I think you are missing a couple of important points here, Richard. One is that I am *not* looking for the ideal thickness/geometry of an H-containing moderator here at all; if I were, I already know the answers, found out the same way as you describe -- try a few things with a known reasonable starting range, which of course I got from information here (including yours) and from the literature.

What I *am* interested in here is something where very little, if any, work has *ever* been done (or at least published). Most all moderator work has been for fission reactors, where this energy region of a few eV is one they avoid like the plague, to keep from getting resonance neutron captures in the U238, so very little work has ever been done on the best design for producing this *particular* energy level from a fast neutron source -- it's normally an avoided range of energy for very good reasons if one is making a fission pile, unless you are working with the "leftover neutrons" to do breeding of plutonium in one (in which case the desired range would be determined by the U238 resonance cross section plot, not the ones shown above). After considerable search of the literature I was unable to find out some of the things Carl has generously offered to help investigate.

So I'm thanking him again in advance.

Why do I care? Again, I am trying to get a more sensitive measurement of neutron production for run to run comparisons. A cursory look at the plots above shows that compared to thermal (what everyone goes for in moderator design, more or less) is maybe almost 1000 times less effective to activate the common silver or indium we use for this than a narrow neutron energy distribution centered on the few eV range would be -- that's doggone significant, and implies that perhaps 1/10 of that factor may be possible to improve the sensitivity we get, or in other words, a lot more activation per neutron input. It may not be as good as 10x or it could be better than that, depending on the resulting energy distribution width we can achieve right where we (or the silver or indium) want it.

If I can get more good data for less radiation exposure for me and my lab, isn't that a good thing?
Getting it faster seems good too, as with other data logging going on, there's more chance that we find those good spots rather than average over a bunch of varying conditions -- so we know what to try and stabilize on "for the real thing". Far better info than "conditions changed all over during this 5 minutes and some average output from them all was X".

The implications of that are better statistical accuracy (at both activation and measurement ends of the chain), need for less neutrons in a run to get a decent reading, and just plain old making of a decent incremental advance in the art, which is always gratifying to certain types (me, at least).

Being able to make shorter runs and quicker measurements is a good thing, I believe, and as someone who runs a fusor often and long, you must know that keeping one right on the "sweet spot" for long periods is problematic even with DC input and a good pilot at the controls, so an activation test is more a measure of that -- than how good the sweet spot was for the couple moments you were right on it.

Further, despite the 3He and BF3 tubes being more or less ideal for a situation where the neutrons come out at random times (as in normal DC/equilibrium fusor operation), they fail miserably if the source is doing short, intense bursts, which mine happens to do in some modes I've found desirable while making *another* significant advance in our art -- much higher Q. The tubes mentioned (we have both here) fail miserably when all the neutrons come out in a couple of microseconds at most -- we measure some of the pulses as sub uS of power input. The simply detect the pulse rate, not the number of neutrons per pulse, as they cannot time-resolve that well. The net result is that in normal fusor operation, the tubes do track activation numbers really well here, but when in pulse mode, the activation numbers indicate that far more neutrons were produces than the tubes counted. In pulse modes, the BTI's track the activation results far better than the tubes do and probably for the same reasons -- fast bursts don't fool them like they can a proportional gas tube with long drift times.

In the pulsed case, the "tracking error" between the tubes and activation goes from something better than a couple percent to factors of 10 off and more, utterly unacceptable, the tubes are useless for this mode of operation, other than to count the pulse rate which is easily measured in other ways (like counting the pulse generator output which needs no radiation detector at all to do, but tells me nothing of the success rate of fusion).

So this is a fairly significant and new line of inquiry, and I think, worth doing, if it can give us the hoped for result, or any decent fraction thereof. Heck, it's worth doing even if it proves we can't get any improvement, as one of my mantras is "why guess, when you can know". If it's a dead end, why not prove it and be done with it? I don't think so, but that's what Carl will reveal when he gets a chance to do the math and enlighten us. Which I will then test experimentally (and which will probably agree with Carl's results) and it will then be a locked-down closed subject.

But not before -- until then it's just guessing.

As an adjunct to this, I am also constructing a beam-selecting one pixel neutron camera (unmoderated fast neutron detector with limited angular purview) just in case I'm seeing non-isotropy from the main reaction zone, which would explain some of the bizarre BTI results we've seen even in DC modes. But it would also indicate that a lot of the theory is wrong (or that some ions are getting to MeV energies someway), both of which I kind of doubt -- but mean to find out once and for all.

Besides, being able to see where the neutrons are really being produced in a fusor might be enlightening as its own thing -- so again, why guess?

[Richard Hull]

Being able to hum in a fast neutron beam to a useful rather precise area of activation to within a few EV would be like shooting, off hand, a 22 bullet through a keyhole at 1000 yards. Just not really doable. yes you can optimise to a horribly generalized degree, but you will never warrant a truly hot usable zone so narrow, especially with the output of a fusor.

I attach a combo plot of silver 107 1nd 109 from the endif 300k library.

Viewing this plot we see you are getting decent, tight activation from 5000ev all the way to 10ev. Instead of trying to hum in on a tiny peak in the single units ev range, it would be interesting to use maybe 1/2" to 1" of HDPE with decent side and rear backing to compare against a deeper and more complete moderation. There appears to be thousands of resonances in a tight wad in the 3kev to 100ev range all at 100++ barns. Lots of slop room for all kinds of decent activation at a vast range of energies rather that trying for a 10X better cross section at the biggest, narrow peak. Only experiment would tell here.

Richard Hull

[Chris Bradley]

OK - so , fantastic neutron detecting invention #2; this invention depends on maybeunobtanium, so I'm not sure if you can get it, but let's say there is a high-H content scintillating material with a reasonable light output in response to betas within the thickness of the maybeunobtanium layers.

So, what you do is build up a sandwich of layers of silver foil with thin sheets of this maybeunobtanium between them. As you have a block, the probability of neutrons bouncing around and slowing down to just the right levels for neutron activation of the silver foil will be very high (they will just keep on bouncing till they hit the right energy, or leave the block). The measurement of activation is taken from the light emitted from the polished edges of the sheets of maybeunobtanium, where the emissions from the many-surfaces of the silver foils will stimulate light emissions in the sheets which'll be guided, waveguide-like, through those sheets to their edges.

...A silver-foil-scintillation sandwich, with as many layers of silver as possible between sheets of maybeunobtanium. Is there any such thing as maybeunobtanium?

[Richard Hull]

A better mouse trap....... attached image. Click image to view details

In this iteration you can actually work for best moderation and thus activation on the fly while running allowing a minute or two for the activation to build as you move the slider detector tank back and forth in the moderator.

I gotta' build this.

Richard Hull

[attachment=0]7422.activationtank.JPG[/attachment]

[Doug Coulter]

Thanks for the plot, which btw agrees with your idea that absolute measurement is hard. That first peak you show at ~15 eV, a K barn or so -- look at the plots I got from the other literature and see just how much they disagree....20 something barns vs a thousand? Wonder which one is the correct number?

Both are no doubt from the "vetted" literature (mine is from a conference in honor of T Bonner), which should inculcate a healthy disrespect for "authority", I think.

But of course it's area under the curve intersections (between the curve of distribution of neutron energies and that of cross sections) that matters here -- a bunch of 10-100 barn stuff merely means an average of 10+ barns with random distribution of energies, which has been good enough for most people in the past....

Your plot doesn't show down to the ~10k barn peak at 5 ev, though, since it starts at double that energy.

I very much realize that no way we can get a "narrow & tight" distribution right there. But better is better if we can have it, so why not try a little harder, that's all I'm asking.

I regularly shoot .223 bullets into chicken eggs at 500 & 1000 meters (it's somewhat easier with .308 of course, but a lot harder on my skinny shoulder at the bigger calibers) -- I am a competitive shooter, and last I checked held the all time record at hunter benchrest at Roanoke Rifle and Revolver club, having shot a perfect "possible" at 100,200, 300 yards, 120 shots total through X rings the size of pencil erasers at all three ranges -- less than one caliber absolute error in the whole string. Never been done before in 160 years of club history. But I did it. I admit, I didn't do that off-hand, I used a rest...and the best gun and ammo I could build here in the same shop I now make fusors in.
Why use inferior tools if you have a choice?

And I should add, I didn't stand out all that much at that rifle match, it's not that I'm all that special.
It was the X's not the score that decided. A couple other guys were right there with me.
I don't win them all either, just many of them. Impressions of difficulty formed a long time ago may not stay valid over long time periods, is the point of that anecdote.

Which, by the way, is better accuracy than needed to make one nucleus hit another in a beam on target device...in a fusor or a beam on target, there need be no wind, no barrel vibrations, no unevenness in the bullet weight or balance, case capacity, muzzle blast turbulence (which gets ahead of the bullet for a little while), and so on.... this fusion thing *should* be easier than that was, "in theory". Vastly oversimplified, I know, but clearly, advantage => fusion.

A lot of times the impossible gets done because someone said, "there's no way you can do that" to someone else who then said, oh yeah? Watch my smoke! I will take your various negative remarks on the forum in that spirit -- I know you in person too, so I know you're not actually all that negative! Just very conservative, and like all other humans, sometimes miss the point of the discussion.

And I AM that someone else...Never tell me something can't be done unless you want to see it done.

Hopefully soon ;~)

[Doug Coulter]

Again, I'm trying to skip the H moderator in favor of something hopefully better for getting a more tight and uniform distribution in the results.

I don't see the need for the extra step in what you describe here though. For a fast response fast neutron detector, a simple scintillator plastic does quite well, the main objection being keeping electrical noise out of the signal from the phototube takes some real good care. I have one here that works pretty well, actually, and the tube response is in the 5 ns range, the scintillator and any neutron straggling add to that somewhat, with no need to wait for something activated to decay. It's what I'm planning to use in the neutron "camera" as the detector, as it's small, light, and doesn't need moderation so I can have a definite field of view via a small hole through a large moderator it sits behind. I'm within a few days of testing that lashup, so expect some reports on it soon.

I normally set the threshold of my big 3he tube about double the norm, as I only want to see multiple hits on it, mainly to keep the count rate down to audibility on the audio amp I listen to it on. With a more normal threshold on this device I get about the same count rate -- and it only "sees" about a square inch, rather than the 2 feet by 6" moderator picking up neutrons for the 3He tube. So it's actually not bad at all. This is with 3mm lead wrapping the thing to keep all the X rays out, but it probably also sees the rare high energy gammas from that 1:1000 reaction path of DD->He + y.
For my purposes, that doesn't matter -- still means fusion when I get a count.

For grins, since I have it all torn down to finish it up, here's a pic of the device parts. Sigh, so many almost-finished projects around here, I'm sure you can relate.

The first is a look down the bore, which is 1/2" hole surrounded by borated wax and interspersed with cast Cd washers, about 6" long for that part. The back half of the lead tubing holds the phototube and scint plastic shown in the next pic, with a little UV present to show the plastic glowing. In front of the tube is 1" length of cast cerroshield, which also surrounds the tube about 3/8"thick to keep stray X rays out of it all - including the ones from the boron and Cd. There will of course be a lead end cover to keep light from going down the bore, and to stop low energy (eg power supply volts level) X rays from the fusor out of things.

While this will have no where near the spatial resolution that John Hendron was wishing for on another, older, thread, it should actually *work* -- get enough neutrons to get an idea where they are coming from to some extent as it is aimed around. Should I find that it is sensitive enough or has some to spare, I'll just make another tube with a smaller hole in the moderator/neutron eater. At the neutron rates coming out of the current fusor lashup, it should in fact have plenty sensitivity to spare, but it pays to be conservative at first -- to keep the poor guy who builds all this highly motivated (eg me ;~).

But even this isn't fast enough to resolve the bursts in burst mode...they come faster than 10ns or so per during a pulse. Hence the desire that started this thread in the first place.

[Doug Coulter]

Yes, you definitely should build that -- I am glad I'm getting things started here, that is always some of my intent when I post on the forum. I am an enthusiasm-junkie, I suppose.

Now if we could figure out how to build that with some non-H substance as the moderator, it would have answered all my questions, including the original one I asked, which was -- "how does the type of moderating substance affect the energy distribution of the resulting neutrons", rather than how much water or hydrocarbon would be ideal -- I knew that answer already and so did you -- you told me, and it was correct, I did check!

Heck, all I want is more....you know.

[Richard Hull]

Here is the endif plot for Ag109 which includes the wonderment spot you are agog about. To actually realize the value of the 10,000 barn peak you will have to pack as many neuts as possible into a totally warranted range of 4.5-5.5 ev window. If you are willing to suffer a loss range where most activations would be in the 1000 barn range with a few in the 10kilobarn range you will have to warrant a tight window of only 2 ev from about 4 ev to 6 ev. Tighter than a 26 inch girdle on a 250 pound lady. All the best on this using any scheme imaginable or at least cost effective for the amateur.

With no such thing or even possibility of a neutron beam from the fusor, the classic chopper is just not possible.

Richard Hull

[Doug Coulter]

Oops, no file attached, but yes, I get you on this. I still think it may be worth a try -- as you are wont to say (and so am I) experiment rules, so I will do just that. Check up-thread for a picture of the other current project here -- a directional neutron sensor....hoping to get that tested in the next week or so.
Now, down to the shop to finish that puppy - can't do the experiment without making the gear first!

Never did I say I thought we could pack all the neuts into a 1v range around 5v, but we only have to do it a little better than now to see some of that 10k vs 100 barn improvement, eh? I was hoping at best to get 1/10 of that -- and if say, the rest are up there in the other 100 barn X section range, all the better. I was just trying to avoid the crummy 10 barn thermal range in the process. Should be quite easy to improve on that one. All those little peaks -- they do go well below 100 barns between them too, right?
Nothing is perfect, but I don't let the perfect be the enemy of the better than it is now.

And I do think you should try your last picture as a fun thing -- would make a great video demo like the one Carl did on moderation, would add nicely to that. I'd guess that you'd want really thin silver for that, as activating just one side of a thick piece, you'd never see the resulting betas on the back, or am I missing something there?

For thin, I'm really liking that real pure In from rotometals you turned us all on to (they are having a sale now, I may get more). It can be made very thin, it's amazingly ductile. Nice stuff. Also that's where I got the cadmium for the camera project and it seems great too. I made a Cd/Bi alloy, about 44% Cd for that -- safer than dealing with casting pure Cd by far, and I wanted it to stop X rays anyway. You have to chill it fast, though, or you get a mixture of large crystals of each element, eg places where there's no Cd or no Bi in thin castings. Very pretty but....not what you want I think.

I find that 2-3 mm of lead sheet will stop about all the X rays at 40-50kv just fine, it doesn't take that much, so that's a good start for your water box idea. In my case I just bit the bullet and also covered the entire tank with that stuff, a lot of work -- once, but now you can walk around the thing close while in full operation at the e6 n/sec or so range and above, and the geiger counter stays nice and quiet -- good for peace of mind if nothing else.

[Richard Hull]

Activation occurs throughout the sheet's thickness up to a point, but the thinner the sheet the better, again, up to a point. It is all related to you cross section and how close you are to the ideal energy. This why thermals activate better, traditionally in ultra thin sections.

I doubt that you will ever get even 1/1000th of the moderator impinging neutrons in any form of a 1 ev window.

Tests will tell.

Richard Hull

[Doug Coulter]

Yup, tests will tell.

On your water bath rig, which I hope you'll build. Not sure how it is now, but awhile back 5 or 10 gallon rectangular glass fishtanks were dirt cheap even new, and often show up at yard sales too. This would get rid of the carbon in the plexiglass so you'd be testing with a pure substance, or more so. The small ones were very thin soda glass in a metal or plastic frame.

Also, back in the day, I found it fairly easy to cut thin window glass (or you can have it cut at most full service hardware stores to pretty good precision) and make my own small tanks using RTV sealant, and rubber bands to hold together until the RTV cured. Although there is a runnier version of RTV meant for glass windshield seals, the plain old stuff was what I used and it was fine -- though that is becoming hard to find -- the new Ge "Silicon II" stuff -- avoid! It half cures in the tube on the shelf and that's it, it is just a defective product. The stuff in toothpaste tubes (autoparts store at higher price) is still the good stuff and you can sometimes find the old RTV in caulk tubes at the hardware store cheap. Of course, you should pay the $.50 extra for the kind with the cap you can replace...
I used this technique and still do to make tanks for nasty things like PCB etchant.


Sure, I'll never get the bulk of neutrons into a 1 ev window -- I said that already -- I also note that success for what I want doesn't require that amount of extreme perfection, either -- not even close. I might get more activation per input than now though, and I find it intellectually interesting to see just what the effect of using different moderator materials may have on the resulting distribution width at whatever energy. Seems to my intuition that materials that take less energy off per scattering event would have an effect on that resulting distribution width. You may lose more than you gain because with more scattering events needed to get down there, more may scatter out of the thing entirely.

That's what Carl's simulation should help to answer.

One might take a page from the nuke weapons designers here -- while I'm sure most of us don't agree with all their goals, I don't think any of us would claim they are stupid. And they often use a heavy substance on the outside layer of the device to reflect fast neutrons back into the reaction zone.
And various other techniques to control neutrons.

We know that works. Why ignore things that *have* been tested and are known to work?

This leads to the idea of a meta material, or composite design for *our* activation ovens, but optimized for our different goals. Perhaps certain substances work better in certain parts of the oven?

For example -- high H content in the area where most of the neutrons are way too fast, then once you get "closer" to desired energy, some heavier moderator (carbon or similar) to pull them down the rest of the way, or nearly. Behind the sample, a very heavy substance to "reflect" without taking much further energy off. At each stage, match the material to the situation and desired result.

I don't believe this has been explored even in the theoretical literature, much less experimental, and I see no theory/math that says this wouldn't be a possible large improvement on the current "simple and dumb" design. Whether it could be a *significant* improvement is something to be checked out before making some blanket statement. It doesn't need to be perfect -- I don't need 1000x gain, but I bet that 10x is pretty easy and 100x is possible.

Just another example (from a paid-professional inventor, me) == not long ago, things having negative index of refraction were beneath consideration -- only a couple natural substance exhibited this at all, in small amounts and over very narrow bands -- a curiosity with no real uses. Now metamaterials that exhibit far better NIR properties are all over the place in science news, and are finding real world applications. Never safe to bet something can't be done or won't be useful if you try it.

So far, in the literature on narrow band neutron work, all I see is what amounts to diffraction gratings that effectively select a narrow band, and simply ditch the rest (or choppers with time of flight, same result). Useful if you've got a very loud source, not so much useful for us, but even that technique might be useful to help keep medium energy neutrons from scattering completely out of the oven....It is already used "in reverse" to get sub thermal (~18k if i recall right) neutrons by scattering the hotter ones *out* of the moderator column using poly-crystalline graphite with random crystal orientation to accomplish that. The super slow neutrons have too long a DeBroglie wavelength to see this effect and just continue on through, kind of a cool effect (pun intended ;~).

I see work here as a potential "low hanging fruit" that has been skipped over by others who have different goals than we do -- something where a person like one of us might make a really new contribution to the art.

This could be important. Radio Hams got (and still do, but to a lesser extent now since they no longer lead the way) a lot of special treatment from the regulating authorities because it was recognized that they were improving the art in ways that had uses for others. We could certainly benefit from that kind of thing, as a group. But first we have to lead the way.

Edit:

After about 24 more hours of intense searching, I AM finding evidence that yes, indeed you can moderate neutrons into fairly narrow bands above thermal, using some non-conventional materials and constructions. The magic term is "BNCT" for boron neutron capture therapy, a newly hot topic, and sadly, most of the info is in pay-for papers, but thank god for grad students who publish theses. One paper describes using a combination of AlF3 and teflon, and another a combo of MgF, Al2O3, and teflon again. Who'da thunk it. The need for above thermal neutrons there is to precompensate for further moderation in the patient on the way to the boron laced tumor.

Along the way I realized that putting boron in my wax and Cd washers in my fast neutron camera was a big mistake. One that will be easily fixed by melting the wax etc back out of there and replacing with plain old wax or HDPE as a moderator for the neutrons I don't want to detect (here we go to McMaster again, the 4th or 5th time this week) -- thermal neutrons won't make light in the plastic scintillator, but the very high energy gammas off the boron and Cd will, so no point even testing that in its current form. Sigh.

[Carl Willis]

As promised I have some results from this calculation. Please refer to the post above for details on the geometry. Keep in mind that sweeping conclusions can't be drawn from this kind of calculation; the specific geometry matters a lot.

I'll upload the "pretty pictures" when I get home from the trip I am on, and I'll upload the data graphs now.

There are three graphs with results. Two illustrate the relative neutron capture rates, in units of capture probability per metal (Ag, In) atom per source neutron, as a function of depth in various moderators. One compares Ag and In capture rates in HDPE with slightly more accessible units on the vertical axis of capture rate per gram metal per source neutron. The moderators considered are HDPE, graphite, beryllium, heavy water, and Teflon. The densities and compositions in the model may deviate significantly from some real products. Most notably, commercial graphite has a very wide range of densities and equivalent boron concentrations. The model considers graphite to have a density of 1.88 g / cc and no boron, high-end material that costs a few thousand dollars for a standard bar.

What can be said from this? Well, for starters, the HDPE is hands-down the best of these five choices for a moderator in this application. The ideal depth in the moderator is 4.0 cm for indium, very slightly less for silver. As an example, with a one-gram, 2"-diameter sheet of indium properly located 4 cm from the irradiated face of the moderator, a fusor source rate of 1,000,000 n / s would generate 15 nCi of total activity (In-116, In-116m1, In-116m2) after several hours of irradiation. Why is HDPE the best? Because it can slow down the most neutrons in the shortest path length to the useful epithermal and thermal energies. In the other moderators, which are actually less lossy than the hydrogen in HDPE, the energy decrement per collision is so much lower that the neutrons tend to escape before they reach useful energies. The source is very close to the activation material, so a small change in distance from the source comes at the price of a large loss in flux just due to the divergence of the neutron field around the source. The right way to mitigate the material losses and the geometric losses in this case is to use the hydrogenous moderator.

Image of the as-modeled geometry and the "mesh tallies" will show up when I get the opportunity.

EDIT: A couple images added.

The second-to-last image shows the geometry, with the source represented in blue and the moderator in red. At left is the view toward the moderator front face in the "YZ" plane; at right is the view of the side of the moderator and the line source in the "XY" plane. For reference, the source is 4" long.

The last image is a color contour plot of neutron capture rate in Ag in an HDPE moderator in the "XZ" plane. Coloring in logarithmic, with red being highest. The length of moderator in the X direction in this depiction is 20 cm.

-Carl

[attachment=4]7445.geom1.jpg[/attachment]


[attachment=3]7430.doug_3.jpg[/attachment]
[attachment=1]7428.doug_1.jpg[/attachment]
[attachment=0]dc_1.doc[/attachment]

[Frank Sanns]

Very nice Carl!

So the conclusion seems clear. 4 cm of HDPE on the source side of the sample and 1 cm to 2 cm on the far side of the sample with a higher Z backing like graphite behind it to kick some neutrons back in the direction of the sample. A cup or V shape adding a bit more gain from the high Z backing.

Frank Sanns