Neutron rad safety

[Doug Coulter]

We've been having some pretty good success with fusors lately, and the question of are we getting hammered too badly by fast neutrons came up. I'm starting a stub here, and I am hoping that someone will come along on this (and other threads) here and fill in perhaps with some unit conversions -- numbers, rads, rems, sieverts, and all that kind of thing. I've read up on most of it, and the "biological relative effectiveness" of this and that, but it's daunting to someone who doesn't want to make a life work of it, and just wants to be reasonably safe while working this stuff.

At any rate, the book Fast Neutron Physics, Part 1 (1960) has a couple chapters on fast neutron biological effects I will share here, as most aren't likely to get the two volume set for $250 -- used.
It wouldn't surprise me if the dosage recommendations and limits had been tightened since that book was published, but I will have to go with data I have.

After studying data from actual injuries in accidents, the bombs in Japan, and also people who either had no troubles or non severe troubles, they came up with a number for 40 hr work weeks for 40 years with no observed troubles from exposure which was that 60 neutrons per sq centimeter per second was tolerable and didn't give eye cataracts over that time span, so it seems a usable number. They separated this data from data that also involved other rad exposure. They wind up recommending a lower number of 20 n/sec/sq cm for long term daily dosages, but note that no symptoms occur in their studies below 60 or more. Unfortunately, they flip units from rems to rads to....you name it throughout, and no conversions are given hat would make modern gear more useful in measureing. They are assuming RBE of 10 for fast neutrons (eg for the same energy deposited, they're ten times worse than gammas for us biological critters).

So, I worked that backwards to a fusor making 1 million neutrons per second -- how far away from the grid would you have to be to be safe in that situation if you were going to do it all day every day. The answer really surprised (and relieved) me. FWIW, elsewhere I worked out the total fusion energy from 1m n/s fusion rate, and it's about 1.1 uW -- not much. Most of the radiation from a fusor is X rays, mostly those of power supply volts (400 watt DC input vs 1.1 uW fusion output -- makes sense). The tank does get real hot, so a lot of that just goes to heat one way or another.

Ok, so assuming we have a point source and the neutrons are isotropic (have to begin somewhere) then what radius sphere do you have to have for the neutrons per second on the surface to be 60 n/s? The answer was a pleasent surprise. I worked this by dividing 1e6 by 60, which is 16,666.66....that's kind of the ratio we need to reduce by to get down to 60 n/s/sq-cm.
A sphere with 16,666 sq cm area is about 35-36 cm in radius, so I'd been far enough away the entire time, in fact perhaps never even closer than twice that during a run, as this is happening at the back end of a large tank I sit in front of. My face is perhaps 3 feet from the grid when I peer in through the lead glass window.

Now, during that same time, a geiger counter (not known for hyper sensitivity to X rays) that reads 42 cpm on my lab background, reads up to about 1000 cpm, or 10-20 times background, right between me and the lashup. Cut that ratio in half to compare to the background that existed when we were growing up from nuclear weapons tests, but still -- that's getting to a point where I worry a little bit. This is with my tank coated with 2-3 mm of lead all over, and 1/4" lead back near "the action". Without that the numbers would be much higher (they were). And at this point I suspect that a bunch of this is scatter from a couple of leaks near the "business end" where we put the neutron oven, and X rays and gammas that get out then scatter off the ceiling and so on to get into the geiger counter, so obviously I have more work to do there to be some sort of "safe". We plan to put a gamma spec head around this to see how much of that is the rare reaction that makes the 16 mev gammas, which would be one heck of a job to get stopped. My floor wouldn't support the lead needed to do that for the large tank, I'd have to go to a much smaller one to keep it down in the sub-ton range at the thickness required.

Now, if anyone knows how to change Sieverts into neutron/second counts, let me know so I can back check this with a BTI calibrated in Sieverts (actually micro Sieverts). The BTI's only respond to "fast" neutrons, which from my other readings are the most "biologically effective" and it seems to say that the 2.5 mev ones we make are about the worst there is -- perhaps 5-10 times worse than thermal ones. As the RBE goes down with neutron energy (so it seems) then perhaps a safety nut would want to use some moderator around the tank?

This isn't much concern right now, it seems, as it seems we are well under the doses that one worries about, or did in 1960, and we don't do this 8 hours a day. But as we do better (we're approaching 10 m neuts/second) it will become an issue it would seems, so it's a topic of some interest here.

Note, our Henny Penny scintillator survey meter goes absolutely nuts when the fusor is on, even at some distance. I think it must be seeing and counting some awful low energy stuff to do that -- it reads a couple k counts second off a feista ware plate. This is making BillF worry, even though it also does that inside a car simply driving past a shale outcropping...Maybe I should take it away from him? I'm kind of willing to trust a new 2" pancake geiger to tell me if anything is there that will hurt me. When we bought it from Geo, it came with a cal sheet for X rays so you could test in X ray facilities, so it has a known sensitivity to the ones of energies that can do bad things to you. Maybe I can find some old CRT TV's to let Bill put Henny in front of and watch it go nuts there too -- if that would kill you, we'd all be dead! I suspect, but haven't yet proven, that it responds all the way down to single digit kv X ray energies, and as a counter cannot tell the difference between those and the real zingers.

[chrismb]

Does this answer your question? (from "Nuclear Reactor Engineering", Glasstone/Sesonske, 1967):

[Doug Coulter]

Yes, that helps a lot, as I have some rem to other conversions (perhaps we should collect all of them here).

I processed the chart a little further so it's easier to read... a flat scan would be better of course (but I know it's a pain, because I do it here a lot).

[chrismb]

Sorry, I don't have a scanner.

But bottom line [literally!] is that according to this, and text, it looks like fusion neutrons are worth;

___1mrem/hr = 6neuts/cm^2/s

and max continuous rate for a 40hr/week industry worker is 2.5mrems/hr.

So 15neuts/cm^2/s is the max continuous rating which is 1Mneut/s at a distance of 72 cm.

Or in rateable DD power, the unshielded distance from which represents the safe continuous 2.5mrem/hr level;

___1 uW = 850,000 neut/s == 67 cm

___1 mW = 850,000,000 neut/s == 21 m

___1 W = 850E9 neut/s == 670 m

___1 kW == 21 km

___1 MW == 670 km

___1 GW == 21,000 km (...so...how close is your nearest nuclear power station??...)

The text also describes 'occupational exposures'; 3 rems in any one period of 13 weeks, which would be the same as being 1 m away from a source that has emitted a grand total of 1E12 neutrons (~1 Joules worth).

From wikipedia; "A dose of under 100 rems is subclinical and will produce nothing other than blood changes. 100 to 200 rems will cause illness but will rarely be fatal. Doses of 200 to 1000 rems will likely cause serious illness with poor outlook at the upper end of the range. Doses of more than 1000 rems are almost invariably fatal."

...So final conclusion is... don't produce more than 1E14 neutrons in your [own] lab, in your lifetime.....

[edit: sorry, originally missed a '0' above 1mW..oopps.... ]

[Doug Coulter]

Good show Chris! Well, in real life there are things like shielding, and capture. The thing that is most frustrating is that most captures result in fairly high energy gamma rays, which themselves become hard to stop. And of course if you wait long enough, neutrons go spung and become H atoms (also with a gamma). I've been looking into that at this US gov site:

http://www.nndc.bnl.gov/capgam/indexbyn.html

Which shows how energetic the gammas are when a neutron is captured by something. Bottom line is that the lead you need to stop those winds up as the bulk of shielding in most practical cases. Things with nice high capture rates like Cd (555 KeV) or B (4 Mev) are fairly nasty gamma emitters when they do (and silver and gold are out of my budget!). The actual picture is somewhat worse, those numbers are merely the most common gammas, not the whole story. Even H capture is like 2 MeV. A couple of mm of lead easily stops the mere 50 KeV X rays from the power supply (in fact, it's rated by radiologists to about twice that) but...

Sure would be nice to find a simultaneous large capture cross section and low energy gamma output, but I'd guess those are sort of mutually exclusive? Can't tell from that one site, it's a laborious set of matching info from here and there to find things like that. Of what I've found, Cd looks like one of the better ones, actually, with "mere" half a meg gammas (mostly).
I can buy it fairly cheap (rotometals) but....not fun to work with -- very poisonous and tends to catch fire in the casting melt with resulting fumes one breath of which is fatal. I sorta solved that by making alloys of it that dissolve the Cd at lower temperatures, but when they cool they tend to segregate, so you can't use them thin, or you risk holes in the coverage. Very pretty, but not that useful. The piece pictured is about 5" diameter and about 1" thick. It's very brittle in this form, so you couldn't cast or pound it into sheets and form it around things.

Semi eutectic CD alloy

I used some alloy design from the Rare Metals Handbook here (though at the moment I forget which). This has Bi, Tin, Lead as well as about 40% Cd. The thing is the crystals are so large it has to be used pretty thick or you've got holes in the actual Cd part of it. Or in other words, it's only a solution when molten.

Which becomes the problem here -- gammas created by various neutron reactions with things (which is taken into account when you capture neutrons in tissue as well). It seems as though I'm going to be very glad I got the thing to nearly pure hands-off operation, and have a nice cave in the back I can put a future one into for remote operation and observation. I'd think 50 ft thick granite would do fine for shielding (and it's a bit radioactive anyway already). But no way my floor is going to handle 3-4" thick lead around that big tank even if I could afford that much!

FWIW, with the shielding I have in place now, my trusty old geiger counter reads about 1 mrem/hr with the fusor at full input, between me and it. I don't think that's seeing any of the neutrons though. Of course, it's not running 40 hour weeks, no where near that. More like once in awhile, we'll make 3-4 5 minute timed runs in a day, maybe one day a week at this point.

I am considering making a much smaller version I could shield due to the small size...but it would still take mechanical advantage of some kind to lift into place. After all, if you take a piece of 6" diameter pipe a foot long, and encase it in 3-4" thick lead, it's going to weigh more than some auto engines.

Even a couple of the lead isotopes produce MeV range gammas when hit with neutrons! So it seems if you want a shield, you need a composite of moderator/capture material, then something that eats the resulting gammas, and those can be hot indeed. This is what caused my first attempt at a neutron one pixel camera to be an epic failure, BTW. I'll have to rebuild that without the cadmium and boron in the neutron shadow part, and just depend on a moderator there and the fact that the plastic scintillator won't see slow neutrons well.

But in fact, this is the problem you want to have -- success! As long as you manage to live through it somehow. We are actually looking into how to scale this down as our results get better, but other aspects of the physics may not allow it, the jury is still out on that one.

[Joe Jarski]

Doug,
Did you try using chills (large heatsink) when that was cast? It might help with the grain size somewhat (faster cooling/less time to segregate and form large crystals), but at that size it's probably tough to get a reasonable grain size anyway.

[Doug Coulter]

Yes, I did try a little quicker cooling, attempting to get to say, 1/4" thick without voids -- failed but it looked better than the picture which was just natural cooling in a parts tray. It helps but my method wasn't so good. I suppose I would have to make more of a custom mold to pull that off well, so I could dunk the whole mess into a bucket of water or similar. Can't start with a cold mold for thin things -- it hardens and stops the flow into the rest. This is weird stuff to cast -- it gives off a lot of heat of solidification right at that temperature, and expands some, sticking to the mold, it wets everything all too well. To even get it out of the parts tray (which has a lot of "draw") I had to use silicone mold release. A flat bottom Al pan means casting in a pull-out bolt or similar or the only way to get it out is to break it. The stuff isn't terribly heat conductive (or more correctly, has a big ratio of thermal mass to conductivity) so even if you chill the outer part the inner part takes its time cooling.

The big issue though, is that when this stuff "eats" neutrons (which you have to have slowed down first with a moderator) it then emits really hot gammas that take very thick lead to absorb (inches). So, given the above data and X ray dose data, it just moves the problem around, without solving it so much -- X rays are already worse in rems than the neutrons. Which kind of makes sense -- we put in ~ 400 watts, and most of that becomes heat and low energy (50kv) X rays, only a few micro-watts of fusion happen and that makes neutrons and also some hot X rays. The 50kv x rays are no sweat (well, I sweated putting lead all over the tank!) but this really hot stuff in the megavolts is a real bear. So a shield design looks like an inner layer of moderator (HDPE about 2" thick) then the Cd alloy, then lead a couple inches thick. On my rig that would amount so some tons of stuff due to the sheer size of the tank. I can make (and will) a smaller one, which will help some, I think, but I'm trying to stay with the big, convenient version as long as possible -- all those doors and feedthroughs are nice for experiments.

I plan at some point to go with a-neutronic reactions and fuels or closer (for example, the p->Li reaction), which are harder to make fuse and harder to instrument as they make hot alpha rays, but there's a sea of hot protons to try and measure that in, so I've been putting that off. I just now got really good neutron measuring gear going!

[Joe Jarski]

Yeah, it sounds like strange stuff to work with. Trying to shield all of the high energy stuff is a massive task in itself - find a way to stop one and it transforms into something new and keeps on going. It sounds like you're making good progress with your fusor... kind of a double edged sword, except the complications grow exponentially with fusion output!

[johnf]

All of this is probably why our accelerator @ work uses 0.75m of concrete

good for Gammas and neutrons up 3MeV.
I know quite a few years ago they did a weekend experiment with the 6MV tandem not sure if it was 12MeV P- Be or D - Be but the neutron flux was detectable many 10's of miles from our site with mobile detectors.

[Starfire]

John - thats scary.