Fast neutron spectroscopy

[Jon Rosenstiel]

Carl Willis first brought up the possibility of fast neutron spectroscopy being within reach of the amateur fusor operator in this post describing the setup and operation of a 1” x 22” Reuter-Stokes RS-P4-0820-103 aluminum tubed He3 detector…
viewtopic.php?f=13&t=5569#p34013
And again in this post describing the setup and operation of a 1” x 22” Texlium “special” stainless-steel tubed He3 detector…
viewtopic.php?f=13&t=5690#p34134

Over the past 4 or 5 years I’ve made several attempts at doing the experiment, first with the R-S tube, and then with the Texlium tube, but for various reasons (too low of a neutron flux at the detector, no shielding, data collection time too short, noisy preamps, etc) all attempts came up negative. With the dawn of the new decade I decided it was once again time to take a stab at this fast neutron spectroscopy stuff. Evidently the planets were all in proper alignment as I was finally able to crack this nut… with both the R-S tube and the Texlium tube. Because of it’s slightly better resolution the focus of this report will be on the Texlium tube.

First image: From Knolls “Radiation Detection and Measurement.” Shows the main features of an He3 tubes fast neutron pulse height spectrum. If the full energy of our fusor’s D+D neutrons (2.45 MeV) were deposited in the detector the full energy peak will have a value of 3.214 MeV. (2.45 + 0.764)

Second image: Fast neutron pulse height spectrum from my fusor. The value I arrived at for the full energy peak is 3.132 MeV. Subtracting 0.764 MeV from 3.132 MeV gives a measured neutron energy (En) of 2.37 MeV. The vertical line labeled 1.776 MeV is 0.75 of the measured neutron energy. (2.37 x 0.75)

Setup…
Detector: 1” x 22” stainless-steel Texlium He-3 ‘special’. Tube was first wrapped with 0.025” cadmium sheet and then wrapped with 0.040” lead sheet.
Detector to poissor: 12.2 cm at closest point.
Detector bias: + 1825 V.
Preamp: Tennelec TC-175.
Main amp: Ortec 472, 3 μs shaping.
MCA: Canberra Series 35+.
Live time: 3000 seconds.
Energy calibration: Tennelec TC-800 precision pulser.

Fusor…
Input power: 60 ~ 63 kV, 10 ~ 13 mA.
Pressure: 10 ~ 11 mTorr.
Neutron output: 3.6E+06 n/s.

Interesting count-rate trivia. (From 0.3 to 0.86 MeV)
Cd + Pb shielded tube: ~40 cps.
Pb shielded tube: ~ 250 cps.
Tube in paraffin moderator: ~ 9000 cps.

Jon R

[Carl Willis]

Holy s..t.

That's phenomenal.

In looking at this, the recoil spectrum might seem to extend out quite a ways past 0.75E almost to E, I suspect because a major component of the gas fill is usually an organic (hydrogenous) quenchant. I do see a bit of a knee there at ~2.4 MeV.

I'm impressed.

-Carl

[Starfire]

Once again incredable work Jon - congrats

[Richard Hull]

Jon, Thanks for once again laying it all out in detail. It is precise and concise. Just the sort of thing a real researcher reports to fellow colleagues, then stands back for cogent comments.

As Jack Webb used to say of Dragnet...."just the facts, please, just the facts"

I remain impressed at the thorough nature of your work.

While we sort of knew the results of our D-D fusion, it is nice to see the text books verfied, especially at an amateur level. (very advanced amateur level)

I can't wait until I can retire and pass some days away doing this stuff. Maybe in a couple of more years.

Richard Hull

[Jon Rosenstiel]

Carl, all-

Here’s the comparison between the R-S tube and the Texlium tube. Note the tube bias values, +1900 V on the Texlium and +1680 V on the R-S. Shortly after these two runs I tried lowering the bias on both tubes. Here is a brief description of the result of that.

R-S tube +1680 to +1500 V:
1. Wall effect “stair-step” on the low energy side of the 0.764 MeV peak almost completely gone.
2. FWHM of the 0.764 peak considerably improved. (Narrower peak)
3. Tube not as sensitive to x-rays.
4. En + 0.764 peak almost disappeared, not a good thing for this experiment.
+1600 V seemed to be a good all around compromise.

Texlium tube +1900 V to +1650 V:
1. Wall effect “stair-step” on the low energy side of the 0.764 MeV peak completely gone.
2. FWHM of the 0.764 peak considerably improved. (Narrower peak)
3. Tube not as sensitive to x-rays.
4. En + 0.764 peak almost disappeared, not a good thing for this experiment.
+1825 V seemed to be a good all around compromise.

The above seems to indicate that to improve detector sensitivity in the En + 0.764 area one should raise tube voltage… but I was too “chicken” to go that direction. Carl, do you have any idea how much HV these tubes can take?

Jon R

[Doug Coulter]

Well, I'm not Carl, but we've run our 1" dia Texlium tube up to 2700-2800 v at times, and it's fine, but starts being not so proportional at some point and the pulses get too big for a lot of other gear to eat without some indigestion. Our tube is about 3 feet long. What we think we saw there was more sensitivity to low energy neutrons, and more tube gain, a little faster risetimes.

I will suggest that if you try that, you start out with about a 1 meg resistor in series and watch for any current to be drawn at DC -- if you do, shut down, and stay below where that limit happened from then on.

We never tried it high enough for that to happen, or to get into a geiger mode (all pulses same size).

[Carl Willis]

Hi Jon,

Unfortunately I don't have the RS data anymore on maximum voltage for the RS tube. It would have to be obtained from them again. I looked long and hard. I must have asked while I was still at Guilford...all that email is lost.

The disappearance of the stair-step on the low-energy side of the capture peak may have less to do with wall effect than with charge collection time. What is your amplifier time constant?

Do you use BNCs anywhere between HVPS and the tube? If so, the maximum safe voltage isn't more than about 1600V. A spark can toast the preamp. That would be the more immediate concern with pushing too hard on the sauce.

You will get better statistics in the fast capture peak if you turn down the gain so that you only fill about a third of the channels (or less) you use now. This does come at the expense of precision for the lower-energy features of the spectrum.

Do you have any cadmium? You could probably eliminate a lot of pileup counts from the high-energy features by killing off the contribution from room-moderated neutrons getting upside that tube.

Fun stuff.

-Carl

[Jon Rosenstiel]

Doug-

Thanks for the info, maybe one of these days I'll try pushing my tube a kV or two and see what happens. (Or maybe not).


Carl-

I'm sure you're right about the charge collection time. Amp time constant for the RS tube is 2 μs. (I believe that’s max for the Ortec 471 I’m using).

Negatory on the BNC's, my He-3 stuff is all SHV. (And I sure wish SHV connectors would show up more often on ebay).

You're right; there are way too many channels in on the act.

I've got cadmium, but it consists of a bunch of ~1.5" wide strips which I have to tape on... I don't mind taping them onto the stainless steel tube, but am somewhat leery of taping them onto the aluminum tube. Know where I could score some Cd sheet?

Jon R

[Doug Coulter]

You are most welcome Jon.

We found this out just trying to figure out how to make the tube go without any info on that particular one, and just being careful with a big R current limiter to be safe, it's not like these tubes grow on trees. What we found with increasing voltage was kind of like what you see on a geiger tube. At first it's proportional, then it has "gas gain" (still pretty proportional but real sensitive to supply volts) like some old gas-gain phototubes did, then....well, don't go to geiger mode if you value your tube living a long time. I don't think these were designed to "quench" in that mode at all, and high current discharge into them (like even a small cap in the power supply) is a big no-no.

I am also interested in finding Cd sheet, so those who look for things -- please do! I have literally a ton of wet NiCad locomotive batteries, but it looks pretty daunting and dangerous to recover that, unless I can just use it in the form it's already in -- coated on carbon particles in a screen-bag holder in the plates, it seems from just looking at it.

On another note, I'd love to see you interpose some things between the fusor and detector with this setup -- say some thin moderator type things and other things of variable Z to see what that does to the curve shapes. Eg not enough moderator to make all the neutrons thermal, but something less to see what happens in between the thermal and full energy neutron peaks you show with varying substances.

It seems your setup is pretty clean to have such little continuum between the major peaks, not much accidental moderation going on nearby. Very nice!

I won't predispose with what my theory thinks we'd see here, but it would be interesting as a "blind test" of it. I'd like to see thin HDPE, graphite, Al, things like that both crystalline and amorphous.

Could one semi-moderate fast neutrons down just some, to get a lot of them into a reasonably small energy band but not down to thermal -- somewhat above that, say in the tens of eV range? That's a question to which the answer would be pretty valuable if positive. At least to me.

I would guess from the overall shape of your plots you're not correcting for the 1/v presumed response these have?

Good show!