Calibration standard for neutron counting.

[KJNW]

Hello all, and particularly to Carl Willis.

The other day I came to have in my possession (at the same time) 3 Reuter Stokes neutron tubes of the same make and model. One after another I attached them to my Ludlum 2221 rate meter, dialed in each tube using the approved "Carl Willis" calibration procedure and my trusty collection of "Big Ben" radium alarm clock hands, then fired up my Fusor and dialed in to the same settings for each Fusor run. While I am sure my controls don't allow me to duplicate the same output precisely each time I run my Fusor, I am confident I can hit the same order of magnitude. The delta between two of these tubes was 2X with reliably repeatable results. I kept the hottest one and returned the others. Now even with my neophyte intelligence I am able to understand that this tells me that I now possess the hottest of three neutron tubes! That's all it tells me. As far as telling me anything about the output of my Fusor, it is worthless. The CPS\\NV for this tube is 26, but after the comparison that number is also meaningless to me.

I wish to purpose a solution that I thought might work for my situation and others as well, insofar as passing a calibration standard among us for neutron counting. Carl posted a video on You-Tube over 2 years ago utilizing a P2042-1000 leased polonium source and a small brick of beryllium. Now it just so happens that I have those same two items here in my lab. If Carl were to be so kind as to measure the output of that source configuration in the same manner he demonstrated on the video, and post the value, distance to the tube and the date of origin for the polonium, I could then calculate the delta in the decay between our sources, and come up with a factor that would allow me to calibrate my various BF3, HE3 and B10 tubes, cables and configurations, and hopefully arrive within 10% of reality/sanity.

You see, I like to tinker, and sending all of my various neutron counting configurations to Ludlum to dial in would cost more than my Fusor. I would love to own a reliable process that would get me within 10% of reality. Now perhaps someone else has a far cleaner and easier approach to this than what I am suggesting, and by all means I want to hear it. I just figured this might be a good place to start this conversation.

Carl Greninger

[Doug Coulter]

Your suggestion is good, but I have one I like a little better. Using a standardized "neutron oven" activate a 2" sq piece of silver sheet, then measure the resulting activity with a beta sensitive geiger counter (usually a pancake of 2" diameter).

There are a few reasons I like that way better. The size of the moderator in the neutron oven smooths things out, and the results don't vary much with minor changes to the "oven".

You cannot ever fool it with a wrong threshold or EMI - nothing but neutrons will activate silver.
This is probably the main advantage to the technique.

If you run 5 min at a million neuts per second, the calibration comes out fairly close to 1k cpm/million neuts/second during the run, though the end of the run matters more than the beginning as some of the counts come from real short half life silver, and you gotta be fast getting it to the counter, or have a way to extrapolate back to turn-off.

The way I do this is through my standard data aq hardware, using a log plot, the silver decay is linear, so you can easily draw a line back to where the fusor was shut off and get a delay-independent assessment of the original activation at that instant.

Here's a run of Richard's fusor at HEAS 2009 where the consensus was right around 1m neuts/second during the run, as measured by us, Carl, and a few others.
You can see at a little over 325 seconds where the fusor was shut off, then the silver put on the geiger counter and allowed to decay long enough to get a good extrapolation. Here the geiger counter is on the left, and our neutron counter (a fairly numb hornyak, but real stable) is on the right, so it's real obvious when he shut off the fusor.

There's more on my site - the project is open source, both hard and software, so if you want to build your own - go for it. We'll also sell you one if you don't have the time or skills to make your own.
More on my site here:
http://www.coultersmithing.com/forums/v ... f=52&t=467

We've since built on this design, and are now calling it standard data aq. Same deal, open source, it's what I used to show those cool 4d plots in a post I made recently here of the results of a bunch of runs with different parameters. The new version doesn't have a built in geiger pancake and power supply (limited availablity on ebay for the pancakes since Fukishima) but does have two counter inputs and 8 a/d inputs for things like power input and gas pressure. That one's not for sale just yet, we are in the "put it in a quality box and get all those BNC's in stock" phase, but it will be around soon.

Note, while I used "cross platform" tools for this, it really is mainly a linux program set. You can make it go on windows, but you have to install a bunch of things from perl and some modules, to the GTK2 runtimes - it's a lot of work and we don't yet have an installer of the "put in this CD and click yes a bunch of times" type.

[Carl Willis]

Hi Carl,

My recommended approach to calibrating a He-3 detector (or really any other kind of detector) to calculate flux from a count rate is as follows:

1. Buy or borrow a dose-equivalent detector (a BTI dosimeter, Bonner ball, Landauer CR-39 badge, etc.) The BTIs are the cheapest option at ~$150, while the Bonner detectors are probably the easiest to borrow for a day from nearby nuclear facilities.

2. Operate your fusion neutron source and simultaneously measure the dose rate with the dosimetric instrument and the count rate with your other instrument in a controlled geometry.

3. Calculate source rate from equivalent dose rate using the conversions routinely discussed on the forum.

4. You can now calculate the efficiency of the unknown detector in counts / neutron, keeping in mind that the measurement ONLY applies to the specific geometry (source, detector distance, presence of a particular moderator.

The idea to make a standard PoBe source isn't a bad one, but it's plagued by several unfortunate realities. Firstly, these Nuclespots are not calibrated Po-210 sources. They contain not more than 5 mCi, but it could be considerably less, and I really have no idea what the variability is like on those things. Secondly, the neutron source made by such a method has an unknown conversion efficiency. To use it as a calibration source, one would have to establish that a certain activity of polonium contributes to a certain neutron yield, and due to the materials and geometry involved this is not reliably calculated, simulated, or measured. And lastly, the neutron spectrum from the Be(a,n) reaction differs considerably from the 2.5-MeV DD fusion spectrum, meaning that detector response will be different. The best you can hope for is to be able to compare your detector to a detector of mine, for Be(a,n) neutrons. That strikes me as being knowledge of very limited applicability, but at least it would be easy to do.

A final idea is to resort to modelling of the geometry and materials and neutron transport through them, using MCNP. My sense is that this is a usefully-accurate approach based on experience over the years.

-Carl

[Richard Hull]

Carl has recommended the method I have used and have expounded upon at length in this forum. Calibrate against a freshly binned BTI bubble detector from dose to flux or isotropic emission. You need to do this over a wide range of regimes, time periods and voltages for many different runs to accumulate a good sample of data points. From here you have a direct count rate to isotropic emission factor constant to use in future countings.

Again, to be even 10% accurate, and I doubt a 10% accuracy, a fixed geometry source-to-detector must be established and maintained. Moving stuff around once done is a no-no. My calibrated setup has not moved 0.1 inch in 7 years.

I must continue to point out that it is nearly impossible in amateur hands to obtain extremely accurate "absolute" neutron numbers from an unknown source to even 5%!
10% is a best normal effort and 15% is more normal. However 2% or even 1% accurate differentials are easily detected from run to run. This allows one to observe, accurately, the effects of changes made within the source.

What ever counting device or mechanism is used you want large individual count numbers within a data set for obvious reasons and a properly functioning and discriminated 3He tube is going to give you that.

Richard Hull