38B57 quantitative (food) measurents

[Springdream]

Hi folks,
my initial goal when stepping into gamma spectrometry was to make quantitative measurents of Cs137 in food samples (mushrooms).
Actually, it was harder than expected. I got a 38B57 tube and a GS-1100 Pro together with theremino MCA.

At the end I think I did it and would like to share the results to anyone interested:
Steps solved:
- estimate efficiency function over keV from sources with known ratio of different peaks (actually only Ra226, Cs137 and Th232 are halfways useful?!, I do not have other sources)
- measure efficiency from a point source around max 10cm of the probe (currently only in front of it at arbitrary radius from the middle) => making a regression of the 2D field =>now I can calculate the geometric influence of an arbitrary probe dimension with any distance from the scintillator
- making a food sample from mushroom (ash) and have it tested by a lab => now the total efficiency can be calibrated against the professional measurement. Together with the information above any other size of sample can be used and measured.

Now I am x2 away from efficiencies given in the literature (for a 3" only) and could be happy so far, but it is only one comparison.

Did anyone of you also try to do quantitative measurements (at least for Cs137)? I'd be happy to share some experience and provide my results if somebody is interested.

Greetings from Nuremberg,
Martin

PS: If you have some idea where to get nice sources from (e.g. from Europe) please let me know. I do not need and like hot sources or sources that produce radioactive dust. Best is something less than 500Bq. Best would be a calibrated source, but it is almost impossible to get one over here for private usage as well as only weak sources....

[Peter-1]

Hello Martin,

you can find by ebay a small probe of Lutetium. No problem in germany.
Nuremberg in germany / Bayern ? Then you are < 50km from me.

Peter

[lodovico]

Hi Martin,
I did some quantitative measurements (at least I tried) :
http://physicsopenlab.org/2017/06/22/ce ... ated-soil/
http://physicsopenlab.org/2017/07/23/an ... il-sample/

Lodovico

[Springdream]

Thanks you for the response, I like to prepare some input but need more time.
@Peter: If you are interested to get in contact with me just send me a PM.

[Springdream]

Hello Lodovico,
thank you for your response and *wow* good essay on physicsopenlab.
Regarding your formula:
Ax = A ( πr2 / 4πd2 ) fγ ε
the geometric factor is given as ( πr2 / 4πd2 ), but I'd say it is only valid for larger distances.
At least in my case for food analysis (but as your soil sample is also not too hot) I need to get as close to the sensor as possible.
And then this geometry factor does not work at all?!
I measured the near field at any radius 0...20mm from the center and from a distance 1...110mm (1mm is assumed to be the distance to the crystal).
The near field geometric factors are attached as xls to allow anyone to compute the geometric factor for an arbitrary shape of the sample.
I also compared the geometric factor given in the formula above to my measurements and they differ 10times at 50mm distance?
That would mean the activity of the soil sample would be much lower than expected?!
Best regards,
Martin

[Peter-1]

Hello Martin,

i have an old book in german from Fünfer-Neunert "Zählrohre und Szintillationszähler" Verlag G. Braun 1945. There is one complex formula for the calculation.

Glockenzähler-2016.pdf

(495.63 KiB) Downloaded 84 times

I hope this will help you.
Peter

[Springdream]

Hello Peter, thank you, but please not that this formula is for GM tubes not scintillators.
There is simulation software available but no results for the 38B57 type.
As I did measure the field already I can measure samples with any small shape and distance already. Anyone with the same sensor may use it. For sensors of other shapes the geometric efficiency field needs to be measured by experiment with a low radiation point source.
Best,
Martin

[lodovico]

Hi Martin,
yes, the short distance and the "shape" of the sample could change a lot the final calculation, I did just a try with some simple equipment and simple maths. The ideal would be a software simulation that can take into account the distance between sample and detector and also the shape of the sample.
Lodovico