Gamma Spectrometry Forum

Hi Folks,
Following our conversation (here https://www.gammaspectacular.com/phpBB3 ... a5fcda4c45) on Cs134 gamma peak at 604 keV detection in Fukushima soil sample I attempted a further experiment.
Steven suggested to use a Cs137 check source as a background in order to remove its gamma peak at 662 keV and get Cs134 first peak out of its slope.
That made me think about another possibility. I have no lead shield therefore every measurement I do involves a background subtraction, so when I test a sample I actually measure the sample + the background and then I subtract the latter.

So I thought “what if I subtracted the 662 keV peak together with my natural background”? In order to do so I needed to “build” an artificial composite background, adding to the natural background a Cs137 gamma peak of the same number of counts, relative to the natural peaks, of the sample + background spectrum. For instance, in the sample + background spectrum the counts ratio between Cs137 gamma peak centre and K40 gamma peak centre is approximately 15 to 1, so in my composite background I aimed to have the same ratio.
In order to have the right counts ratio between the 662 keV peak and the rest every 10 seconds of my composite background was made of roughly 9 seconds of just background and 1 second of Cs137 check source (9.25 kBq).

In the end I subtracted this artificial background from the sample+background spectrum and the 662 keV peak completely disappeared , leaving both Cs134 peaks well visible.
Both peaks are 5-6 keV lower than they should be, that might depend on a calibration drift (the whole test lasted about 24 hours). Anyway, I guess given the kind of measurement that’s acceptable.
The base of the 604 keV peak is slightly stretched towards lower energies, I wonder if it might be because of the presence of the other Cs134 peak at 569 keV (and in theory there's another one at 563).

The 604 keV ROI contains 28k counts, the 795 keV ROI contains 34k counts.
For obvious reasons they stand out more in the energy per bin view.

I have “something” in K40 zone. This is the third spectrum of this sample and something looking like K40 materialised the first and the third time around, but not the second (where the very weak peak was around 1400 keV, a bit too low even in case of a calibration drift).

Massimo

Massimo,

Where there is a will there is a way !

That's a great result, almost perfect separation of Cs134 and Cs137.

Steven

Hi Massimo,
Nice spectra, that is a great solution you found for it, well done!
Luuk

Thanks guys.
After all this hide and seek seeing that peak finally emerging was tasty!
Normally when you do the background you just leave it there on its own for hours, here I was constantly checking the ratios, putting the check source in and out. Basicaly I stayed there the whole time.
I used the Cs137/K40 ratio as a main raference and my aim was between 15 and 15.10, but during the six hours I made that ratio go as low as 13 and then as high as 17 to see how the subtraction would look like and with 17 the Cs137 Compton continium disappears almost completely so it would have maybe looked better, but 15 was the ratio I wanted to have the best accuracy around 662 keV so I got it back there at the end.
The check source at close distance is so much more active than the background that just a few hundreds seconds were enough to move the ratio pretty quickly.

The composite background is below in counts per bin.

Massimo

Hi again folks,
I am starting to do some analysis on my spectra using excel spreadsheets and data from PRA software.

Below there's a recap chart of the Fukushima sample test comparing data from:
- The Background.
- The Background + The Sample.
- The Sample alone.

For each of them there is the total energy of the spectrum, the energy per unit time (in different units of energy) the total number of counts plus CPM and CPS.
Then, I took it a step further (probably one step too far) and I calculated the dose on the crystal assuming a weight of 300 grams. They are really µJ/Kgh so µGy/h but since we are talking about just gamma radiation the weight factor is 1.

All this is actually from May 18 experiment while this post is about May 25 spectrum.
For the background+sample there are two sets of data because I saved at 10 hours before countinuing data acquisition and getting to 18 hours. The average values are practically identical.

Feedbacks and suggestions are more than welcome!

Massimo

Massimo,

Nice analysis, showing that the contaminated soil is around 42% higher than your normal background.

Keep in mind that normal background varies from place to place.

Steven

Steven Sesselmann wrote: ↑

27 May 2019, 09:47

Keep in mind that normal background varies from place to place.

Steven

Yeah, it also varies from time in the same place.
I am taking many backgrounds in my house before every measurement and there are differences from day to day.

I will now use this spreadsheet every time I take a new measurement to start to putting some quantitative data in the picture.
I also want to get a spectroscopic dosimeter in order to check how my estimated value for the equivalent dose compares to it.
Yesterday I tested my hottest sample so far...2500 counts per second on the scintillator. You will read about it soon!

Massimo

I couldn't help to make a test too, in the same way Massimo did :-) .
This is the first spectrum with my faint Cs137 source (a Spark Gap TG36), 37000s And this is the Fukushima Soil's spectrum + background subtraction.
82000s (45000s net.) The Cs134's peak at 794Kev is shown very clearly with a good linearity and resolution, but there is no trace of the peak at 604 Kev!
I don't know why...

pilgrim wrote: ↑

29 May 2019, 03:00

I couldn't help to make a test too, in the same way Massimo did :-) .
This is the first spectrum with my faint Cs137 source (a Spark Gap TG36), 37000s
01-Cs137, SparkGap TG36, 36000s Background.JPG
And this is the Fukushima Soil's spectrum + background subtraction.
82000s (45000s net.)
03-Fukushima Soil 45000s with background subtracted.JPG
The Cs134's peak at 794Kev is shown very clearly with a good linearity and resolution, but there is no trace of the peak at 604 Kev!
I don't know why...

I don't know why either.
I will try to guess, but I could be totally wrong: coud it be that there was too much Cs137 in the composite background and that killed everything lower than 662 keV, where most of the acrivity of Cs137 spectrum is?
In my composite background I "dosed" Cs137 in such a way that the counts-ratio between the 662 keV Cs137 's peak from my check source and K40's peak from my background at 1460 keV was between 15 and 15.10 for reasons explained above. When the background was in progress I made that ratio fluctuate between 13 and 17, just in order to see how the result reacted, and I noticed that at 17 both Cs134 peaks were still there, but the low energy part of the result spectrum was almost zeroed.
Probably with an even higher ratio the 604 keV peak would be the first to go, being Cs137 spectrum more active in that region than in the 795 keV region.
Off the top of my head I would check if the Cs137 contruibution was balanced. I mean, I would check that its activity compared with the natural background peaks is the same in both the composite background spectrum and the sample+natural background spectrum.

Luuk showed us the peak is there so it must emerge once Cs137 slope is out of the way, at least that's what I kept repeating myself while spending six hours balancing my composite background! :)

Massimo

Hi Massimo, I think you're right but I'm not sure what you mean about the ratio.
You have a PM :-) !