Coincident spectra from Compton scattering Cs-137

[toand]

Using Cs-137, plastic detector (unknown spec), NaI detector, PRA and external sound card Asus Xonar U7 (24 bit & 192 kHz), I studied - with a lot of support from my professor Pieter Kuiper - the coincident spectra of electron and photon energies at angles 0, 30, 60, 90, 130 and 160 degrees. The following files are attached (pics from top to bottom):

- Set-up
- Cs137_Angle_0: spectra at 0 degree, left and right channels (full spectra), as well as left if right and right if left (coincident spectra), showing independent spectra from left and right channels
- Cs137_5_Angles: coincident spectra at 30, 60, 90, 130 and 160 degrees, showing clear correlations in energy peaks
- coinc: 2D-spectrum showing height of coincident energy pulses plotted against each other

The coincident spectra are complementary and distinct, demonstrating that coincident spectra and proper energy calibration of one detector (NaI) can be the point of departure for calibrating the other one (plastic detector). The coincident spectra could also be used for efficiency calibration.

The dedectors are fixed, shielded, small openings facing each other. The source Cs-137 is positioned around the plastic detector, with a small opening facing the source. The angle is set by varying the position of the source around the plastic detector.

Angles are approximate (+/- 5 degrees). Energy calibration was simple and rough, just the peak 662 keV extrapolating to 0 linearly. Despite this, the theoretical values and the measurements of the photon energy were pretty good.

In theory for 30, 60, 90, 130 and 160 degrees: 564, 402, 289, 212 respektive 189 keV. Spectral values, linear extrapolation from peak at 662 keV: 575, 409, 289, 213 and 187 keV.

Using a linear energy extrapolation also for the electron energy spectra (peak at 478 keV, maximum energy at 180 degrees), the electron energies at 30, 60, 90, 130 and 160 degrees are 93, 254, 409, 471 and 499 keV.

The spectral totals: 668, 663, 698, 684 and 686 keV.

The last 2D-spectrum illustrates the conservative, additive nature of the Compton scattering. The height of coincident energy pulses are plotted against each other:

- At zero degree, the coincident pattern is completely random since zero degree implies no significant energy transfer from Cs-137 gamma photons to the electrons in the plastic detector.
- At angles above zero, the coincident pulses form a linear negative relation, increasing energy transfer to electrons implies decreasing gamma photon energies in the NaI-detector.

The 2D-spectrum has been generated by using the function "Export Coincident Pulses Height" in PRA, exporting data to the statistical software R to plot the 2D-relations, representing the different angles by different color codings (symbols).

Considering it is the first attempt, it is not too bad.

[Sesselmann]

Tom,

Thanks for posting, nice interesting experiment, one which I have not yet attempted.

It is not 100% clear from the photo how you arranged your detectors during the experiment. When you describe the angle as 0˚, does it mean your detectors were facing each other?

And consequently did you change the angle of only one detector?

Do we assume your Cs137 source was at the origin?

[toand]

Yes. The source is Cs-137. The dedectors are fixed, shielded, small openings facing each other. The source is positioned around the plastic detector, with a small opening facing the source.
Tom

[pietkuip]

Beautiful data! I hade never done this before, estimated the chance that it would work to 50 % or so, but Tom was eager to try. I suggested to try first with the detectors really close to each other, and this gave immediately coincident spectra with good statistics. So then Tom increased distances for better angle resolution.

In the photo, the plastic scintillator is on the left. You can see the Cs-137 source above it: three sources with their plastic handles taped together. So the photons that get detected first go down vertically, are scattered inside the plastic detector where the electron energy is measured, and then the scattered photon goes horizontally to the NaI detector on the right. Scattering by 90 degrees in this case. For other angles, the lead blocks are rearranged, and the source is moved, and the gammas are coming in from a different direction.

We know that the .pls files contain all the data. The PRA manual describes the file format. So it should be possible to make a 2-dimensional histogram, with energies on both axes. Does anybody here have a routine to read such data from the .pls files?

[Sesselmann]

We know that the .pls files contain all the data. The PRA manual describes the file format. So it should be possible to make a 2-dimensional histogram, with energies on both axes. Does anybody here have a routine to read such data from the .pls files?

Yes, thanks for clarifying, in the mean time I more or less figured out how the experiment was conducted, I had to draw the attached diagram to visualise it.

As for PRA I suggest asking Marek Dolleiser (software author) he is the technical expert in the year 3 physics laboratory at University of Sydney, and is usually more than happy to help when it comes to educational pursuits. His email is on the download page here.

https://www.gammaspectacular.com/marek/pra/index.html

Compton Experiment

scatter.png (81.53 KiB) Viewed 14620 times

[toand]

Yes. A very good, instructive visualisation.

[Sesselmann]

Tom,

There is one thing that puzzles me with your experiment. In the first image where you are showing a spectrum of Cs137 with your plastic scintillator, the resolution of the 662 peak is virtually non existent, the peak and Compton scatter flows together like one big peak, but in the consequent spectra, the plastic scintillator shows a well defined photo peak.

How do you explain this?

[toand]

Yes. I asked myself the same thing in the beginning. Why this huge difference between the spectra at 0 degrees and the other ones? If I remember right, I and Pieter talked about it briefly, but I am not sure what ge said. Anyway, this is how I see it now.

At 0 degree, gamma-photons pass (almost) right through the plastic detector with no or minor energy change. So, there is in principle no coincident electrons, since they lack sufficient energy to generate pulses. Small angles of 5 degrees are sufficient to change this qualitative pattern, that is, a scattering where a significant energy is transferred to electrons in the plastic detector.

This can also be verified by calculating STN (signal-to-noise ratio) for total coincidence rate and random coincidence rate at 0 and 30 degrees respectively. At 0 degree the STN is 2.8. At 30 degrees, it is 11.5. In other Words, with increasing angle, there is an increase of coincident pulses between electron and scattered gamma.

Consequently, coincident spectra at zero and close to zero degree reflect the lack of concident pattern due to too low energy transfer to electrons in the plastic detector.

[pietkuip]

Steven, the peaks in the spectra from the plastic scintillator are not photopeaks. They are parts of the Compton continuum, which is a continuum because normally one sees contributions from all angles. Imposing a coincidence condition for a certain range of angles (defined by sized and distances) makes that one sees different parts of the continuum for different geometries. This would be more clear with better resolution (if we had a germanium detector). It would also be nice to generate from the .pls file a 2D-histogram like this one:


Tom answered about the first coincidence spectrum at 0 degrees (forward scattering). I agree that both these spectra are probably dominated by random coincidences. We might test this by trying lower count rates.

[Sesselmann]

Steven, the peaks in the spectra from the plastic scintillator are not photopeaks. They are parts of the Compton continuum, which is a continuum because normally one sees contributions from all angles. Imposing a coincidence condition for a certain range of angles (defined by sized and distances) makes that one sees different parts of the continuum for different geometries.

I read what you write above, but my mind is struggling to understand it, a 662 kev photon enters the plastic scintillator, colliding with an electron inside the plastic scintillator, part of the energy is absorbed by the electron and the other part exits the crystal and is absorbed by the second detector. The sum of the energies from detector 1 and detector 2 should equal 662 keV, which seems to be correct to a good approximation in Tom's experiment.

The puzzling part is how your plastic scintillator has a well defined peak when operated in coincidence mode, but had it been operating in normal mode by a source with the same gamma energy as the coincidence peak the spectrum would have been much less defined.

This being the case, the sum of two spectra from plastic scintillators with poor resolution could be used to obtain one spectrum of high resolution.