Gamma-Gamma Coincidence

[nastaran-s]

Hello everyone,

I am working on my gamma-gamma coincidence analysis and I have some issues. In fact, I see some peak broadenings and for each gamma, at least two of the coincident peaks are broadened. Here, I attached a graph to clarify. This is a y-projection of my 2D gamma-gamma plot. I chose the energy range of 992 to 1009 to include the gamma ray at 1003 keV. The peaks which are marked with red asterisk are broadened. Sometimes these broadened peaks are actual gammas, but sometimes they are not listed among the known gammas of my source. Can you please comment on this issue? Does it mean if there is a problem in the code that causes these broadenings?

Thanks, Nastaran

[mrocchini]

Hi Nastaran! I'm wondering if what you see is the result of the Compton diagonals. A few questions:

• Do the broad peaks moves if you slightly move your energy gate?
• Can you plot here the same gamma-gamma matrix on which you set the gate? I mean the 2D matrix.
• Are you using add back or not?
• How exactly you did this projection? Do you use TBGSubtraction or it is just a simple projection? Thanks.

[nastaran-s]

Hi @mrocchini,

• yes, they are moving with slight changes in the range.
• Sure, I attached the 2D gamma gamma here. (it's a 10000 by 10000 2D, I made it smaller here to see the diagonals)
• no, I haven't done Addbacks
• The photo I attached was a simple y-projection of 2D plot and I didn't apply any background subtraction. However, I wrote a root macro which selects the left and right backgrounds and subtracts them from the main peak region. Left background and right background are different(!) and the peak is broadened for that reason. In this example, you can see the first two broadened peaks after background subtraction. I defined left background from 992 to 997 keV, Peak region from 998 to 1009 keV and Right background 1010 to 1015 keV.

• 

[mrocchini]

Hi! From what you say I'm pretty sure these are Compton diagonals and not real peaks. However, just to be 100% sure, and also to better explain this to you, could you post the matrix you just showed with the "colz" option? Just do Draw("colz") instead of Draw(). Also, it would be very useful if you plot in logarithmic scale on the z axis and, please, zoom on both axes from 0 keV to 2000 keV. Another thing. We have a tool to do gamma-gamma projections within grsisort that is actually working very well. You just need to open the file where you have your matrix with grsisort and type new TBGSubtraction(ggE), where ggE is the name of your gamma-gamma matrix. This tool will allows you to easily change both the gate and the background so that it is easy to check these things for you.

[AdamGarnsworthy]

Marco is correct that these are Compton components. In the raseare spectrum they have the characteristic wings to them.

[AdamGarnsworthy]

They will also take the width of the gate you make. Make a wider gate and these features will get wider by the corresponding amount whereas real peaks will stay the width of the detector resolution.

[nastaran-s]

Thank you Marco. Perfect, I will use the grsisort new TBGSubtraction. I attached the new matrix.

[nastaran-s]

Thank you Adam for your explanation. Yes, I will make the gates wider to check those peaks behaviour.

[mrocchini]

Hi Nastaran!

Here is the matrix you plotted where I outlined the gate you set. You can see that at the energies you showed in your first spectrum you have a crossing of some intense Compton diagonals. These originate when a gamma-ray is going through Compton scattering in a detector and then you have the full absorption in another one. The diagonals point out the energy of the gamma-ray originating the effect (for instance, the stronger diagonal you have starts/ends at 1800 keV) and the sum of the energies in the two axes at each point of the diagonal will give you, by definition, the full energy.

A way of reducing these diagonals (a lot) is to use add back. Add back will reconstruct Compton events, as those originating the diagonal, in the full energy peak. Of course, add back then has other issues (worse energy resolution, need to implement the cross-talk correction, etc...), so you should decide what is best for your specific analysis.

You should do this check also on the other gates you set but 99.9% when you see very broad peaks those result from this effect.

Cheers!

[nastaran-s]

Thank you so much Marco for your complete explanation. I really appreciate it. so, I will deeply look into the diagonals for all the gates. 

Cheers, Nastaran