Uncertainty, Average Energy, USRBIN and DETECT Cards

Scoring
1

Dear FLUKA Experts, with regards to this issue about Uncertainty propagation and average energy from the DETECT card vs that from the USRBIN card (Uncertainty Propagation Vs USRBIN Energy Deposition) that was already treated, I however thought I should write this for the purpose of completeness and future reference.

I scored energy deposition in Tissue Equivalent gasses using FLUKA to calculate radiation damage to simulated tissue in Microdosimetry, using Tissue Equivalent Proportional Counters (TEPCs). This is done by calculating lineal energy (y) and specific energy (z) distributions and their respective averages from the recorded spectra of the DETECT card, and then using this information to calculate radiation quality and associated uncertainty.

Since we are calculating radiation damage, it means there must be some energy deposited in the target. Thus, the above specific and lineal energy distributions are for (E>0 or z>0, or y>0).

I also understand that the zero energy deposition events are ONLY included (z=0) for the purpose of completeness based on poison assumption of the statistics.
See screenshot from this paper: Microdosimetry: Principles and applications - PubMed
image

As such, this often presents a large peak at zero in the recorded distributions which researchers get rid of by either deleting the first bin in downstream processing or setting E_min>0 in the DETECT card .

This therefore means that the average energies recorded by the USRBIN which takes into account the zero energy deposition events as we established here (Uncertainty Propagation Vs USRBIN Energy Deposition) and the one calculated from the DTECT Card could indeed be different if a e_min =/ 0 is set on the DETECT card (which is what I do). I set E_min ~ 1e-9 or 1e-12.

I choose this approach because I think that the first bin will always differ in energy width, depending on the E_max and so, setting E_min = 0 and deleting it later might result to some loss of information (especially when dealing with very low density gases where sparse ionizations are predominant, my case) Please, Correct me if I am wrong!

So, after you last reply, I settled on calculating the average energy and uncertainty form the DETECT with appropriated settings and propagating the associated uncertainty. I then used the USRBIN values just for confirmation since both will still be close enough.

Kindly let me know if this helps and I’m open for any corrections.

Thanks.

3

Dear @zavier.ndum.ndum,

I’m looking now your question and will be back to you soon with an answer.
Thanks for your patience.
Samuel

4

Dear @zavier.ndum.ndum,

Sorry for the late response.
The DETECT card will register each event and place them in the histogram according to their energy. It’s possible that an event results in no energy deposition, and you are free to handle these events as appropriate for your case study (for instance, using E_min > 0).

The USRBIN card records energy deposition, and by definition, it does not account for events that do not deposit energy as explained here. My personal advice would be to always use the most appropriate FLUKA card for your goal. In your case, if you want to know the energy deposition in your element, I recommend using USRBIN.

This statement is wrong:

I choose this approach because I think that the first bin will always differ in energy width, depending on the E_max and so, setting E_min = 0

All energy bins will be equally spaced. However, if you set a very small value for E_min, the output may not display the exact bin edges due to rounding errors. To avoid this, you can set a DETECT card in coincidence mode, using the detector region as the trigger, while setting E_min = 0 and using the previous low value for E_cut.

Cheers,
Samuel

5

Dear Samuel,
Thank you for your response. However, I wish to request more clarifications on these;
“…The USRBIN card records energy deposition, and by definition, it does not account for events that do not deposit energy…” Doesn’t this contradict the response from Giuseppe on this same issue (see screenshot and link)

image
image775×309 31.2 KB

Secondly about the bin widths from the DETECT’s *tab.lis file: “…I choose this approach because I think that the first bin will always differ in energy width, depending on the E_max and so, setting E_min = 0…”

Maybe this wasn’t clear enough but what I meant here is that for a given constant number (e.g 1024) of the bins, Setting different values for E_max results to different bin widths (Even though they might all be equally spaced). So, if an even results in energy 0<E<UB1, where UB1 if the Upper bound of the first bin, and another event results in E = 0, then these will all be placed in the histogram of the first bin. So, setting E_min = 0 and later deleting the first bin downstream might result to a loss of theses events. Correct?
This is why I thought that a really small value (Maybe E_min ~ 1E-09 GeV) might conserve these non-zero energy events especially when dealing with thin targets. Please correct me if I’m wrong.

That seemed to work for me (see pictures)
With E_min = 0
image
Another Simulation where I set E_min >0
image

Sincerely,

Zavier

6

Hi @zavier.ndum.ndum,

Regarding your first point about the USRBIN, you’re absolutely right, and I’ve corrected my post accordingly.

However, I stand by my statement about which card to use: if you want to determine the energy deposition in your element, I recommend using USRBIN.

So, setting E_min = 0 and later deleting the first bin downstream might result in the loss of these events. Correct?
This is why I thought that a very small value (maybe E_min ~ 1E-09 GeV) could help conserve these non-zero energy events, especially when dealing with thin targets. Please correct me if I’m wrong.

Yes, you can even choose a smaller E_min to lose fewer events. And if you want to preserve the bin width, you can set E_min = 1e-12 and E_max = 1e-5 + 1e-12 (assuming your E_max value is 1e-5).

Cheers,
Samuel