Impact of detector efficiency on spectrum and dose

[ColoRad-o]

Howdy all,

As I mentioned in a previous post, I have a decent (made by others) efficiency curve (vs. energy) for a 2"x2" NaI:Tl detector (from Steven), and finally got around to applying it to a nice, clean spectrum from natural uranium (238U plus clearly some peaks from 235U). The corrected spectum AND the energy-weighted curve (which is directly proportional to gamma dose) are shown in the attached graphic.

Has anyone else gone through this process? I can't see any obvious blunders but may have missed something.

curves for natural uranium ore

Thanks, all!

PS: Marek may change the labeling of the 'Sum quantity' to make clear that it multiplies the bin counts by the (calibrated) energy, or whatever the 'x' values are. The International Commission on Radiological Protection assigns a relative biological effectiveness factor of 1 to gammas (basically, because they are so penetrating that they are a low linear energy transfer radiation type). For gammas you can generally replace the whole-body dose in Gray (energy/kg absorbed by tissue) by Sieverts (a whole-body effective dose which becomes simple if all tissues are equally irradiated). Hence: dose rate (energy delivery rate) is proportional to the energy-weighted actual spectrum.

[Sesselmann]

Nice work...

I agree, Marek's Sum-Qty is not a very clear description. I was the one who initially suggested the feature, but my preference would have been to call it "energy per bin" but Marek didn't think it was right for a non calibrated spectrum so he settled on Sum-Qty which frankly doesn't mean much.

Energy per bin is an interesting feature, one thing that stood out to me was how the Compton plateau becomes a straight horizontal line (ignoring backscatter) when viewed as Energy per Bin, which to me makes a lot of sense, because Compton scatter is a function of crystal shape and size, therefore one would expect it to radiate energy (heat) at a constant rate. I haven't seen any paper explaining this, but would be interested to hear if anyone else has noticed this.

Steven

[ColoRad-o]

Here's the same process for 60Co, for which the Compton plateau should be more prominent

[Sesselmann]

Interesting, Compton seems to taper off a lot more in the case of Co60, which probably proves my hypothesis is wrong. Maybe it was a combination of Compton energy and crystal size that resulted in the even energy distribution I saw. Needs more investigation...

From a theoretical point of view I still believe energy in must equal energy out, i.e if a gamma ray with 1 MeV enters a crystal, 1 MeV must escape albeit by many photons of lower energy.

Steven

[ColoRad-o]

From a theoretical point of view I still believe energy in must equal energy out, i.e if a gamma ray with 1 MeV enters a crystal, 1 MeV must escape albeit by many photons of lower energy.

Steven,

Amen to that, but there are a number of loss mechanisms (wish I knew more detector theory)--electron energies probably end up as heat (via 'phonons' in the crystal which warm up the detector. I'm trying to re-define my intuition so that it focuses on the detector efficiency-corrected spectrum. After that it's a straight path to energy-weighting to get dose. Within the International Commission on Radiological Protection framework, the "relative biological efficiency" (at doing radiation damage to tissue) is taken as 1 for X-rays and gammas and 20 for alpha particles. There are refinements even beyond that: recent work suggests the RBE for alphas is closer to 17, and the following curve, for those interested, is relevant for isotropically incident gamma rays. These days the ICRP uses "computational phantoms" which very closely mimic human tissue, and the Monte Carlo tools for predicting dose by brute force use extremely well developed particle accelerator codes such as GEANT4 (and several others) which directly compute the loss of gamma ray energies to matter along its path. They agree incredibly well both with each other and with experiment.

The graphic shown shows that the RBE of 1 somewhat overestimates reality, especially at energies below 50 keV.
I'm hoping any misstatements I make will provoke health physicist lurkers to respond :) --DMW