Answer to Question #5855 Submitted to "Ask the Experts"

Q

Does the energy resolution depend on gamma-ray energy?

A

Yes, in most radiation detectors intended for gamma-radiation energy measurements, the energy resolution is dependent on gamma-ray energy. Commonly, the percent resolution decreases as energy increases. For a given detector, the largest single cause of this decrease is the result of decreases in the statistical uncertainty at the higher energies. This can be demonstrated by example for a NaI(Tl) scintillation detector. (You can review this and more aspects of detector resolution in more detail in Knoll's textbook, Radiation Detection and Measurement, Wiley, 2000.)

If a 1 MeV gamma ray deposits all of its energy in the detector, we might expect about 5 x 10⁴ electron-hole pairs to be produced. About 80% of these charge pairs result in emission of a photon of potentially detectable light, about 4 x 10⁴ such photons being produced. The ultimate signal from the photoconverter, usually a photomultiplier tube, is proportional to this number. The one sigma Poisson uncertainty in this number is (4 x 10⁴)^0.5 = 200 electron-hole pairs; thus, the relative percent standard deviation is 200(100)/4 x 10⁴ = 0.50%. If one did an analogous calculation for a 100 keV photon, one would obtain 1.6% relative standard deviation. These percentage uncertainties in the numbers of light photons emitted are not themselves the percent energy resolutions but are proportional to such resolutions. Thus, the percent resolution at 1 MeV is predicted to be about three times better than that at 100 keV. Similar rationale can be applied to semiconductor detectors, such as germanium and silicon, and even to gas detectors that are sometimes used for low-energy photon measurements.

While the statistical uncertainty noted above contributes notably to the broadening of the photopeaks observed in gamma-ray spectrometry, it is not the only source of peak broadening. Electronic noise and various fluctuations in responses of all associated electronic components may also impact resolution. Noise effects often impact low-energy results more than high energy, making resolution worse at the lower energies.

I hope this answers your question.

George Chabot, PhD, CHP