ORTEC How to Choose the Right Photon Detector

How to Choose the Right Photon Detector

How Do You Select the Detector?

First, let us present some measurements made on different types of detectors to show several features. We will show you how to select the right detector for your application. The right detector is the detector that produces the most analyzable data in the shortest time for the lowest cost. Most spectroscopy problems can be solved with simple detectors. There is no need to have exotic or overly complex designs.

The Analyzable Spectrum: Good Data vs. BAD Data

"Good data" is defined as being spectral data in which the peaks of interest are well shaped and have good "signal to noise." This is a key consideration; just having more data doesn’t make the data better.

One measure of the quality of a spectrum is the minimum detectable activity (MDA) of the detector system. The resolution, background and efficiency of the detector are related to the MDA. This relationship may be simply stated as (Ref. 2):

The MDA varies with energy because the quantities on which it depends vary with energy. Here we have separated out all the factors in the MDA that only depend on the detector itself. The gamma rays per decay, the shield and count time affect the MDA, but will do so in the same way for all detectors.

R(E) is the energy resolution of the detector as a function of energy; B(E) is the background counts per keV (unit energy) as a function of energy and å(E) is the absolute efficiency of the detector as a function of energy.

This simple formula is highly significant in guiding us towards the right choice of detector. Let us examine it in more detail.

You can see that the MDA is linear in efficiency, but proportional to the square root of the resolution and the background. So you would expect that the biggest detector will give the best MDA for a low-activity sample. Is it always the case that "Bigger is Better"? Yes and no! More efficiency will always improve the detection limit reached in a given count time. However, you should consider the sample to be counted:

Figure 3. Background Counts vs. Detector Volume
for a Large Number of Detectors

Figure 4. Peak to Compton ratio vs. relative efficiency
for coaxial P-type detectors

Figure 5. Comparison of 137Cs/241Am spectra obtained
with 18 and 98% relative efficiency GMX detectors, showing
the effect of increasing P:C, improving MDA.

• Does the spectrum have interferences (multiplets) in which a gamma-ray peak of interest is obscured by a peak from another nuclide? Equation 1 is correct, but the resolution of a larger detector is typically worse than the resolution of a smaller detector. This could mean that a good resolution detector will give better MDAs than a larger efficiency detector.

• Does background increase as relative efficiency increases? Certainly, as the efficiency increases, the background increases, but data from a large number of all sizes of detectors shows clearly that the background increases less rapidly than the efficiency. Thus MDA improves on larger detectors. (Fig. 3) Cosmic background will also increase with increasing detector size, but will increase no faster than efficiency and thus MDA will improve. This background is the general background in the detector when no sample activity is present. As soon as a sample source with non-zero activity is presented to the detector, this will also add to the general background in the form of source induced background.

As detectors increase in size (efficiency), the peak-to- Compton ratio (p/C) increases, (Fig. 4) which means that the ratio of source related signal to source induced background in the spectrum will increase, that is improve (Fig. 4). Figure 5 shows an example of this. Two GMX detectors were used to count the same sources in the same geometry. The peak areas for 241Am and 137Cs are shown. The ratios of the counts in the spectra are not as large as the stated efficiency ratio because the stated efficiency is for 1.3 MeV only. The source-induced background is higher for the larger detector (except in the 100 keV region), but the ratio of the two backgrounds is never as high as the efficiency ratio for the peaks. We talked earlier about cosmic and other non-source background. If the Compton background has been produced in the spectrum because of a dense sample matrix, a high p/C detector will not reduce this Compton background. For example, in plutonium-in-human lung measurements, a high contributor to Compton background is the natural 40K gamma rays scattering from the person’s bones. This cannot be reduced by the detector.