ORTEC How to Choose the Right Photon Detector

How to Choose the Right Photon Detector

Sample Presentation

Samples for gamma-ray spectrometry come in all shapes, sizes, chemical and physical forms. The activity you need to measure may be very low in a large sample, or it may be very high in a small sample or anywhere in between. The matrix of the sample may be dense and have a high atomic number, therefore making accurate measurements difficult due to attenuation of the gamma rays.

You may be able to position the sample relative to the Germanium detector in a way to optimize the spectrum gathered, and therefore the results. You may have external reasons which define or restrict the choice of how the sample is presented to the detector. Some reasons you may see are:

A human being in a bioassay measurement is a fixed-format sample. There is no opportunity to change the presentation of the subject into another geometry.
A wide-area, uncollimated soil survey is a very different counting geometry than a waste drum.
While a 2 L Marinelli beaker and large detector might be the best choice, you may have already standardized on 1 L beakers, so be sure any new detector will actually fit inside your existing Marinelli beakers.

Filter vs. Bottle vs. Marinelli Beaker

Remember the MDA (Eq. 1) depends on the absolute efficiency and the absolute efficiency depends on the geometry of sample and detector. You may select the sample geometry from several different containers. Lets look at some different samples counted on a single detector. In Fig. 7, the filter paper was placed directly on the endcap and the filter active area diameter is slightly smaller than the diameter of the detector. Would a smaller diameter detector or a larger diameter detector be better for this filter paper? The best detector diameter for a disk source on endcap (that is, in "close" geometry to the crystal) is about 1.2 times the diameter of the disk (Refs. 5, 6, and 7). A larger crystal does not increase the efficiency significantly and a smaller detector reduces the efficiency. The form of the sample also has an impact on the efficiency. Three different geometries are shown in Fig. 10 and you can see the filter geometry is, by far, the best of the three examples. So if you can, you should make disk samples rather than use the larger sample containers.

Figure 10. Filter paper, Marinelli beaker and
bottle geometries compared.

Figure 11. Comparison of 1L and 2L Marinelli beakers
on the same detector.

Figure 12. 1L and 2L Marinelli Beakers compared to
1L bottle on endcap for a GMX detector.

Figure 13. Disk and "wrap-around" geometries compared.

The 1 L bottle is a larger diameter than the filter paper. The situation changes if you want to determine specific activity or activity per unit sample, such as µCi/kg. In such practical cases, you should consider, if it is possible to use the entire sample in that geometry. If only 1% of the sample could be put on the filter, but 100% of the sample could be put in the Marinelli beaker, then using the Marinelli beaker to count the whole sample would be more efficient over-all in terms of counts in the spectrum per unit activity in the original source.

In Fig. 11, 1 L and 2 L Marinelli beakers are compared on the same detector. It may seem at first surprising, but the 1 L beaker has a higher efficiency than the 2 L. The reason is back to simple geometry. The 1 L beaker puts a greater proportion of the sample closer to the detector. Thus 1000 Bq of activity in the 1 L beaker will produce more counts in the spectrum than 1000 Bq in the 2 L beaker. However, and it is important, if there is enough sample to fill the 2 L Marinelli, then the 2 L beaker will produce lower MDC (minimum detectable concentration MDA/volume) because of the larger sample.

Marinelli Beaker or a Bottle?

Figure 12 shows that a Marinelli beaker has about 3 times the efficiency of a bottle geometry. The Marinelli beaker utilizes the sides of the detector thereby gaining efficiency. At low energies, however the aluminum endcap wall, (replaced by beryllium or carbon fiber on the face of the GMX detector), will attenuate the gamma rays, thus reducing the advantage of the Marinelli.

What About "Wrap-Around" Geometries?

Figure 13 shows that a small disk on endcap has a higher efficiency than a sample wrapped around the curved surface of the detector. This initially surprising result can be explained as follows. Imagine a point source placed on the curved endcap surface. Directly below the source, the germanium is as close to the sample as if it is on the face of the endcap. However, when you consider gamma rays emitted at an angle, the curved surface puts the sensitive Ge further away from the source than it would be on the flat endcap face. However, as in the case of the 1 L and 2 L Marinellis in Fig. 12, if you can make the sample as large as the area of curved surface (much larger than the amount on the front flat surface), the curved surface has the highest efficiency in terms of counts in the spectrum per unit activity of the source. The cylindrical surface area in the detector in Fig. 10 was 15 times that of the flat disk on the end face of the crystal, which would more than offset the differences shown in the curves.