Answer to Question #7530 Submitted to "Ask the Experts"

As you likely know, one of the major advantages of the gas proportional detector over some other detector types lies in its ability to differentiate between high-density ionization pulses, such as those induced by alpha-particle interactions, and lower-density ionization pulses, such as many of those from beta-particle interactions. This ability to discriminate lies in the fact that, for a gas proportional detector operating at a particular high voltage, the alpha-induced pulses are generally noticeably larger than the beta-induced pulses because the alpha particles readily deposit much of their energy in the gas, yielding a large number of ion pairs, while many fewer ionization events result from the typically lower specific ionization produced by the beta radiation.

The success in discrimination between such pulses depends on a number of factors, among the most important being the operating voltage of the detector, the electronic discriminator level setting(s) that is intended to reject pulses below a particular level—e.g., to reject beta-induced pulses when counting alpha pulses. Other factors such as the window thickness, the sample mass and geometry, and the sample-detector geometry also influence the outcome. While for typical alpha and beta emitters the pulses from alpha particles will be mostly larger than those from the beta interactions, there is generally an area of overlap where the largest of the beta pulses overlap with the smallest of the alpha pulses.

Many of the current alpha-beta proportional counters use two channels for collection of respective pulses from the alpha and beta radiation. If a single discriminator is used there will be some spillover of beta pulses into the alpha channel and some spillover of alpha pulses into the beta channel. Usually the degree of spillover (also called crosstalk) of alpha pulses into the beta region is considerably greater than the beta contribution to the alpha region.

Some systems use a two discriminator method in which the discriminators are set to reject most of the pulses that fall in the overlap region. This technique reduces spillover but also somewhat reduces counting efficiency, since more pulses of both types are rejected. Some proportional detectors are windowless, and this has the advantage that low-penetrating radiation, most notably the alpha radiation, is not attenuated in the window material; this often results in somewhat fewer low-energy alpha pulses, thus reducing the degree of spillover of alpha counts into the beta channel. Windowless counters have the disadvantage that they can become contaminated internally and require cleaning.

When a manufacturer specifies a particular degree of spillover between channels, it is based on the use of specific radionuclides, often ²¹⁰Po as the alpha emitter and ⁹⁰Sr-Y as the beta emitter, frequently as near weightless sources. If you are analyzing alpha and/or beta emitters that have significantly different energies or that have sufficient mass so as to present noticeable self attenuation of the emitted radiations, the degree of spillover may be different from that specified. You can do some evaluation of this by using separate standards of the beta and alpha emitters of interest and preparing them in a fashion that closely simulates the actual samples to be counted. Counting each of the prepared standards will provide a more realistic measure of the extent of spillover between channels.

You can learn a lot about the operating characteristics of proportional counters by using a completely manual system in which you can control the high voltage and the discriminator level. By obtaining counts as a function of applied voltage for an alpha source, you will obtain a counting plateau that extends over a fairly wide voltage interval. You will also see that the plateau is not completely flat but has a small positive slope with increasing voltage. The implication of this is that at a higher voltage, as will be required to produce beta pulses above the discriminator level, the alpha count rate will be increased beyond what it might be at a lower voltage selected for alpha counting. If you generate a curve of counts vs. high voltage for the beta emitter, you will most likely see a plateau that begins at a considerably higher voltage than did the alpha plateau and that is much less well defined—shorter in length and with a larger slope than the alpha plateau. It is also worthwhile to use a mixed alpha-beta standard to generate a complete curve of counts vs. high voltage to observe possible changes associated with the mixed radiations compared to curves from the individual radionuclides. There are a variety of measurements that can be made with such a system that can provide some insights into practical implications of certain variables, such as sample geometry, sample mass, operating voltage, discriminator level, alpha- and beta-particle energies, etc. While such information may not be directly transferable to the operation of an automatic system that is designed for simultaneous alpha and beta counting, it can be helpful in predicting and/or evaluating the influence of various factors on the operation.

I hope this is helpful.

George Chabot, PhD, CHP