Posted on by maxfomitchev

Eberline AC-3 is a common ZnAs(Ag) area scintillator probe similar to Canberra/Mirion SA-100. The probe can be easily used with MCA-LITE and requires about 1000V of bias voltage. I decided to use this ALPHA-LITE system to check alpha-activity of thorium uranium, bismuth, and of course background. Although a ZnAs(Ag) will produce a pulse-height spectrum, the spectrum reveals no energy data of the impinging alpha particles because the alpha particles dissipate their energy as they travel through air. This is why alpha spectrometry is always done in vacuum; area probes produce indistinguishable pulse-height spectra regardless of alpha particle energy and regardless of substance being measured. Typical pulse-height spectrum acquired with the ALPHA-LITE system is shown on Fig. 1.

Fig. 1. Typical pulse-height spectrum acquired with the ALPHA-LITE system.

The detector is sensitive enough to be overwhelmed with counts when measuring uranium samples and easily picks up counts when measuring small quantities (mg) of thorium. But what about background?

Background counts can vary widely from zero 60 CPM. This variability appear totally unexpected and baffling to a novice experementalist. In fact, it is easy to take ‘background’ counts and get a reading of 1 CPM on average. Then acquire the ‘experiment’ counts and get a figure that is almost 2 orders of magnitude higher than the ‘background’. How could this be?

The answer is simple – natural radon is the most likely culprit. Radon sips through the ground and radon levels can vary throughout the day. Some places are worse than others, and basements are known to accumulate radon because it is heavier than air. In fact in U.S. radon is recognized as nuisance and health hazard and depending on where one lives one may be required to have a monitoring (e.g. alpha-counting) and management (e.g. an exhaust fan) system to keep radon levels low.

So, when you take background counts using an area probe your background may vary significantly and the counts can be all over the place. For example, when I logged background alpha counts for 3 hours using my ALPHA-LITE system I have gotten a count history that exhibited a significant systematic drift – Fig. 2.

Fig. 2. Background alpha counts history of 3 hours (180 minutes).

Inspection of the count rate histogram using the PulseCounter software revealed the absence normal distribution. The count rate histogram must have the shape of the Gaussian (or, for very low count rates, Poisson) distribution under ‘normal’ circumstances corresponding to stable background.

Fig. 3. Alpha counts (CPS) histogram showing the the absence of normal distribution.

12-hour count is shown on Fig. 4. This time the background was stable and the count rate histogram matched Poisson distribution – Fig. 5.

Fig. 4. Stable 12-hour background alpha count history.
Fig. 5. 12-hour background alpha counts histogram, Poisson distribution.

So, the presence of natural radon very much complicates detection of a weak alpha emission. At the very least one has to enclose an area probe in an air-tight container to prevent natural radon from entering the probe during counting. However, it will be impossible to control radon when one has to open the air-tight enclosure in order to remove or replace the sample. This difficulty makes area probes ill-suited for scientific experimentation – they were designed to check for surface contamination and they work well for that purpose.

Alpha counting for scientific purposes has to be done using vacuum based systems such as those made by Ortec / Ametek. Ortec 576, Ortec 676, Ortec Soloist, and Tennelec TC-256 vacuum-pump backed NIM BIN modules come to mind.