Answer to Question #6163 Submitted to "Ask the Experts"

Q

Question 5204—"Efficiency for a well counter"—details how to determine the counter detector efficiency for a particular window when detection windows are set up around the various ⁵⁷Co and ^99mTc peaks. How would you modify this answer if a single-wide window is used (20-1,500 keV)? I am interested in this because using a wide window is what is done in practice for wipe and leak test samples for package receipt and sealed source checks where potential contaminant energy is initially unknown.

It is useful in determining if a contaminant approaches the 2,200 dpm (disintegrations per minute) regulatory limit level to know detector efficiencies at various energy levels such as low for 140+/- keV range, medium for 365+/- range, and high for 662+/- range.

Once a contaminant is detected, a glance at the spectrum would allow selecting a low, medium, or high energy. Then, applying the correct energy-dependent factor would give the desired dpm estimate. I recognize that when I use a wide window it allows into the detected count range Compton scatters, certain x-ray peaks, and coincident detections which would show up at multiples of the actual energy. I would appreciate a discussion of the errors that occur when I use this approach.

A

There is no reason why you cannot use a wide window for general assessment purposes, as you presently do. You do have to have some decision criteria as to when contamination is present. Normally such a decision requires knowing the background count (rate) for the system you are using. The background for a well counter, especially when a wide window is used, may vary depending on what sources are in the vicinity, which depends on what activities are ongoing. In general, for a given background count rate, R_b, the decision level is often given by the critical level, which is specified at a particular confidence and is given by

where T_b is the background counting time, T_s is the gross sample counting time, and the value of k depends on the specified confidence level. For the commonly used 95 percent confidence level, k has a value of 1.645 and, if T_s = T_b, then L_c equals 2.33 (R_b/T_b)^1/2. For example, if you did a five-minute background count and a five-minute sample count and obtained a background count of 750 counts in the window of interest, the value of L_c would be 2.33 (150/5)^1/2 = 12.8 cpm.

This is interpreted to mean that, when we counted our sample (wipe), if we obtained a net count rate greater than 12.8 cpm, we would assume that activity was present.

Note, this value will change if the background rate changes, so the background should be obtained shortly before the wipe is counted. If either the background or sample counting time changes, the value of L_c will also change as per the above equation. Thus, if you counted the above background for five minutes but your sample counting time was only one minute, the value of L_c would have been 22.1 cpm instead of 12.8 cpm.

Once you have concluded that activity is present, you must then determine how much. This now requires knowing the efficiency of your system. You said that when you determine activity is present you would evaluate the spectrum to determine what energy was involved, and then you could select the correct efficiency of your system. This means, naturally, that you would have to have evaluated the counting efficiency for gamma rays of different energies. This is fine, and you have the option of basing such efficiencies on counting in a wide window or a narrow window (around the photopeak).

If more than one radionuclide that emitted different gamma-ray energies was present, you would pretty much have to use photopeak efficiencies rather than wide window, but this is unlikely to be a common occurrence for the situations in which you are interested.

In cases where net count rates are low, you may conclude that activity is present, based on a count rate above your decision level, but you may not be able to discern a clear photopeak. In such an instance you could either count for a longer time until you could identify the radionuclide, or you could apply your most conservative (lowest) counting efficiency to provide a maximum estimate of the activity. The actual process for determining the counting efficiency for a particular radionuclide would be pretty much as outlined in ATE question 5204, to which you referred.

George Chabot, PhD, CHP