Answer to Question #11283 Submitted to "Ask the Experts"
Category: Instrumentation and Measurements
The following question was answered by an expert in the appropriate field:
I am interested in verifying my perception about the inadequacy of available instruments for package receipt wipe tests. I expect that there are thousands of routine package wipe tests performed daily just in the United States both before shipping and upon receipt. The range of radionuclides that could be encountered is more than most of the people performing these tests would be able to recognize. Without the identity of the radionuclide, converting count rates to disintegrations per second, becquerels (Bq), is not possible, so distinguishing whether the immediate reporting threshold of 120 Bq has been exceeded, on a wipe of assumed 300 centimeters2 (cm2), becomes a wild guess at best.
It is not always a good assumption that the source of external contamination is from the known contents of the package. When the contamination is known to be other than the contents (thus, cross-contaminated), unless one has already modeled every possible scenario using something like a multichannel analyzer (MCA) sodium iodide (NaI) well counter, assessing 120 Bq isn't possible, in my level of understanding of this situation.
I know how to calibrate a well counter with a known National Institute of Standards and Technology standard for measuring that radionuclide. But I don't know how to know for sure what window widths to select and how to convert gamma yield, energy efficiency, summation losses, peak shifts, and other contributing counts in a particular window into a theoretical conversion efficiency for a range of radionuclides with no available calibration standards. Does anyone know how to do this? That is part of my question.
The rest of my question is why isn't someone merging radionuclide identification spectrum analysis software with MCA NaI well counter technology to create an instrument that could be calibrated with a few common standards to address the package wipe and routine leak test market? Or is there such an instrument that I haven't heard of? I recently received a cross-contaminated package with a radionuclide that I hate to admit I didn't initially recognize. It turned out to be indium-111 (111In), I'd just never looked at it with an MCA. I was measuring about 100,000 counts with a geometric efficiency around 95%. I was not sure what the composite efficiency was, but I could assume 100% with a wide window except for some particular complicating factors related to 111In (and really anything else). What if it had been radium-223 (223Ra)? I would have had no clue where to start. It could have been almost anything. This could happen to anyone and I expect that almost no one is prepared for it. I'm planning to present this at a scientific meeting soon. I feel a little in the dark so I wanted to ask the experts whether I'm on track or missing something obvious or important. I believe someone needs to build a new meter for a really important market. It just doesn't seem like I should be the first person facing this situation. Where have the experts been who left us without an adequate instrument for something as fundamental as package wipe tests?
Your concerns regarding the assessment of contamination on received packages are legitimate, and I suspect that you are correct in your inference that many such assessments are in error because of not having properly identified the radionuclide(s) of interest and/or not having an appropriate efficiency value to convert counts to disintegrations.
To my knowledge, no manufacturers have marketed a counting system specifically aimed at the quantitative assessment of removable contamination on package surfaces. Some justifications for not doing so lie in the facts that (1) more than one type of counting system would likely be required by facilities that may use alpha emitters, pure beta emitters, and gamma emitters; and (2) there are already counting systems available, as well as software or acceptable in-house methodologies, not designed specifically for this application but that can be applied to it. Your major interest appears to be in radionuclides that emit gamma radiation so I will confine my additional remarks to assessment of such radionuclides. To address your concern regarding not knowing the identity of the radionuclide that is responsible for the contamination, it becomes necessary to use a method that first identifies the radionuclide(s). This is the case because the gamma detection efficiency for the kinds of detectors that would be most useful varies rather markedly with gamma ray energy.
You have cited that you are familiar with the use and limited calibration of a NaI well counter. Such a detector connected to an MCA constitutes a powerful system for assessing gamma-emitting radionuclides present on wipes taken of package surfaces. The high atomic number and high mass density of solid NaI offers high intrinsic gamma detection efficiency, and the well geometry offers a great advantage in that the geometric detection efficiency is high and, for a particular gamma ray energy, does not change markedly with small changes in sample geometries. Thus if individual wipes are folded into small geometries that can be inserted into the well of the detector, preferably deep in the well, small changes in the geometries associated with the folded samples will not be a significant problem insofar as possible variations in efficiency for a given energy are concerned. There is various NaI gamma spectrometry/MCA software that will perform radionuclide identification as well as quantification, following proper calibration; such software could be useful, especially if you have the need to perform a relatively large number of analyses on a regular basis. A couple of commercial vendors that provide such software are Canberra and Ortec. There is also some free software available—e.g., FitzPeaks and Theremino. (Citing of these sources for software does not imply a recommendation of such.)
If you are dealing with a rather small number of packages, you can set up the well counter and associated MCA to handle any radionuclide likely to be encountered without the need for additional software. A wide-window approach generally is unsuitable unless the identity of the radionuclide has already been established. Even the ionization chamber types of dose calibrators that are used in nuclear medicine require adjustments for different radionuclides to account for the facts that ionization density is proportional to the photon energy and to the mass energy absorption coefficient, which itself is energy dependent (but much less for a typical gas than for NaI), as well as to the gamma ray yield of the radionuclide, so calibrator settings are changed when doses of different radionuclides are being calibrated. Use of dose calibrators presupposes knowledge of the radionuclide's identity. Typical dose calibrators do not have sufficient sensitivity to measure reliably the small amounts of radioactivity that would be associated with likely package contamination samples of concern, even if the radionuclide is known.
As you have noted, to use the NaI system one must determine the appropriate efficiencies to cover the range of gamma ray energies that might be encountered. Unfortunately, it is difficult (not impossible but likely not worth the investment of time and money) to accomplish this without the use of calibrated standard radionuclides. The approach to doing this is relatively simple, but it does require obtaining a few rather low-activity sources that emit discrete gamma energies of interest. The fact that you are using a well counter allows use of various prepared sources, such as rod sources, each with a known amount of radioactivity of a specific radionuclide present in a small volume embedded near the end of a plastic rod. For example if you wanted to cover the energy range from around 120 kiloelectronvolts (keV) to about 1 megaelectronvolt (MeV), you might obtain sources of cobalt-57 (57Co, energy about 124 keV) or technetium-99m (99mTc, energy about 140 keV), barium-133 (133Ba, mixed energies with three gamma rays from about 300 keV to 384 keV; effective energy about 348 keV), cesium-137 (137Cs, energy about 662 keV), and cobalt-60 (60Co, energy about 1.25 MeV if both the 1.17 MeV and 1.33 MeV gamma rays are taken together). If 99mTc were used it would likely be as a calibrated liquid solution (perhaps 1 milliliter [ml] or so). The approach is to determine the counting efficiency for each effective gamma ray energy associated with the respective radionuclides. These efficiencies would be in dimensions of net counts in the respective photopeak region per gamma ray of the specified energy emitted (c γ-1).
A typical approach to getting the number of counts in the photopeak for a given counting time is to look at the pulse height distribution (this is often improperly called the gamma ray spectrum) on the MCA display and to define the photopeak region of interest (ROI) by taking three or four channels to both the immediate left and right of the full peak—channels that lie in the "background region" of the distribution (the "background" level may be noticeably affected by pulses from Compton scattered photons interacting in the detector). Draw a line through these channels to define the background line under the peak. The net counts in the photopeak are then calculated by subtracting the counts under the line extending between the end channels selected from the gross counts in the photopeak ROI. For a given radionuclide used in this process, the net counts in the photopeak ROI are divided by the number of gamma rays emitted during the counting time to get the gamma detection efficiency for the well detector at that energy. For example, if one counted a 137Cs source of 1.92 × 103 Bq, for which the 662 keV gamma yield is 0.85 gammas per disintegration (γ dis-1), for a period of one minute, the number of 662 keV gamma rays emitted during the one minute would be 9.81 × 104. If one observed 10,115 net counts in the photopeak ROI, the gamma detection efficiency at 0.662 MeV would then be (10,115 counts)/(9.81 × 104 gammas) = 0.103 c γ-1. Similar efficiency determinations would be carried out for each of the other pertinent gamma ray energies for the other radionuclides.
Often the gamma detection efficiencies can be plotted against the photon energy, and a simple mathematical function can be fitted to the curve. Alternatively, one may simply read off the gamma efficiency from the plot at a particular energy when dealing with an energy that has not been explicitly calibrated. The advantage of using the gamma detection efficiencies is that they may be applied to any radionuclides that emit the energy or energies of interest. For example, if the efficiency determined at 140 keV is 0.79 c γ-1, and a contaminated wipe counted in the well produces 4,250 counts in the 140 keV photopeak ROI for a one-minute count, implying 99mTc contamination, the calculated 99mTc activity, based on these results and a 140 keV gamma ray yield of 0.889 γ dis-1, would be (70.83 c s-1)/[(0.79 c γ-1)(0.889 γ dis-1)] = 101 Bq.
Once you have gone through this energy calibration/efficiency procedure you will find it is rather easy to accomplish. It should be repeated at least annually to ensure that no significant changes have occurred in the system.
I hope this is of some help to you.
George Chabot, PhD