Answer to Question #8227 Submitted to "Ask the Experts"
The following question was answered by an expert in the appropriate field:
If we use a NaI detector to detect the radiation from a radiological dispersal device (RDD), how would we deal with the distribution of the radiation on the ground? In some papers it has been taken as plane source. However, we must consider the natural radiation from the soil since the radiation that the NaI detector sees is the combination of the RDD and natural radiation. How can we compute the net radiation from the RDD since the radiation from the RDD and the soil are very similar? I think the most simple solution is that we neglect the counts from the soil if the RDD's activity is rather stronger than natural radiation, but we also do not know what the real situation is.
You state that you are using an NaI detector to detect radiation from an RDD, but you do not say what the intent of the measurement is—e.g., are you trying to evaluate the extent of ground contamination (e.g., Bq m-2) or are you trying to evaluate dose rate? I'll consider both, but I'll spend most of the discussion dealing with the evaluation of the extent of surface contamination.
When an RDD has been detonated, the distribution of radioactivity on the ground may not be well defined, especially in the early times following the event. At least for initial activity evaluations, it is not uncommon to assume that the activity distribution may be estimated as a planar surface source of uniform activity distribution. The areal extent of the source may be known from preliminary bounding measurements, or one may assume, from a practical standpoint, that the source is infinite in extent. This assumption of an infinite planar source may be reasonable when the measurement point is relatively close to the ground surface and the radioactivity is spread over a relatively large area. Clearly, these assumptions may not be very good in areas, such as downtown city areas, where there are numerous buildings and other objects that may also be contaminated and/or possibly provide some shielding. Additionally, depending on the nature of the RDD, chunks of the device may be spread around and may result in very nonuniform activity distribution. Such geometries may be very difficult to model effectively. The most critical early measurements are often dose (rate) measurements necessary to evaluate the immediate external radiation threat in the area. Instruments such as ionization chambers (micro dose meters may be helpful if dose rates are less than 10 µSv h-1) are generally more suitable for such measurements than are NaI detectors, since the latter have a marked energy-dependent response and conversion of count rate to dose rate requires corrections that depend on the specific photon energies involved.
You state that you are using an NaI detector, although you do not specify whether you are using the detector to do gamma-energy spectrometry or whether you are simply measuring a gross count rate above the surface. Since you are concerned about not being able to distinguish the RDD radiation from background, I assume that you are simply doing gross counting with no energy spectrometry, since spectrometry would allow you to look at specific gamma rays associated with the RDD nuclides. It is true that some gamma-ray energies from natural background activity overlap with gamma rays from possible RDD nuclides, but peak shapes and ratios among various peak heights can also be used to identify RDD nuclides among the background radionuclides.
When doing gross gamma counting with the NaI detector you have a couple of options for dealing with the background response of the detector. The first is, as you have noted, simply to ignore the background contribution and assume that any elevated readings are associated with the contribution from the RDD radionuclide(s). If you do this you should have at least some idea of the normal background to be expected on the NaI detector—for example, for a 5.08 cm × 5.08 cm NaI detector the typical background count rate is on the order of 170 cps. Alternatively, for a better estimate, you could measure the background in an uncontaminated area that is similar in ground characteristics to the RDD-affected area and subtract this from the gross reading obtained in the area of interest.
One of the difficulties in making simple count-rate measurements with a NaI detector is that the detector exhibits a strong energy-dependent response, and in order to correlate count rate with activity distribution (or with dose rate) you should have some knowledge of what radionuclides are involved so that appropriate counting efficiencies may be used. If samples of the contamination can be taken and subjected to gamma spectrometry, the identities of the involved radionuclides may be determined.
It is also possible to perform in-situ gamma spectrometry using a portable system. There are relatively simple handheld systems that allow gamma spectrometric measurements. Common instruments of this type use a NaI detector and provide a liquid crystal display to display the pulse height distribution that allows identification of the pertinent radionuclide(s). The quantitative determination of ground contamination would then require the assumption of a specific geometric source model and detector response functions to convert count rates to activity distribution. Here are links to a couple of companies that offer handheld instruments for gamma spectrometric measurements—the first is SAIC, which offers the GR-135 instrument, and the second is Ludlum Measurements, Inc., which offers several models (different NaI detector sizes), this link being to a 2.54 cm × 2.54 cm size, the Model 701. There are numerous companies that offer similar devices, and the examples here are not intended as endorsements of either product.
There has been quite a bit of work done in the development of more sophisticated systems that, for a given assumed source geometry, use predetermined counting efficiencies (from Monte Carlo simulations) for the gamma energy of interest, to evaluate the activity distribution. One such system is produced by Canberra (now under AREVA) under the acronym name ISOCS (In Situ Object Counting System). It is primarily available for use with germanium detectors, and you can find a description on the Canberra Web site. There has also been some work done in the ISOCS approach to establish efficiencies for NaI detectors (5.08 cm × 5.08 cm and 7.62 cm × 7.62 cm sizes); here is a link to an IEEE paper by Canberra authors S.A. Phillips and colleagues. The ISOCS system is very convenient because it allows the user to make changes in the source geometry to evaluate the impact on activity estimations. It is portable but is not a handheld system.
In summary, the approach you take will depend on what instrumentation you have available and the knowledge you have about the specific radionuclides involved and the ability to model the geometry of the contamination. If you are not using a system that performs activity-response correlations automatically, in order to perform rapid assessments shortly following an event, you should have worked out in advance the relationships between radionuclide activity concentration (e.g., Bq m-2) and instrument reading for the likely radionuclides that might be used in RDDs and for common geometries.
I hope this helps.
George Chabot, PhD, CHP