Answer to Question #8288 Submitted to "Ask the Experts"

I work in a nuclear propulsion program as a health physicist. I have worked at prototype reactor complexes contaminated with high-energy beta emitters (e.g, ⁹⁰Sr) and mixed isotopes, however, the bulk of my experience is only with ⁶⁰Co. We had a fixed contamination problem I was not directly involved with and this is where the question comes in. I am tasked to write a paper for this problem for presentation by my nuclear director at a director's conference. Our lab uses the typical germanium detector to do isotope identification. The germanium counter revealed no gamma emitters. Fixed radioactivity on this particular apparatus was indicated with a frisker and was suspected to be a low-energy beta emitter. No alpha was detected. We do not have a liquid scintillation counter to do beta identification. Somehow (the lab has not walked me through the exact procedure as of yet) our lab determined it was ¹⁴C based on an "attenuation kit." I'm guessing it is an antiquated (as in the '50s kind of antiquated) process. Supposedly the lab folks compared the sample with the beta emitter to a standard ¹⁴C source and used the attenuation kit to show the proper reduction factor for ¹⁴C. Is this procedure standardized somewhere? It obviously couldn't be an exact scientific process or the use would be widespread. Is this a historical health physics procedure phased out with the advent of accurate electronic meters? Any assistance you can provide would be appreciated.

You are correct in some of your inferences regarding the methodology used in attempting to identify the radionuclide of concern. I shall address this issue shortly, but I will first make a few comments regarding some other considerations that affect conclusions as to the radionuclide identity.

Whether ¹⁴C is a likely candidate for the radionuclide in question depends on a few factors, perhaps the most pertinent of which is whether ¹⁴C was a radionuclide in use at your facility and whether it could have ended up as a contaminant on the affected area/object. Since the laboratory staff has made a likely identity as ¹⁴C, I shall assume that this was a nuclide in use and that its presence as a contaminant is feasible.

A second question is what other radionuclides have been in use that might be confused with ¹⁴C in terms of the deterministic measurements that were made. For example, ³⁵S is also a pure beta emitter with end point and average beta energies quite similar to the respective values for ¹⁴C. If it has been in relatively recent use (it has a half-life of about 90 days), ³⁵S might require consideration.

A third consideration relates, at least in part, to the characteristics of the germanium detector used for radionuclide identification by gamma and/or x-ray energy analysis. You state that it is a "typical germanium detector." My judgment would then be that the detector is probably something like a moderate volume coaxial detector with an encapsulating aluminum cap that allows energy evaluation above approximately 50 keV. If this is the case, it could also be possible that your radionuclide of concern emitted some low-energy photons that were not detected. Such might be the case, for example, if ⁵⁵Fe were a radionuclide available at your facility. This radionuclide decays by electron capture and emits 5 keV Auger electrons in about 60% of its disintegrations and 6 keV x rays in about 28% of the disintegrations. These electrons and photons are readily attenuated and it is possible that attenuation measurements of the type you allude to could confuse this radionuclide with some low-energy beta emitters.

If the likelihood of some of these complicating factors can be ruled out, and the nuclide of concern is, indeed, a pure beta emitter, then attenuation measurements, using sequential absorber films (usually fabricated of aluminum or plastic) of increasing thickness can provide useful information. These absorber kits were standard fare in many student and professional laboratories for many decades and still are used when alternative energy assessment techniques are not available or, in the case of educational facilities, when certain radiation attenuation characteristics are being demonstrated. It is possible to perform a series of attenuation measurements with the beta source held in a fixed position and attenuators of known thicknesses inserted between the source and the detector. The count rate is recorded as a function of attenuator thickness, usually in mg cm^-2, with increasing thicknesses until the count rate is reduced to the background level. The background count rate may be due completely to natural background, independent of the source, or it may be due to natural background and photons that might be emitted by the source (in the event that the radionuclide emits some photons) or bremsstrahlung radiation produced by interactions of the beta particles in the attenuators and/or in the source material or holder. A simple visual inspection of the curve can be made to estimate where the beta attenuation portion intersects the background portion, this intersection point approximating the range of the maximum-energy beta particles. This range can then be looked up in appropriate range tables to estimate the maximum beta particle energy. Such an analysis is subject to some error, especially because as the added thickness of attenuator gets close to the maximum range the angle of approach of the beta attenuation curve to the background portion is very shallow, beta count rates are low, and it is difficult to make an accurate estimate of where the two portions of the curve intersect.

Generally the data are plotted with count rate on a log scale, the background portion of the curve extrapolated back to zero thickness, and count rates on this extrapolated line are subtracted from the respective gross count rates to obtain a net count rate attenuation curve. Much of the curve on the semilog plot will be close to a straight line (not the tail, however). Similar data can also be generated for a source of a known radionuclide that is suspected to have the same identity as the unknown. Comparisons of the curves can provide evidence that the radionuclides are the same or not. Some consideration must be paid to possible differences in the extent of beta-particle self attenuation that might be present in the unknown compared to the standard.

Undoubtedly, the most well-known and among the best of the attenuation methods for identifying beta-particle energies is that ascribed to N. Feather, who did the original work in the late 1930s. The "Feather plot" method requires having a beta source of known energy characteristics (for many years ²¹⁰Bi was the best characterized and most used, but others may be used). The net attenuation curve, obtained as noted above, is plotted on semilog paper. A similar attenuation curve is also generated for the unknown sample, and results are normalized to the same count rate as the standard at zero added attenuator thickness. Both curves are plotted on the same paper. A thickness corresponding to a specified fractional range for the standard (e.g., 0.10 of the maximum range) is selected, and the count rate read from the ordinate. The same count rate is then located on the unknown curve, the corresponding thickness is read, and the assumption is made that the same fractional range applies (in this case 0.10); the range estimate is made by dividing the thickness by the fractional range. This process is repeated for several more selected fractional ranges. Because of differences in spectral shapes between the standard and unknown, the unknown beta range estimates made for each selected fractional range will be different, but when these ranges are plotted vs. fractional range they typically will fall on a smooth curve that can easily be extrapolated to a fractional range of 1.0, which provides the best estimate of the actual range for the unknown. One advantage of the attenuation measurement technique is that it requires no very special detector. Any beta-sensitive detector with a sufficiently thin window to allow beta-particle penetration will suffice. The method is not very useful when more than one beta-emitting radionuclide is present, although visual inspection of the transmission curves may still yield helpful information. You can find the "Feather plot" method described in many textbooks that deal  with nuclear physics, radiochemistry, etc., especially some of the earlier books—e.g., R. Evans, The Atomic Nucleus, McGraw-Hill, 1955; Friedlander and Kennedy, Nuclear and Radiochemistry, Wiley and Sons, 1955; Lapp and Andrews, Nuclear Radiation Physics, 4th ed., Prentice-Hall, 1972.

There are other more sophisticated instruments available at some facilities that allow direct measurement of beta-particle energy spectra. These include such devices as magnetic and electrostatic energy spectrometers and surface barrier type semiconductor detectors with thick depleted regions. Liquid scintillation systems can also be used to obtain energy spectral data, especially when outputs are directed to a multichannel analyzing system.

I hope this is helpful to you.

George Chabot, PhD, CHP