Answer to Question #7823 Submitted to "Ask the Experts"
The following question was answered by an expert in the appropriate field:
I have been told (but not seen any objective supporting evidence yet) that commercial laboratory measurements of 226Ra using the ingrowth method consistently underreport the "true" concentration of 226Ra. As a result, measurement by gamma spectrometry, using the 186 KeV line, is advocated. Can you point me to published data, pro or con, with respect to this issue?
I cannot vouch for your cited observation that 226Ra analyses consistently underestimate the amount of 226Ra present when the ingrowth method is used, although underestimation has been observed. The fact is that the ingrowth method is often used for 226Ra analysis with good results. Variability in results and underestimates or overestimates may occur for a number of reasons. I shall assume initially that the analytical technique involving ingrowth of the short-lived progeny involves counting of the gamma rays from 214Pb/Bi.
In many instances, especially when dealing with low-activity water samples, the radium is first chemically separated from other sample constituents by dissolving the sample and coprecipitating the radium with BaSO4. The radium yield is determined by filtering, drying, and weighing the BaSO4 to determine the recovery of barium. The BaSO4 is a quite insoluble crystalline material that generally is quite effective in trapping radon gas as it grows into the radium in the precipitate. Depending on the physical and chemical conditions when the precipitate was formed and the aging of the precipitate in the solution prior to filtration, the sizes and other physical characteristics of crystals may vary, and this can affect the likelihood of escape of radon from the crystalline matrix. In unsealed preparations of BaSO4 radon can simply diffuse out of the crystal and/or effuse through crystalline defects and be lost. Some radon is also inevitably lost through the process of alpha particle recoil; when a 226Ra atom decays close to the surface of a crystal of BaSO4, the radon daughter may recoil with sufficient energy to leave the crystalline matrix and be lost. In general, losses of 222Rn upset the equilibrium ingrowth of the short-lived progeny 218Po, 214Pb, 214Bi, and 218Po. Thus, radon loss is expected to decrease the interpreted 226Ra content, inferred from counting the deficient progeny, as might be done in gamma counting of the 214Pb/Bi.
If the radium is not first separated from the rest of the sample matrix, the daughters may simply be allowed to grow into the sample, which could be a solid or liquid if gamma counting is being used. If radium concentrations are low, as is often the case, liquid samples might be evaporated, often to dryness to allow more efficient detection. Again, if samples are not properly sealed, some radon loss may occur with a possible result possibly worse than that noted above. Samples such as soil to be analyzed may also be counted directly, and care should also be taken that the samples are in containers that can be effectively sealed to minimize radon escape; also the void volume above the sample should be minimized so as to prevent radon from accumulating in that space and changing the ultimate counting efficiency for the photons of interest.
If one opts not to use the ingrowth approach, and the sample contains any significant uranium, when counting the 186 keV photons from 226Ra there is the possibility for added counts from the 185 keV photons emitted by 235U. The latter would lead to overestimation of the 226Ra content of the sample and must be corrected for in order to obtain correct results (not always easy to do).
We should also note that other “ingrowth” methods are also used for 226Ra determination. One of the most notable of these is the radon emanation technique that involves allowing the ingrowth of 222Rn into a prepared sample and then purging of the radon gas from the sample into an appropriate scintillation cell. After a few hours of ingrowth time, the sample in the scintillation cell is counted using a photomultiplier tube to detect the alpha-induced scintillation events. This method is quite accurate if carried out properly but, as is the case for many analytical procedures, there are various occasions for error. For example, one that I have observed involves a reduction in the expected amount of radon as a consequence of some radon being trapped in insoluble material prior to purging. This has been most noticeable in water samples in which the radium has been concentrated by BaSO4 precipitation; the water-insoluble BaSO4 has to be converted into a suitably soluble form so that the radon that grows in is available for purging. A common mistake I have seen is in the inadequacy of this chemical conversion step so that some of the Ba(Ra)SO4 remains insoluble with some 222Rn occluded within the crystalline material. This naturally leads to a reduced value for the 226Ra.
There are numerous articles in the literature that discuss various problems with analysis of samples for 226Ra content, and you can access some of them on the Internet. A recent article by Arndt and West appeared in Health Physics ("An Experimental Analysis of the Contribution of 224Ra and 226Ra and Progeny to the Gross Alpha-Particle Activity of Water Samples," 94:5; 2008) and is available in full text to Health Physics Society (HPS) members on the HPS site (www.hps.org). The article relates to gross alpha counting but discusses some of the problems of analysis that apply in other situations as well. You can find numerous other articles related to 226Ra analysis by typing 226Ra in Health Physics search box. The problem of contribution of 235U to 226Ra gamma counting has been known for many years, and many instances can also be found on the Internet. Here is one link to a 1998 report from Idaho National Environmental Laboratory that addresses the problem.
I hope this is helpful.
George Chabot, PhD, CHP