The instantaneous values of the airborne activity concentrations of radon progeny and thoron progeny have been determined 34 times in a closed and windowless room in a cellar using two independent grab-sampling methods in order to compare the performance of the methods. The activity concentration of radon ( 222Rn) was also measured and it varied between 200 and 650 Bq m −3. Two samples of radon and thoron progeny were collected simultaneously from roughly the same air volume by filtering. For the first method, the isotopes were collected on membrane filter and gross α counting was applied over several successive time intervals. This method was a slightly improved version of the methods that are applied generally for this reason for decades. For the second method, the isotopes were collected on glass-fibre filter and gross β counts were registered over several time intervals. This other method was developed a few years ago and the above series of measurements was the first opportunity to make an intercomparison between it and another similar method based on α counting. Individual radon progeny and thoron progeny activity concentrations (for the isotopes 218Po, 214Pb, 214Bi and 212Pb) were evaluated by both methods. The detailed investigation of the results showed that the systematic deviation of the methods is small but significant and isotope-dependent. The weighted averages of the β/α activity concentration ratios for 218Po, 214Pb, 214Bi, EEDC 222 (Equilibrium-Equivalent Decay-product Concentration of radon progeny) and 212Pb were 0.99±0.03, 0.90±0.02, 1.03±0.02, 0.96±0.02 and 0.80±0.03, respectively. The source of the systematic deviation is probably the inaccurate knowledge of the counting efficiencies mainly in the case of the α-counting method. A significant random-type difference between the results obtained with the two methods has also been revealed. For example, the β/α ratio for EEDC 222 varied between 0.81±0.01 and 1.22±0.03, where the errors derive from the random errors of the independent β and α results. The detailed investigation of the results showed that the α method is responsible for the overwhelming part of this random deviation that was caused mainly by the lower accuracy, stability and reliability of this method. The performance of the β method proved to be better mainly because of its higher accuracy, stability and reliability.
Read full abstract