Standard measurement procedure for soil radon exhalation rate and its uncertainty

Radon contributing about 42% of annual average dose, mainly comes from soil. In this paper, standard measurement procedures for soil radon exhalation rate are suggested and their measurement uncertainties are analyzed. We used accumulation method for estimating surface exhalation rate. The closed-loop measurement system was made up with a RAD7 detector and a surface chamber. Radon activity concentrations in the system were observed as a function of time, with data collection of 5 and 15-minute and the measurement time of 4 hours. Linear and exponential fittings were used to obtain radon exhalation rates from observed data. Standard deviations of measurement uncertainties for two approaches were estimated using usual propagation rules. The exhalation rates (E) from linear approach, with 30 minutes measurement time were 44.8-48.6 mBq∙m^-2∙ s^-1 or 2.14-2.32 atom∙cm^-2∙ s^-1 with relative measurement uncertainty of about 10%. The contributions of fitting parameter A, volume (V) and surface (S) to the estimated measurement uncertainty of E were 59.8%, 30.1% and 10.1%, in average respectively. In exponential fitting, at 3-hour measurement we had E ranged of 51.6-69.2 mBq∙m^-2∙ s^-1 or 2.46-3.30 atom∙cm^-2∙ s^-1 with about 15% relative uncertainty. Fitting with 4-hour measurement resulted E about 51.3-68.2 mBq∙m^-2 ∙ s^-1 or 2.45-3.25 atom∙ cm^-2 ∙ s^-1 with 10% relative uncertainty. The uncertainty contributions in exponential approach were 75.1%, 13.4%, 8.7%, and 2.9% for total decay constant k, fitting parameter B, V, and S, respectively. In obtaining exhalation rates, the linear approach is easy to apply, but by saturation feature of radon concentrations, the slope tends to decrease away from the expected slope for extended measurement time. For linear approach, measurement time of 1-hour or less was suggested. For exponential approach, the obtained exhalation rates showed similar values for any measurement time, but measurement time of 3-hour or more was suggested for about 10% relative uncertainty