"Exact" and approximate methods for analysis of total error in radiation measurements.

Radiation measurement results are typically stated as an estimated activity concentration with an associated 95% error bound. Very frequently, however, the error bound cited at the "95% confidence level" is nothing more than counting error and, especially at higher levels of activity, counting error may constitute a gross understatement of the total error that should reasonably be attributed to the measurement. This paper compares the measurement confidence limits obtained using a typical approximate error propagation procedure with the "exact" confidence limits. A Monte Carlo error propagation method is also considered. The results of the three methods are compared using 222Rn measurement procedures for illustrative purposes.

Detection limit concepts: foundations, myths, and utilization.

Detection limit parameters such as the minimum detectable concentration have been widely discussed in the literature for more than 30 y. Misunderstanding and misapplication of these parameters continue to be widespread and, indeed, even encoded into computer programs, especially those developed in recent years for use with PC-based analyzer boards. This brief review of the principal concepts related to the most common types of detection limits provides guidance for correctly using these parameters. In particular, several myths concerning issues such as the "sample-specific minimum detectable concentration" and the "unreliability" of measurements made below this concentration are discussed.

A performance evaluation study of three types of alpha-track detector radon monitors.

Three models of alpha-track detector (ATD) Rn monitors were exposed in Environmental Protection Agency (EPA) Rn chambers to obtain estimates of precision and bias for the National Residential Radon Survey (NRRS). Exposures in this study ranged from 37 to 740 Bq y m-3 (1 to 20 pCi y L-1), plus blanks. These exposures correspond to the range expected in most U.S. residences. All detectors were purchased through a Rn mitigation firm to assure that the vendors did not give special attention to the ATDs used in this study. Ten ATDs of each model were studied at 12 exposures. The mean and standard deviation of the reported values for each model were calculated and compared with the continuously monitored chamber concentrations to determine the bias and precision at each exposure. Results of this analysis were discussed with the vendors, who took corrective actions. Changes in track counting procedures and calibrations improved detector performance. Readings of one detector were adjusted based on a regression of the monitored values on the reported values.

Calibration of Scintillation Cells for Radon-222 Measurements at the U.S. Environmental Protection Agency.

Zinc sulfide coated scintillation cells are the primary method for measuring radon-222 at the U.S. Environmental Protection Agency (EPA), Office of Radiation Programs (ORP), Eastern Environmental Radiation Facility (EERF). These cells are used to measure concentrations of radon in exposure chambers that are used to calibrate or test other devices or instruments. Individual cells are calibrated by analyzing samples of air with known concentrations of radon produced by emanation of radon from standard radium-226 solutions obtained from the National Institute of Standards and Technology. The calibration procedure includes ingrowth of radon-222 into equilibrium with the radium in the standard solution, transfer from the solution into an evacuated container, and dilution with a measured volume of air. Samples of the radon in air mixture are transferred to evacuated scintillation cells and sealed for 4 h prior to counting, which allows secular equilibrium to be established between the radon and its decay products. Calibration factors for each individual cell are computed by decay correcting the radon to the time of collection and calculating the ratio of count rate (cpm), corrected for background, to radon activity (Bq) for the specific volume of the cell. Four or more calibration factors are determined for each cell and aver-aged to provide the calibration factor used for measurements. Calibrations are repeated at 6-mo intervals, and the results of each calibration are compared to the previous averages. When calibration factors fall outside the 95% confidence interval, they are rejected and the cell is checked for defects prior to recalibration.