Evaluation of variability in radionuclide measurements in drinking water

Abstract

A blind interlaboratory study of radium‐226 (Ra‐226), radium‐228 (Ra‐228), and gross alpha and beta activity was conducted among five multistate certified laboratories to examine concerns regarding analytical variability for compliance with the revised Radionuclides Rule. Results of the interlaboratory study demonstrated that although Ra‐226 and gross beta measurements appeared to be both accurate and precise, even at low levels, Ra‐228 and gross alpha measurements were much less robust both within and among laboratories, which could result in potentially incorrect compliance decisions. There are several steps that utilities can take to improve reliability of compliance data, including mandating specific quality control (QC) steps by laboratories and ensuring that they follow the QC guidance in the US Environmental Protection Agency's laboratory certification manual and the Multi‐Agency Radiological Laboratory Analytical Protocols Manual, even though these may not always be enforced by laboratory certification authorities.

Interlaboratory Comparison of ²²⁶Ra and ²²⁸Ra Activity Concentrations in Groundwater and Surface Water

“…The results obtained from this ILC of Site groundwater and surface water samples are generally consistent with the findings from the prior 2011 study [5] with respect to observed variability in 226 Ra compared to 228 Ra. The current ILC indicate that 228 Ra data reported by the existing contract laboratory (Lab A) are biased high relative to those reported by Lab B, Lab C, Lab D, and this may be due to differences in laboratory skill and technique among the various laboratory technicians [5]. However, the resulting upward trends in 228 Ra/ 226 Ra ratios (observed in both impacted wells and the background location) after the 228 Ra method was changed from 904.0 to Ra-05 are more difficult to explain.…”

“…Lab A also reported the highest measured 228 Ra concentration of all field blanks (2 pCi/L), and consequently those data for total 228 Ra were qualified as estimated with high bias. A prior study relating to radionuclide variability in water samples [5] showed that radionuclide activity in groundwater can vary up to four-fold due to seasonal effects, but also noted that the variability could reflect issues related to method and laboratory variability. The greatest source of error for 226 Ra is counting uncertainty, but the overall accuracy and precision for 226 Ra results was acceptable among various laboratories.…”

“…For 228 Ra, however, the methods are generally not as robust as those used for 226 Ra. No single laboratory was able to provide consistently accurate 228 Ra results, and due to the much greater relative standard deviation (RSD) values, it was concluded that the accuracy of 228 Ra results are dependent on laboratory technique and skill and "may also reflect less rugged analytical methods" [5].…”

“…Even though this study did not indicate that M903.0 results in consistent over-reporting of 226 Ra concentrations, use of M903.1 would better ensure that the appropriate analysis for 226 Ra is being conducted. Alternatively, it could be requested that the laboratory employ a 14-day waiting time which will effectively reduce the contribution of 224 Ra to the reported 226 Ra [5].…”

Interlaboratory Comparison of 226Ra and 228Ra Activity Concentrations in Groundwater and Surface Water

“…For 228 Ra, two of the laboratories used Method 904.0 (M904.0) (Labs B, C), while Lab A and Lab D used Method Ra-05. Each Ra method in Table 2 has been approved for drinking water compliance analysis by the USEPA and therefore should not produce significantly different results [6].…”