Unexpected bias in NIST 4πγ ionization chamber measurements.

In January of 2010, it was discovered that the source holder used for calibrations in the NIST 4πγ ionization chamber (IC) has not been stable. The positioning ring that determines the height of the sample in the reentrant tube of the IC has slowly shifted during 35 years of use. This has led to a slow change in the calibration factors for the various radionuclides measured by this instrument. The changes are dependent on γ-ray energy and the time the IC was calibrated for a given radionuclide. A review of the historic data with regard to when the calibrations were done has enabled us to approximate the magnitude of the changes with time. This requires a number of assumptions, and corresponding uncertainty components, including whether the changes in height were gradual or in steps as will be shown in drawings of sample holder. For calibrations the changes in calibration factors have been most significant for low energy gamma emitters such as (133)Xe, (241)Am, (125)I and (85)Kr. The corrections to previous calibrations can be approximated and the results corrected with an increase in the overall uncertainty. At present we are recalibrating the IC based on new primary measurements of the radionuclides measured on the IC. Likewise we have been calibrating a new automated ionization-chamber system. A bigger problem is the significant number of half-life results NIST has published over the last 35 years that are based on IC measurements. The effect on half-life is largest for long-lived radionuclei, especially low-energy γ-ray emitters. This presentation will review our results and recommend changes in values and/or uncertainties. Any recommendation for withdrawal of any results will also be undertaken.