[en] We have developed a method to test powder leakage that has passed O-ring seals. To validate this method we have spiked a test fixture with 98 ng of U and recovered 130 +- 25 ng of U. We did not detect U at a detection limit of 26 ng in a fixture which was treated as a blank. This method has been applied to the leakage of UO₂ powder passing the type of EPDM O-ring seals used in a SNM shipping cask belonging to PNC. Considering the three experimental tests in which no or very small quantities of U were detected as effective blank test, it appears that the level of external contamination is negligible. Therefore, we believe that the U quantities greater than 26 ng (6 tests) passed the primary O-ring seal. From this limited quantity of data, we observe no apparent correlation between the amount of U measured and either helium leak rate or equivalent tube diameter. The data for the 130⁰C tests indicate the possibility of a U/time relationship; however, more data are needed for verification