Recently ASTM adopted a Test Method for the Determination of Reference Temperature, T_o, for Ferritic Steels in the Transition Range (E1921-97). This test method quantifies the variation of fracture toughness with temperature throughout the transition regime based on limited replicate testing at a single temperature. Advances in elastic plastic fracture mechanics enabled adoption of size requirements in E1921-97 that permit testing of specimens as small as 10 mm cross section. Thus, E1921-97 presents the first opportunity to measure directly the fracture toughness of samples removed from surveillance capsules, rather than inferring toughness using RT_NDT indexed K_IC curve.

We assemble a database of fracture toughness values for reactor pressure vessel steels, including over 1,600 E1921-97 valid fracture toughness values for both irradiated and unirradiated steels. In this paper, we use this database to address concerns regarding the bias and pecision of T_o values determined by ASTM E1921-97, and regarding the use of T_o as an index temperature for the ASME K_IC curve. When the requirements of E1921-97 are satisfied, neither the temperature at which the fracture toughness tests are conducted nor the level of deformation prior to fracture bias the estimate of T_o. Furthermore, the recently proposed T_o-based index for the K_IC curve (RT_To ≡ T_o+ 35°F) reduces considerably the degree of scatter in facture toughness data, and contains implicit margins on toughness that are consistent for every steel considered. RT_To is superior to RT_NDT as an index temperature for the K_IC curve, and also provides for a more conservative assessment of vessel safety than does RT_NDT.